From 39baeb53fe254d8ac506e8ed07fd8a4094c72932 Mon Sep 17 00:00:00 2001 From: Diogo Ribeiro Date: Sun, 13 Oct 2024 00:05:14 +0100 Subject: [PATCH 1/7] fix --- _posts/2024-06-13-Stepwise_regression.md | 1 - 1 file changed, 1 deletion(-) diff --git a/_posts/2024-06-13-Stepwise_regression.md b/_posts/2024-06-13-Stepwise_regression.md index 9946b6a2..417afd32 100644 --- a/_posts/2024-06-13-Stepwise_regression.md +++ b/_posts/2024-06-13-Stepwise_regression.md @@ -25,7 +25,6 @@ tags: - Python - R - Julia -- '[''mathematics'']' - Statistics - Data science title: 'Stepwise Regression: Methodology, Applications, and Concerns' From 060a4c488052a263ea5951e90cbdceaed3436d70 Mon Sep 17 00:00:00 2001 From: Diogo Ribeiro Date: Sun, 13 Oct 2024 00:06:00 +0100 Subject: [PATCH 2/7] fixes --- _posts/-_ideas/2030-01-01-climate_change.md | 9 +++---- .../2030-01-01-ideas_statistical_tests.md | 9 +++---- ...LP and Data Science Article Topic Ideas.md | 9 +++---- _posts/-_ideas/numerical_methods_fortran.md | 2 ++ .../2020-01-01-causality_and_correlation.md | 11 +++----- ...elihood_estimation_statistical_modeling.md | 15 ++++++----- ...ing_goodness-of-fit_non-parametric_data.md | 13 +++------- ...:_bonferroni_correction_other_solutions.md | 15 ++++------- ..._anova_vs._two-way_anova_when_use_which.md | 12 +++------ ...1-08-heteroscedascity_statistical_tests.md | 9 +++---- ...20-01-30-cox_proportional_hazards_model.md | 13 +++++----- _posts/2020-02-01-anova_kruskal_walis.md | 11 +++----- ...istical_testing:_null_hypothesis_beyond.md | 11 +++----- _posts/2020-02-17-arimax_time_series.md | 13 ++++------ _posts/2020-03-01-type_one_type_two_erros.md | 12 +++------ ...ow_data_science_drives_green_innovation.md | 11 +++----- _posts/2020-04-01-the_friedman_test.md | 12 +++------ ...7-prediction_errors_bias_variance_model.md | 14 ++++------- _posts/2020-05-01-shapiro_wilk_test.md | 12 +++------ _posts/2020-05-26-false_positive_rate.md | 13 ++++------ ...-06-01-ordinary_least_square_regression.md | 16 +++--------- _posts/2020-06-10-arima_time_series.md | 16 ++++++------ _posts/2020-07-01-cocharan_q_test.md | 16 ++++-------- ...dent_t_test_non_parametric_alternatives.md | 16 +++--------- _posts/2020-07-26-measurement_errors.md | 11 +++----- ...-understanding_markov_chain_monte_carlo.md | 16 ++++++------ ...lassification_zero_inflated_time_series.md | 18 ++++--------- ...ival_analysis_comparing_survival_curves.md | 16 +++--------- ...2020-09-24-demand_forecast_supply_chain.md | 18 +++++-------- ...time_series_models_predicting_emergency.md | 17 ++++--------- ...-01-predictive_maintenance_data_science.md | 12 +++------ _posts/2020-12-30-ordinal_regression.md | 14 ++++------- _posts/2021-01-01-pde_data_science.md | 14 +++-------- _posts/2021-02-01-bayesian.md | 13 +++------- _posts/2021-02-17-traffic_safety_kde.md | 22 ++++++---------- _posts/2021-03-01-polynomial_regression.md | 17 +++---------- _posts/2021-03-01-type_1_type_2_errors.md | 15 +++-------- ...21-04-01-asymmetric_confidence_interval.md | 14 +++++------ _posts/2021-04-27-forest_fires_kde.md | 23 +++++++---------- ...4-30-big_data_climate_change_mitigation.md | 11 +++----- ...2021-05-01-rare_labels_machine_learning.md | 14 ++++------- ...ural_networks_using_monte_carlo_dropout.md | 13 +++------- ...coefficient_variation_health_monitoring.md | 13 +++------- _posts/2021-05-26-kernel_math.md | 17 +++---------- _posts/2021-06-01-customer_segmentation.md | 14 ++++------- _posts/2021-07-26-regression_tasks.md | 11 +++----- ...building_linear_regression_from_scratch.md | 13 ++++------ _posts/2021-09-24-crime_analysis.md | 13 ++++------ _posts/2021-12-24-linear_programming.md | 15 +++-------- _posts/2021-12-25-suply_chain.md | 8 ++---- _posts/2021-12-31-FDM.md | 22 +++++++--------- _posts/2022-01-02-OLS.md | 7 ++---- _posts/2022-02-17-staff_schedulling.md | 14 +++++------ ...tts_test_checking_homogeneity_variances.md | 22 ++++++---------- _posts/2022-03-15-bayesian_ab_testing.md | 9 +++---- _posts/2022-03-23-degrees_freedom.md | 9 ++----- _posts/2022-05-26-networks.md | 7 ++---- _posts/2022-07-23-statistical_tests.md | 12 +++------ _posts/2022-07-26-features.md | 13 +++------- ..._hypothesis_testing_regression_analysis.md | 12 +++------ _posts/2022-08-15-linear_relashionships.md | 12 +++------ .../2022-09-27-entropy_information_theory.md | 12 +++------ _posts/2022-10-31-Jacknife.md | 7 ++---- _posts/2022-11-30-Bootstrap.md | 13 ++++------ ...2022-12-25-probability_machine_learning.md | 6 ++--- _posts/2022-12-30-simpsons_paradox.md | 10 +++----- _posts/2022-12-31-PCA_explained.md | 16 ++++-------- _posts/2023-01-01-error_coefficientes.md | 12 +++------ _posts/2023-01-08-crownd_behaviour.md | 11 +++----- _posts/2023-02-17-ab_testing.md | 14 ++++++----- .../2023-05-05-Mean_Time_Between_Failures.md | 11 ++++---- _posts/2023-07-23-VAR.md | 7 ++---- _posts/2023-07-26-customer-life-time-value.md | 21 ++++++---------- _posts/2023-08-12-guassian_processes.md | 8 +++--- _posts/2023-08-13-shared_nearest_neighbors.md | 14 ++++------- .../2023-08-21-demystifying_data_science.md | 13 +++------- _posts/2023-08-21-large_languague_models.md | 17 +++---------- _posts/2023-08-22-Paul-Erdos.md | 16 ++++++------ ...multivariate_analysis_variance_vs_anova.md | 16 +++--------- _posts/2023-08-25-runnning_windows.md | 13 ++++------ _posts/2023-08-30-Data_Science.md | 7 ++---- _posts/2023-09-01-regression_path_analysis.md | 15 +++-------- _posts/2023-09-03-binary_classification.md | 8 ++---- _posts/2023-09-04-Fears-Surrounding.md | 8 ++---- _posts/2023-09-08-trafic_dynamics.md | 10 +++----- _posts/2023-09-20-rolling_windows.md | 8 +++--- _posts/2023-09-26-Innumeracy.md | 10 +++----- _posts/2023-09-27-Data_communication.md | 8 ++---- _posts/2023-09-27-sample_size.md | 7 ++---- ...tiple_regression_vs_stepwise_regression.md | 23 +++++++---------- _posts/2023-10-01-coverage_probability.md | 11 +++----- ...atural_language_processing_data_science.md | 11 +++----- ...10-31-detecting_trends_time-series_data.md | 16 ++++++------ _posts/2023-11-01-linear_vs_logistic_model.md | 9 +++---- ...hip_between_continuous_binary_variables.md | 19 ++++---------- ...tric_comparison_two_independent_samples.md | 16 ++++++------ _posts/2023-11-30-math_fundamentals.md | 6 ++--- _posts/2023-12-01-managing_data_science.md | 12 +++------ ...023-12-30-data_engineering_introduction.md | 10 +++----- _posts/2023-12-30-expected_shortfall.md | 10 +++----- ...2024-01-01-mathematics_machine_learning.md | 8 ++---- ...eprocessing_techniques_nlp_data_science.md | 14 +++-------- _posts/2024-01-28-normal_distribution.md | 9 +++---- .../2024-01-29-probabilistic_programming.md | 8 +++--- _posts/2024-01-30-Monte_Carlo.md | 12 ++++----- _posts/2024-02-01-customer_life_value.md | 10 +++----- _posts/2024-02-02-topology_data_science.md | 11 +++----- _posts/2024-02-08-Clustering.md | 8 ++---- _posts/2024-02-09-spectral_clustering.md | 7 ++---- _posts/2024-02-10-pingenhole_principle.md | 18 ++++++------- _posts/2024-02-11-Ergodicity.md | 13 ++++------ _posts/2024-02-11-combinatorics_python.md | 10 +++----- .../2024-02-12-combinatorics_probability.md | 11 +++----- ...-12-ethical_considerations_elderly_care.md | 14 +++-------- ...-02-14-advanced_sequential_change-point.md | 13 ++++------ _posts/2024-02-17-climate_var.md | 9 +++---- _posts/2024-02-20-validate_models.md | 10 +++----- _posts/2024-05-09-kernel_clustering_r.md | 2 ++ _posts/2024-05-09-understanding_t-sne.md | 1 + _posts/2024-05-10-data_analysis_gdp.md | 3 +-- _posts/2024-05-10-survival_analysis.md | 15 ++++------- _posts/2024-05-14-Kullback.md | 1 + _posts/2024-05-14-P_value.md | 4 +-- _posts/2024-05-15-Feature_Engineering.md | 1 + ...24-05-15-detect_multivariate_data_drift.md | 9 +++---- _posts/2024-05-17-Markov_Chain.md | 11 +++----- _posts/2024-05-19-gini_coefficiente.md | 4 +-- ...24-05-21-Probability_integral_transform.md | 1 + _posts/2024-05-22-Peer_review.md | 3 +-- .../2024-06-03-g-test_vs_chi-square_test.md | 12 +++------ _posts/2024-06-04-poisson_distribution.md | 1 + .../2024-06-05-data_science_in_health_tech.md | 8 ++---- .../2024-06-05-sensor_activations_models.md | 10 +++----- _posts/2024-06-06-wine_sensory_evaluation.md | 3 +-- _posts/2024-06-07-z-score.md | 10 +++----- _posts/2024-06-11-survival_analysis.md | 1 + _posts/2024-06-13-Stepwise_regression.md | 3 +++ _posts/2024-06-14-matthew_correlation.md | 25 +++++++++++-------- _posts/2024-06-15-EMI_RSSI_SIGNAL.md | 3 +-- _posts/2024-06-29-GLM.md | 2 ++ _posts/2024-06-30-RSSI_body_effects.md | 10 +++----- _posts/2024-06-30-RSSI_humanbody.md | 11 +++----- _posts/2024-07-02-monitoring_drift.md | 11 +++----- _posts/2024-07-03-ancova.md | 2 ++ _posts/2024-07-04-Logram_test.md | 2 ++ _posts/2024-07-05-savitzky_golay.md | 12 ++++----- _posts/2024-07-07-logistic-model.md | 4 +-- _posts/2024-07-08-PSOD.md | 1 + _posts/2024-07-09-error_bars.md | 8 ++---- .../2024-07-10-prob_distributions_clinical.md | 7 ++---- _posts/2024-07-11-pre_commit.md | 2 ++ _posts/2024-07-13-CLT.md | 5 ++-- _posts/2024-07-14-confidence-intervales.md | 3 +-- _posts/2024-07-14-copulas.md | 1 + _posts/2024-07-15-outlier_detection_doping.md | 10 +++----- _posts/2024-07-16-Einstein.md | 8 ++---- _posts/2024-07-19-clt_revisited.md | 18 ++++--------- _posts/2024-07-20-FPOF.md | 1 + _posts/2024-07-20-sequential_change.md | 1 + _posts/2024-07-21-iknn.md | 1 + _posts/2024-07-31-Custom_libraries.md | 1 + _posts/2024-08-01-Data_leakeage.md | 1 + _posts/2024-08-03-feature_engineering.md | 13 ++++------ _posts/2024-08-15-structural_equations.md | 12 +++------ _posts/2024-08-16-utility_functions_python.md | 14 ++++------- _posts/2024-08-19-pre_comit_tutorial.md | 17 ++++++------- _posts/2024-08-24-circular_economy.md | 14 ++++------- _posts/2024-08-24-kruskal_wallis.md | 17 ++++++------- _posts/2024-08-25-Vehicle_Routing_Problem.md | 17 ++++++------- _posts/2024-08-26-energie.md | 15 ++++------- _posts/2024-08-27-coeeficient_variation.md | 15 ++++------- _posts/2024-08-28-mathematics.md | 13 +++------- _posts/2024-08-31-PAPE.md | 13 +++------- _posts/2024-08-31-pedestrian_movement.md | 19 +++++++------- _posts/2024-09-01-graph_theory.md | 13 +++------- _posts/2024-09-01-math_and_music.md | 13 +++------- _posts/2024-09-03-climate_change.md | 19 +++----------- _posts/2024-09-03-fundamentals_matter.md | 17 +++---------- _posts/2024-09-04-outlier_detection.md | 15 ++++------- _posts/2024-09-05-detecting_drift.md | 16 +++--------- _posts/2024-09-05-real_time_data_streaming.md | 17 ++++++------- _posts/2024-09-06-covariate_shift.md | 13 +++------- _posts/2024-09-06-normality.md | 13 +++------- ...024-09-06-sequential_detection_switches.md | 13 +++------- _posts/2024-09-07-energie_efficiency.md | 15 ++++------- _posts/2024-09-08-nonparametric_tests.md | 21 ++++++++-------- _posts/2024-09-09-kmeans.md | 13 +++------- _posts/2024-09-10-wilcoxon.md | 20 ++++++--------- _posts/2024-09-11-cross_validation.md | 13 +++------- _posts/2024-09-12-importance_sampling.md | 21 ++++++++-------- _posts/2024-09-13-multi_colinearity.md | 13 +++------- _posts/2024-09-14-ML_supply_chain.md | 14 +++-------- _posts/2024-09-15-forest_fiers.md | 16 +++--------- _posts/2024-09-16-ML_and_forest_fires.md | 11 +++----- _posts/2024-09-17-feature_engenniring.md | 10 +++----- _posts/2024-09-17-ml_healthcare.md | 20 +++------------ _posts/2024-09-18-baysean_statistics.md | 18 +++---------- _posts/2024-09-19-build_ds_team.md | 25 +++++-------------- _posts/2024-09-20-model_customer_behaviour.md | 17 ++++--------- _posts/2024-09-21-data_design.md | 20 +++------------ _posts/2024-09-21-data_drift_example.md | 14 +++-------- _posts/2024-09-22-randomized_inference.md | 17 +++---------- _posts/2024-09-23-improving_decision_trees.md | 13 ++++------ _posts/2024-09-24-sample_size_clinical.md | 16 +++--------- _posts/2024-09-25-simuled_anneling.md | 15 ++++------- _posts/2024-09-27-entropy_data_science.md | 15 ++++------- _posts/2024-09-28-roc.auc.md | 9 +++---- ...-business_intelligence_machine_learning.md | 13 +++------- _posts/2024-09-29-causal_inference.md | 18 ++++--------- _posts/2024-09-30-ds_projects.md | 6 ++--- ...oratory_data_analysis_techniques_pandas.md | 16 +++++------- ...2024-10-01-automated_prompt_engineering.md | 17 +++++-------- _posts/2024-10-01-edge_machine_learning.md | 16 ++++-------- ...-building_data_driven_business_strategy.md | 19 ++++---------- _posts/2024-10-02-entropy.md | 7 ++---- ...ting_machine_learning_engineering_mlops.md | 18 ++++--------- ...-10-04-guide_arima_time_series_modeling.md | 17 ++++++------- _posts/2024-10-05-simple_distribution.md | 14 +++-------- _posts/2024-10-06-evaluating_distributions.md | 15 +++-------- _posts/2024-10-07-extending_simple_model.md | 12 +++------ _posts/2024-10-08-implementing_time_series.md | 17 +++++-------- ...10-09-magnitude_matter_machine_learning.md | 15 ++++------- ...drift_what_why_matters_machine_learning.md | 13 +++------- ..._machine_learning_models_fail_over_time.md | 13 +++------- ..._business_strategy_age_machine_learning.md | 15 +++-------- ...oothing_methods_time_series_forecasting.md | 21 ++++++---------- ...tion_seasonal_decomposition_time_series.md | 17 ++++++------- ...-10-31-machine_learning_fall_prediction.md | 12 +++------ _posts/2024-11-01-data_driven_elderly_care.md | 13 +++------- _posts/2024-11-30-outliers.md | 3 +-- ...24-12-01-remote_monitoring_elderly_care.md | 16 +++--------- ...-12-30-predicting_hospital_readmissions.md | 13 +++------- 232 files changed, 902 insertions(+), 1858 deletions(-) diff --git a/_posts/-_ideas/2030-01-01-climate_change.md b/_posts/-_ideas/2030-01-01-climate_change.md index 67c9436e..6cc2565b 100644 --- a/_posts/-_ideas/2030-01-01-climate_change.md +++ b/_posts/-_ideas/2030-01-01-climate_change.md @@ -5,8 +5,7 @@ categories: - Data Science classes: wide date: '2030-01-01' -excerpt: Explore how data science, machine learning, and big data are critical tools - in addressing climate change and promoting sustainability. +excerpt: Explore how data science, machine learning, and big data are critical tools in addressing climate change and promoting sustainability. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_9.jpg @@ -20,12 +19,10 @@ keywords: - Machine learning - Sustainability - Big data -seo_description: An in-depth look at how data science, big data, and machine learning - can help solve climate change and sustainability challenges. +seo_description: An in-depth look at how data science, big data, and machine learning can help solve climate change and sustainability challenges. seo_title: 'Climate Change and Data Science: Solving Global Problems' seo_type: article -summary: This article provides a comprehensive list of potential topics at the intersection - of climate change, sustainability, and data science. +summary: This article provides a comprehensive list of potential topics at the intersection of climate change, sustainability, and data science. tags: - Climate change - Sustainability diff --git a/_posts/-_ideas/2030-01-01-ideas_statistical_tests.md b/_posts/-_ideas/2030-01-01-ideas_statistical_tests.md index 21df8dd1..323ee28c 100644 --- a/_posts/-_ideas/2030-01-01-ideas_statistical_tests.md +++ b/_posts/-_ideas/2030-01-01-ideas_statistical_tests.md @@ -6,8 +6,7 @@ categories: - Hypothesis Testing classes: wide date: '2030-01-01' -excerpt: A list of 15 article ideas covering statistical tests, ranging from ANOVA - and Kruskal-Wallis to non-parametric tests and power analysis. +excerpt: A list of 15 article ideas covering statistical tests, ranging from ANOVA and Kruskal-Wallis to non-parametric tests and power analysis. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_2.jpg @@ -21,12 +20,10 @@ keywords: - Kruskal-wallis - Data analysis - Hypothesis testing -seo_description: Explore 15 ideas for writing articles on various statistical tests, - including their differences, assumptions, and applications in data analysis. +seo_description: Explore 15 ideas for writing articles on various statistical tests, including their differences, assumptions, and applications in data analysis. seo_title: '15 Article Ideas: Writing about Statistical Tests' seo_type: article -summary: This article provides 15 ideas for articles on statistical tests, including - their use cases, assumptions, and applications in real-world data analysis. +summary: This article provides 15 ideas for articles on statistical tests, including their use cases, assumptions, and applications in real-world data analysis. tags: - Anova - Hypothesis testing diff --git a/_posts/-_ideas/NLP and Data Science Article Topic Ideas.md b/_posts/-_ideas/NLP and Data Science Article Topic Ideas.md index c5040fcb..91214bb6 100644 --- a/_posts/-_ideas/NLP and Data Science Article Topic Ideas.md +++ b/_posts/-_ideas/NLP and Data Science Article Topic Ideas.md @@ -4,19 +4,16 @@ categories: - NLP - Data Science classes: wide -excerpt: Explore in-depth article topics combining NLP and Data Science, from text - preprocessing to deep learning models, sentiment analysis, and chatbots. +excerpt: Explore in-depth article topics combining NLP and Data Science, from text preprocessing to deep learning models, sentiment analysis, and chatbots. keywords: - Nlp - Data science - Machine learning - Topic modeling - Sentiment analysis -seo_description: Explore in-depth article topics combining Natural Language Processing - and Data Science, covering a range of tasks, models, and techniques. +seo_description: Explore in-depth article topics combining Natural Language Processing and Data Science, covering a range of tasks, models, and techniques. seo_title: 'NLP and Data Science: Article Topics' -summary: This article provides a list of topic ideas for writing detailed articles - about NLP and Data Science, suitable for technical and practical discussions. +summary: This article provides a list of topic ideas for writing detailed articles about NLP and Data Science, suitable for technical and practical discussions. tags: - Nlp - Data science diff --git a/_posts/-_ideas/numerical_methods_fortran.md b/_posts/-_ideas/numerical_methods_fortran.md index 072e9fc5..14120f7e 100644 --- a/_posts/-_ideas/numerical_methods_fortran.md +++ b/_posts/-_ideas/numerical_methods_fortran.md @@ -2,6 +2,8 @@ tags: - Plaintext - Fortran +- plaintext +- fortran --- # Numerical Methods Using Fortran Repository diff --git a/_posts/2020-01-01-causality_and_correlation.md b/_posts/2020-01-01-causality_and_correlation.md index e625f77d..dbfff3dc 100644 --- a/_posts/2020-01-01-causality_and_correlation.md +++ b/_posts/2020-01-01-causality_and_correlation.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2020-01-01' -excerpt: Understand how causal reasoning helps us move beyond correlation, resolving - paradoxes and leading to more accurate insights from data analysis. +excerpt: Understand how causal reasoning helps us move beyond correlation, resolving paradoxes and leading to more accurate insights from data analysis. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_1.jpg @@ -19,14 +18,10 @@ keywords: - Berkson's paradox - Correlation - Data science -seo_description: Explore how causal reasoning, through paradoxes like Simpson's and - Berkson's, can help us avoid the common pitfalls of interpreting data solely based - on correlation. +seo_description: Explore how causal reasoning, through paradoxes like Simpson's and Berkson's, can help us avoid the common pitfalls of interpreting data solely based on correlation. seo_title: 'Causality Beyond Correlation: Understanding Paradoxes and Causal Graphs' seo_type: article -summary: An in-depth exploration of the limits of correlation in data interpretation, - highlighting Simpson's and Berkson's paradoxes and introducing causal graphs as - a tool for uncovering true causal relationships. +summary: An in-depth exploration of the limits of correlation in data interpretation, highlighting Simpson's and Berkson's paradoxes and introducing causal graphs as a tool for uncovering true causal relationships. tags: - Simpson's paradox - Berkson's paradox diff --git a/_posts/2020-01-02-maximum_likelihood_estimation_statistical_modeling.md b/_posts/2020-01-02-maximum_likelihood_estimation_statistical_modeling.md index 67c0e652..8280db8f 100644 --- a/_posts/2020-01-02-maximum_likelihood_estimation_statistical_modeling.md +++ b/_posts/2020-01-02-maximum_likelihood_estimation_statistical_modeling.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-01-02' -excerpt: Discover the fundamentals of Maximum Likelihood Estimation (MLE), its role - in data science, and how it impacts businesses through predictive analytics and - risk modeling. +excerpt: Discover the fundamentals of Maximum Likelihood Estimation (MLE), its role in data science, and how it impacts businesses through predictive analytics and risk modeling. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -22,13 +20,12 @@ keywords: - Mle - Bash - Python -seo_description: Explore Maximum Likelihood Estimation (MLE), its importance in data - science, machine learning, and real-world applications. +- python +- bash +seo_description: Explore Maximum Likelihood Estimation (MLE), its importance in data science, machine learning, and real-world applications. seo_title: 'MLE: A Key Tool in Data Science' seo_type: article -summary: This article covers the essentials of Maximum Likelihood Estimation (MLE), - breaking down its mathematical foundation, importance in data science, practical - applications, and limitations. +summary: This article covers the essentials of Maximum Likelihood Estimation (MLE), breaking down its mathematical foundation, importance in data science, practical applications, and limitations. tags: - Statistical modeling - Bash @@ -36,6 +33,8 @@ tags: - Data science - Mle - Python +- python +- bash title: 'Maximum Likelihood Estimation (MLE): Statistical Modeling in Data Science' --- diff --git a/_posts/2020-01-03-assessing_goodness-of-fit_non-parametric_data.md b/_posts/2020-01-03-assessing_goodness-of-fit_non-parametric_data.md index 582b6ebb..83e794e1 100644 --- a/_posts/2020-01-03-assessing_goodness-of-fit_non-parametric_data.md +++ b/_posts/2020-01-03-assessing_goodness-of-fit_non-parametric_data.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-01-03' -excerpt: The Kolmogorov-Smirnov test is a powerful tool for assessing goodness-of-fit - in non-parametric data. Learn how it works, how it compares to the Shapiro-Wilk - test, and explore real-world applications. +excerpt: The Kolmogorov-Smirnov test is a powerful tool for assessing goodness-of-fit in non-parametric data. Learn how it works, how it compares to the Shapiro-Wilk test, and explore real-world applications. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -20,15 +18,10 @@ keywords: - Non-parametric statistics - Distribution fitting - Shapiro-wilk test -seo_description: This article introduces the Kolmogorov-Smirnov test for assessing - goodness-of-fit in non-parametric data, comparing it with other tests like Shapiro-Wilk, - and exploring real-world use cases. +seo_description: This article introduces the Kolmogorov-Smirnov test for assessing goodness-of-fit in non-parametric data, comparing it with other tests like Shapiro-Wilk, and exploring real-world use cases. seo_title: 'Kolmogorov-Smirnov Test: A Guide to Non-Parametric Goodness-of-Fit Testing' seo_type: article -summary: This article explains the Kolmogorov-Smirnov (K-S) test for assessing the - goodness-of-fit of non-parametric data. We compare the K-S test to other goodness-of-fit - tests, such as Shapiro-Wilk, and provide real-world use cases, including testing - whether a dataset follows a specific distribution. +summary: This article explains the Kolmogorov-Smirnov (K-S) test for assessing the goodness-of-fit of non-parametric data. We compare the K-S test to other goodness-of-fit tests, such as Shapiro-Wilk, and provide real-world use cases, including testing whether a dataset follows a specific distribution. tags: - Kolmogorov-smirnov test - Goodness-of-fit tests diff --git a/_posts/2020-01-04-multiple_comparisons_problem:_bonferroni_correction_other_solutions.md b/_posts/2020-01-04-multiple_comparisons_problem:_bonferroni_correction_other_solutions.md index 3f399e67..cae8baaf 100644 --- a/_posts/2020-01-04-multiple_comparisons_problem:_bonferroni_correction_other_solutions.md +++ b/_posts/2020-01-04-multiple_comparisons_problem:_bonferroni_correction_other_solutions.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-01-04' -excerpt: The multiple comparisons problem arises in hypothesis testing when performing - multiple tests increases the likelihood of false positives. Learn about the Bonferroni - correction and other solutions to control error rates. +excerpt: The multiple comparisons problem arises in hypothesis testing when performing multiple tests increases the likelihood of false positives. Learn about the Bonferroni correction and other solutions to control error rates. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -21,15 +19,11 @@ keywords: - False discovery rate - Hypothesis testing - Python -seo_description: This article explains the multiple comparisons problem in hypothesis - testing and discusses solutions such as Bonferroni correction, Holm-Bonferroni, - and FDR, with practical applications in fields like medical studies and genetics. +- python +seo_description: This article explains the multiple comparisons problem in hypothesis testing and discusses solutions such as Bonferroni correction, Holm-Bonferroni, and FDR, with practical applications in fields like medical studies and genetics. seo_title: 'Understanding the Multiple Comparisons Problem: Bonferroni and Other Solutions' seo_type: article -summary: This article explores the multiple comparisons problem in hypothesis testing, - discussing solutions like the Bonferroni correction, Holm-Bonferroni method, and - False Discovery Rate (FDR). It includes practical examples from experiments involving - multiple testing, such as medical studies and genetics. +summary: This article explores the multiple comparisons problem in hypothesis testing, discussing solutions like the Bonferroni correction, Holm-Bonferroni method, and False Discovery Rate (FDR). It includes practical examples from experiments involving multiple testing, such as medical studies and genetics. tags: - Multiple comparisons problem - Bonferroni correction @@ -37,6 +31,7 @@ tags: - False discovery rate (fdr) - Multiple testing - Python +- python title: 'Multiple Comparisons Problem: Bonferroni Correction and Other Solutions' --- diff --git a/_posts/2020-01-05-one-way_anova_vs._two-way_anova_when_use_which.md b/_posts/2020-01-05-one-way_anova_vs._two-way_anova_when_use_which.md index bdf93477..d3192691 100644 --- a/_posts/2020-01-05-one-way_anova_vs._two-way_anova_when_use_which.md +++ b/_posts/2020-01-05-one-way_anova_vs._two-way_anova_when_use_which.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-01-05' -excerpt: One-way and two-way ANOVA are essential tools for comparing means across - groups, but each test serves different purposes. Learn when to use one-way versus - two-way ANOVA and how to interpret their results. +excerpt: One-way and two-way ANOVA are essential tools for comparing means across groups, but each test serves different purposes. Learn when to use one-way versus two-way ANOVA and how to interpret their results. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_1.jpg @@ -20,14 +18,10 @@ keywords: - Interaction effects - Main effects - Hypothesis testing -seo_description: This article explores the differences between one-way and two-way - ANOVA, when to use each test, and how to interpret main effects and interaction - effects in two-way ANOVA. +seo_description: This article explores the differences between one-way and two-way ANOVA, when to use each test, and how to interpret main effects and interaction effects in two-way ANOVA. seo_title: 'One-Way ANOVA vs. Two-Way ANOVA: When to Use Which' seo_type: article -summary: This article discusses one-way and two-way ANOVA, focusing on when to use - each method. It explains how two-way ANOVA is useful for analyzing interactions - between factors and details the interpretation of main effects and interactions. +summary: This article discusses one-way and two-way ANOVA, focusing on when to use each method. It explains how two-way ANOVA is useful for analyzing interactions between factors and details the interpretation of main effects and interactions. tags: - One-way anova - Two-way anova diff --git a/_posts/2020-01-08-heteroscedascity_statistical_tests.md b/_posts/2020-01-08-heteroscedascity_statistical_tests.md index 33e4f265..f5779002 100644 --- a/_posts/2020-01-08-heteroscedascity_statistical_tests.md +++ b/_posts/2020-01-08-heteroscedascity_statistical_tests.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2020-01-08' -excerpt: Heteroscedasticity can affect regression models, leading to biased or inefficient - estimates. Here's how to detect it and what to do when it's present. +excerpt: Heteroscedasticity can affect regression models, leading to biased or inefficient estimates. Here's how to detect it and what to do when it's present. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -19,12 +18,10 @@ keywords: - White test - Heteroscedasticity - Breusch-pagan test -seo_description: Learn about heteroscedasticity, the statistical tests to detect it, - and steps to take when it is present in regression analysis. +seo_description: Learn about heteroscedasticity, the statistical tests to detect it, and steps to take when it is present in regression analysis. seo_title: 'Heteroscedasticity: Statistical Tests and What to Do When Detected' seo_type: article -summary: Explore heteroscedasticity in regression analysis, its consequences, how - to test for it, and practical solutions for correcting it when detected. +summary: Explore heteroscedasticity in regression analysis, its consequences, how to test for it, and practical solutions for correcting it when detected. tags: - Regression analysis - Econometrics diff --git a/_posts/2020-01-30-cox_proportional_hazards_model.md b/_posts/2020-01-30-cox_proportional_hazards_model.md index 3c120af3..10ce40f4 100644 --- a/_posts/2020-01-30-cox_proportional_hazards_model.md +++ b/_posts/2020-01-30-cox_proportional_hazards_model.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2020-01-30' -excerpt: The Cox Proportional Hazards Model is a vital tool for analyzing time-to-event - data in medical studies. Learn how it works and its applications in survival analysis. +excerpt: The Cox Proportional Hazards Model is a vital tool for analyzing time-to-event data in medical studies. Learn how it works and its applications in survival analysis. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -24,12 +23,12 @@ keywords: - Proportional hazards assumption - R - Python -seo_description: Explore the Cox Proportional Hazards Model and its application in - survival analysis, with examples from clinical trials and medical research. +- r +- python +seo_description: Explore the Cox Proportional Hazards Model and its application in survival analysis, with examples from clinical trials and medical research. seo_title: Understanding Cox Proportional Hazards Model for Medical Survival Analysis seo_type: article -summary: A comprehensive guide to the Cox Proportional Hazards Model, its assumptions, - and applications in survival analysis and clinical trials. +summary: A comprehensive guide to the Cox Proportional Hazards Model, its assumptions, and applications in survival analysis and clinical trials. tags: - Cox proportional hazards model - Survival analysis @@ -39,6 +38,8 @@ tags: - Censored data - R - Python +- r +- python title: 'Cox Proportional Hazards Model: A Guide to Survival Analysis in Medical Studies' --- diff --git a/_posts/2020-02-01-anova_kruskal_walis.md b/_posts/2020-02-01-anova_kruskal_walis.md index a81e4dd4..b4182d4f 100644 --- a/_posts/2020-02-01-anova_kruskal_walis.md +++ b/_posts/2020-02-01-anova_kruskal_walis.md @@ -6,8 +6,7 @@ categories: - Hypothesis Testing classes: wide date: '2020-02-01' -excerpt: Learn the key differences between ANOVA and Kruskal-Wallis tests, and understand - when to use each method based on your data's assumptions and characteristics. +excerpt: Learn the key differences between ANOVA and Kruskal-Wallis tests, and understand when to use each method based on your data's assumptions and characteristics. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_5.jpg @@ -21,14 +20,10 @@ keywords: - Anova - Non-parametric test - Hypothesis testing -seo_description: Explore the differences between ANOVA and Kruskal-Wallis tests. Learn - when to use parametric (ANOVA) and non-parametric (Kruskal-Wallis) methods for comparing - multiple groups. +seo_description: Explore the differences between ANOVA and Kruskal-Wallis tests. Learn when to use parametric (ANOVA) and non-parametric (Kruskal-Wallis) methods for comparing multiple groups. seo_title: 'ANOVA vs Kruskal-Wallis: Key Differences and When to Use Them' seo_type: article -summary: This article explores the fundamental differences between ANOVA and Kruskal-Wallis - tests, with a focus on their assumptions, applications, and when to use each method - in data analysis. +summary: This article explores the fundamental differences between ANOVA and Kruskal-Wallis tests, with a focus on their assumptions, applications, and when to use each method in data analysis. tags: - Kruskal-wallis - Non-parametric methods diff --git a/_posts/2020-02-02-understanding_statistical_testing:_null_hypothesis_beyond.md b/_posts/2020-02-02-understanding_statistical_testing:_null_hypothesis_beyond.md index b99b7f45..ea3ef896 100644 --- a/_posts/2020-02-02-understanding_statistical_testing:_null_hypothesis_beyond.md +++ b/_posts/2020-02-02-understanding_statistical_testing:_null_hypothesis_beyond.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2020-02-02' -excerpt: A detailed look at hypothesis testing, the misconceptions around the null - hypothesis, and the diverse methods for detecting data deviations. +excerpt: A detailed look at hypothesis testing, the misconceptions around the null hypothesis, and the diverse methods for detecting data deviations. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -19,14 +18,10 @@ keywords: - Data non-normality - Statistical methods - Hypothesis rejection -seo_description: An in-depth exploration of the complexities behind hypothesis testing, - the null hypothesis, and multiple testing methods that detect data deviations from - theoretical patterns. +seo_description: An in-depth exploration of the complexities behind hypothesis testing, the null hypothesis, and multiple testing methods that detect data deviations from theoretical patterns. seo_title: 'Statistical Testing: Exploring the Complexities of the Null Hypothesis' seo_type: article -summary: This article delves into the core principles of hypothesis testing, the nuances - of the null hypothesis, and the various statistical tools used to test data compatibility - with theoretical distributions. +summary: This article delves into the core principles of hypothesis testing, the nuances of the null hypothesis, and the various statistical tools used to test data compatibility with theoretical distributions. tags: - Hypothesis testing - Null hypothesis diff --git a/_posts/2020-02-17-arimax_time_series.md b/_posts/2020-02-17-arimax_time_series.md index e55ed6c3..d160caa8 100644 --- a/_posts/2020-02-17-arimax_time_series.md +++ b/_posts/2020-02-17-arimax_time_series.md @@ -4,8 +4,7 @@ categories: - Time Series Analysis classes: wide date: '2020-02-17' -excerpt: The ARIMAX model extends ARIMA by integrating exogenous variables into time - series forecasting, offering more accurate predictions for complex systems. +excerpt: The ARIMAX model extends ARIMA by integrating exogenous variables into time series forecasting, offering more accurate predictions for complex systems. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -20,14 +19,11 @@ keywords: - Forecasting - Time series - Arimax -seo_description: Explore the ARIMAX model, a powerful statistical tool for time series - forecasting that incorporates exogenous variables. Learn how ARIMAX builds on ARIMA - to improve predictive performance. +- r +seo_description: Explore the ARIMAX model, a powerful statistical tool for time series forecasting that incorporates exogenous variables. Learn how ARIMAX builds on ARIMA to improve predictive performance. seo_title: 'ARIMAX Time Series Model: An In-Depth Guide' seo_type: article -summary: This article explores the ARIMAX time series model, which enhances ARIMA - by incorporating external variables. We'll dive into the model's structure, assumptions, - applications, and how it compares to ARIMA. +summary: This article explores the ARIMAX time series model, which enhances ARIMA by incorporating external variables. We'll dive into the model's structure, assumptions, applications, and how it compares to ARIMA. tags: - R - Statistical modeling @@ -35,6 +31,7 @@ tags: - Arima - Time series forecasting - Arimax +- r title: 'ARIMAX Time Series: Comprehensive Guide' --- diff --git a/_posts/2020-03-01-type_one_type_two_erros.md b/_posts/2020-03-01-type_one_type_two_erros.md index b3aa021e..fe2b097e 100644 --- a/_posts/2020-03-01-type_one_type_two_erros.md +++ b/_posts/2020-03-01-type_one_type_two_erros.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-03-01' -excerpt: Explore Type I and Type II errors in hypothesis testing. Learn how to balance - error rates, interpret significance levels, and understand the implications of statistical - errors in real-world scenarios. +excerpt: Explore Type I and Type II errors in hypothesis testing. Learn how to balance error rates, interpret significance levels, and understand the implications of statistical errors in real-world scenarios. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_4.jpg @@ -20,14 +18,10 @@ keywords: - False negative - Hypothesis testing - Type i error -seo_description: A comprehensive guide to understanding Type I (false positive) and - Type II (false negative) errors in hypothesis testing, including balancing error - rates, significance levels, and power. +seo_description: A comprehensive guide to understanding Type I (false positive) and Type II (false negative) errors in hypothesis testing, including balancing error rates, significance levels, and power. seo_title: 'Understanding Type I and Type II Errors: Hypothesis Testing Explained' seo_type: article -summary: This article provides an in-depth exploration of Type I and Type II errors - in hypothesis testing, explaining their importance, the trade-offs between them, - and how they impact decisions in various domains, from clinical trials to business. +summary: This article provides an in-depth exploration of Type I and Type II errors in hypothesis testing, explaining their importance, the trade-offs between them, and how they impact decisions in various domains, from clinical trials to business. tags: - Type ii error - False positive diff --git a/_posts/2020-03-30-sustainability_analytics_how_data_science_drives_green_innovation.md b/_posts/2020-03-30-sustainability_analytics_how_data_science_drives_green_innovation.md index 2835da49..dad2baff 100644 --- a/_posts/2020-03-30-sustainability_analytics_how_data_science_drives_green_innovation.md +++ b/_posts/2020-03-30-sustainability_analytics_how_data_science_drives_green_innovation.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2020-03-30' -excerpt: Data science is a key driver of sustainability, offering insights that help - optimize resources, reduce waste, and improve the energy efficiency of supply chains. +excerpt: Data science is a key driver of sustainability, offering insights that help optimize resources, reduce waste, and improve the energy efficiency of supply chains. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -19,14 +18,10 @@ keywords: - Green innovation - Resource optimization - Supply chain efficiency -seo_description: This article explores how companies and organizations are using data - science to enhance sustainability practices in areas like resource optimization, - waste reduction, and energy efficiency. +seo_description: This article explores how companies and organizations are using data science to enhance sustainability practices in areas like resource optimization, waste reduction, and energy efficiency. seo_title: How Data Science is Driving Green Innovation through Sustainability Analytics seo_type: article -summary: In this article, we explore the role of data science in driving green innovation - through sustainability analytics, examining how companies use data to optimize resources, - cut waste, and enhance supply chain efficiency. +summary: In this article, we explore the role of data science in driving green innovation through sustainability analytics, examining how companies use data to optimize resources, cut waste, and enhance supply chain efficiency. tags: - Sustainability analytics - Data science diff --git a/_posts/2020-04-01-the_friedman_test.md b/_posts/2020-04-01-the_friedman_test.md index d86f2739..4b307cb2 100644 --- a/_posts/2020-04-01-the_friedman_test.md +++ b/_posts/2020-04-01-the_friedman_test.md @@ -4,9 +4,7 @@ categories: - Data Analysis classes: wide date: '2020-04-01' -excerpt: The Friedman test is a non-parametric alternative to repeated measures ANOVA, - designed for use with ordinal data or non-normal distributions. Learn how and when - to use it in your analyses. +excerpt: The Friedman test is a non-parametric alternative to repeated measures ANOVA, designed for use with ordinal data or non-normal distributions. Learn how and when to use it in your analyses. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_8.jpg @@ -19,14 +17,10 @@ keywords: - Non-parametric test - Friedman test - Ordinal data -seo_description: Learn about the Friedman test, its application as a non-parametric - alternative to repeated measures ANOVA, and its use with ordinal data or non-normal - distributions. +seo_description: Learn about the Friedman test, its application as a non-parametric alternative to repeated measures ANOVA, and its use with ordinal data or non-normal distributions. seo_title: 'The Friedman Test: A Non-Parametric Alternative to Repeated Measures ANOVA' seo_type: article -summary: This article provides an in-depth explanation of the Friedman test, including - its use as a non-parametric alternative to repeated measures ANOVA, when to use - it, and practical examples in ranking data and repeated measurements. +summary: This article provides an in-depth explanation of the Friedman test, including its use as a non-parametric alternative to repeated measures ANOVA, when to use it, and practical examples in ranking data and repeated measurements. tags: - Non-parametric tests - Repeated measures anova diff --git a/_posts/2020-04-27-prediction_errors_bias_variance_model.md b/_posts/2020-04-27-prediction_errors_bias_variance_model.md index 230324cb..f7de7dd8 100644 --- a/_posts/2020-04-27-prediction_errors_bias_variance_model.md +++ b/_posts/2020-04-27-prediction_errors_bias_variance_model.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2020-04-27' -excerpt: Learn about different methods for estimating prediction error, addressing - the bias-variance tradeoff, and how cross-validation, bootstrap methods, and Efron - & Tibshirani's .632 estimator help improve model evaluation. +excerpt: Learn about different methods for estimating prediction error, addressing the bias-variance tradeoff, and how cross-validation, bootstrap methods, and Efron & Tibshirani's .632 estimator help improve model evaluation. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -16,14 +14,11 @@ header: twitter_image: /assets/images/data_science_6.jpg keywords: - Python -seo_description: An in-depth look at prediction error, bias-variance tradeoff, and - model evaluation techniques like cross-validation and bootstrap methods, with insights - into the .632 estimator. +- python +seo_description: An in-depth look at prediction error, bias-variance tradeoff, and model evaluation techniques like cross-validation and bootstrap methods, with insights into the .632 estimator. seo_title: 'Understanding Prediction Error: Bias, Variance, and Evaluation Techniques' seo_type: article -summary: This article explores methods for estimating prediction error, including - cross-validation, bootstrap techniques, and their variations like the .632 estimator, - focusing on balancing bias, variance, and model evaluation accuracy. +summary: This article explores methods for estimating prediction error, including cross-validation, bootstrap techniques, and their variations like the .632 estimator, focusing on balancing bias, variance, and model evaluation accuracy. tags: - Bias-variance tradeoff - Model evaluation @@ -32,6 +27,7 @@ tags: - Bootstrap methods - Prediction error - Python +- python title: 'Understanding Prediction Error: Bias, Variance, and Model Evaluation Techniques' --- diff --git a/_posts/2020-05-01-shapiro_wilk_test.md b/_posts/2020-05-01-shapiro_wilk_test.md index 17095222..bc5f90a5 100644 --- a/_posts/2020-05-01-shapiro_wilk_test.md +++ b/_posts/2020-05-01-shapiro_wilk_test.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-05-01' -excerpt: Learn about the Shapiro-Wilk and Anderson-Darling tests for normality, their - differences, and how they guide decisions between parametric and non-parametric - statistical methods. +excerpt: Learn about the Shapiro-Wilk and Anderson-Darling tests for normality, their differences, and how they guide decisions between parametric and non-parametric statistical methods. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_8.jpg @@ -20,14 +18,10 @@ keywords: - Shapiro-wilk test - Normality test - Parametric methods -seo_description: Explore the differences between the Shapiro-Wilk and Anderson-Darling - tests for checking normality in data. Learn when to use each test and how to interpret - the results. +seo_description: Explore the differences between the Shapiro-Wilk and Anderson-Darling tests for checking normality in data. Learn when to use each test and how to interpret the results. seo_title: 'Shapiro-Wilk Test vs. Anderson-Darling Test: Normality Tests Explained' seo_type: article -summary: "This article explores two common normality tests\u2014the Shapiro-Wilk test\ - \ and the Anderson-Darling test\u2014discussing their differences, when to use each,\ - \ and how to interpret their results to determine the appropriate statistical method." +summary: This article explores two common normality tests—the Shapiro-Wilk test and the Anderson-Darling test—discussing their differences, when to use each, and how to interpret their results to determine the appropriate statistical method. tags: - Anderson-darling test - Normality tests diff --git a/_posts/2020-05-26-false_positive_rate.md b/_posts/2020-05-26-false_positive_rate.md index c1dc2694..5793a1c3 100644 --- a/_posts/2020-05-26-false_positive_rate.md +++ b/_posts/2020-05-26-false_positive_rate.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2020-05-26' -excerpt: Learn what the False Positive Rate (FPR) is, how it impacts machine learning - models, and when to use it for better evaluation. +excerpt: Learn what the False Positive Rate (FPR) is, how it impacts machine learning models, and when to use it for better evaluation. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_7.jpg @@ -20,20 +19,18 @@ keywords: - Machine learning - Fpr - Binary classification metrics -seo_description: A comprehensive analysis of the False Positive Rate (FPR), including - its role in machine learning, strengths, weaknesses, use cases, and alternative - metrics. +- r +seo_description: A comprehensive analysis of the False Positive Rate (FPR), including its role in machine learning, strengths, weaknesses, use cases, and alternative metrics. seo_title: Understanding the False Positive Rate in Machine Learning seo_type: article -summary: This article provides a detailed examination of the False Positive Rate (FPR) - in binary classification, its calculation, interpretation, and the contexts in which - it plays a crucial role. +summary: This article provides a detailed examination of the False Positive Rate (FPR) in binary classification, its calculation, interpretation, and the contexts in which it plays a crucial role. tags: - R - False positive rate - Model evaluation - Machine learning metrics - Binary classification +- r title: Analysis of the False Positive Rate (FPR) in Machine Learning --- diff --git a/_posts/2020-06-01-ordinary_least_square_regression.md b/_posts/2020-06-01-ordinary_least_square_regression.md index e8d63e36..7e616ad7 100644 --- a/_posts/2020-06-01-ordinary_least_square_regression.md +++ b/_posts/2020-06-01-ordinary_least_square_regression.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-06-01' -excerpt: Discover the foundations of Ordinary Least Squares (OLS) regression, its - key properties such as consistency, efficiency, and maximum likelihood estimation, - and its applications in linear modeling. +excerpt: Discover the foundations of Ordinary Least Squares (OLS) regression, its key properties such as consistency, efficiency, and maximum likelihood estimation, and its applications in linear modeling. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_3.jpg @@ -21,16 +19,10 @@ keywords: - Gauss-markov - Ols - Maximum likelihood estimator -seo_description: A detailed exploration of Ordinary Least Squares (OLS) regression, - its properties like consistency, efficiency, and minimum variance, and its applications - in statistics, machine learning, and data science. -seo_title: 'Ordinary Least Squares (OLS) Regression: Understanding Properties and - Applications' +seo_description: A detailed exploration of Ordinary Least Squares (OLS) regression, its properties like consistency, efficiency, and minimum variance, and its applications in statistics, machine learning, and data science. +seo_title: 'Ordinary Least Squares (OLS) Regression: Understanding Properties and Applications' seo_type: article -summary: This article covers Ordinary Least Squares (OLS) regression, one of the most - commonly used techniques in statistics, data science, and machine learning. Learn - about its key properties, how it works, and its wide range of applications in modeling - linear relationships between variables. +summary: This article covers Ordinary Least Squares (OLS) regression, one of the most commonly used techniques in statistics, data science, and machine learning. Learn about its key properties, how it works, and its wide range of applications in modeling linear relationships between variables. tags: - Homoscedasticity - Ols regression diff --git a/_posts/2020-06-10-arima_time_series.md b/_posts/2020-06-10-arima_time_series.md index 7ede6979..4e93ff7a 100644 --- a/_posts/2020-06-10-arima_time_series.md +++ b/_posts/2020-06-10-arima_time_series.md @@ -4,9 +4,7 @@ categories: - Time Series Analysis classes: wide date: '2020-06-10' -excerpt: Learn the fundamentals of ARIMA modeling for time series analysis. This guide - covers the AR, I, and MA components, model identification, validation, and its comparison - with other models. +excerpt: Learn the fundamentals of ARIMA modeling for time series analysis. This guide covers the AR, I, and MA components, model identification, validation, and its comparison with other models. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_1.jpg @@ -22,20 +20,20 @@ keywords: - Sarima - R - Arma -seo_description: Explore the fundamentals of ARIMA (AutoRegressive Integrated Moving - Average) model, its components, parameter identification, validation, and applications. - Comparison with ARIMAX, SARIMA, and ARMA. +- r +- python +seo_description: Explore the fundamentals of ARIMA (AutoRegressive Integrated Moving Average) model, its components, parameter identification, validation, and applications. Comparison with ARIMAX, SARIMA, and ARMA. seo_title: 'Comprehensive ARIMA Model Guide: Time Series Analysis' seo_type: article -summary: This guide provides an in-depth exploration of ARIMA modeling for time series - data, discussing its core components, parameter estimation, validation, and comparison - with models like ARIMAX, SARIMA, and ARMA. +summary: This guide provides an in-depth exploration of ARIMA modeling for time series data, discussing its core components, parameter estimation, validation, and comparison with models like ARIMAX, SARIMA, and ARMA. tags: - Arima - Time series - Forecasting - R - Python +- r +- python title: A Comprehensive Guide to ARIMA Time Series Modeling --- diff --git a/_posts/2020-07-01-cocharan_q_test.md b/_posts/2020-07-01-cocharan_q_test.md index 0599b109..4c90699d 100644 --- a/_posts/2020-07-01-cocharan_q_test.md +++ b/_posts/2020-07-01-cocharan_q_test.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-07-01' -excerpt: "Understand Cochran\u2019s Q test, a non-parametric test for comparing proportions\ - \ across related groups, and its applications in binary data and its connection\ - \ to McNemar's test." +excerpt: Understand Cochran’s Q test, a non-parametric test for comparing proportions across related groups, and its applications in binary data and its connection to McNemar's test. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_8.jpg @@ -21,20 +19,16 @@ keywords: - Machine learning - Logistic regression - Data science -seo_description: "Learn about Cochran\u2019s Q test, its use for comparing proportions\ - \ across related groups, and its connection with McNemar\u2019s test and logistic\ - \ regression." -seo_title: "Cochran\u2019s Q Test: Comparing Proportions in Related Groups" +seo_description: Learn about Cochran’s Q test, its use for comparing proportions across related groups, and its connection with McNemar’s test and logistic regression. +seo_title: 'Cochran’s Q Test: Comparing Proportions in Related Groups' seo_type: article -summary: "This article explores Cochran\u2019s Q test, a non-parametric method for\ - \ comparing proportions in related groups, particularly in binary data. It also\ - \ covers the relationship between Cochran's Q, McNemar's test, and logistic regression." +summary: This article explores Cochran’s Q test, a non-parametric method for comparing proportions in related groups, particularly in binary data. It also covers the relationship between Cochran's Q, McNemar's test, and logistic regression. tags: - Logistic regression - Mcnemar's test - Non-parametric tests - Cochran's q test -title: "Cochran\u2019s Q Test: Comparing Three or More Related Proportions" +title: 'Cochran’s Q Test: Comparing Three or More Related Proportions' --- In the realm of statistical analysis, there are many situations where we need to compare proportions across **related groups**, particularly when the data is **binary** (e.g., success/failure, yes/no). For such cases, **Cochran’s Q test** provides an effective way to determine whether there are significant differences in proportions across three or more related samples. diff --git a/_posts/2020-07-02-mann-whitney_u_test_vs._independent_t_test_non_parametric_alternatives.md b/_posts/2020-07-02-mann-whitney_u_test_vs._independent_t_test_non_parametric_alternatives.md index c071aae5..e3d48b77 100644 --- a/_posts/2020-07-02-mann-whitney_u_test_vs._independent_t_test_non_parametric_alternatives.md +++ b/_posts/2020-07-02-mann-whitney_u_test_vs._independent_t_test_non_parametric_alternatives.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2020-07-02' -excerpt: The Mann-Whitney U test and independent t-test are used for comparing two - independent groups, but the choice between them depends on data distribution. Learn - when to use each and explore real-world applications. +excerpt: The Mann-Whitney U test and independent t-test are used for comparing two independent groups, but the choice between them depends on data distribution. Learn when to use each and explore real-world applications. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -20,16 +18,10 @@ keywords: - Non-parametric tests - Parametric tests - Hypothesis testing -seo_description: This article compares the parametric independent t-test and the non-parametric - Mann-Whitney U test, explaining when to use each based on data distribution, with - practical examples. -seo_title: 'Mann-Whitney U Test vs. Independent T-Test: When to Use Non-Parametric - Tests' +seo_description: This article compares the parametric independent t-test and the non-parametric Mann-Whitney U test, explaining when to use each based on data distribution, with practical examples. +seo_title: 'Mann-Whitney U Test vs. Independent T-Test: When to Use Non-Parametric Tests' seo_type: article -summary: This article provides a comprehensive comparison between the Mann-Whitney - U test and the independent t-test. It explains when and why the non-parametric Mann-Whitney - U test is preferred over the parametric t-test, especially in the case of non-normal - distributions, and provides practical examples of both tests. +summary: This article provides a comprehensive comparison between the Mann-Whitney U test and the independent t-test. It explains when and why the non-parametric Mann-Whitney U test is preferred over the parametric t-test, especially in the case of non-normal distributions, and provides practical examples of both tests. tags: - Mann-whitney u test - Independent t-test diff --git a/_posts/2020-07-26-measurement_errors.md b/_posts/2020-07-26-measurement_errors.md index 4aac20d0..817a8270 100644 --- a/_posts/2020-07-26-measurement_errors.md +++ b/_posts/2020-07-26-measurement_errors.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2020-07-26' -excerpt: Explore the different types of observational errors, their causes, and their - impact on accuracy and precision in various fields, such as data science and engineering. +excerpt: Explore the different types of observational errors, their causes, and their impact on accuracy and precision in various fields, such as data science and engineering. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_8.jpg @@ -13,14 +12,10 @@ header: show_overlay_excerpt: false teaser: /assets/images/data_science_3.jpg twitter_image: /assets/images/data_science_8.jpg -seo_description: Understand the types of observational errors, their causes, and how - to estimate and reduce their effects for better accuracy and precision in scientific - and data-driven fields. +seo_description: Understand the types of observational errors, their causes, and how to estimate and reduce their effects for better accuracy and precision in scientific and data-driven fields. seo_title: 'Observational Error: A Deep Dive into Measurement Accuracy and Precision' seo_type: article -summary: A comprehensive guide to understanding observational and measurement errors, - covering random and systematic errors, their statistical models, and methods to - estimate and mitigate their effects. +summary: A comprehensive guide to understanding observational and measurement errors, covering random and systematic errors, their statistical models, and methods to estimate and mitigate their effects. tags: - Statistical bias - Statistical methods diff --git a/_posts/2020-08-01-understanding_markov_chain_monte_carlo.md b/_posts/2020-08-01-understanding_markov_chain_monte_carlo.md index 8b54ebcd..012dea18 100644 --- a/_posts/2020-08-01-understanding_markov_chain_monte_carlo.md +++ b/_posts/2020-08-01-understanding_markov_chain_monte_carlo.md @@ -4,9 +4,7 @@ categories: - Algorithms classes: wide date: '2020-08-01' -excerpt: This article delves into the fundamentals of Markov Chain Monte Carlo (MCMC), - its applications, and its significance in solving complex, high-dimensional probability - distributions. +excerpt: This article delves into the fundamentals of Markov Chain Monte Carlo (MCMC), its applications, and its significance in solving complex, high-dimensional probability distributions. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_5.jpg @@ -22,14 +20,12 @@ keywords: - Python - Bayesian inference - Bash -seo_description: An in-depth exploration of Markov Chain Monte Carlo (MCMC), its algorithms, - and its applications in statistics, probability theory, and numerical approximations. +- bash +- python +seo_description: An in-depth exploration of Markov Chain Monte Carlo (MCMC), its algorithms, and its applications in statistics, probability theory, and numerical approximations. seo_title: Comprehensive Guide to Markov Chain Monte Carlo (MCMC) seo_type: article -summary: Markov Chain Monte Carlo (MCMC) is an essential tool in probabilistic computation, - used for sampling from complex distributions. This article explores its foundations, - algorithms like Metropolis-Hastings, and various applications in statistics and - numerical integration. +summary: Markov Chain Monte Carlo (MCMC) is an essential tool in probabilistic computation, used for sampling from complex distributions. This article explores its foundations, algorithms like Metropolis-Hastings, and various applications in statistics and numerical integration. tags: - Markov chain monte carlo - Probability distributions @@ -37,6 +33,8 @@ tags: - Bash - Bayesian statistics - Numerical methods +- bash +- python title: Understanding Markov Chain Monte Carlo (MCMC) --- diff --git a/_posts/2020-09-01-threshold_classification_zero_inflated_time_series.md b/_posts/2020-09-01-threshold_classification_zero_inflated_time_series.md index 89af4d38..9f5f019d 100644 --- a/_posts/2020-09-01-threshold_classification_zero_inflated_time_series.md +++ b/_posts/2020-09-01-threshold_classification_zero_inflated_time_series.md @@ -4,8 +4,7 @@ categories: - Time Series Analysis classes: wide date: '2020-09-01' -excerpt: This article explores the use of stationary distributions in time series - models to define thresholds in zero-inflated data, improving classification accuracy. +excerpt: This article explores the use of stationary distributions in time series models to define thresholds in zero-inflated data, improving classification accuracy. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_7.jpg @@ -18,23 +17,16 @@ keywords: - Zero-inflated data - Threshold classification - Statistical modeling -seo_description: A methodology for threshold classification in zero-inflated time - series data using stationary distributions and parametric modeling to enhance classification - accuracy. -seo_title: Threshold Classification for Zero-Inflated Time Series Using Stationary - Distributions +seo_description: A methodology for threshold classification in zero-inflated time series data using stationary distributions and parametric modeling to enhance classification accuracy. +seo_title: Threshold Classification for Zero-Inflated Time Series Using Stationary Distributions seo_type: article -summary: A novel approach for threshold classification in zero-inflated time series - data using stationary distributions derived from time series models. This method - addresses the limitations of traditional techniques by leveraging parametric distribution - quantiles for better accuracy and generalization. +summary: A novel approach for threshold classification in zero-inflated time series data using stationary distributions derived from time series models. This method addresses the limitations of traditional techniques by leveraging parametric distribution quantiles for better accuracy and generalization. tags: - Statistical modeling - Zero-inflated data - Stationary distribution - Time series -title: A Generalized Approach to Threshold Classification for Zero-Inflated Time Series - Data Using Stationary Distributions +title: A Generalized Approach to Threshold Classification for Zero-Inflated Time Series Data Using Stationary Distributions --- ## Abstract diff --git a/_posts/2020-09-02-log_rank_test_survival_analysis_comparing_survival_curves.md b/_posts/2020-09-02-log_rank_test_survival_analysis_comparing_survival_curves.md index 539f587e..771d36ba 100644 --- a/_posts/2020-09-02-log_rank_test_survival_analysis_comparing_survival_curves.md +++ b/_posts/2020-09-02-log_rank_test_survival_analysis_comparing_survival_curves.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2020-09-02' -excerpt: The log-rank test is a key tool in survival analysis, commonly used to compare - survival curves between groups in medical research. Learn how it works and how to - interpret its results. +excerpt: The log-rank test is a key tool in survival analysis, commonly used to compare survival curves between groups in medical research. Learn how it works and how to interpret its results. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_7.jpg @@ -20,16 +18,10 @@ keywords: - Survival curves - Kaplan-meier curves - P-values -seo_description: This article explores the log-rank test used in survival analysis, - its applications in medical studies to compare survival times, and how to interpret - survival curves and p-values. -seo_title: 'Understanding the Log-Rank Test in Survival Analysis: Comparing Survival - Curves' +seo_description: This article explores the log-rank test used in survival analysis, its applications in medical studies to compare survival times, and how to interpret survival curves and p-values. +seo_title: 'Understanding the Log-Rank Test in Survival Analysis: Comparing Survival Curves' seo_type: article -summary: This article provides a comprehensive guide to the log-rank test in survival - analysis, focusing on its use in medical studies to compare survival curves between - two or more groups. We explain how to interpret Kaplan-Meier curves, p-values from - the log-rank test, and real-world applications in clinical trials. +summary: This article provides a comprehensive guide to the log-rank test in survival analysis, focusing on its use in medical studies to compare survival curves between two or more groups. We explain how to interpret Kaplan-Meier curves, p-values from the log-rank test, and real-world applications in clinical trials. tags: - Log-rank test - Survival analysis diff --git a/_posts/2020-09-24-demand_forecast_supply_chain.md b/_posts/2020-09-24-demand_forecast_supply_chain.md index 4b7f03d9..6c787afb 100644 --- a/_posts/2020-09-24-demand_forecast_supply_chain.md +++ b/_posts/2020-09-24-demand_forecast_supply_chain.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2020-09-24' -excerpt: Leveraging customer behavior through predictive modeling, the BG/NBD model - offers a more accurate approach to demand forecasting in the supply chain compared - to traditional time-series models. +excerpt: Leveraging customer behavior through predictive modeling, the BG/NBD model offers a more accurate approach to demand forecasting in the supply chain compared to traditional time-series models. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_7.jpg @@ -21,23 +19,19 @@ keywords: - Demand forecasting - Python - Python -seo_description: Explore how using customer behavior and predictive models can improve - demand forecasting in the supply chain industry, leveraging the BG/NBD model for - better accuracy. +- python +seo_description: Explore how using customer behavior and predictive models can improve demand forecasting in the supply chain industry, leveraging the BG/NBD model for better accuracy. seo_title: Demand Forecasting in Supply Chain Using Customer Behavior seo_type: article -summary: This article explores the use of customer behavior modeling to improve demand - forecasting in the supply chain industry. We demonstrate how the BG/NBD model and - the Lifetimes Python library are used to predict repurchases and optimize sales - predictions over a future period. +summary: This article explores the use of customer behavior modeling to improve demand forecasting in the supply chain industry. We demonstrate how the BG/NBD model and the Lifetimes Python library are used to predict repurchases and optimize sales predictions over a future period. tags: - Customer behavior - Python - Demand forecasting - Repurchase models - Python -title: A Predictive Approach for Demand Forecasting in the Supply Chain Using Customer - Behavior Modeling +- python +title: A Predictive Approach for Demand Forecasting in the Supply Chain Using Customer Behavior Modeling --- ## Introduction diff --git a/_posts/2020-10-01-time_series_models_predicting_emergency.md b/_posts/2020-10-01-time_series_models_predicting_emergency.md index fae07a54..cd2cddee 100644 --- a/_posts/2020-10-01-time_series_models_predicting_emergency.md +++ b/_posts/2020-10-01-time_series_models_predicting_emergency.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2020-10-01' -excerpt: A comparison between machine learning models and univariate time series models - for predicting emergency department visit volumes, focusing on predictive accuracy. +excerpt: A comparison between machine learning models and univariate time series models for predicting emergency department visit volumes, focusing on predictive accuracy. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_8.jpg @@ -19,23 +18,17 @@ keywords: - Gradient boosted machines - Resource allocation - Random forest -seo_description: This study examines machine learning and univariate time series models - for predicting emergency department visit volumes, highlighting the superior predictive - accuracy of random forest models. -seo_title: Comparing Machine Learning and Time Series Models for Predicting ED Visit - Volumes +seo_description: This study examines machine learning and univariate time series models for predicting emergency department visit volumes, highlighting the superior predictive accuracy of random forest models. +seo_title: Comparing Machine Learning and Time Series Models for Predicting ED Visit Volumes seo_type: article -summary: A study comparing machine learning models (random forest, GBM) with univariate - time series models (ARIMA, ETS, Prophet) for predicting emergency department visits. - Results show machine learning models perform better, though not substantially so. +summary: A study comparing machine learning models (random forest, GBM) with univariate time series models (ARIMA, ETS, Prophet) for predicting emergency department visits. Results show machine learning models perform better, though not substantially so. tags: - Emergency department - Time series forecasting - Machine learning - Gradient boosted machines - Random forest -title: Machine Learning vs. Univariate Time Series Models in Predicting Emergency - Department Visit Volumes +title: Machine Learning vs. Univariate Time Series Models in Predicting Emergency Department Visit Volumes --- ## 1. Introduction diff --git a/_posts/2020-12-01-predictive_maintenance_data_science.md b/_posts/2020-12-01-predictive_maintenance_data_science.md index 4f003e85..1ce6d0c7 100644 --- a/_posts/2020-12-01-predictive_maintenance_data_science.md +++ b/_posts/2020-12-01-predictive_maintenance_data_science.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2020-12-01' -excerpt: Learn how data science revolutionizes predictive maintenance through key - techniques like regression, anomaly detection, and clustering to forecast machine - failures and optimize maintenance schedules. +excerpt: Learn how data science revolutionizes predictive maintenance through key techniques like regression, anomaly detection, and clustering to forecast machine failures and optimize maintenance schedules. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_6.jpg @@ -21,14 +19,10 @@ keywords: - Regression - Machine learning - Data science -seo_description: Explore the impact of data science on predictive maintenance, including - techniques like regression, anomaly detection, and clustering for failure forecasting - and optimization of maintenance schedules. +seo_description: Explore the impact of data science on predictive maintenance, including techniques like regression, anomaly detection, and clustering for failure forecasting and optimization of maintenance schedules. seo_title: 'Data Science in Predictive Maintenance: Techniques and Applications' seo_type: article -summary: This article delves into the role of data science in predictive maintenance - (PdM), explaining how methods such as regression, anomaly detection, and clustering - help forecast equipment failures, reduce downtime, and optimize maintenance strategies. +summary: This article delves into the role of data science in predictive maintenance (PdM), explaining how methods such as regression, anomaly detection, and clustering help forecast equipment failures, reduce downtime, and optimize maintenance strategies. tags: - Data science - Machine learning diff --git a/_posts/2020-12-30-ordinal_regression.md b/_posts/2020-12-30-ordinal_regression.md index 13dc226d..58c63281 100644 --- a/_posts/2020-12-30-ordinal_regression.md +++ b/_posts/2020-12-30-ordinal_regression.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2020-12-30' -excerpt: Explore the architecture of ordinal regression models, their applications - in real-world data, and how marginal effects enhance the interpretability of complex - models using Python. +excerpt: Explore the architecture of ordinal regression models, their applications in real-world data, and how marginal effects enhance the interpretability of complex models using Python. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_9.jpg @@ -21,20 +19,18 @@ keywords: - Ordinal regression - Marginal effects - Python -seo_description: This article covers the principles of ordinal regression, its applications - in real-world data, and how to interpret the results using marginal effects. We - provide detailed examples to help you implement this model effectively in Python. +- python +seo_description: This article covers the principles of ordinal regression, its applications in real-world data, and how to interpret the results using marginal effects. We provide detailed examples to help you implement this model effectively in Python. seo_title: 'Ordinal Regression Explained: Models, Marginal Effects, and Applications' seo_type: article -summary: This article explains ordinal regression models, from their mathematical - structure to real-world applications, including how marginal effects make model - outputs more interpretable in Python. +summary: This article explains ordinal regression models, from their mathematical structure to real-world applications, including how marginal effects make model outputs more interpretable in Python. tags: - Statistical models - Data analysis - Ordinal regression - Marginal effects - Python +- python title: 'Understanding Ordinal Regression: A Comprehensive Guide' --- diff --git a/_posts/2021-01-01-pde_data_science.md b/_posts/2021-01-01-pde_data_science.md index 1de7891a..459ed80a 100644 --- a/_posts/2021-01-01-pde_data_science.md +++ b/_posts/2021-01-01-pde_data_science.md @@ -4,10 +4,7 @@ categories: - Mathematics classes: wide date: '2021-01-01' -excerpt: PDEs offer a powerful framework for understanding complex systems in fields - like physics, finance, and environmental science. Discover how data scientists can - integrate PDEs with modern machine learning techniques to create robust predictive - models. +excerpt: PDEs offer a powerful framework for understanding complex systems in fields like physics, finance, and environmental science. Discover how data scientists can integrate PDEs with modern machine learning techniques to create robust predictive models. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_2.jpg @@ -21,15 +18,10 @@ keywords: - Data science - Numerical solutions - Physics-informed neural networks -seo_description: Explore the importance of Partial Differential Equations (PDEs) in - data science, including their role in machine learning, physics-informed models, - and numerical methods. +seo_description: Explore the importance of Partial Differential Equations (PDEs) in data science, including their role in machine learning, physics-informed models, and numerical methods. seo_title: Partial Differential Equations for Data Scientists seo_type: article -summary: This article explores the role of Partial Differential Equations (PDEs) in - data science, including their applications in machine learning, finance, image processing, - and environmental modeling. It covers basic classifications of PDEs, solution methods, - and why data scientists should care about them. +summary: This article explores the role of Partial Differential Equations (PDEs) in data science, including their applications in machine learning, finance, image processing, and environmental modeling. It covers basic classifications of PDEs, solution methods, and why data scientists should care about them. tags: - Physics-informed models - Machine learning diff --git a/_posts/2021-02-01-bayesian.md b/_posts/2021-02-01-bayesian.md index 3aae45b0..df5f81b5 100644 --- a/_posts/2021-02-01-bayesian.md +++ b/_posts/2021-02-01-bayesian.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2021-02-01' -excerpt: Bayesian data science offers a powerful framework for incorporating prior - knowledge into statistical analysis, improving predictions, and informing decisions - in a probabilistic manner. +excerpt: Bayesian data science offers a powerful framework for incorporating prior knowledge into statistical analysis, improving predictions, and informing decisions in a probabilistic manner. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_8.jpg @@ -21,15 +19,10 @@ keywords: - Probabilistic modeling - Posterior distribution - Bayesian inference -seo_description: Explore the principles of Bayesian data science, its importance in - modern analytics, and how it differs from traditional methods. Learn how Bayesian - inference improves decision-making and model reliability. +seo_description: Explore the principles of Bayesian data science, its importance in modern analytics, and how it differs from traditional methods. Learn how Bayesian inference improves decision-making and model reliability. seo_title: 'Understanding Bayesian Data Science: What, Why, and How' seo_type: article -summary: Bayesian data science is a statistical approach that incorporates prior knowledge - with observed data using Bayes' theorem. It provides a more intuitive and flexible - framework for modeling uncertainty and improving decision-making, especially in - complex or small data scenarios. +summary: Bayesian data science is a statistical approach that incorporates prior knowledge with observed data using Bayes' theorem. It provides a more intuitive and flexible framework for modeling uncertainty and improving decision-making, especially in complex or small data scenarios. tags: - Inference - Statistical modeling diff --git a/_posts/2021-02-17-traffic_safety_kde.md b/_posts/2021-02-17-traffic_safety_kde.md index 34c48731..504b1d05 100644 --- a/_posts/2021-02-17-traffic_safety_kde.md +++ b/_posts/2021-02-17-traffic_safety_kde.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2021-02-17' -excerpt: A deep dive into using Kernel Density Estimation (KDE) for identifying traffic - accident hotspots and improving road safety, including practical applications and - case studies from Japan. +excerpt: A deep dive into using Kernel Density Estimation (KDE) for identifying traffic accident hotspots and improving road safety, including practical applications and case studies from Japan. header: image: /assets/images/traffic_kde_2.png og_image: /assets/images/data_science_1.jpg @@ -25,17 +23,12 @@ keywords: - Gis - Bash - Python -seo_description: This article explores how Kernel Density Estimation (KDE) can be - used for detecting traffic accident hotspots and improving urban traffic safety, - with case studies from Japan. +- bash +- python +seo_description: This article explores how Kernel Density Estimation (KDE) can be used for detecting traffic accident hotspots and improving urban traffic safety, with case studies from Japan. seo_title: Using KDE for Traffic Accident Hotspots Detection seo_type: article -summary: Traffic safety in urban areas remains a significant challenge globally. This - article discusses how Kernel Density Estimation (KDE), a statistical tool used in - spatial analysis, can help identify accident hotspots. The use of KDE provides urban - planners with a proactive approach to reducing traffic accidents, addressing the - limitations of traditional methods, and offering practical solutions for real-world - applications. +summary: Traffic safety in urban areas remains a significant challenge globally. This article discusses how Kernel Density Estimation (KDE), a statistical tool used in spatial analysis, can help identify accident hotspots. The use of KDE provides urban planners with a proactive approach to reducing traffic accidents, addressing the limitations of traditional methods, and offering practical solutions for real-world applications. tags: - Traffic safety - Traffic accident hotspots @@ -45,8 +38,9 @@ tags: - Kde - Bash - Python -title: 'Traffic Safety with Data: A Comprehensive Approach Using Kernel Density Estimation - (KDE) to Detect Traffic Accident Hotspots' +- bash +- python +title: 'Traffic Safety with Data: A Comprehensive Approach Using Kernel Density Estimation (KDE) to Detect Traffic Accident Hotspots' --- ![Example Image](/assets/images/traffic_kde_3.png) diff --git a/_posts/2021-03-01-polynomial_regression.md b/_posts/2021-03-01-polynomial_regression.md index 22c15588..c7190468 100644 --- a/_posts/2021-03-01-polynomial_regression.md +++ b/_posts/2021-03-01-polynomial_regression.md @@ -4,10 +4,7 @@ categories: - Machine Learning classes: wide date: '2021-03-01' -excerpt: Polynomial regression is a popular extension of linear regression that models - nonlinear relationships between the response and explanatory variables. However, - despite its name, polynomial regression remains a form of linear regression, as - the response variable is still a linear combination of the regression coefficients. +excerpt: Polynomial regression is a popular extension of linear regression that models nonlinear relationships between the response and explanatory variables. However, despite its name, polynomial regression remains a form of linear regression, as the response variable is still a linear combination of the regression coefficients. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_2.jpg @@ -23,16 +20,10 @@ keywords: - Machine learning regression - Nonlinear regression models - Linear regression -seo_description: Explore why polynomial regression, despite modeling nonlinear relationships - between the response and explanatory variables, is mathematically considered a form - of linear regression. -seo_title: "Polynomial Regression: Why It\u2019s Still Linear Regression" +seo_description: Explore why polynomial regression, despite modeling nonlinear relationships between the response and explanatory variables, is mathematically considered a form of linear regression. +seo_title: 'Polynomial Regression: Why It’s Still Linear Regression' seo_type: article -summary: Polynomial regression models the relationship between the response variable - and explanatory variables using a pth-order polynomial. Although this suggests a - nonlinear relationship between the response and explanatory variables, it is still - linear regression, as the linearity pertains to the relationship between the response - variable and the regression coefficients. +summary: Polynomial regression models the relationship between the response variable and explanatory variables using a pth-order polynomial. Although this suggests a nonlinear relationship between the response and explanatory variables, it is still linear regression, as the linearity pertains to the relationship between the response variable and the regression coefficients. tags: - Polynomial regression - Regression analysis diff --git a/_posts/2021-03-01-type_1_type_2_errors.md b/_posts/2021-03-01-type_1_type_2_errors.md index dfb0944b..ebf9be23 100644 --- a/_posts/2021-03-01-type_1_type_2_errors.md +++ b/_posts/2021-03-01-type_1_type_2_errors.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2021-03-01' -excerpt: Learn how to avoid false positives and false negatives in hypothesis testing - by understanding Type I and Type II errors, their causes, and how to balance statistical - power and sample size. +excerpt: Learn how to avoid false positives and false negatives in hypothesis testing by understanding Type I and Type II errors, their causes, and how to balance statistical power and sample size. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -20,22 +18,17 @@ keywords: - Type i error - Data science - Hypothesis testing -seo_description: Explore the differences between Type I and Type II errors in statistical - testing, learn how to minimize them, and understand their impact on data science, - clinical trials, and AI model evaluation. +seo_description: Explore the differences between Type I and Type II errors in statistical testing, learn how to minimize them, and understand their impact on data science, clinical trials, and AI model evaluation. seo_title: 'Type I vs. Type II Errors in Statistical Testing: How to Avoid False Conclusions' seo_type: article -summary: This article explains the fundamental concepts behind Type I and Type II - errors in statistical testing, covering their causes, how to minimize them, and - the critical role of statistical power and sample size in data science. +summary: This article explains the fundamental concepts behind Type I and Type II errors in statistical testing, covering their causes, how to minimize them, and the critical role of statistical power and sample size in data science. tags: - Statistical testing - Type ii error - Type i error - Data science - Hypothesis testing -title: 'Understanding Type I and Type II Errors in Statistical Testing: How to Minimize - False Conclusions' +title: 'Understanding Type I and Type II Errors in Statistical Testing: How to Minimize False Conclusions' --- ## Introduction: The Importance of Understanding Type I and Type II Errors diff --git a/_posts/2021-04-01-asymmetric_confidence_interval.md b/_posts/2021-04-01-asymmetric_confidence_interval.md index 428d46ae..5ddbbc2a 100644 --- a/_posts/2021-04-01-asymmetric_confidence_interval.md +++ b/_posts/2021-04-01-asymmetric_confidence_interval.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2021-04-01' -excerpt: Discover the reasons behind asymmetric confidence intervals in statistics - and how they impact research interpretation. +excerpt: Discover the reasons behind asymmetric confidence intervals in statistics and how they impact research interpretation. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_5.jpg @@ -20,13 +19,12 @@ keywords: - Bash - Data distribution - Statistical methods -seo_description: Learn why confidence intervals can be asymmetric, the factors that - contribute to this phenomenon, and how to interpret them in statistical analysis. +- python +- bash +seo_description: Learn why confidence intervals can be asymmetric, the factors that contribute to this phenomenon, and how to interpret them in statistical analysis. seo_title: 'Asymmetric Confidence Intervals: Causes and Understanding' seo_type: article -summary: Asymmetric confidence intervals can result from the nature of your data or - the statistical method used. This article explores the causes and implications of - these intervals for interpreting research results. +summary: Asymmetric confidence intervals can result from the nature of your data or the statistical method used. This article explores the causes and implications of these intervals for interpreting research results. tags: - Asymmetric ci - Confidence intervals @@ -34,6 +32,8 @@ tags: - Data distribution - Statistical tests - Python +- python +- bash title: 'Understanding Asymmetric Confidence Intervals: Causes and Implications' --- diff --git a/_posts/2021-04-27-forest_fires_kde.md b/_posts/2021-04-27-forest_fires_kde.md index 742c1a30..ea08e67d 100644 --- a/_posts/2021-04-27-forest_fires_kde.md +++ b/_posts/2021-04-27-forest_fires_kde.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2021-04-27' -excerpt: A study using GIS-based techniques for forest fire hotspot identification - and analysis, validated with contributory factors like population density, precipitation, - elevation, and vegetation cover. +excerpt: A study using GIS-based techniques for forest fire hotspot identification and analysis, validated with contributory factors like population density, precipitation, elevation, and vegetation cover. header: image: /assets/images/forest_fire_kde_1.jpg og_image: /assets/images/data_science_4.jpg @@ -26,26 +24,23 @@ keywords: - Belait district - Bash - Python -seo_description: "Explore GIS techniques like KDE, Getis-Ord Gi*, and Anselin Local\ - \ Moran\u2019s I for identifying forest fire hotspots in Southeast Asia, validated\ - \ by contributory factors." +- bash +- python +seo_description: Explore GIS techniques like KDE, Getis-Ord Gi*, and Anselin Local Moran’s I for identifying forest fire hotspots in Southeast Asia, validated by contributory factors. seo_title: GIS-Based Forest Fire Hotspot Identification Using Contributory Factors seo_type: article -summary: This article explores the application of GIS-based techniques, such as Kernel - Density Estimation (KDE), Getis-Ord Gi*, and Anselin Local Moran's I, in identifying - forest fire hotspots. The study focuses on Belait District, Brunei Darussalam, and - validates hotspot results using contributory factors like population density, precipitation, - elevation, and vegetation cover. +summary: This article explores the application of GIS-based techniques, such as Kernel Density Estimation (KDE), Getis-Ord Gi*, and Anselin Local Moran's I, in identifying forest fire hotspots. The study focuses on Belait District, Brunei Darussalam, and validates hotspot results using contributory factors like population density, precipitation, elevation, and vegetation cover. tags: -- "Anselin local moran\u2019s i" +- Anselin local moran’s i - Gis - Forest fires - Getis-ord gi* - Python - Kernel density estimation - Bash -title: 'GIS-Based Forest Fire Hotspot Identification: A Comprehensive Approach Using - Contributory Factors' +- bash +- python +title: 'GIS-Based Forest Fire Hotspot Identification: A Comprehensive Approach Using Contributory Factors' --- ![Example Image](/assets/images/forest_fire_kde_3.png) diff --git a/_posts/2021-04-30-big_data_climate_change_mitigation.md b/_posts/2021-04-30-big_data_climate_change_mitigation.md index 41e3a983..4256798b 100644 --- a/_posts/2021-04-30-big_data_climate_change_mitigation.md +++ b/_posts/2021-04-30-big_data_climate_change_mitigation.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2021-04-30' -excerpt: Big data is revolutionizing climate science, enabling more accurate predictions - and helping formulate effective mitigation strategies. +excerpt: Big data is revolutionizing climate science, enabling more accurate predictions and helping formulate effective mitigation strategies. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -19,14 +18,10 @@ keywords: - Environmental monitoring - Satellite data - Predictive analytics -seo_description: This article explores how big data is being used to monitor and predict - climate change, utilizing tools like satellite data, sensors, and environmental - monitoring systems. +seo_description: This article explores how big data is being used to monitor and predict climate change, utilizing tools like satellite data, sensors, and environmental monitoring systems. seo_title: How Big Data Can Help Mitigate Climate Change seo_type: article -summary: In this article, we examine the intersection of big data and climate science, - focusing on how large-scale data collection and analysis are transforming our ability - to monitor, predict, and mitigate climate change. +summary: In this article, we examine the intersection of big data and climate science, focusing on how large-scale data collection and analysis are transforming our ability to monitor, predict, and mitigate climate change. tags: - Big data - Climate change diff --git a/_posts/2021-05-01-rare_labels_machine_learning.md b/_posts/2021-05-01-rare_labels_machine_learning.md index 3fdfef6c..92a955c9 100644 --- a/_posts/2021-05-01-rare_labels_machine_learning.md +++ b/_posts/2021-05-01-rare_labels_machine_learning.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2021-05-01' -excerpt: Rare labels in categorical variables can cause significant issues in machine - learning, such as overfitting. This article explains why rare labels can be problematic - and provides examples on how to handle them. +excerpt: Rare labels in categorical variables can cause significant issues in machine learning, such as overfitting. This article explains why rare labels can be problematic and provides examples on how to handle them. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_1.jpg @@ -23,14 +21,11 @@ keywords: - Overfitting - Mercedes-benz challenge - Python -seo_description: Explore the impact of rare labels in categorical variables on machine - learning models, particularly their tendency to cause overfitting, and learn how - to handle rare values using feature engineering. +- python +seo_description: Explore the impact of rare labels in categorical variables on machine learning models, particularly their tendency to cause overfitting, and learn how to handle rare values using feature engineering. seo_title: Handling Rare Labels in Categorical Variables for Machine Learning seo_type: article -summary: This article covers how rare labels in categorical variables can impact machine - learning models, particularly tree-based methods, and why it's important to address - these rare labels during preprocessing. +summary: This article covers how rare labels in categorical variables can impact machine learning models, particularly tree-based methods, and why it's important to address these rare labels during preprocessing. tags: - Mercedes-benz greener manufacturing challenge - Categorical variables @@ -39,6 +34,7 @@ tags: - Rare labels - Feature engineering - Python +- python title: Handling Rare Labels in Categorical Variables in Machine Learning --- diff --git a/_posts/2021-05-10-estimating_uncertainty_neural_networks_using_monte_carlo_dropout.md b/_posts/2021-05-10-estimating_uncertainty_neural_networks_using_monte_carlo_dropout.md index 4ed37791..6f503d75 100644 --- a/_posts/2021-05-10-estimating_uncertainty_neural_networks_using_monte_carlo_dropout.md +++ b/_posts/2021-05-10-estimating_uncertainty_neural_networks_using_monte_carlo_dropout.md @@ -4,9 +4,7 @@ categories: - Neural Networks classes: wide date: '2021-05-10' -excerpt: This article discusses Monte Carlo dropout and how it is used to estimate - uncertainty in multi-class neural network classification, covering methods such - as entropy, variance, and predictive probabilities. +excerpt: This article discusses Monte Carlo dropout and how it is used to estimate uncertainty in multi-class neural network classification, covering methods such as entropy, variance, and predictive probabilities. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -20,15 +18,10 @@ keywords: - Multi-class classification - Neural networks - Entropy -seo_description: Explore how Monte Carlo dropout can estimate uncertainty in neural - networks for multi-class classification, examining various methods to derive uncertainty - scores. +seo_description: Explore how Monte Carlo dropout can estimate uncertainty in neural networks for multi-class classification, examining various methods to derive uncertainty scores. seo_title: Estimating Uncertainty with Monte Carlo Dropout in Neural Networks seo_type: article -summary: In this article, we explore how to estimate uncertainty in neural network - predictions using Monte Carlo dropout. We explain the mechanism of Monte Carlo dropout - and dive into methods like entropy, predictive probabilities, and error-function-based - uncertainty estimation. +summary: In this article, we explore how to estimate uncertainty in neural network predictions using Monte Carlo dropout. We explain the mechanism of Monte Carlo dropout and dive into methods like entropy, predictive probabilities, and error-function-based uncertainty estimation. tags: - Monte carlo dropout - Uncertainty quantification diff --git a/_posts/2021-05-12-understanding_heart_rate_variability_through_lens_coefficient_variation_health_monitoring.md b/_posts/2021-05-12-understanding_heart_rate_variability_through_lens_coefficient_variation_health_monitoring.md index 77deb746..2fec968f 100644 --- a/_posts/2021-05-12-understanding_heart_rate_variability_through_lens_coefficient_variation_health_monitoring.md +++ b/_posts/2021-05-12-understanding_heart_rate_variability_through_lens_coefficient_variation_health_monitoring.md @@ -4,8 +4,7 @@ categories: - Health Monitoring classes: wide date: '2021-05-12' -excerpt: Discover the significance of heart rate variability (HRV) and how the coefficient - of variation (CV) provides a more nuanced view of cardiovascular health. +excerpt: Discover the significance of heart rate variability (HRV) and how the coefficient of variation (CV) provides a more nuanced view of cardiovascular health. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -19,19 +18,15 @@ keywords: - Cardiovascular health - Fitness monitoring - Stress assessment -seo_description: Explore how the coefficient of variation offers deeper insights into - heart rate variability and health monitoring. +seo_description: Explore how the coefficient of variation offers deeper insights into heart rate variability and health monitoring. seo_title: Understanding HRV and Coefficient of Variation seo_type: article -summary: This article delves into heart rate variability (HRV), focusing on the coefficient - of variation (CV) as a critical metric for understanding cardiovascular health and - overall well-being. +summary: This article delves into heart rate variability (HRV), focusing on the coefficient of variation (CV) as a critical metric for understanding cardiovascular health and overall well-being. tags: - Heart rate variability - Coefficient of variation - Health metrics -title: Understanding Heart Rate Variability Through the Lens of the Coefficient of - Variation in Health Monitoring +title: Understanding Heart Rate Variability Through the Lens of the Coefficient of Variation in Health Monitoring --- Heart rate variability (HRV) is one of the most important indicators of cardiovascular health and overall well-being. It reflects the body’s ability to adapt to stress, rest, exercise, and environmental stimuli. Traditionally, HRV has been measured using several statistical tools, including standard deviation, root mean square of successive differences (RMSSD), and the low-frequency to high-frequency (LF/HF) ratio, to name a few. diff --git a/_posts/2021-05-26-kernel_math.md b/_posts/2021-05-26-kernel_math.md index 33ddf367..ecf0cae8 100644 --- a/_posts/2021-05-26-kernel_math.md +++ b/_posts/2021-05-26-kernel_math.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2021-05-26' -excerpt: Explore the foundations, concepts, and mathematics behind Kernel Density - Estimation (KDE), a powerful tool in non-parametric statistics for estimating probability - density functions. +excerpt: Explore the foundations, concepts, and mathematics behind Kernel Density Estimation (KDE), a powerful tool in non-parametric statistics for estimating probability density functions. header: excerpt: false image: /assets/images/kernel_math.jpg @@ -26,19 +24,10 @@ keywords: - Machine learning - Kernel density estimation - Bandwidth selection -seo_description: A deep dive into the math, theory, and practical considerations of - Kernel Density Estimation (KDE), covering its core components, bandwidth selection, - kernel functions, multivariate KDE, and real-world applications. +seo_description: A deep dive into the math, theory, and practical considerations of Kernel Density Estimation (KDE), covering its core components, bandwidth selection, kernel functions, multivariate KDE, and real-world applications. seo_title: Exploring the Math Behind Kernel Density Estimation seo_type: article -summary: Kernel Density Estimation (KDE) is a non-parametric method used to estimate - the probability density function of data without assuming a specific distribution. - This article explores the mathematical foundations behind KDE, including the role - of kernel functions, bandwidth selection, and their impact on bias and variance. - The article also covers multivariate KDE, efficient computational techniques, and - applications of KDE in fields such as data science, machine learning, and statistics. - With a focus on practical insights and theoretical rigor, the article offers a comprehensive - guide to understanding KDE. +summary: Kernel Density Estimation (KDE) is a non-parametric method used to estimate the probability density function of data without assuming a specific distribution. This article explores the mathematical foundations behind KDE, including the role of kernel functions, bandwidth selection, and their impact on bias and variance. The article also covers multivariate KDE, efficient computational techniques, and applications of KDE in fields such as data science, machine learning, and statistics. With a focus on practical insights and theoretical rigor, the article offers a comprehensive guide to understanding KDE. tags: - Non-parametric statistics - Multivariate kde diff --git a/_posts/2021-06-01-customer_segmentation.md b/_posts/2021-06-01-customer_segmentation.md index 6dbea409..0b0ba7f1 100644 --- a/_posts/2021-06-01-customer_segmentation.md +++ b/_posts/2021-06-01-customer_segmentation.md @@ -4,9 +4,7 @@ categories: - Customer Analytics classes: wide date: '2021-06-01' -excerpt: RFM Segmentation (Recency, Frequency, Monetary Value) is a widely used method - to segment customers based on their behavior. This article provides a deep dive - into RFM, showing how to apply clustering techniques for effective customer segmentation. +excerpt: RFM Segmentation (Recency, Frequency, Monetary Value) is a widely used method to segment customers based on their behavior. This article provides a deep dive into RFM, showing how to apply clustering techniques for effective customer segmentation. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_1.jpg @@ -22,14 +20,11 @@ keywords: - Rfm segmentation - Machine learning - Python -seo_description: Learn about RFM Segmentation, a customer segmentation technique used - to increase retention, improve marketing strategies, and enhance customer experiences. - Discover how to implement RFM clustering using unsupervised learning. +- python +seo_description: Learn about RFM Segmentation, a customer segmentation technique used to increase retention, improve marketing strategies, and enhance customer experiences. Discover how to implement RFM clustering using unsupervised learning. seo_title: 'RFM Segmentation: Understanding Customer Value with Machine Learning' seo_type: article -summary: This article provides an in-depth exploration of RFM segmentation, explaining - how businesses can use Recency, Frequency, and Monetary Value to identify customer - groups, improve marketing, and enhance retention strategies using clustering techniques. +summary: This article provides an in-depth exploration of RFM segmentation, explaining how businesses can use Recency, Frequency, and Monetary Value to identify customer groups, improve marketing, and enhance retention strategies using clustering techniques. tags: - Clustering - Unsupervised learning @@ -38,6 +33,7 @@ tags: - Data science - Rfm segmentation - Python +- python title: 'RFM Segmentation: A Powerful Customer Segmentation Technique' --- diff --git a/_posts/2021-07-26-regression_tasks.md b/_posts/2021-07-26-regression_tasks.md index e03ff316..a344e5ed 100644 --- a/_posts/2021-07-26-regression_tasks.md +++ b/_posts/2021-07-26-regression_tasks.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2021-07-26' -excerpt: Regression tasks are at the heart of machine learning. This guide explores - methods like Linear Regression, Principal Component Regression, Gaussian Process - Regression, and Support Vector Regression, with insights on when to use each. +excerpt: Regression tasks are at the heart of machine learning. This guide explores methods like Linear Regression, Principal Component Regression, Gaussian Process Regression, and Support Vector Regression, with insights on when to use each. header: image: /assets/images/regression-analysis-2.jpg og_image: /assets/images/data_science_8.jpg @@ -29,10 +27,8 @@ keywords: - Dimensionality reduction - Machine learning - Gaussian process regression -seo_description: A comprehensive guide to selecting the best regression algorithm - for your dataset, based on complexity, dimensionality, and the need for probabilistic - output. Explore traditional machine learning methods with detailed explanations - and code examples. +- python +seo_description: A comprehensive guide to selecting the best regression algorithm for your dataset, based on complexity, dimensionality, and the need for probabilistic output. Explore traditional machine learning methods with detailed explanations and code examples. seo_title: 'Choosing the Right Regression Task: From Linear Models to Advanced Techniques' seo_type: article tags: @@ -43,6 +39,7 @@ tags: - Machine learning algorithms - Principal component regression - Python +- python title: 'A Guide to Regression Tasks: Choosing the Right Approach' --- diff --git a/_posts/2021-08-01-building_linear_regression_from_scratch.md b/_posts/2021-08-01-building_linear_regression_from_scratch.md index 3916165b..194315f6 100644 --- a/_posts/2021-08-01-building_linear_regression_from_scratch.md +++ b/_posts/2021-08-01-building_linear_regression_from_scratch.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2021-08-01' -excerpt: A step-by-step guide to implementing Linear Regression from scratch using - the Normal Equation method, complete with Python code and evaluation techniques. +excerpt: A step-by-step guide to implementing Linear Regression from scratch using the Normal Equation method, complete with Python code and evaluation techniques. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -19,19 +18,17 @@ keywords: - Python - Data science interviews - Python -seo_description: Learn how to build a Linear Regression model from scratch using the - Normal Equation approach. This article covers the theoretical foundations, algorithm - design, and Python implementation. +- python +seo_description: Learn how to build a Linear Regression model from scratch using the Normal Equation approach. This article covers the theoretical foundations, algorithm design, and Python implementation. seo_title: Building Linear Regression from Scratch Using the Normal Equation seo_type: article -summary: This article provides a detailed algorithmic approach to building a Linear - Regression model from scratch, covering theory, Python code implementation, and - performance evaluation. +summary: This article provides a detailed algorithmic approach to building a Linear Regression model from scratch, covering theory, Python code implementation, and performance evaluation. tags: - Linear regression - Python - Normal equation - Python +- python title: 'Building Linear Regression from Scratch: A Detailed Algorithmic Approach' --- diff --git a/_posts/2021-09-24-crime_analysis.md b/_posts/2021-09-24-crime_analysis.md index 5db0f948..f69ecce1 100644 --- a/_posts/2021-09-24-crime_analysis.md +++ b/_posts/2021-09-24-crime_analysis.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2021-09-24' -excerpt: This article explores the use of K-means clustering in crime analysis, including - practical implementation, case studies, and future directions. +excerpt: This article explores the use of K-means clustering in crime analysis, including practical implementation, case studies, and future directions. header: image: /assets/images/machine_learning/machine_learning_3.jpeg og_image: /assets/images/data_science_9.jpg @@ -20,20 +19,18 @@ keywords: - K-means clustering - Data mining - Python -seo_description: Explore how K-means clustering can enhance crime analysis by identifying - patterns, predicting trends, and improving crime prevention through data mining. +- python +seo_description: Explore how K-means clustering can enhance crime analysis by identifying patterns, predicting trends, and improving crime prevention through data mining. seo_title: Crime Analysis Using K-Means Clustering seo_type: article -summary: This article delves into the application of K-means clustering in crime analysis, - showing how law enforcement agencies can uncover crime patterns, allocate resources, - and predict criminal activity. The article includes a detailed exploration of data - mining, clustering methods, and practical use cases. +summary: This article delves into the application of K-means clustering in crime analysis, showing how law enforcement agencies can uncover crime patterns, allocate resources, and predict criminal activity. The article includes a detailed exploration of data mining, clustering methods, and practical use cases. tags: - Data mining - K-means clustering - Machine learning - Crime analysis - Python +- python title: 'Crime Analysis Using K-Means Clustering: Enhancing Security through Data Mining' --- diff --git a/_posts/2021-12-24-linear_programming.md b/_posts/2021-12-24-linear_programming.md index 46b8e52a..973cdcea 100644 --- a/_posts/2021-12-24-linear_programming.md +++ b/_posts/2021-12-24-linear_programming.md @@ -4,10 +4,7 @@ categories: - Operations Research classes: wide date: '2021-12-24' -excerpt: Linear Programming is the foundation of optimization in operations research. - We explore its traditional methods, challenges in scaling large instances, and introduce - PDLP, a scalable solver using first-order methods, designed for modern computational - infrastructures. +excerpt: Linear Programming is the foundation of optimization in operations research. We explore its traditional methods, challenges in scaling large instances, and introduce PDLP, a scalable solver using first-order methods, designed for modern computational infrastructures. header: image: /assets/images/linear_program.jpeg og_image: /assets/images/data_science_4.jpg @@ -31,11 +28,8 @@ keywords: - Scalable lp solutions - First-order methods - Computational optimization -seo_description: A detailed exploration of linear programming, its traditional methods - like Simplex and interior-point methods, and the emergence of scalable first-order - methods such as PDLP, a revolutionary solver for large-scale LP problems. -seo_title: 'Classic Linear Programming and PDLP: Scaling Solutions for Modern Computational - Optimization' +seo_description: A detailed exploration of linear programming, its traditional methods like Simplex and interior-point methods, and the emergence of scalable first-order methods such as PDLP, a revolutionary solver for large-scale LP problems. +seo_title: 'Classic Linear Programming and PDLP: Scaling Solutions for Modern Computational Optimization' seo_type: article tags: - Primal-dual hybrid gradient method @@ -43,8 +37,7 @@ tags: - Computational optimization - Linear programming - Or-tools -title: 'Exploring Classic Linear Programming (LP) Problems and Scalable Solutions: - A Deep Dive into PDLP' +title: 'Exploring Classic Linear Programming (LP) Problems and Scalable Solutions: A Deep Dive into PDLP' --- ## Introduction diff --git a/_posts/2021-12-25-suply_chain.md b/_posts/2021-12-25-suply_chain.md index e7b0b520..c9c03960 100644 --- a/_posts/2021-12-25-suply_chain.md +++ b/_posts/2021-12-25-suply_chain.md @@ -4,9 +4,7 @@ categories: - Optimization classes: wide date: '2021-12-25' -excerpt: Discover how data science enhances supply chain optimization and industrial - network analysis, leveraging techniques like predictive analytics, machine learning, - and graph theory to optimize operations. +excerpt: Discover how data science enhances supply chain optimization and industrial network analysis, leveraging techniques like predictive analytics, machine learning, and graph theory to optimize operations. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_2.jpg @@ -29,9 +27,7 @@ keywords: - Resource allocation - Supply chain optimization - Data science in supply chain -seo_description: Explore how data science drives supply chain optimization and industrial - network analysis, focusing on predictive analytics, IoT, and graph theory for improved - efficiency. +seo_description: Explore how data science drives supply chain optimization and industrial network analysis, focusing on predictive analytics, IoT, and graph theory for improved efficiency. seo_title: Data-Driven Supply Chain Optimization and Industrial Network Analysis seo_type: article tags: diff --git a/_posts/2021-12-31-FDM.md b/_posts/2021-12-31-FDM.md index ab07c2ab..0ecf097f 100644 --- a/_posts/2021-12-31-FDM.md +++ b/_posts/2021-12-31-FDM.md @@ -4,9 +4,7 @@ categories: - Mathematics classes: wide date: '2021-12-31' -excerpt: Explore how Finite Difference Methods and the Black-Scholes-Merton differential - equation are used to solve option pricing problems numerically, with a focus on - explicit and implicit schemes. +excerpt: Explore how Finite Difference Methods and the Black-Scholes-Merton differential equation are used to solve option pricing problems numerically, with a focus on explicit and implicit schemes. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_1.jpg @@ -26,15 +24,12 @@ keywords: - Stability analysis - Bash - Python -seo_description: Learn how Finite Difference Methods (FDM) are used in solving the - Black-Scholes-Merton equation for option pricing, using explicit and implicit schemes, - and stability analysis. -seo_title: 'Finite Difference Methods in Option Pricing: The Black-Scholes-Merton - Equation' +- bash +- python +seo_description: Learn how Finite Difference Methods (FDM) are used in solving the Black-Scholes-Merton equation for option pricing, using explicit and implicit schemes, and stability analysis. +seo_title: 'Finite Difference Methods in Option Pricing: The Black-Scholes-Merton Equation' seo_type: article -summary: This article explains how Finite Difference Methods (FDM) are applied to - solve the Black-Scholes-Merton equation for option pricing, focusing on explicit - and implicit schemes, as well as stability analysis. +summary: This article explains how Finite Difference Methods (FDM) are applied to solve the Black-Scholes-Merton equation for option pricing, focusing on explicit and implicit schemes, as well as stability analysis. tags: - Numerical analysis - Financial engineering @@ -48,8 +43,9 @@ tags: - Numerical methods - Bash - Python -title: 'Finite Difference Methods and the Black-Scholes-Merton Equation: A Numerical - Approach to Option Pricing' +- bash +- python +title: 'Finite Difference Methods and the Black-Scholes-Merton Equation: A Numerical Approach to Option Pricing' --- ### Introduction: Numerical Methods in Financial Engineering diff --git a/_posts/2022-01-02-OLS.md b/_posts/2022-01-02-OLS.md index 21eaca38..0cd303a8 100644 --- a/_posts/2022-01-02-OLS.md +++ b/_posts/2022-01-02-OLS.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2022-01-02' -excerpt: A deep dive into the relationship between OLS and Theil-Sen estimators, revealing - their connection through weighted averages and robust median-based slopes. +excerpt: A deep dive into the relationship between OLS and Theil-Sen estimators, revealing their connection through weighted averages and robust median-based slopes. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -24,9 +23,7 @@ keywords: - Median-based slope - Ols estimator - Econometrics -seo_description: Explore the mathematical connection between OLS and Theil-Sen estimators - in regression analysis, highlighting their similarities, differences, and implications - for data analysis. +seo_description: Explore the mathematical connection between OLS and Theil-Sen estimators in regression analysis, highlighting their similarities, differences, and implications for data analysis. seo_title: 'OLS and Theil-Sen Estimators: Understanding Their Connection' seo_type: article tags: diff --git a/_posts/2022-02-17-staff_schedulling.md b/_posts/2022-02-17-staff_schedulling.md index b7c0a687..e7b6ee6f 100644 --- a/_posts/2022-02-17-staff_schedulling.md +++ b/_posts/2022-02-17-staff_schedulling.md @@ -4,8 +4,7 @@ categories: - Optimization classes: wide date: '2022-02-17' -excerpt: Discover how linear programming and Python's PuLP library can efficiently - solve staff scheduling challenges, minimizing costs while meeting operational demands. +excerpt: Discover how linear programming and Python's PuLP library can efficiently solve staff scheduling challenges, minimizing costs while meeting operational demands. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_8.jpg @@ -30,13 +29,12 @@ keywords: - Python - Bash - Python -seo_description: Learn how to use linear programming with the PuLP library in Python - to optimize staff scheduling and minimize costs in a 24/7 operational environment. +- bash +- python +seo_description: Learn how to use linear programming with the PuLP library in Python to optimize staff scheduling and minimize costs in a 24/7 operational environment. seo_title: Staff Scheduling Optimization with Linear Programming in Python seo_type: article -summary: "This article discusses using linear programming and Python\u2019s PuLP library\ - \ to optimize staff scheduling, focusing on cost minimization and meeting operational\ - \ requirements efficiently." +summary: This article discusses using linear programming and Python’s PuLP library to optimize staff scheduling, focusing on cost minimization and meeting operational requirements efficiently. tags: - Linear programming - Scheduling @@ -44,6 +42,8 @@ tags: - Python - Bash - Python +- bash +- python title: Optimizing Staff Scheduling with Linear Programming --- diff --git a/_posts/2022-03-14-levenes_test_vs._bartletts_test_checking_homogeneity_variances.md b/_posts/2022-03-14-levenes_test_vs._bartletts_test_checking_homogeneity_variances.md index 3f5f517c..312cac47 100644 --- a/_posts/2022-03-14-levenes_test_vs._bartletts_test_checking_homogeneity_variances.md +++ b/_posts/2022-03-14-levenes_test_vs._bartletts_test_checking_homogeneity_variances.md @@ -4,9 +4,7 @@ categories: - Hypothesis Testing classes: wide date: '2022-03-14' -excerpt: Levene's Test and Bartlett's Test are key tools for checking homogeneity - of variances in data. Learn when to use each test, based on normality assumptions, - and how they relate to tests like ANOVA. +excerpt: Levene's Test and Bartlett's Test are key tools for checking homogeneity of variances in data. Learn when to use each test, based on normality assumptions, and how they relate to tests like ANOVA. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -16,27 +14,21 @@ header: twitter_image: /assets/images/data_science_4.jpg keywords: - Levene's test -- "Bartlett\u2019s test" +- Bartlett’s test - Homogeneity of variances - Anova - Hypothesis testing -seo_description: This article compares Levene's Test and Bartlett's Test for checking - homogeneity of variances, discussing when to use each test based on data normality, - and their application in conjunction with ANOVA. -seo_title: "Levene's Test vs. Bartlett\u2019s Test: A Comparison for Testing Homogeneity\ - \ of Variances" +seo_description: This article compares Levene's Test and Bartlett's Test for checking homogeneity of variances, discussing when to use each test based on data normality, and their application in conjunction with ANOVA. +seo_title: 'Levene''s Test vs. Bartlett’s Test: A Comparison for Testing Homogeneity of Variances' seo_type: article -summary: "This article provides a detailed comparison between Levene's Test and Bartlett\u2019\ - s Test for assessing the homogeneity of variances in data. It explains the differences\ - \ in when to use these tests\u2014parametric vs. non-parametric data, normal vs.\ - \ non-normal data\u2014and their applications alongside statistical tests like ANOVA." +summary: This article provides a detailed comparison between Levene's Test and Bartlett’s Test for assessing the homogeneity of variances in data. It explains the differences in when to use these tests—parametric vs. non-parametric data, normal vs. non-normal data—and their applications alongside statistical tests like ANOVA. tags: - Levene's test -- "Bartlett\u2019s test" +- Bartlett’s test - Homogeneity of variances - Anova - Parametric and non-parametric tests -title: "Levene's Test vs. Bartlett\u2019s Test: Checking for Homogeneity of Variances" +title: 'Levene''s Test vs. Bartlett’s Test: Checking for Homogeneity of Variances' --- ## Introduction to Homogeneity of Variances diff --git a/_posts/2022-03-15-bayesian_ab_testing.md b/_posts/2022-03-15-bayesian_ab_testing.md index 8db55ff5..de7f3b00 100644 --- a/_posts/2022-03-15-bayesian_ab_testing.md +++ b/_posts/2022-03-15-bayesian_ab_testing.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2022-03-15' -excerpt: Explore Bayesian A/B testing as a powerful framework for analyzing conversion - rates, providing more nuanced insights than traditional frequentist approaches. +excerpt: Explore Bayesian A/B testing as a powerful framework for analyzing conversion rates, providing more nuanced insights than traditional frequentist approaches. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_9.jpg @@ -28,15 +27,15 @@ keywords: - Marketing optimization - Credible intervals - Python -seo_description: Learn how Bayesian A/B testing provides nuanced insights into conversion - rates, offering a robust alternative to traditional frequentist methods in data - analysis. +- python +seo_description: Learn how Bayesian A/B testing provides nuanced insights into conversion rates, offering a robust alternative to traditional frequentist methods in data analysis. seo_title: 'Bayesian A/B Testing: Enhancing Conversion Rate Analysis' seo_type: article tags: - A/b testing - Bayesian methods - Python +- python title: A Guide to Bayesian A/B Testing for Conversion Rates --- diff --git a/_posts/2022-03-23-degrees_freedom.md b/_posts/2022-03-23-degrees_freedom.md index 61564f4f..790afcdc 100644 --- a/_posts/2022-03-23-degrees_freedom.md +++ b/_posts/2022-03-23-degrees_freedom.md @@ -26,15 +26,10 @@ keywords: - Model monitoring - Artificial intelligence - Technology -seo_description: Explore advanced methods for machine learning monitoring by moving - beyond univariate data drift detection. Learn about direct loss estimation, detecting - outliers, and addressing alarm fatigue in production AI systems. +seo_description: Explore advanced methods for machine learning monitoring by moving beyond univariate data drift detection. Learn about direct loss estimation, detecting outliers, and addressing alarm fatigue in production AI systems. seo_title: 'Machine Learning Monitoring: Moving Beyond Univariate Data Drift Detection' seo_type: article -summary: This article explores advanced methods for monitoring machine learning models - beyond simple univariate data drift detection. It covers direct loss estimation, - outlier detection, and strategies to mitigate alarm fatigue, ensuring robust model - performance in production environments. +summary: This article explores advanced methods for monitoring machine learning models beyond simple univariate data drift detection. It covers direct loss estimation, outlier detection, and strategies to mitigate alarm fatigue, ensuring robust model performance in production environments. tags: - Data drift - Direct loss estimation diff --git a/_posts/2022-05-26-networks.md b/_posts/2022-05-26-networks.md index c24bede1..d9ea389a 100644 --- a/_posts/2022-05-26-networks.md +++ b/_posts/2022-05-26-networks.md @@ -4,8 +4,7 @@ categories: - Optimization classes: wide date: '2022-05-26' -excerpt: Learn how graph theory is applied to network analysis in production systems - to optimize processes, identify bottlenecks, and improve supply chain efficiency. +excerpt: Learn how graph theory is applied to network analysis in production systems to optimize processes, identify bottlenecks, and improve supply chain efficiency. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_2.jpg @@ -25,9 +24,7 @@ keywords: - Network models in production - Operational optimization - Industrial network analysis -seo_description: Explore how graph theory enhances network analysis in production - systems, improving efficiency in processes such as bottleneck identification, resource - allocation, and supply chain optimization. +seo_description: Explore how graph theory enhances network analysis in production systems, improving efficiency in processes such as bottleneck identification, resource allocation, and supply chain optimization. seo_title: 'Graph Theory in Production Systems: Network Analysis and Optimization' seo_type: article tags: diff --git a/_posts/2022-07-23-statistical_tests.md b/_posts/2022-07-23-statistical_tests.md index 3194565e..c1c37e54 100644 --- a/_posts/2022-07-23-statistical_tests.md +++ b/_posts/2022-07-23-statistical_tests.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2022-07-23' -excerpt: Discover the universal structure behind statistical tests, highlighting the - core comparison between observed and expected data that drives hypothesis testing - and data analysis. +excerpt: Discover the universal structure behind statistical tests, highlighting the core comparison between observed and expected data that drives hypothesis testing and data analysis. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_3.jpg @@ -27,14 +25,10 @@ keywords: - Common statistical test structure - Hypothesis comparison - Statistical methodologies -seo_description: Explore the underlying structure common to most statistical tests, - revealing how the comparison of observed versus expected data forms the basis of - hypothesis testing. +seo_description: Explore the underlying structure common to most statistical tests, revealing how the comparison of observed versus expected data forms the basis of hypothesis testing. seo_title: Understanding the Universal Structure of Statistical Tests seo_type: article -summary: This article explains the universal structure of statistical tests, focusing - on the comparison between observed and expected data that forms the foundation of - hypothesis testing and statistical inference. +summary: This article explains the universal structure of statistical tests, focusing on the comparison between observed and expected data that forms the foundation of hypothesis testing and statistical inference. tags: - Statistical tests - Data analysis diff --git a/_posts/2022-07-26-features.md b/_posts/2022-07-26-features.md index 74af110e..4edb3cb0 100644 --- a/_posts/2022-07-26-features.md +++ b/_posts/2022-07-26-features.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2022-07-26' -excerpt: Explore feature discretization as a powerful technique to enhance linear - models, bridging the gap between linear precision and non-linear complexity in data - analysis. +excerpt: Explore feature discretization as a powerful technique to enhance linear models, bridging the gap between linear precision and non-linear complexity in data analysis. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_8.jpg @@ -27,15 +25,10 @@ keywords: - Linear model optimization - Categorical features - Data binning techniques -seo_description: Learn how feature discretization transforms linear models, enabling - them to capture non-linear patterns and provide deeper insights in data analysis - and machine learning. +seo_description: Learn how feature discretization transforms linear models, enabling them to capture non-linear patterns and provide deeper insights in data analysis and machine learning. seo_title: 'Feature Discretization: Enhancing Linear Models for Non-Linear Insights' seo_type: article -summary: This article delves into feature discretization as a technique to enhance - linear models by enabling them to capture non-linear patterns. It explains how discretizing - continuous variables can optimize data analysis and machine learning models, offering - improved interpretability and performance in predictive tasks. +summary: This article delves into feature discretization as a technique to enhance linear models by enabling them to capture non-linear patterns. It explains how discretizing continuous variables can optimize data analysis and machine learning models, offering improved interpretability and performance in predictive tasks. tags: - Feature engineering - Linear models diff --git a/_posts/2022-08-14-wald_test_hypothesis_testing_regression_analysis.md b/_posts/2022-08-14-wald_test_hypothesis_testing_regression_analysis.md index 278abf45..d49f4ea3 100644 --- a/_posts/2022-08-14-wald_test_hypothesis_testing_regression_analysis.md +++ b/_posts/2022-08-14-wald_test_hypothesis_testing_regression_analysis.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2022-08-14' -excerpt: Explore the Wald test, a key tool in hypothesis testing for regression models, - its applications, and its role in logistic regression, Poisson regression, and beyond. +excerpt: Explore the Wald test, a key tool in hypothesis testing for regression models, its applications, and its role in logistic regression, Poisson regression, and beyond. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -19,15 +18,10 @@ keywords: - Regression analysis - Logistic regression - Poisson regression -seo_description: A comprehensive guide to the Wald test for hypothesis testing in - regression models, its applications in logistic regression, Poisson regression, - and more. +seo_description: A comprehensive guide to the Wald test for hypothesis testing in regression models, its applications in logistic regression, Poisson regression, and more. seo_title: 'Wald Test in Regression Analysis: An In-Depth Guide' seo_type: article -summary: The Wald test is a fundamental statistical method used to evaluate hypotheses - in regression analysis. This article provides an in-depth discussion on the theory, - practical applications, and interpretation of the Wald test in various regression - models. +summary: The Wald test is a fundamental statistical method used to evaluate hypotheses in regression analysis. This article provides an in-depth discussion on the theory, practical applications, and interpretation of the Wald test in various regression models. tags: - Wald test - Logistic regression diff --git a/_posts/2022-08-15-linear_relashionships.md b/_posts/2022-08-15-linear_relashionships.md index abe3762d..1da3351a 100644 --- a/_posts/2022-08-15-linear_relashionships.md +++ b/_posts/2022-08-15-linear_relashionships.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2022-08-15' -excerpt: In machine learning, linear models assume a direct relationship between predictors - and outcome variables. Learn why understanding these assumptions is critical for - model performance and how to work with non-linear relationships. +excerpt: In machine learning, linear models assume a direct relationship between predictors and outcome variables. Learn why understanding these assumptions is critical for model performance and how to work with non-linear relationships. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_9.jpg @@ -21,14 +19,10 @@ keywords: - Logistic regression - Lda - Feature transformation -seo_description: Exploring machine learning models that assume linear relationships, - including linear regression, logistic regression, and LDA, and why understanding - these assumptions is crucial for better model performance. +seo_description: Exploring machine learning models that assume linear relationships, including linear regression, logistic regression, and LDA, and why understanding these assumptions is crucial for better model performance. seo_title: 'Linear Relationships in Machine Learning: Understanding Their Importance' seo_type: article -summary: This article covers the importance of understanding linear assumptions in - machine learning models, which models assume linearity, and what steps can be taken - when the assumption is not met. +summary: This article covers the importance of understanding linear assumptions in machine learning models, which models assume linearity, and what steps can be taken when the assumption is not met. tags: - Linear models - Logistic regression diff --git a/_posts/2022-09-27-entropy_information_theory.md b/_posts/2022-09-27-entropy_information_theory.md index 6a9a8903..3de9b227 100644 --- a/_posts/2022-09-27-entropy_information_theory.md +++ b/_posts/2022-09-27-entropy_information_theory.md @@ -4,8 +4,7 @@ categories: - Information Theory classes: wide date: '2022-09-27' -excerpt: Explore entropy's role in thermodynamics, information theory, and quantum - mechanics, and its broader implications in physics and beyond. +excerpt: Explore entropy's role in thermodynamics, information theory, and quantum mechanics, and its broader implications in physics and beyond. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -21,15 +20,10 @@ keywords: - Quantum mechanics - Statistical mechanics - Maximum entropy principle -seo_description: An in-depth exploration of entropy in thermodynamics, statistical - mechanics, and information theory, from classical formulations to quantum mechanics - applications. +seo_description: An in-depth exploration of entropy in thermodynamics, statistical mechanics, and information theory, from classical formulations to quantum mechanics applications. seo_title: 'Entropy and Information Theory: A Comprehensive Analysis' seo_type: article -summary: This article provides an in-depth exploration of entropy, tracing its roots - from classical thermodynamics to its role in quantum mechanics and information theory. - It discusses entropy's applications across various fields, including physics, data - science, and cosmology. +summary: This article provides an in-depth exploration of entropy, tracing its roots from classical thermodynamics to its role in quantum mechanics and information theory. It discusses entropy's applications across various fields, including physics, data science, and cosmology. tags: - Entropy - Information theory diff --git a/_posts/2022-10-31-Jacknife.md b/_posts/2022-10-31-Jacknife.md index 9b8cb256..4deadda6 100644 --- a/_posts/2022-10-31-Jacknife.md +++ b/_posts/2022-10-31-Jacknife.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2022-10-31' -excerpt: Explore the jackknife technique, a robust resampling method used in statistics - for estimating bias, variance, and confidence intervals, with applications across - various fields. +excerpt: Explore the jackknife technique, a robust resampling method used in statistics for estimating bias, variance, and confidence intervals, with applications across various fields. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_3.jpg @@ -27,8 +25,7 @@ keywords: - Jackknife vs bootstrapping - Bias correction - Jackknife benefits -seo_description: Learn about the jackknife technique, a resampling method for estimating - bias and variance in statistical analysis, including its applications and benefits. +seo_description: Learn about the jackknife technique, a resampling method for estimating bias and variance in statistical analysis, including its applications and benefits. seo_title: 'The Jackknife Technique: Applications and Benefits in Statistical Analysis' seo_type: article tags: diff --git a/_posts/2022-11-30-Bootstrap.md b/_posts/2022-11-30-Bootstrap.md index 2d53bdb5..a46ae746 100644 --- a/_posts/2022-11-30-Bootstrap.md +++ b/_posts/2022-11-30-Bootstrap.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2022-11-30' -excerpt: Delve into bootstrapping, a versatile statistical technique for estimating - the sampling distribution of a statistic, offering insights into its applications - and implementation. +excerpt: Delve into bootstrapping, a versatile statistical technique for estimating the sampling distribution of a statistic, offering insights into its applications and implementation. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_7.jpg @@ -29,18 +27,17 @@ keywords: - Variance estimation - Python - Python -seo_description: Explore bootstrapping, a resampling method in statistics used to - estimate sampling distributions. Learn about its applications, implementation, and - limitations. +- python +seo_description: Explore bootstrapping, a resampling method in statistics used to estimate sampling distributions. Learn about its applications, implementation, and limitations. seo_title: 'Understanding Bootstrapping: A Resampling Method in Statistics' seo_type: article -summary: An overview of bootstrapping, its significance as a resampling method in - statistics, and how it is used to estimate the sampling distribution of a statistic. +summary: An overview of bootstrapping, its significance as a resampling method in statistics, and how it is used to estimate the sampling distribution of a statistic. tags: - Bootstrapping - Resampling - Python - Python +- python title: 'Understanding Bootstrapping: A Resampling Method in Statistics' --- diff --git a/_posts/2022-12-25-probability_machine_learning.md b/_posts/2022-12-25-probability_machine_learning.md index 6e75ebd5..883e7f4c 100644 --- a/_posts/2022-12-25-probability_machine_learning.md +++ b/_posts/2022-12-25-probability_machine_learning.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2022-12-25' -excerpt: Understand key probability distributions in machine learning and their applications, - including Bernoulli, Gaussian, and Beta distributions. +excerpt: Understand key probability distributions in machine learning and their applications, including Bernoulli, Gaussian, and Beta distributions. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_6.jpg @@ -26,8 +25,7 @@ keywords: - Data analysis with probability distributions - Distribution types in machine learning - Modeling uncertainty in ai -seo_description: An in-depth exploration of key probability distributions in machine - learning, including Bernoulli, Multinoulli, Gaussian, Exponential, and Beta distributions. +seo_description: An in-depth exploration of key probability distributions in machine learning, including Bernoulli, Multinoulli, Gaussian, Exponential, and Beta distributions. seo_title: Probability Distributions in Machine Learning seo_type: article tags: diff --git a/_posts/2022-12-30-simpsons_paradox.md b/_posts/2022-12-30-simpsons_paradox.md index 96316e93..8bdb4134 100644 --- a/_posts/2022-12-30-simpsons_paradox.md +++ b/_posts/2022-12-30-simpsons_paradox.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2022-12-30' -excerpt: Simpson's Paradox shows how aggregated data can lead to misleading trends. - Learn the theory behind this paradox, its practical implications, and how to analyze - data rigorously. +excerpt: Simpson's Paradox shows how aggregated data can lead to misleading trends. Learn the theory behind this paradox, its practical implications, and how to analyze data rigorously. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_6.jpg @@ -14,9 +12,7 @@ header: show_overlay_excerpt: false teaser: /assets/images/data_science_8.jpg twitter_image: /assets/images/data_science_6.jpg -seo_description: "Explore the theoretical foundations of Simpson\u2019s Paradox, its\ - \ role in data analysis, and how lurking variables and data aggregation lead to\ - \ contradictory statistical conclusions." +seo_description: Explore the theoretical foundations of Simpson’s Paradox, its role in data analysis, and how lurking variables and data aggregation lead to contradictory statistical conclusions. seo_title: 'Simpson''s Paradox: Theory, Lurking Variables, and Data Aggregation' seo_type: article tags: @@ -25,7 +21,7 @@ tags: - Data aggregation - Statistical paradoxes - Data visualization -title: "Simpson\u2019s Paradox: Theoretical Foundations and Implications in Data Analysis" +title: 'Simpson’s Paradox: Theoretical Foundations and Implications in Data Analysis' --- Simpson’s Paradox is a fascinating statistical phenomenon where the relationship between two variables can drastically change when a third variable is introduced. This paradox is widely misunderstood and can lead to erroneous conclusions if data is not analyzed carefully. It reveals the complexities of data aggregation and emphasizes the necessity of considering lurking variables to avoid false interpretations. diff --git a/_posts/2022-12-31-PCA_explained.md b/_posts/2022-12-31-PCA_explained.md index e9774892..428b97c1 100644 --- a/_posts/2022-12-31-PCA_explained.md +++ b/_posts/2022-12-31-PCA_explained.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2022-12-31' -excerpt: Learn about Principal Component Analysis (PCA) and how it helps in feature - extraction, dimensionality reduction, and identifying key patterns in data. +excerpt: Learn about Principal Component Analysis (PCA) and how it helps in feature extraction, dimensionality reduction, and identifying key patterns in data. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_4.jpg @@ -28,21 +27,16 @@ keywords: - Pattern recognition - Data compression - Python -seo_description: A comprehensive guide to Principal Component Analysis (PCA), covering - feature selection, dimension reduction, explained variance, and outlier detection. +- python +seo_description: A comprehensive guide to Principal Component Analysis (PCA), covering feature selection, dimension reduction, explained variance, and outlier detection. seo_title: Principal Component Analysis (PCA) Guide seo_type: article -summary: Principal Component Analysis (PCA) is a powerful technique in data science, - used for reducing the dimensionality of large datasets while preserving essential - patterns. This article offers a step-by-step guide to understanding PCA, from its - core mathematical concepts to practical applications in feature extraction, outlier - detection, and multivariate data analysis. Whether you're using PCA for data compression - or to improve machine learning models, this guide will help you grasp its key principles, - including how to interpret explained variance and identify significant components. +summary: Principal Component Analysis (PCA) is a powerful technique in data science, used for reducing the dimensionality of large datasets while preserving essential patterns. This article offers a step-by-step guide to understanding PCA, from its core mathematical concepts to practical applications in feature extraction, outlier detection, and multivariate data analysis. Whether you're using PCA for data compression or to improve machine learning models, this guide will help you grasp its key principles, including how to interpret explained variance and identify significant components. tags: - Pca - Dimensionality reduction - Python +- python title: 'Understanding PCA: A Step-by-Step Guide to Principal Component Analysis' --- diff --git a/_posts/2023-01-01-error_coefficientes.md b/_posts/2023-01-01-error_coefficientes.md index f3c7a712..31fb55b2 100644 --- a/_posts/2023-01-01-error_coefficientes.md +++ b/_posts/2023-01-01-error_coefficientes.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2023-01-01' -excerpt: Delve into how multiple linear regression and binary logistic regression - handle errors. Learn about explicit and implicit error terms and their impact on - model performance. +excerpt: Delve into how multiple linear regression and binary logistic regression handle errors. Learn about explicit and implicit error terms and their impact on model performance. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_2.jpg @@ -28,14 +26,10 @@ keywords: - Error analysis in statistics - Predictive model accuracy - Linear vs logistic regression errors -seo_description: Explore the differences in error handling between multiple linear - regression and binary logistic regression. Understand the explicit and implicit - roles of error terms in these statistical models. +seo_description: Explore the differences in error handling between multiple linear regression and binary logistic regression. Understand the explicit and implicit roles of error terms in these statistical models. seo_title: 'Error Terms in Regression Models: Linear vs. Logistic Regression' seo_type: article -summary: This article explores how error terms are handled in both multiple linear - regression and binary logistic regression, emphasizing their roles in statistical - model performance and accuracy. +summary: This article explores how error terms are handled in both multiple linear regression and binary logistic regression, emphasizing their roles in statistical model performance and accuracy. tags: - Regression models - Error terms diff --git a/_posts/2023-01-08-crownd_behaviour.md b/_posts/2023-01-08-crownd_behaviour.md index c5059b7c..be8b1461 100644 --- a/_posts/2023-01-08-crownd_behaviour.md +++ b/_posts/2023-01-08-crownd_behaviour.md @@ -4,9 +4,7 @@ categories: - Mathematics classes: wide date: '2023-01-08' -excerpt: Dive into the fascinating world of pedestrian behavior through mathematical - models like the Social Force Model. Learn how these models inform urban planning, - crowd management, and traffic control for safer and more efficient public spaces. +excerpt: Dive into the fascinating world of pedestrian behavior through mathematical models like the Social Force Model. Learn how these models inform urban planning, crowd management, and traffic control for safer and more efficient public spaces. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_6.jpg @@ -25,11 +23,8 @@ keywords: - Fluid dynamics in traffic - Public space safety - Transportation systems -seo_description: Explore the mathematical modeling of pedestrian behavior, focusing - on the Social Force Model, statistical methods, and fluid dynamics to enhance urban - planning, crowd management, and traffic control. -seo_title: 'Mathematical Models of Pedestrian Behavior: Insights into Urban Planning - and Crowd Management' +seo_description: Explore the mathematical modeling of pedestrian behavior, focusing on the Social Force Model, statistical methods, and fluid dynamics to enhance urban planning, crowd management, and traffic control. +seo_title: 'Mathematical Models of Pedestrian Behavior: Insights into Urban Planning and Crowd Management' seo_type: article subtitle: Understanding Pedestrian Behavior through Mathematical Models tags: diff --git a/_posts/2023-02-17-ab_testing.md b/_posts/2023-02-17-ab_testing.md index 0219f2d0..24ad2ff7 100644 --- a/_posts/2023-02-17-ab_testing.md +++ b/_posts/2023-02-17-ab_testing.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2023-02-17' -excerpt: An in-depth exploration of sequential testing and its application in A/B - testing. Understand the statistical underpinnings, advantages, limitations, and - practical implementations in R, JavaScript, and Python. +excerpt: An in-depth exploration of sequential testing and its application in A/B testing. Understand the statistical underpinnings, advantages, limitations, and practical implementations in R, JavaScript, and Python. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_3.jpg @@ -32,9 +30,10 @@ keywords: - R - Javascript - Python -seo_description: Explore advanced statistical concepts behind sequential testing in - A/B testing. Learn about SPRT, error control, practical implementation, and potential - pitfalls. +- r +- javascript +- python +seo_description: Explore advanced statistical concepts behind sequential testing in A/B testing. Learn about SPRT, error control, practical implementation, and potential pitfalls. seo_title: 'In-Depth Sequential Testing in A/B Testing: Advanced Statistical Methods' seo_type: article tags: @@ -47,6 +46,9 @@ tags: - R - Javascript - Python +- r +- javascript +- python title: Advanced Statistical Methods for Efficient A/B Testing --- diff --git a/_posts/2023-05-05-Mean_Time_Between_Failures.md b/_posts/2023-05-05-Mean_Time_Between_Failures.md index e4845830..5092034d 100644 --- a/_posts/2023-05-05-Mean_Time_Between_Failures.md +++ b/_posts/2023-05-05-Mean_Time_Between_Failures.md @@ -4,8 +4,7 @@ categories: - Predictive Maintenance classes: wide date: '2023-05-05' -excerpt: Explore the key concepts of Mean Time Between Failures (MTBF), how it is - calculated, its applications, and its alternatives in system reliability. +excerpt: Explore the key concepts of Mean Time Between Failures (MTBF), how it is calculated, its applications, and its alternatives in system reliability. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_7.jpg @@ -20,17 +19,17 @@ keywords: - System maintenance - Predictive maintenance - Python -seo_description: An in-depth explanation of Mean Time Between Failures (MTBF), its - importance, strengths, weaknesses, and related metrics like MTTR and MTTF. +- python +seo_description: An in-depth explanation of Mean Time Between Failures (MTBF), its importance, strengths, weaknesses, and related metrics like MTTR and MTTF. seo_title: What is Mean Time Between Failures (MTBF)? seo_type: article -summary: A comprehensive guide on Mean Time Between Failures (MTBF), covering its - calculation, use cases, strengths, and weaknesses in reliability engineering. +summary: A comprehensive guide on Mean Time Between Failures (MTBF), covering its calculation, use cases, strengths, and weaknesses in reliability engineering. tags: - Mtbf - Reliability metrics - Predictive maintenance - Python +- python title: Understanding Mean Time Between Failures (MTBF) --- diff --git a/_posts/2023-07-23-VAR.md b/_posts/2023-07-23-VAR.md index 53a109d0..6a7e3320 100644 --- a/_posts/2023-07-23-VAR.md +++ b/_posts/2023-07-23-VAR.md @@ -4,8 +4,7 @@ categories: - Finance classes: wide date: '2023-07-23' -excerpt: A detailed exploration of Value at Risk (VaR), covering its different types, - methods of calculation, and applications in modern portfolio management. +excerpt: A detailed exploration of Value at Risk (VaR), covering its different types, methods of calculation, and applications in modern portfolio management. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_3.jpg @@ -28,9 +27,7 @@ keywords: - Var in portfolio management - Var types - Financial risk metrics -seo_description: Explore the key concepts, types, and applications of Value at Risk - (VaR) in portfolio management, including Parametric VaR, Historical VaR, and Monte - Carlo VaR. +seo_description: Explore the key concepts, types, and applications of Value at Risk (VaR) in portfolio management, including Parametric VaR, Historical VaR, and Monte Carlo VaR. seo_title: Comprehensive Guide to Value at Risk (VaR) and Its Types seo_type: article tags: diff --git a/_posts/2023-07-26-customer-life-time-value.md b/_posts/2023-07-26-customer-life-time-value.md index 471346bb..b0922bfa 100644 --- a/_posts/2023-07-26-customer-life-time-value.md +++ b/_posts/2023-07-26-customer-life-time-value.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2023-07-26' -excerpt: A detailed exploration of Customer Lifetime Value (CLV) for data practitioners - and marketers, including its calculation, prediction, and integration with other - business data. +excerpt: A detailed exploration of Customer Lifetime Value (CLV) for data practitioners and marketers, including its calculation, prediction, and integration with other business data. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_7.jpg @@ -31,24 +29,19 @@ keywords: - Clv metrics - Python - Python -seo_description: Explore an in-depth guide to Customer Lifetime Value (CLV), covering - calculation, prediction, integration with business data, and its role in data-driven - marketing strategies. -seo_title: 'Customer Lifetime Value (CLV): A Comprehensive Guide for Data Science - and Marketing' +- python +seo_description: Explore an in-depth guide to Customer Lifetime Value (CLV), covering calculation, prediction, integration with business data, and its role in data-driven marketing strategies. +seo_title: 'Customer Lifetime Value (CLV): A Comprehensive Guide for Data Science and Marketing' seo_type: article -summary: This article provides a comprehensive exploration of Customer Lifetime Value - (CLV), detailing its calculation methods, predictive models, and its importance - in data-driven marketing strategies. It also covers how CLV can be integrated with - other business data to optimize customer retention and enhance profitability. +summary: This article provides a comprehensive exploration of Customer Lifetime Value (CLV), detailing its calculation methods, predictive models, and its importance in data-driven marketing strategies. It also covers how CLV can be integrated with other business data to optimize customer retention and enhance profitability. tags: - Clv - Predictive analytics - Marketing strategy - Python - Python -title: 'Customer Lifetime Value: An In-Depth Exploration for Data Practitioners and - Marketers' +- python +title: 'Customer Lifetime Value: An In-Depth Exploration for Data Practitioners and Marketers' --- ![Customer Lifetime Value](https://unsplash.com/photos/BJaqPaH6AGQ) diff --git a/_posts/2023-08-12-guassian_processes.md b/_posts/2023-08-12-guassian_processes.md index 6135029b..4473ff02 100644 --- a/_posts/2023-08-12-guassian_processes.md +++ b/_posts/2023-08-12-guassian_processes.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2023-08-12' -excerpt: Dive into Gaussian Processes for time-series analysis using Python, combining - flexible modeling with Bayesian inference for trends, seasonality, and noise. +excerpt: Dive into Gaussian Processes for time-series analysis using Python, combining flexible modeling with Bayesian inference for trends, seasonality, and noise. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_3.jpg @@ -15,8 +14,8 @@ header: twitter_image: /assets/images/data_science_3.jpg keywords: - Python -seo_description: Explore Gaussian Processes and their application in time-series analysis. - Learn the theory, mathematical background, and practical implementations in Python. +- python +seo_description: Explore Gaussian Processes and their application in time-series analysis. Learn the theory, mathematical background, and practical implementations in Python. seo_title: 'Gaussian Processes for Time Series: A Deep Dive in Python' seo_type: article tags: @@ -25,6 +24,7 @@ tags: - Bayesian inference - Python - Python +- python title: Gaussian Processes for Time-Series Analysis in Python --- diff --git a/_posts/2023-08-13-shared_nearest_neighbors.md b/_posts/2023-08-13-shared_nearest_neighbors.md index 3c6020f7..2c637842 100644 --- a/_posts/2023-08-13-shared_nearest_neighbors.md +++ b/_posts/2023-08-13-shared_nearest_neighbors.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2023-08-13' -excerpt: SNN is a distance metric that enhances traditional methods like k Nearest - Neighbors, especially in high-dimensional, variable-density datasets. +excerpt: SNN is a distance metric that enhances traditional methods like k Nearest Neighbors, especially in high-dimensional, variable-density datasets. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_9.jpg @@ -24,15 +23,11 @@ keywords: - Machine learning - Python - Python -seo_description: An exploration of Shared Nearest Neighbors (SNN) as a distance metric, - and its application in outlier detection, clustering, and density-based algorithms. +- python +seo_description: An exploration of Shared Nearest Neighbors (SNN) as a distance metric, and its application in outlier detection, clustering, and density-based algorithms. seo_title: Shared Nearest Neighbors in Outlier Detection seo_type: article -summary: Shared Nearest Neighbors (SNN) is a distance metric designed to enhance outlier - detection, clustering, and predictive modeling in datasets with high dimensionality - and varying density. This article explores how SNN mitigates the weaknesses of traditional - metrics like Euclidean and Manhattan, providing robust performance in complex data - scenarios. +summary: Shared Nearest Neighbors (SNN) is a distance metric designed to enhance outlier detection, clustering, and predictive modeling in datasets with high dimensionality and varying density. This article explores how SNN mitigates the weaknesses of traditional metrics like Euclidean and Manhattan, providing robust performance in complex data scenarios. tags: - Machine learning - Outlier detection @@ -42,6 +37,7 @@ tags: - K-nearest neighbors - Python - Python +- python title: Exploring Shared Nearest Neighbors (SNN) for Outlier Detection --- diff --git a/_posts/2023-08-21-demystifying_data_science.md b/_posts/2023-08-21-demystifying_data_science.md index e63363bf..fd47977a 100644 --- a/_posts/2023-08-21-demystifying_data_science.md +++ b/_posts/2023-08-21-demystifying_data_science.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2023-08-21' -excerpt: Discover how data science, a multidisciplinary field combining statistics, - computer science, and domain expertise, can drive better business decisions and - outcomes. +excerpt: Discover how data science, a multidisciplinary field combining statistics, computer science, and domain expertise, can drive better business decisions and outcomes. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_8.jpg @@ -30,16 +28,11 @@ keywords: - Ai in business - Data science for revenue growth - Data science trends in business -seo_description: Learn what data science is and how it can transform your business - through improved decision-making, cost savings, and increased revenue. +seo_description: Learn what data science is and how it can transform your business through improved decision-making, cost savings, and increased revenue. seo_title: 'Demystifying Data Science: A Guide to Its Benefits for Business' seo_type: article subtitle: What It Is and How It Can Help Your Business -summary: This article explores the role of data science in business, highlighting - its potential to enhance decision-making, optimize operations, and drive revenue - growth. It delves into key applications such as customer behavior analysis, supply - chain optimization, and predictive analytics, showcasing how companies can leverage - data science for competitive advantage. +summary: This article explores the role of data science in business, highlighting its potential to enhance decision-making, optimize operations, and drive revenue growth. It delves into key applications such as customer behavior analysis, supply chain optimization, and predictive analytics, showcasing how companies can leverage data science for competitive advantage. tags: - Data science - Business intelligence diff --git a/_posts/2023-08-21-large_languague_models.md b/_posts/2023-08-21-large_languague_models.md index 81b1a53f..e513e9d7 100644 --- a/_posts/2023-08-21-large_languague_models.md +++ b/_posts/2023-08-21-large_languague_models.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2023-08-21' -excerpt: An in-depth exploration of how the closure of open-source data platforms - threatens the growth of Large Language Models and the vital role humans play in - this ecosystem. +excerpt: An in-depth exploration of how the closure of open-source data platforms threatens the growth of Large Language Models and the vital role humans play in this ecosystem. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_5.jpg @@ -25,24 +23,17 @@ keywords: - Ethical ai development - Open data impact on ai - Future of machine learning -seo_description: Explore the vulnerability of Large Language Models like GPT when - open-source data platforms such as Stack Overflow close, and the potential impact - on AI's evolution. +seo_description: Explore the vulnerability of Large Language Models like GPT when open-source data platforms such as Stack Overflow close, and the potential impact on AI's evolution. seo_title: The Fragility of Large Language Models in a World Without Open-Source Data seo_type: article subtitle: Exploring the Fragility and Future of Machine Learning Without Open Data -summary: An exploration into the challenges faced by Large Language Models (LLMs) - like GPT in the absence of open-source data platforms. The article discusses the - consequences of platforms like Stack Overflow closing, the fragility of AI systems - dependent on these data sources, and the broader implications for ethical AI development - and the future of machine learning. +summary: An exploration into the challenges faced by Large Language Models (LLMs) like GPT in the absence of open-source data platforms. The article discusses the consequences of platforms like Stack Overflow closing, the fragility of AI systems dependent on these data sources, and the broader implications for ethical AI development and the future of machine learning. tags: - Llm - Open-source data - Machine learning models - Ai ethics -title: The Vulnerability of Large Language Models to the Closure of Open-Source Data - Platforms +title: The Vulnerability of Large Language Models to the Closure of Open-Source Data Platforms --- ![Example Image](/assets/images/stackoverflow.jpg) diff --git a/_posts/2023-08-22-Paul-Erdos.md b/_posts/2023-08-22-Paul-Erdos.md index 5077ffde..acaaac47 100644 --- a/_posts/2023-08-22-Paul-Erdos.md +++ b/_posts/2023-08-22-Paul-Erdos.md @@ -5,8 +5,7 @@ categories: - Biographies classes: wide date: '2023-08-22' -excerpt: "Delve into the fascinating life of Paul Erd\u0151s, a wandering mathematician\ - \ whose love for numbers and collaboration reshaped the world of mathematics." +excerpt: Delve into the fascinating life of Paul Erdős, a wandering mathematician whose love for numbers and collaboration reshaped the world of mathematics. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_3.jpg @@ -15,27 +14,26 @@ header: teaser: /assets/images/data_science_9.jpg twitter_image: /assets/images/data_science_3.jpg keywords: -- "Paul erd\u0151s biography" +- Paul erdős biography - Number theory contributions - Mathematical collaboration -- "Erd\u0151s number" +- Erdős number - Combinatorics - Graph theory - Hungarian mathematicians - Mathematical prodigies - Collaborative mathematics - Famous mathematicians -seo_description: "Explore the life and legacy of Paul Erd\u0151s, a nomadic mathematician\ - \ who made groundbreaking contributions to number theory and collaborative science." -seo_title: "Paul Erd\u0151s: The Mathematical Prodigy Who Changed Mathematics Forever" +seo_description: Explore the life and legacy of Paul Erdős, a nomadic mathematician who made groundbreaking contributions to number theory and collaborative science. +seo_title: 'Paul Erdős: The Mathematical Prodigy Who Changed Mathematics Forever' seo_type: article subtitle: A Mathematician for the Ages tags: -- "Paul erd\u0151s" +- Paul erdős - Mathematical genius - Number theory - Collaboration in science -title: "The Life and Legacy of Paul Erd\u0151s" +title: The Life and Legacy of Paul Erdős --- ![Example Image](/assets/images/Erdos_Paul.jpg) diff --git a/_posts/2023-08-23-multivariate_analysis_variance_vs_anova.md b/_posts/2023-08-23-multivariate_analysis_variance_vs_anova.md index ac9accb0..84b72695 100644 --- a/_posts/2023-08-23-multivariate_analysis_variance_vs_anova.md +++ b/_posts/2023-08-23-multivariate_analysis_variance_vs_anova.md @@ -4,8 +4,7 @@ categories: - Multivariate Analysis classes: wide date: '2023-08-23' -excerpt: Learn the key differences between MANOVA and ANOVA, and when to apply them - in experimental designs with multiple dependent variables, such as clinical trials. +excerpt: Learn the key differences between MANOVA and ANOVA, and when to apply them in experimental designs with multiple dependent variables, such as clinical trials. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_8.jpg @@ -19,24 +18,17 @@ keywords: - Experimental design - Clinical trials - Multivariate analysis -seo_description: A detailed exploration of the differences between MANOVA and ANOVA, - and when to use them in experimental designs, such as in clinical trials with multiple - outcome variables. +seo_description: A detailed exploration of the differences between MANOVA and ANOVA, and when to use them in experimental designs, such as in clinical trials with multiple outcome variables. seo_title: 'MANOVA vs. ANOVA: Differences and Use Cases in Experimental Design' seo_type: article -summary: Multivariate Analysis of Variance (MANOVA) and Analysis of Variance (ANOVA) - are statistical methods used to analyze group differences. While ANOVA focuses on - a single dependent variable, MANOVA extends this to multiple dependent variables. - This article explores their differences and application in experimental designs - like clinical trials. +summary: Multivariate Analysis of Variance (MANOVA) and Analysis of Variance (ANOVA) are statistical methods used to analyze group differences. While ANOVA focuses on a single dependent variable, MANOVA extends this to multiple dependent variables. This article explores their differences and application in experimental designs like clinical trials. tags: - Manova - Anova - Multivariate statistics - Experimental design - Clinical trials -title: 'Multivariate Analysis of Variance (MANOVA) vs. ANOVA: When to Analyze Multiple - Dependent Variables' +title: 'Multivariate Analysis of Variance (MANOVA) vs. ANOVA: When to Analyze Multiple Dependent Variables' --- In the world of experimental design and statistical analysis, **Analysis of Variance (ANOVA)** and **Multivariate Analysis of Variance (MANOVA)** are essential tools for comparing groups and determining whether differences exist between them. While ANOVA is designed to analyze a single dependent variable across groups, MANOVA extends this capability to multiple dependent variables, making it particularly useful in complex experimental designs. Understanding when to use ANOVA versus MANOVA can significantly impact the robustness and interpretability of statistical results, especially in fields like psychology, clinical trials, and educational research, where multiple outcomes are common. diff --git a/_posts/2023-08-25-runnning_windows.md b/_posts/2023-08-25-runnning_windows.md index 1e43b83f..29850876 100644 --- a/_posts/2023-08-25-runnning_windows.md +++ b/_posts/2023-08-25-runnning_windows.md @@ -4,8 +4,7 @@ categories: - R Programming classes: wide date: '2023-08-25' -excerpt: Explore the `runner` package in R, which allows applying any R function to - rolling windows of data with full control over window size, lags, and index types. +excerpt: Explore the `runner` package in R, which allows applying any R function to rolling windows of data with full control over window size, lags, and index types. header: image: /assets/images/Rolling-window.jpg og_image: /assets/images/data_science_4.jpg @@ -24,20 +23,18 @@ keywords: - Dplyr runner integration - Rolling regression r - R -seo_description: Learn how to use the `runner` package in R to apply any function - on rolling windows of data. Supports custom window sizes, lags, and flexible indexing - using dates, ideal for time series analysis. +- r +seo_description: Learn how to use the `runner` package in R to apply any function on rolling windows of data. Supports custom window sizes, lags, and flexible indexing using dates, ideal for time series analysis. seo_title: Apply Any R Function on Rolling Windows with the `runner` Package seo_type: article -summary: This article explores the `runner` package in R, detailing how to apply functions - to rolling windows of data with custom window sizes, lags, and indexing, particularly - useful for time series and cumulative operations. +summary: This article explores the `runner` package in R, detailing how to apply functions to rolling windows of data with custom window sizes, lags, and indexing, particularly useful for time series and cumulative operations. tags: - Rolling windows - Time series analysis - Data manipulation - Statistical modeling - R +- r title: Applying R Functions on Rolling Windows Using the `runner` Package --- diff --git a/_posts/2023-08-30-Data_Science.md b/_posts/2023-08-30-Data_Science.md index 547cbde9..ab97bf2b 100644 --- a/_posts/2023-08-30-Data_Science.md +++ b/_posts/2023-08-30-Data_Science.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2023-08-30' -excerpt: A deep dive into the ethical challenges of data science, covering privacy, - bias, social impact, and the need for responsible AI decision-making. +excerpt: A deep dive into the ethical challenges of data science, covering privacy, bias, social impact, and the need for responsible AI decision-making. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_7.jpg @@ -25,9 +24,7 @@ keywords: - Fairness in machine learning - Algorithmic bias - Ethical challenges in ai -seo_description: Explore the ethical challenges in data science, including privacy - protection, bias, social impact, and responsible decision-making. A comprehensive - guide for ethical AI. +seo_description: Explore the ethical challenges in data science, including privacy protection, bias, social impact, and responsible decision-making. A comprehensive guide for ethical AI. seo_title: 'Ethics in Data Science: Privacy, Bias, Social Impact & Responsible AI' seo_type: article subtitle: A Comprehensive Guide to Privacy, Bias, Social Impact and Responsible Decision-Making diff --git a/_posts/2023-09-01-regression_path_analysis.md b/_posts/2023-09-01-regression_path_analysis.md index d2e158c0..f573335a 100644 --- a/_posts/2023-09-01-regression_path_analysis.md +++ b/_posts/2023-09-01-regression_path_analysis.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2023-09-01' -excerpt: Regression and path analysis are two statistical techniques used to model - relationships between variables. This article explains their differences, highlighting - key features and use cases for each. +excerpt: Regression and path analysis are two statistical techniques used to model relationships between variables. This article explains their differences, highlighting key features and use cases for each. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_5.jpg @@ -20,17 +18,10 @@ keywords: - Statistical modeling - Structural equation models - Multivariate analysis -seo_description: Explore the key differences between regression analysis and path - analysis, two important techniques in statistical modeling. Understand their applications, - advantages, and limitations. +seo_description: Explore the key differences between regression analysis and path analysis, two important techniques in statistical modeling. Understand their applications, advantages, and limitations. seo_title: 'Regression vs. Path Analysis: A Comprehensive Comparison' seo_type: article -summary: Regression and path analysis are both important statistical methods, but - they differ in terms of their complexity, scope, and purpose. While regression focuses - on predicting dependent variables from independent variables, path analysis allows - for the modeling of more complex, multivariate relationships between variables. - This comprehensive article delves into the theoretical and practical distinctions - between these two methods. +summary: Regression and path analysis are both important statistical methods, but they differ in terms of their complexity, scope, and purpose. While regression focuses on predicting dependent variables from independent variables, path analysis allows for the modeling of more complex, multivariate relationships between variables. This comprehensive article delves into the theoretical and practical distinctions between these two methods. tags: - Regression analysis - Path analysis diff --git a/_posts/2023-09-03-binary_classification.md b/_posts/2023-09-03-binary_classification.md index ee799f3a..bf19ce27 100644 --- a/_posts/2023-09-03-binary_classification.md +++ b/_posts/2023-09-03-binary_classification.md @@ -5,9 +5,7 @@ categories: - Data Science classes: wide date: '2023-09-03' -excerpt: Learn the core concepts of binary classification, explore common algorithms - like Decision Trees and SVMs, and discover how to evaluate performance using precision, - recall, and F1-score. +excerpt: Learn the core concepts of binary classification, explore common algorithms like Decision Trees and SVMs, and discover how to evaluate performance using precision, recall, and F1-score. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_8.jpg @@ -26,9 +24,7 @@ keywords: - Model evaluation metrics - Classification problems - Machine learning applications -seo_description: Explore the fundamentals of binary classification in machine learning, - including key algorithms, evaluation metrics like precision and recall, and real-world - applications. +seo_description: Explore the fundamentals of binary classification in machine learning, including key algorithms, evaluation metrics like precision and recall, and real-world applications. seo_title: 'Binary Classification in Machine Learning: Methods, Metrics, and Applications' seo_type: article tags: diff --git a/_posts/2023-09-04-Fears-Surrounding.md b/_posts/2023-09-04-Fears-Surrounding.md index 3b6ddfef..121b0866 100644 --- a/_posts/2023-09-04-Fears-Surrounding.md +++ b/_posts/2023-09-04-Fears-Surrounding.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2023-09-04' -excerpt: Delve into the fears and complexities of artificial intelligence and automation, - addressing concerns like job displacement, data privacy, ethical decision-making, - and the true capabilities and limitations of AI. +excerpt: Delve into the fears and complexities of artificial intelligence and automation, addressing concerns like job displacement, data privacy, ethical decision-making, and the true capabilities and limitations of AI. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_7.jpg @@ -25,9 +23,7 @@ keywords: - Ethical dilemmas in ai - Ai in automation - Future of ai -seo_description: Explore the fears and challenges surrounding artificial intelligence, - including job displacement, data privacy, ethical dilemmas, and the limitations - of AI and machine learning. +seo_description: Explore the fears and challenges surrounding artificial intelligence, including job displacement, data privacy, ethical dilemmas, and the limitations of AI and machine learning. seo_title: The Fears and Challenges of Artificial Intelligence and Automation seo_type: article subtitle: Automation and Machine Learning diff --git a/_posts/2023-09-08-trafic_dynamics.md b/_posts/2023-09-08-trafic_dynamics.md index facd406f..461ce8c0 100644 --- a/_posts/2023-09-08-trafic_dynamics.md +++ b/_posts/2023-09-08-trafic_dynamics.md @@ -4,9 +4,7 @@ categories: - Science and Engineering classes: wide date: '2023-09-08' -excerpt: This article explores the complex interplay between traffic control, pedestrian - movement, and the application of fluid dynamics to model and manage these phenomena - in urban environments. +excerpt: This article explores the complex interplay between traffic control, pedestrian movement, and the application of fluid dynamics to model and manage these phenomena in urban environments. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -21,8 +19,7 @@ keywords: - Intelligent traffic systems - Mathematical models in traffic flow - Crowd management -seo_description: An in-depth analysis of how traffic control systems and pedestrian - dynamics can be modeled using principles of fluid dynamics. +seo_description: An in-depth analysis of how traffic control systems and pedestrian dynamics can be modeled using principles of fluid dynamics. seo_title: Traffic Control, Pedestrian Dynamics, and Fluid Dynamics seo_type: article tags: @@ -30,8 +27,7 @@ tags: - Pedestrian dynamics - Fluid dynamics - Urban planning -title: Exploring the Dynamics of Traffic Control and Pedestrian Behavior Through the - Lens of Fluid Dynamics +title: Exploring the Dynamics of Traffic Control and Pedestrian Behavior Through the Lens of Fluid Dynamics --- ## Overview diff --git a/_posts/2023-09-20-rolling_windows.md b/_posts/2023-09-20-rolling_windows.md index 21936137..b991b815 100644 --- a/_posts/2023-09-20-rolling_windows.md +++ b/_posts/2023-09-20-rolling_windows.md @@ -5,8 +5,7 @@ categories: - Data Analysis classes: wide date: '2023-09-20' -excerpt: Explore the diverse applications of rolling windows in signal processing, - covering both the underlying theory and practical implementations. +excerpt: Explore the diverse applications of rolling windows in signal processing, covering both the underlying theory and practical implementations. header: image: /assets/images/download.png og_image: /assets/images/data_science_8.jpg @@ -26,8 +25,8 @@ keywords: - Filtering techniques - Data smoothing - Python -seo_description: Learn how rolling windows can be applied in signal processing for - smoothing, feature extraction, and time-frequency analysis. +- python +seo_description: Learn how rolling windows can be applied in signal processing for smoothing, feature extraction, and time-frequency analysis. seo_title: Unlock the Power of Rolling Windows in Signal Processing seo_type: article social_image: /assets/images/rollingwindow.png @@ -37,6 +36,7 @@ tags: - Signal smoothing - Time-frequency analysis - Python +- python title: Rolling Windows in Signal Processing --- diff --git a/_posts/2023-09-26-Innumeracy.md b/_posts/2023-09-26-Innumeracy.md index c0bb0530..0d34de3a 100644 --- a/_posts/2023-09-26-Innumeracy.md +++ b/_posts/2023-09-26-Innumeracy.md @@ -4,9 +4,7 @@ categories: - Mathematics classes: wide date: '2023-09-26' -excerpt: Innumeracy is becoming the new illiteracy, with far-reaching implications - for decision-making in various aspects of life. Discover how the inability to understand - numbers affects our world and what can be done to address this growing issue. +excerpt: Innumeracy is becoming the new illiteracy, with far-reaching implications for decision-making in various aspects of life. Discover how the inability to understand numbers affects our world and what can be done to address this growing issue. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_5.jpg @@ -25,16 +23,14 @@ keywords: - Cognitive bias - Public policy - Critical thinking -seo_description: "Explore the growing issue of innumeracy\u2014our inability to understand\ - \ and work with numbers. Learn how this new illiteracy impacts decision-making in\ - \ society, from corporate boardrooms to public policy." +seo_description: Explore the growing issue of innumeracy—our inability to understand and work with numbers. Learn how this new illiteracy impacts decision-making in society, from corporate boardrooms to public policy. seo_title: 'Innumeracy: The New Illiteracy Crippling Decision-Making' seo_type: article tags: - Numeracy - Data literacy - Decision making -title: "The New Illiteracy That\u2019s Crippling Our Decision-Making" +title: The New Illiteracy That’s Crippling Our Decision-Making --- ![Example Image](/assets/images/inumeracy.jpg) diff --git a/_posts/2023-09-27-Data_communication.md b/_posts/2023-09-27-Data_communication.md index b49cfd74..6911777e 100644 --- a/_posts/2023-09-27-Data_communication.md +++ b/_posts/2023-09-27-Data_communication.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2023-09-27' -excerpt: Data and communication are intricately linked in modern business. This article - explores how to balance data analysis with storytelling, ensuring clear and actionable - insights. +excerpt: Data and communication are intricately linked in modern business. This article explores how to balance data analysis with storytelling, ensuring clear and actionable insights. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_1.jpg @@ -25,9 +23,7 @@ keywords: - Data sampling - Effect size - Research methodology -seo_description: Explore the crucial role of communication in data-driven environments, - examining how to balance data analysis with effective storytelling and context to - drive actionable insights. +seo_description: Explore the crucial role of communication in data-driven environments, examining how to balance data analysis with effective storytelling and context to drive actionable insights. seo_title: 'Data and Communication: Orchestrating a Harmonious Future' seo_type: article tags: diff --git a/_posts/2023-09-27-sample_size.md b/_posts/2023-09-27-sample_size.md index dad67b2b..dbbfb5e6 100644 --- a/_posts/2023-09-27-sample_size.md +++ b/_posts/2023-09-27-sample_size.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2023-09-27' -excerpt: Dive into the nuances of sample size in statistical analysis, challenging - the common belief that larger samples always lead to better results. +excerpt: Dive into the nuances of sample size in statistical analysis, challenging the common belief that larger samples always lead to better results. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_5.jpg @@ -24,9 +23,7 @@ keywords: - Data sampling - Effect size - Research methodology -seo_description: Explore the complexities of sample size in statistical analysis. - Learn why bigger isn't always better, and the importance of data quality and experimental - design. +seo_description: Explore the complexities of sample size in statistical analysis. Learn why bigger isn't always better, and the importance of data quality and experimental design. seo_title: The Myth and Reality of Sample Size in Statistical Analysis seo_type: article subtitle: A Nuanced Perspective diff --git a/_posts/2023-09-30-multiple_regression_vs_stepwise_regression.md b/_posts/2023-09-30-multiple_regression_vs_stepwise_regression.md index 3ada4236..24f98058 100644 --- a/_posts/2023-09-30-multiple_regression_vs_stepwise_regression.md +++ b/_posts/2023-09-30-multiple_regression_vs_stepwise_regression.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2023-09-30' -excerpt: Learn the differences between multiple regression and stepwise regression, - and discover when to use each method to build the best predictive models in business - analytics and scientific research. +excerpt: Learn the differences between multiple regression and stepwise regression, and discover when to use each method to build the best predictive models in business analytics and scientific research. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -24,16 +22,12 @@ keywords: - Python - Bash - Python -seo_description: A detailed comparison between multiple regression and stepwise regression, - with insights on when to use each for predictive modeling in business analytics - and scientific research. -seo_title: 'Multiple Regression vs. Stepwise Regression: Choosing the Best Predictive - Model' +- bash +- python +seo_description: A detailed comparison between multiple regression and stepwise regression, with insights on when to use each for predictive modeling in business analytics and scientific research. +seo_title: 'Multiple Regression vs. Stepwise Regression: Choosing the Best Predictive Model' seo_type: article -summary: Multiple regression and stepwise regression are powerful tools for predictive - modeling. This article explains their differences, strengths, and appropriate applications - in fields like business analytics and scientific research, helping you build effective - models. +summary: Multiple regression and stepwise regression are powerful tools for predictive modeling. This article explains their differences, strengths, and appropriate applications in fields like business analytics and scientific research, helping you build effective models. tags: - Multiple regression - Stepwise regression @@ -44,8 +38,9 @@ tags: - Python - Bash - Python -title: 'Multiple Regression vs. Stepwise Regression: Building the Best Predictive - Models' +- bash +- python +title: 'Multiple Regression vs. Stepwise Regression: Building the Best Predictive Models' --- Predictive modeling is at the heart of modern data analysis, helping researchers and analysts forecast outcomes based on a variety of input variables. Two widely used methods for creating predictive models are **multiple regression** and **stepwise regression**. While both approaches aim to uncover relationships between independent (predictor) variables and a dependent (outcome) variable, they differ significantly in their methodology, assumptions, and use cases. diff --git a/_posts/2023-10-01-coverage_probability.md b/_posts/2023-10-01-coverage_probability.md index f4c64e6f..2d090572 100644 --- a/_posts/2023-10-01-coverage_probability.md +++ b/_posts/2023-10-01-coverage_probability.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2023-10-01' -excerpt: 'Understanding coverage probability in statistical estimation and prediction: - its role in constructing confidence intervals and assessing their accuracy.' +excerpt: 'Understanding coverage probability in statistical estimation and prediction: its role in constructing confidence intervals and assessing their accuracy.' header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_1.jpg @@ -19,14 +18,10 @@ keywords: - Nominal confidence level - Prediction intervals - Statistical estimation -seo_description: A detailed explanation of coverage probability, its role in statistical - estimation theory, and its relationship to confidence intervals and prediction intervals. +seo_description: A detailed explanation of coverage probability, its role in statistical estimation theory, and its relationship to confidence intervals and prediction intervals. seo_title: Coverage Probability in Statistical Estimation Theory seo_type: article -summary: In statistical estimation theory, coverage probability measures the likelihood - that a confidence interval contains the true parameter of interest. This article - explains its importance in statistical theory, prediction intervals, and nominal - coverage probability. +summary: In statistical estimation theory, coverage probability measures the likelihood that a confidence interval contains the true parameter of interest. This article explains its importance in statistical theory, prediction intervals, and nominal coverage probability. tags: - Confidence intervals - Statistical theory diff --git a/_posts/2023-10-02-overview_natural_language_processing_data_science.md b/_posts/2023-10-02-overview_natural_language_processing_data_science.md index 6af45967..709fc683 100644 --- a/_posts/2023-10-02-overview_natural_language_processing_data_science.md +++ b/_posts/2023-10-02-overview_natural_language_processing_data_science.md @@ -6,9 +6,7 @@ categories: - Machine Learning classes: wide date: '2023-10-02' -excerpt: Natural Language Processing (NLP) is integral to data science, enabling tasks - like text classification and sentiment analysis. Learn how NLP works, its common - tasks, tools, and applications in real-world projects. +excerpt: Natural Language Processing (NLP) is integral to data science, enabling tasks like text classification and sentiment analysis. Learn how NLP works, its common tasks, tools, and applications in real-world projects. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_1.jpg @@ -24,13 +22,10 @@ keywords: - Nltk - Spacy - Hugging face -seo_description: Explore how Natural Language Processing (NLP) fits into data science, - common NLP tasks, popular libraries like NLTK and SpaCy, and real-world applications. +seo_description: Explore how Natural Language Processing (NLP) fits into data science, common NLP tasks, popular libraries like NLTK and SpaCy, and real-world applications. seo_title: 'Natural Language Processing in Data Science: Tasks, Tools, and Applications' seo_type: article -summary: This article provides an overview of Natural Language Processing (NLP) in - data science, covering its role in the field, common NLP tasks, tools like NLTK - and SpaCy, and real-world applications in various industries. +summary: This article provides an overview of Natural Language Processing (NLP) in data science, covering its role in the field, common NLP tasks, tools like NLTK and SpaCy, and real-world applications in various industries. tags: - Natural language processing (nlp) - Text classification diff --git a/_posts/2023-10-31-detecting_trends_time-series_data.md b/_posts/2023-10-31-detecting_trends_time-series_data.md index 7077211c..6c8656e3 100644 --- a/_posts/2023-10-31-detecting_trends_time-series_data.md +++ b/_posts/2023-10-31-detecting_trends_time-series_data.md @@ -4,9 +4,7 @@ categories: - Time-Series Analysis classes: wide date: '2023-10-31' -excerpt: Learn how the Mann-Kendall Test is used for trend detection in time-series - data, particularly in fields like environmental studies, hydrology, and climate - research. +excerpt: Learn how the Mann-Kendall Test is used for trend detection in time-series data, particularly in fields like environmental studies, hydrology, and climate research. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_7.jpg @@ -25,14 +23,12 @@ keywords: - Python - Bash - Python -seo_description: Explore the Mann-Kendall Test for detecting trends in time-series - data, with applications in environmental studies, hydrology, and climate research. +- bash +- python +seo_description: Explore the Mann-Kendall Test for detecting trends in time-series data, with applications in environmental studies, hydrology, and climate research. seo_title: 'Mann-Kendall Test: A Guide to Detecting Trends in Time-Series Data' seo_type: article -summary: The Mann-Kendall Test is a non-parametric method for detecting trends in - time-series data. This article provides an overview of the test, its mathematical - formulation, and its application in environmental studies, hydrology, and climate - research. +summary: The Mann-Kendall Test is a non-parametric method for detecting trends in time-series data. This article provides an overview of the test, its mathematical formulation, and its application in environmental studies, hydrology, and climate research. tags: - Mann-kendall test - Trend detection @@ -44,6 +40,8 @@ tags: - Python - Bash - Python +- bash +- python title: 'Mann-Kendall Test: Detecting Trends in Time-Series Data' --- diff --git a/_posts/2023-11-01-linear_vs_logistic_model.md b/_posts/2023-11-01-linear_vs_logistic_model.md index a72c0104..0996d14b 100644 --- a/_posts/2023-11-01-linear_vs_logistic_model.md +++ b/_posts/2023-11-01-linear_vs_logistic_model.md @@ -4,8 +4,7 @@ categories: - Probability Modeling classes: wide date: '2023-11-01' -excerpt: Both linear and logistic models offer unique advantages depending on the - circumstances. Learn when each model is appropriate and how to interpret their results. +excerpt: Both linear and logistic models offer unique advantages depending on the circumstances. Learn when each model is appropriate and how to interpret their results. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_1.jpg @@ -19,12 +18,10 @@ keywords: - Statistical modeling - Interpretability - Statistical estimation -seo_description: A comprehensive guide to understanding the advantages and limitations - of linear and logistic probability models in statistical analysis. +seo_description: A comprehensive guide to understanding the advantages and limitations of linear and logistic probability models in statistical analysis. seo_title: 'Linear vs. Logistic Probability Models: Which is Better?' seo_type: article -summary: This article explores the pros and cons of linear and logistic probability - models, highlighting interpretability, computation, and when to use each. +summary: This article explores the pros and cons of linear and logistic probability models, highlighting interpretability, computation, and when to use each. tags: - Linear probability model - Logistic regression diff --git a/_posts/2023-11-15-analyzing_relationship_between_continuous_binary_variables.md b/_posts/2023-11-15-analyzing_relationship_between_continuous_binary_variables.md index c028f396..050a32a7 100644 --- a/_posts/2023-11-15-analyzing_relationship_between_continuous_binary_variables.md +++ b/_posts/2023-11-15-analyzing_relationship_between_continuous_binary_variables.md @@ -4,9 +4,7 @@ categories: - Data Analysis classes: wide date: '2023-11-15' -excerpt: Learn the differences between biserial and point-biserial correlation methods, - and discover how they can be applied to analyze relationships between continuous - and binary variables in educational testing, psychology, and medical diagnostics. +excerpt: Learn the differences between biserial and point-biserial correlation methods, and discover how they can be applied to analyze relationships between continuous and binary variables in educational testing, psychology, and medical diagnostics. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_9.jpg @@ -20,16 +18,10 @@ keywords: - Educational testing - Psychology - Medical diagnostics -seo_description: Explore biserial and point-biserial correlation methods for analyzing - relationships between continuous and binary variables, with applications in educational - testing, psychology, and medical diagnostics. -seo_title: 'Biserial vs. Point-Biserial Correlation: Analyzing Continuous and Binary - Variable Relationships' +seo_description: Explore biserial and point-biserial correlation methods for analyzing relationships between continuous and binary variables, with applications in educational testing, psychology, and medical diagnostics. +seo_title: 'Biserial vs. Point-Biserial Correlation: Analyzing Continuous and Binary Variable Relationships' seo_type: article -summary: Biserial and point-biserial correlation methods are used to analyze relationships - between binary and continuous variables. This article explains the differences between - these two correlation techniques and their practical applications in fields like - educational testing, psychology, and medical diagnostics. +summary: Biserial and point-biserial correlation methods are used to analyze relationships between binary and continuous variables. This article explains the differences between these two correlation techniques and their practical applications in fields like educational testing, psychology, and medical diagnostics. tags: - Biserial correlation - Point-biserial correlation @@ -38,8 +30,7 @@ tags: - Educational testing - Psychology - Medical diagnostics -title: 'Biserial and Point-Biserial Correlation: Analyzing the Relationship Between - Continuous and Binary Variables' +title: 'Biserial and Point-Biserial Correlation: Analyzing the Relationship Between Continuous and Binary Variables' --- In statistical analysis, understanding the relationship between variables is essential for gaining insights and making informed decisions. When analyzing the relationship between **continuous** and **binary** variables, two specialized correlation methods are often employed: **biserial correlation** and **point-biserial correlation**. Both techniques are used to measure the strength and direction of association between these two types of variables, but they are applied in different contexts and are based on distinct assumptions. diff --git a/_posts/2023-11-16-mann-whitney_u_test_non-parametric_comparison_two_independent_samples.md b/_posts/2023-11-16-mann-whitney_u_test_non-parametric_comparison_two_independent_samples.md index 34529c15..889632f6 100644 --- a/_posts/2023-11-16-mann-whitney_u_test_non-parametric_comparison_two_independent_samples.md +++ b/_posts/2023-11-16-mann-whitney_u_test_non-parametric_comparison_two_independent_samples.md @@ -6,9 +6,7 @@ categories: - Data Analysis classes: wide date: '2023-11-16' -excerpt: Learn how the Mann-Whitney U Test is used to compare two independent samples - in non-parametric statistics, with applications in fields such as psychology, medicine, - and ecology. +excerpt: Learn how the Mann-Whitney U Test is used to compare two independent samples in non-parametric statistics, with applications in fields such as psychology, medicine, and ecology. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_8.jpg @@ -25,14 +23,12 @@ keywords: - Medicine - Bash - Python -seo_description: Explore the Mann-Whitney U Test, a non-parametric method for comparing - two independent samples, with applications in fields like psychology, medicine, - and ecology. +- bash +- python +seo_description: Explore the Mann-Whitney U Test, a non-parametric method for comparing two independent samples, with applications in fields like psychology, medicine, and ecology. seo_title: 'Mann-Whitney U Test: Comparing Two Independent Samples' seo_type: article -summary: The Mann-Whitney U Test is a non-parametric method used to compare two independent - samples. This article explains the test's assumptions, mathematical foundations, - and its applications in fields like psychology, medicine, and ecology. +summary: The Mann-Whitney U Test is a non-parametric method used to compare two independent samples. This article explains the test's assumptions, mathematical foundations, and its applications in fields like psychology, medicine, and ecology. tags: - Mann-whitney u test - Non-parametric statistics @@ -41,6 +37,8 @@ tags: - Data analysis - Bash - Python +- bash +- python title: 'Mann-Whitney U Test: Non-Parametric Comparison of Two Independent Samples' --- diff --git a/_posts/2023-11-30-math_fundamentals.md b/_posts/2023-11-30-math_fundamentals.md index 44102628..a83eaf9b 100644 --- a/_posts/2023-11-30-math_fundamentals.md +++ b/_posts/2023-11-30-math_fundamentals.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2023-11-30' -excerpt: A comprehensive exploration of data drift in credit risk models, examining - practical methods to identify and address drift using multivariate techniques. +excerpt: A comprehensive exploration of data drift in credit risk models, examining practical methods to identify and address drift using multivariate techniques. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_9.jpg @@ -24,8 +23,7 @@ keywords: - Drift detection - Predictive modeling - Credit scoring -seo_description: Explore a practical approach to solving data drift in credit risk - models, focusing on multivariate analysis and its impact on model performance. +seo_description: Explore a practical approach to solving data drift in credit risk models, focusing on multivariate analysis and its impact on model performance. seo_title: 'Addressing Data Drift in Credit Risk Models: A Case Study' seo_type: article tags: diff --git a/_posts/2023-12-01-managing_data_science.md b/_posts/2023-12-01-managing_data_science.md index 0410b8ac..3c0e4a29 100644 --- a/_posts/2023-12-01-managing_data_science.md +++ b/_posts/2023-12-01-managing_data_science.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2023-12-01' -excerpt: While engineering projects have defined solutions and known processes, data - science is all about experimentation and discovery. Managing them in the same way - can be detrimental. +excerpt: While engineering projects have defined solutions and known processes, data science is all about experimentation and discovery. Managing them in the same way can be detrimental. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_3.jpg @@ -20,14 +18,10 @@ keywords: - Project management - Ai - Experimentation -seo_description: Managing data science projects like engineering projects sets them - up to fail. Learn the key differences in scope, timelines, and processes between - the two fields. +seo_description: Managing data science projects like engineering projects sets them up to fail. Learn the key differences in scope, timelines, and processes between the two fields. seo_title: 'Managing Data Science Projects vs Engineering: Why It Fails' seo_type: article -summary: This article explores why managing data science projects with the same expectations - as engineering leads to failure, explaining how the unknown nature of data science - solutions differs from engineering's structured approach. +summary: This article explores why managing data science projects with the same expectations as engineering leads to failure, explaining how the unknown nature of data science solutions differs from engineering's structured approach. tags: - Data science - Engineering diff --git a/_posts/2023-12-30-data_engineering_introduction.md b/_posts/2023-12-30-data_engineering_introduction.md index 3f2b6f4f..9011a3fb 100644 --- a/_posts/2023-12-30-data_engineering_introduction.md +++ b/_posts/2023-12-30-data_engineering_introduction.md @@ -4,8 +4,7 @@ categories: - Data Engineering classes: wide date: '2023-12-30' -excerpt: This article explores the fundamentals of data engineering, including the - ETL/ELT processes, required skills, and the relationship with data science. +excerpt: This article explores the fundamentals of data engineering, including the ETL/ELT processes, required skills, and the relationship with data science. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_6.jpg @@ -19,13 +18,10 @@ keywords: - Elt - Data science - Data pipelines -seo_description: An in-depth overview of Data Engineering, discussing the ETL and - ELT processes, data pipelines, and the necessary skills for data engineers. +seo_description: An in-depth overview of Data Engineering, discussing the ETL and ELT processes, data pipelines, and the necessary skills for data engineers. seo_title: 'Understanding Data Engineering: Skills, ETL, and ELT Processes' seo_type: article -summary: Data Engineering is critical for managing and processing large datasets. - Learn about the skills, processes like ETL and ELT, and how they fit into modern - data workflows. +summary: Data Engineering is critical for managing and processing large datasets. Learn about the skills, processes like ETL and ELT, and how they fit into modern data workflows. tags: - Etl - Data pipelines diff --git a/_posts/2023-12-30-expected_shortfall.md b/_posts/2023-12-30-expected_shortfall.md index c8e12f5e..49b4b953 100644 --- a/_posts/2023-12-30-expected_shortfall.md +++ b/_posts/2023-12-30-expected_shortfall.md @@ -5,9 +5,7 @@ categories: - Financial Risk Management classes: wide date: '2023-12-30' -excerpt: A comprehensive comparison of Value at Risk (VaR) and Expected Shortfall - (ES) in financial risk management, with a focus on their performance during volatile - and stable market conditions. +excerpt: A comprehensive comparison of Value at Risk (VaR) and Expected Shortfall (ES) in financial risk management, with a focus on their performance during volatile and stable market conditions. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_9.jpg @@ -28,9 +26,8 @@ keywords: - Risk metrics - Python - Python -seo_description: An in-depth analysis of Value at Risk (VaR) and Expected Shortfall - (ES) as risk assessment models, comparing their performance during different market - conditions. +- python +seo_description: An in-depth analysis of Value at Risk (VaR) and Expected Shortfall (ES) as risk assessment models, comparing their performance during different market conditions. seo_title: 'VaR vs Expected Shortfall: A Data-Driven Analysis' seo_type: article tags: @@ -40,6 +37,7 @@ tags: - Risk models - Python - Python +- python title: 'Comparing Value at Risk (VaR) and Expected Shortfall (ES): A Data-Driven Analysis' --- diff --git a/_posts/2024-01-01-mathematics_machine_learning.md b/_posts/2024-01-01-mathematics_machine_learning.md index cece040b..f07b8657 100644 --- a/_posts/2024-01-01-mathematics_machine_learning.md +++ b/_posts/2024-01-01-mathematics_machine_learning.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-01-01' -excerpt: This article delves into the core mathematical principles behind machine - learning, including classification and regression settings, loss functions, risk - minimization, decision trees, and more. +excerpt: This article delves into the core mathematical principles behind machine learning, including classification and regression settings, loss functions, risk minimization, decision trees, and more. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_5.jpg @@ -30,9 +28,7 @@ keywords: - Machine learning algorithms - Generalization in machine learning - Concentration inequalities in machine learning -seo_description: An extensive look at the mathematical foundations of machine learning, - exploring classification, regression, empirical risk minimization, and popular algorithms - like decision trees and random forests. +seo_description: An extensive look at the mathematical foundations of machine learning, exploring classification, regression, empirical risk minimization, and popular algorithms like decision trees and random forests. seo_title: 'Mathematics of Machine Learning: Key Concepts and Methods' seo_type: article tags: diff --git a/_posts/2024-01-02-text_preprocessing_techniques_nlp_data_science.md b/_posts/2024-01-02-text_preprocessing_techniques_nlp_data_science.md index 43300a66..f96746b6 100644 --- a/_posts/2024-01-02-text_preprocessing_techniques_nlp_data_science.md +++ b/_posts/2024-01-02-text_preprocessing_techniques_nlp_data_science.md @@ -4,9 +4,7 @@ categories: - Natural Language Processing classes: wide date: '2024-01-02' -excerpt: Text preprocessing is a crucial step in NLP for transforming raw text into - a structured format. Learn key techniques like tokenization, stemming, lemmatization, - and text normalization for successful NLP tasks. +excerpt: Text preprocessing is a crucial step in NLP for transforming raw text into a structured format. Learn key techniques like tokenization, stemming, lemmatization, and text normalization for successful NLP tasks. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -21,16 +19,10 @@ keywords: - Stemming - Lemmatization - Text normalization -seo_description: Explore essential text preprocessing techniques for NLP, including - tokenization, stemming, lemmatization, handling stopwords, and advanced text cleaning - using regex. +seo_description: Explore essential text preprocessing techniques for NLP, including tokenization, stemming, lemmatization, handling stopwords, and advanced text cleaning using regex. seo_title: 'Text Preprocessing Techniques for NLP: Tokenization, Stemming, and More' seo_type: article -summary: This article provides an in-depth look at text preprocessing techniques for - Natural Language Processing (NLP) in data science. It covers core concepts like - tokenization, stemming, lemmatization, handling stopwords, text normalization, and - advanced cleaning techniques such as regex for handling misspellings, slang, and - abbreviations. +summary: This article provides an in-depth look at text preprocessing techniques for Natural Language Processing (NLP) in data science. It covers core concepts like tokenization, stemming, lemmatization, handling stopwords, text normalization, and advanced cleaning techniques such as regex for handling misspellings, slang, and abbreviations. tags: - Text preprocessing - Tokenization diff --git a/_posts/2024-01-28-normal_distribution.md b/_posts/2024-01-28-normal_distribution.md index 551cd8ff..8d6932cd 100644 --- a/_posts/2024-01-28-normal_distribution.md +++ b/_posts/2024-01-28-normal_distribution.md @@ -4,8 +4,7 @@ categories: - Mathematics classes: wide date: '2024-01-28' -excerpt: Discover the significance of the Normal Distribution, also known as the Bell - Curve, in statistics and its widespread application in real-world scenarios. +excerpt: Discover the significance of the Normal Distribution, also known as the Bell Curve, in statistics and its widespread application in real-world scenarios. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_9.jpg @@ -25,9 +24,8 @@ keywords: - Standard deviation - Mean and variance - Python -seo_description: An in-depth exploration of the Normal Distribution, often called - the Bell Curve, and its critical role in data science, machine learning, and statistical - analysis. +- python +seo_description: An in-depth exploration of the Normal Distribution, often called the Bell Curve, and its critical role in data science, machine learning, and statistical analysis. seo_title: 'Understanding the Classic Bell Curve: The Normal Distribution' seo_type: article subtitle: The Normal Distribution @@ -39,6 +37,7 @@ tags: - Statistical analysis - Bell curve - Python +- python title: A Closer Look at the Classic Bell Curve --- diff --git a/_posts/2024-01-29-probabilistic_programming.md b/_posts/2024-01-29-probabilistic_programming.md index 38a1382d..602edd80 100644 --- a/_posts/2024-01-29-probabilistic_programming.md +++ b/_posts/2024-01-29-probabilistic_programming.md @@ -4,8 +4,7 @@ categories: - Mathematics classes: wide date: '2024-01-29' -excerpt: Explore Markov Chain Monte Carlo (MCMC) methods, specifically the Metropolis - algorithm, and learn how to perform Bayesian inference through Python code. +excerpt: Explore Markov Chain Monte Carlo (MCMC) methods, specifically the Metropolis algorithm, and learn how to perform Bayesian inference through Python code. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_9.jpg @@ -25,8 +24,8 @@ keywords: - Data science - Machine learning - Python -seo_description: A practical explanation of MCMC and the Metropolis algorithm, focusing - on Bayesian inference with Python code examples to make the concepts accessible. +- python +seo_description: A practical explanation of MCMC and the Metropolis algorithm, focusing on Bayesian inference with Python code examples to make the concepts accessible. seo_title: 'Demystifying MCMC: A Hands-On Guide to Bayesian Inference' seo_type: article subtitle: Understanding the Metropolis Algorithm Through Code @@ -40,6 +39,7 @@ tags: - Probabilistic programming - Bayesian statistics - Python +- python title: 'Demystifying MCMC: A Practical Guide to Bayesian Inference' --- diff --git a/_posts/2024-01-30-Monte_Carlo.md b/_posts/2024-01-30-Monte_Carlo.md index 8b0c5aee..968af1d7 100644 --- a/_posts/2024-01-30-Monte_Carlo.md +++ b/_posts/2024-01-30-Monte_Carlo.md @@ -5,8 +5,7 @@ categories: classes: wide date: '2024-01-30' draft: false -excerpt: Discover how Bayesian inference and MCMC algorithms like Metropolis-Hastings - can solve complex probability problems through real-world examples and Python implementation. +excerpt: Discover how Bayesian inference and MCMC algorithms like Metropolis-Hastings can solve complex probability problems through real-world examples and Python implementation. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_4.jpg @@ -19,15 +18,13 @@ keywords: - Mcmc algorithms - Python - Python +- python math: true -seo_description: Explore Bayesian statistics and the power of Markov Chain Monte Carlo - (MCMC) in handling complex probabilistic models. Learn with practical examples and - Python code. +seo_description: Explore Bayesian statistics and the power of Markov Chain Monte Carlo (MCMC) in handling complex probabilistic models. Learn with practical examples and Python code. seo_title: 'Mastering Bayesian Statistics with MCMC: A Deep Dive into Complex Probabilities' seo_type: article subtitle: Complex Probabilities with Markov Chain Monte Carlo -summary: A comprehensive guide to understanding Bayesian statistics and MCMC methods, - including real-world applications and Python examples. +summary: A comprehensive guide to understanding Bayesian statistics and MCMC methods, including real-world applications and Python examples. tags: - Bayesian statistics - Markov chain monte carlo (mcmc) @@ -39,6 +36,7 @@ tags: - Machine learning algorithms - Python - Python +- python title: 'Mastering Bayesian Statistics: An In-Depth Guide to MCMC' --- diff --git a/_posts/2024-02-01-customer_life_value.md b/_posts/2024-02-01-customer_life_value.md index 03290781..f7c42bcf 100644 --- a/_posts/2024-02-01-customer_life_value.md +++ b/_posts/2024-02-01-customer_life_value.md @@ -5,9 +5,7 @@ categories: - Data Science classes: wide date: '2024-02-01' -excerpt: Discover the importance of Customer Lifetime Value (CLV) in shaping business - strategies, improving customer retention, and enhancing marketing efforts for sustainable - growth. +excerpt: Discover the importance of Customer Lifetime Value (CLV) in shaping business strategies, improving customer retention, and enhancing marketing efforts for sustainable growth. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_2.jpg @@ -27,9 +25,8 @@ keywords: - Loyalty programs - Data analytics - Python -seo_description: Explore Customer Lifetime Value (CLV) and its role in driving business - growth. Learn how CLV influences customer retention, acquisition, and marketing - strategies. +- python +seo_description: Explore Customer Lifetime Value (CLV) and its role in driving business growth. Learn how CLV influences customer retention, acquisition, and marketing strategies. seo_title: 'Understanding Customer Lifetime Value: A Key to Business Growth' seo_type: article subtitle: A Key Metric for Business Growth @@ -44,6 +41,7 @@ tags: - Business growth - Loyalty programs - Python +- python title: Understanding Customer Lifetime Value toc: false toc_label: The Complexity of Real-World Data Distributions diff --git a/_posts/2024-02-02-topology_data_science.md b/_posts/2024-02-02-topology_data_science.md index 99db2911..02df7f71 100644 --- a/_posts/2024-02-02-topology_data_science.md +++ b/_posts/2024-02-02-topology_data_science.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2024-02-02' -excerpt: Dive into Topological Data Analysis (TDA) and discover how its methods, such - as persistent homology and the mapper algorithm, help uncover hidden insights in - high-dimensional and complex datasets. +excerpt: Dive into Topological Data Analysis (TDA) and discover how its methods, such as persistent homology and the mapper algorithm, help uncover hidden insights in high-dimensional and complex datasets. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_1.jpg @@ -25,13 +23,10 @@ keywords: - Network analysis - Interdisciplinary data science - Mathematical foundations -seo_description: Explore Topological Data Analysis (TDA) and its transformative role - in data science, from persistent homology to the mapper algorithm, revealing hidden - structures in complex datasets. +seo_description: Explore Topological Data Analysis (TDA) and its transformative role in data science, from persistent homology to the mapper algorithm, revealing hidden structures in complex datasets. seo_title: 'Convergence of Topology and Data Science: Uncovering Insights with TDA' seo_type: article -subtitle: Exploring Topological Data Analysis and Its Impact on Uncovering Hidden - Insights in Complex Data Sets +subtitle: Exploring Topological Data Analysis and Its Impact on Uncovering Hidden Insights in Complex Data Sets tags: - Topological data analysis (tda) - Data science diff --git a/_posts/2024-02-08-Clustering.md b/_posts/2024-02-08-Clustering.md index bada148f..bf29820e 100644 --- a/_posts/2024-02-08-Clustering.md +++ b/_posts/2024-02-08-Clustering.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2024-02-08' -excerpt: Discover the inner workings of clustering algorithms, from K-Means to Spectral - Clustering, and how they unveil patterns in machine learning, bioinformatics, and - data analysis. +excerpt: Discover the inner workings of clustering algorithms, from K-Means to Spectral Clustering, and how they unveil patterns in machine learning, bioinformatics, and data analysis. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_1.jpg @@ -25,9 +23,7 @@ keywords: - Pattern recognition - Bioinformatics - Data analysis -seo_description: Explore the mysteries of clustering algorithms like K-Means, DBSCAN, - and Spectral Clustering. Learn how these techniques reveal hidden patterns in data - science, machine learning, and bioinformatics. +seo_description: Explore the mysteries of clustering algorithms like K-Means, DBSCAN, and Spectral Clustering. Learn how these techniques reveal hidden patterns in data science, machine learning, and bioinformatics. seo_title: 'Mysteries of Clustering: A Deep Dive into Data''s Inner Circles' seo_type: article subtitle: A Dive into Data's Inner Circles diff --git a/_posts/2024-02-09-spectral_clustering.md b/_posts/2024-02-09-spectral_clustering.md index 66050fc4..b4a5f82b 100644 --- a/_posts/2024-02-09-spectral_clustering.md +++ b/_posts/2024-02-09-spectral_clustering.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2024-02-09' -excerpt: A comprehensive guide to spectral clustering and its role in dimensionality - reduction, enhancing data analysis, and uncovering patterns in machine learning. +excerpt: A comprehensive guide to spectral clustering and its role in dimensionality reduction, enhancing data analysis, and uncovering patterns in machine learning. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_5.jpg @@ -22,9 +21,7 @@ keywords: - Data analysis - Pattern recognition - Unsupervised learning -seo_description: Explore the power of dimensionality reduction through spectral clustering. - Learn how this algorithm enhances data analysis and pattern recognition in machine - learning. +seo_description: Explore the power of dimensionality reduction through spectral clustering. Learn how this algorithm enhances data analysis and pattern recognition in machine learning. seo_title: 'The Power of Dimensionality Reduction: Spectral Clustering Guide' seo_type: article subtitle: A Comprehensive Guide to Spectral Clustering diff --git a/_posts/2024-02-10-pingenhole_principle.md b/_posts/2024-02-10-pingenhole_principle.md index d62dc922..13974a76 100644 --- a/_posts/2024-02-10-pingenhole_principle.md +++ b/_posts/2024-02-10-pingenhole_principle.md @@ -4,9 +4,7 @@ categories: - Mathematics classes: wide date: '2024-02-10' -excerpt: A journey into the Pigeonhole Principle, uncovering its profound simplicity - and exploring its applications in fields like combinatorics, number theory, and - geometry. +excerpt: A journey into the Pigeonhole Principle, uncovering its profound simplicity and exploring its applications in fields like combinatorics, number theory, and geometry. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_3.jpg @@ -29,15 +27,13 @@ keywords: - Python - Python - R -seo_description: Explore the simplicity and power of the Pigeonhole Principle, delving - into its applications across combinatorics, number theory, geometry, and more. +- python +- r +seo_description: Explore the simplicity and power of the Pigeonhole Principle, delving into its applications across combinatorics, number theory, geometry, and more. seo_title: 'The Elegance of the Pigeonhole Principle: Universal Applications in Mathematics' seo_type: article -subtitle: Exploring the Profound Simplicity and Universal Applications of a Foundational - Mathematical Concept -summary: This article delves into the Pigeonhole Principle, illustrating its profound - simplicity and exploring its applications in various mathematical fields such as - combinatorics, number theory, geometry, and data compression. +subtitle: Exploring the Profound Simplicity and Universal Applications of a Foundational Mathematical Concept +summary: This article delves into the Pigeonhole Principle, illustrating its profound simplicity and exploring its applications in various mathematical fields such as combinatorics, number theory, geometry, and data compression. tags: - Pigeonhole principle - Mathematical logic @@ -52,6 +48,8 @@ tags: - Python - Python - R +- python +- r title: 'Elegance of the Pigeonhole Principle: A Mathematical Odyssey' toc: false toc_label: The Complexity of Real-World Data Distributions diff --git a/_posts/2024-02-11-Ergodicity.md b/_posts/2024-02-11-Ergodicity.md index 35389754..387737ed 100644 --- a/_posts/2024-02-11-Ergodicity.md +++ b/_posts/2024-02-11-Ergodicity.md @@ -4,9 +4,7 @@ categories: - Mathematics classes: wide date: '2024-02-11' -excerpt: An in-depth look into ergodicity and its applications in statistical analysis, - mathematical modeling, and computational physics, featuring real-world processes - and Python simulations. +excerpt: An in-depth look into ergodicity and its applications in statistical analysis, mathematical modeling, and computational physics, featuring real-world processes and Python simulations. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_4.jpg @@ -27,11 +25,9 @@ keywords: - Statistical physics - Python - Python -seo_description: Explore ergodic regimes in mathematics, statistical physics, and - data science, with practical insights into processes, Bernoulli trials, and Python-based - simulations. -seo_title: 'Distinguishing Ergodic Regimes: Clarifying Ergodicity in Statistical and - Mathematical Models' +- python +seo_description: Explore ergodic regimes in mathematics, statistical physics, and data science, with practical insights into processes, Bernoulli trials, and Python-based simulations. +seo_title: 'Distinguishing Ergodic Regimes: Clarifying Ergodicity in Statistical and Mathematical Models' seo_type: article subtitle: Clarifying Ergodicity tags: @@ -47,6 +43,7 @@ tags: - Machine learning - Python - Python +- python title: Distinguishing Ergodic Regimes from Processes toc: false toc_label: The Complexity of Real-World Data Distributions diff --git a/_posts/2024-02-11-combinatorics_python.md b/_posts/2024-02-11-combinatorics_python.md index 9521c12c..beca15e7 100644 --- a/_posts/2024-02-11-combinatorics_python.md +++ b/_posts/2024-02-11-combinatorics_python.md @@ -4,9 +4,7 @@ categories: - Mathematics classes: wide date: '2024-02-11' -excerpt: A practical guide to mastering combinatorics with Python, featuring hands-on - examples using the itertools library and insights into scientific computing and - probability theory. +excerpt: A practical guide to mastering combinatorics with Python, featuring hands-on examples using the itertools library and insights into scientific computing and probability theory. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -27,9 +25,8 @@ keywords: - Data analysis techniques - Python - R -seo_description: Learn how to master combinatorial mathematics using Python. Explore - practical applications with the itertools library, scientific computing, and probability - theory. +- python +seo_description: Learn how to master combinatorial mathematics using Python. Explore practical applications with the itertools library, scientific computing, and probability theory. seo_title: 'Mastering Combinatorics with Python: A Practical Guide' seo_type: article subtitle: A Practical Guide @@ -46,6 +43,7 @@ tags: - Python libraries - Python - R +- python title: Mastering Combinatorics with Python toc: false toc_label: The Complexity of Real-World Data Distributions diff --git a/_posts/2024-02-12-combinatorics_probability.md b/_posts/2024-02-12-combinatorics_probability.md index d96f2142..833a5d87 100644 --- a/_posts/2024-02-12-combinatorics_probability.md +++ b/_posts/2024-02-12-combinatorics_probability.md @@ -4,8 +4,7 @@ categories: - Mathematics classes: wide date: '2024-02-12' -excerpt: Dive into the intersection of combinatorics and probability, exploring how - these fields work together to solve problems in mathematics, data science, and beyond. +excerpt: Dive into the intersection of combinatorics and probability, exploring how these fields work together to solve problems in mathematics, data science, and beyond. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_5.jpg @@ -24,15 +23,11 @@ keywords: - Probability models - Educational resources - Applied mathematics -seo_description: Discover the deep connections between combinatorics and probability - theory, exploring their mathematical foundations, applications, and the synergies - that drive statistical analysis and data science. +seo_description: Discover the deep connections between combinatorics and probability theory, exploring their mathematical foundations, applications, and the synergies that drive statistical analysis and data science. seo_title: 'Combinatorics and Probability: Exploring Mathematical Synergies' seo_type: article subtitle: Unveiling Mathematical Synergies -summary: This article explores the intersection of combinatorics and probability theory, - uncovering how their mathematical synergies solve complex problems in data science, - mathematics, and beyond. +summary: This article explores the intersection of combinatorics and probability theory, uncovering how their mathematical synergies solve complex problems in data science, mathematics, and beyond. tags: - Mathematics - Combinatorics diff --git a/_posts/2024-02-12-ethical_considerations_elderly_care.md b/_posts/2024-02-12-ethical_considerations_elderly_care.md index 4a680601..52fca2f5 100644 --- a/_posts/2024-02-12-ethical_considerations_elderly_care.md +++ b/_posts/2024-02-12-ethical_considerations_elderly_care.md @@ -4,9 +4,7 @@ categories: - HealthTech classes: wide date: '2024-02-12' -excerpt: As AI revolutionizes elderly care, ethical concerns around privacy, autonomy, - and consent come into focus. This article explores how to balance technological - advancements with the dignity and personal preferences of elderly individuals. +excerpt: As AI revolutionizes elderly care, ethical concerns around privacy, autonomy, and consent come into focus. This article explores how to balance technological advancements with the dignity and personal preferences of elderly individuals. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_9.jpg @@ -20,16 +18,10 @@ keywords: - Big data privacy - Elderly autonomy - Informed consent -seo_description: This article explores the ethical challenges of using AI, big data, - and machine learning in elderly care, focusing on privacy, autonomy, and informed - consent. +seo_description: This article explores the ethical challenges of using AI, big data, and machine learning in elderly care, focusing on privacy, autonomy, and informed consent. seo_title: 'Ethical Issues in AI-Powered Elderly Care: Privacy, Autonomy, and Consent' seo_type: article -summary: The integration of AI and machine learning in elderly care promises significant - advancements but raises critical ethical concerns. This article examines the challenges - of protecting privacy, maintaining autonomy, and ensuring informed consent in AI-powered - care systems, offering strategies to balance innovation with the dignity of elderly - individuals. +summary: The integration of AI and machine learning in elderly care promises significant advancements but raises critical ethical concerns. This article examines the challenges of protecting privacy, maintaining autonomy, and ensuring informed consent in AI-powered care systems, offering strategies to balance innovation with the dignity of elderly individuals. tags: - Ai in healthcare - Elderly care diff --git a/_posts/2024-02-14-advanced_sequential_change-point.md b/_posts/2024-02-14-advanced_sequential_change-point.md index 61de4e57..e631b633 100644 --- a/_posts/2024-02-14-advanced_sequential_change-point.md +++ b/_posts/2024-02-14-advanced_sequential_change-point.md @@ -6,9 +6,7 @@ categories: - Data Analysis classes: wide date: '2024-02-14' -excerpt: Sequential change-point detection plays a crucial role in real-time monitoring - across industries. Learn about advanced methods, their practical applications, and - how they help detect changes in univariate models. +excerpt: Sequential change-point detection plays a crucial role in real-time monitoring across industries. Learn about advanced methods, their practical applications, and how they help detect changes in univariate models. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_4.jpg @@ -28,17 +26,16 @@ keywords: - Sequential change-point algorithms - Time series analysis - Python -seo_description: Explore advanced methods and practical implementations for sequential - change-point detection in univariate models, covering theoretical foundations, real-world - applications, and key statistical techniques. -seo_title: Advanced Techniques for Sequential Change-Point Detection in Univariate - Models +- python +seo_description: Explore advanced methods and practical implementations for sequential change-point detection in univariate models, covering theoretical foundations, real-world applications, and key statistical techniques. +seo_title: Advanced Techniques for Sequential Change-Point Detection in Univariate Models seo_type: article tags: - Change-point detection - Univariate models - Sequential analysis - Python +- python title: Advanced Sequential Change-Point Detection for Univariate Models --- diff --git a/_posts/2024-02-17-climate_var.md b/_posts/2024-02-17-climate_var.md index d6c0b93c..fcee7787 100644 --- a/_posts/2024-02-17-climate_var.md +++ b/_posts/2024-02-17-climate_var.md @@ -6,8 +6,7 @@ categories: - Financial Risk classes: wide date: '2024-02-17' -excerpt: Exploring Climate Value at Risk (VaR) from a data science perspective, detailing - its role in assessing financial risks associated with climate change. +excerpt: Exploring Climate Value at Risk (VaR) from a data science perspective, detailing its role in assessing financial risks associated with climate change. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_4.jpg @@ -27,9 +26,8 @@ keywords: - Climate finance - Sustainability and risk - Python -seo_description: An in-depth analysis of Climate Value at Risk (VaR) from a data science - perspective, exploring its importance in financial risk assessment amidst climate - change. +- python +seo_description: An in-depth analysis of Climate Value at Risk (VaR) from a data science perspective, exploring its importance in financial risk assessment amidst climate change. seo_title: 'Climate VaR: Data Science and Financial Risk Assessment' seo_type: article tags: @@ -38,6 +36,7 @@ tags: - Data science - Financial risk management - Python +- python title: 'Climate Value at Risk (VaR): A Data Science Perspective' --- diff --git a/_posts/2024-02-20-validate_models.md b/_posts/2024-02-20-validate_models.md index 62a1a2c4..2d3d15d7 100644 --- a/_posts/2024-02-20-validate_models.md +++ b/_posts/2024-02-20-validate_models.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-02-20' -excerpt: Discover critical lessons learned from validating COPOD, a popular anomaly - detection model, through test-driven validation techniques. Avoid common pitfalls - in anomaly detection modeling. +excerpt: Discover critical lessons learned from validating COPOD, a popular anomaly detection model, through test-driven validation techniques. Avoid common pitfalls in anomaly detection modeling. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_3.jpg @@ -27,9 +25,8 @@ keywords: - Scalability in anomaly detection - High-dimensional data - Python -seo_description: Explore how to validate anomaly detection models like COPOD. Learn - the importance of model validation through test-driven development and avoid pitfalls - in high-dimensional data analysis. +- python +seo_description: Explore how to validate anomaly detection models like COPOD. Learn the importance of model validation through test-driven development and avoid pitfalls in high-dimensional data analysis. seo_title: 'Validating COPOD for Anomaly Detection: Key Insights and Lessons' seo_type: article tags: @@ -38,6 +35,7 @@ tags: - Copod - Python - Python +- python title: 'Validating Anomaly Detection Models: Lessons from COPOD' toc: false toc_label: The Complexity of Real-World Data Distributions diff --git a/_posts/2024-05-09-kernel_clustering_r.md b/_posts/2024-05-09-kernel_clustering_r.md index af06f51f..b5f48da3 100644 --- a/_posts/2024-05-09-kernel_clustering_r.md +++ b/_posts/2024-05-09-kernel_clustering_r.md @@ -38,6 +38,8 @@ tags: - Scalable clustering algorithms in r - Unknown - R +- r +- unknown title: Kernel Clustering in R --- diff --git a/_posts/2024-05-09-understanding_t-sne.md b/_posts/2024-05-09-understanding_t-sne.md index c901b700..ad8d5a91 100644 --- a/_posts/2024-05-09-understanding_t-sne.md +++ b/_posts/2024-05-09-understanding_t-sne.md @@ -37,6 +37,7 @@ tags: - Genomics data analysis - Interactive data visualization - Python +- python title: Understanding t-SNE --- diff --git a/_posts/2024-05-10-data_analysis_gdp.md b/_posts/2024-05-10-data_analysis_gdp.md index 78e97cd6..a64846f5 100644 --- a/_posts/2024-05-10-data_analysis_gdp.md +++ b/_posts/2024-05-10-data_analysis_gdp.md @@ -15,8 +15,7 @@ header: teaser: /assets/images/data_science_4.jpg twitter_image: /assets/images/data_science_1.jpg seo_type: article -subtitle: Exploring the Shortcomings of GDP as a Sole Economic Indicator in Data Science - Applications +subtitle: Exploring the Shortcomings of GDP as a Sole Economic Indicator in Data Science Applications tags: - Gdp limitations - Economic analysis diff --git a/_posts/2024-05-10-survival_analysis.md b/_posts/2024-05-10-survival_analysis.md index 07d48dd8..600a4a32 100644 --- a/_posts/2024-05-10-survival_analysis.md +++ b/_posts/2024-05-10-survival_analysis.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2024-05-10' -excerpt: Explore the role of survival analysis in management, focusing on time-to-event - data and techniques like the Kaplan-Meier estimator and Cox proportional hazards - model for business decision-making. +excerpt: Explore the role of survival analysis in management, focusing on time-to-event data and techniques like the Kaplan-Meier estimator and Cox proportional hazards model for business decision-making. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_9.jpg @@ -28,16 +26,12 @@ keywords: - Business analytics - R - Python -seo_description: Learn about survival analysis and its applications in management - for analyzing time-to-event data. Discover key techniques like the Kaplan-Meier - estimator and the Cox model, useful in decision-making for employee retention and - customer churn. +- python +seo_description: Learn about survival analysis and its applications in management for analyzing time-to-event data. Discover key techniques like the Kaplan-Meier estimator and the Cox model, useful in decision-making for employee retention and customer churn. seo_title: 'Survival Analysis in Management: Techniques and Applications' seo_type: article subtitle: Techniques and Applications -summary: This article examines survival analysis in management, detailing its key - concepts like hazard and survival functions, censoring, and applications such as - employee retention, customer churn, and product lifespan modeling. +summary: This article examines survival analysis in management, detailing its key concepts like hazard and survival functions, censoring, and applications such as employee retention, customer churn, and product lifespan modeling. tags: - Survival analysis - Time-to-event data @@ -56,6 +50,7 @@ tags: - Data-driven management - R - Python +- python title: Survival Analysis in Management --- diff --git a/_posts/2024-05-14-Kullback.md b/_posts/2024-05-14-Kullback.md index 06aafbfc..d4b45c4c 100644 --- a/_posts/2024-05-14-Kullback.md +++ b/_posts/2024-05-14-Kullback.md @@ -40,6 +40,7 @@ tags: - Information theory - Data analysis - Python +- python title: Kullback-Leibler and Wasserstein Distances --- diff --git a/_posts/2024-05-14-P_value.md b/_posts/2024-05-14-P_value.md index 256fdba1..ce22994a 100644 --- a/_posts/2024-05-14-P_value.md +++ b/_posts/2024-05-14-P_value.md @@ -14,8 +14,7 @@ header: teaser: /assets/images/data_science_8.jpg twitter_image: /assets/images/data_science_2.jpg seo_type: article -subtitle: A Step-by-Step Guide to Understanding and Calculating the P Value in Statistical - Analysis +subtitle: A Step-by-Step Guide to Understanding and Calculating the P Value in Statistical Analysis tags: - P value - Probability distribution @@ -28,6 +27,7 @@ tags: - Biostatistics - Statistical analysis - Python +- python title: From Data to Probability --- diff --git a/_posts/2024-05-15-Feature_Engineering.md b/_posts/2024-05-15-Feature_Engineering.md index 1680d13e..2fa459c9 100644 --- a/_posts/2024-05-15-Feature_Engineering.md +++ b/_posts/2024-05-15-Feature_Engineering.md @@ -30,6 +30,7 @@ tags: - Genetic algorithms - Model optimization - Python +- python title: Automating Feature Engineering --- diff --git a/_posts/2024-05-15-detect_multivariate_data_drift.md b/_posts/2024-05-15-detect_multivariate_data_drift.md index 0e33a868..83a81eec 100644 --- a/_posts/2024-05-15-detect_multivariate_data_drift.md +++ b/_posts/2024-05-15-detect_multivariate_data_drift.md @@ -29,14 +29,12 @@ keywords: - Statistical methods - Python - Python -seo_description: Learn how to detect multivariate data drift and monitor your machine - learning model's performance using PCA and Reconstruction Error. +- python +seo_description: Learn how to detect multivariate data drift and monitor your machine learning model's performance using PCA and Reconstruction Error. seo_title: Detect Multivariate Data Drift with PCA and Reconstruction Error seo_type: article subtitle: Ensuring Model Accuracy by Monitoring Subtle Changes in Data Structure -summary: A detailed guide on detecting multivariate data drift using Principal Component - Analysis (PCA) and Reconstruction Error to monitor changes in data structure and - ensure model performance in production environments. +summary: A detailed guide on detecting multivariate data drift using Principal Component Analysis (PCA) and Reconstruction Error to monitor changes in data structure and ensure model performance in production environments. tags: - Multivariate data drift - Principal component analysis (pca) @@ -52,6 +50,7 @@ tags: - Production data - Python - Python +- python title: Detect Multivariate Data Drift --- diff --git a/_posts/2024-05-17-Markov_Chain.md b/_posts/2024-05-17-Markov_Chain.md index 6f902319..abbfb1c0 100644 --- a/_posts/2024-05-17-Markov_Chain.md +++ b/_posts/2024-05-17-Markov_Chain.md @@ -22,16 +22,11 @@ keywords: - Parking lot occupancy - Predictive modeling - Markov chains -seo_description: A deep dive into Markov systems, including Markov chains and Hidden - Markov Models, and their applications in real-world scenarios like parking lot occupancy - prediction. +seo_description: A deep dive into Markov systems, including Markov chains and Hidden Markov Models, and their applications in real-world scenarios like parking lot occupancy prediction. seo_title: 'Markov Systems: Foundations and Applications' seo_type: article -subtitle: Exploring the Foundations and Applications of Markov Models in Real-World - Scenarios -summary: This article explores the foundations and real-world applications of Markov - systems, including Markov chains and Hidden Markov Models, in areas such as parking - lot occupancy prediction. +subtitle: Exploring the Foundations and Applications of Markov Models in Real-World Scenarios +summary: This article explores the foundations and real-world applications of Markov systems, including Markov chains and Hidden Markov Models, in areas such as parking lot occupancy prediction. tags: - Markov systems - Markov chains diff --git a/_posts/2024-05-19-gini_coefficiente.md b/_posts/2024-05-19-gini_coefficiente.md index 1a24a58b..65888f10 100644 --- a/_posts/2024-05-19-gini_coefficiente.md +++ b/_posts/2024-05-19-gini_coefficiente.md @@ -15,8 +15,7 @@ header: teaser: /assets/images/data_science_2.jpg twitter_image: /assets/images/data_science_7.jpg seo_type: article -subtitle: Guide to the Normalized Gini Coefficient and Default Rate in Credit Scoring - and Risk Assessment +subtitle: Guide to the Normalized Gini Coefficient and Default Rate in Credit Scoring and Risk Assessment tags: - Gini coefficient - Default rate @@ -36,6 +35,7 @@ tags: - Tensorflow implementation - Loan risk analysis - Python +- python title: Understanding the Normalized Gini Coefficient and Default Rate --- diff --git a/_posts/2024-05-21-Probability_integral_transform.md b/_posts/2024-05-21-Probability_integral_transform.md index eca16f91..91f354cd 100644 --- a/_posts/2024-05-21-Probability_integral_transform.md +++ b/_posts/2024-05-21-Probability_integral_transform.md @@ -28,6 +28,7 @@ tags: - Financial risk management - R - R +- r title: 'Probability Integral Transform: Theory and Applications' --- diff --git a/_posts/2024-05-22-Peer_review.md b/_posts/2024-05-22-Peer_review.md index 70cfa5e3..b3d3fea2 100644 --- a/_posts/2024-05-22-Peer_review.md +++ b/_posts/2024-05-22-Peer_review.md @@ -31,8 +31,7 @@ tags: - Status homophily - Online political behavior - Social media analysis -title: 'Critical Review of ''Bursting the (Filter) Bubble: Interactions of Members - of Parliament on Twitter''' +title: 'Critical Review of ''Bursting the (Filter) Bubble: Interactions of Members of Parliament on Twitter''' --- ## Introduction diff --git a/_posts/2024-06-03-g-test_vs_chi-square_test.md b/_posts/2024-06-03-g-test_vs_chi-square_test.md index 6245bf04..ff955890 100644 --- a/_posts/2024-06-03-g-test_vs_chi-square_test.md +++ b/_posts/2024-06-03-g-test_vs_chi-square_test.md @@ -5,9 +5,7 @@ categories: - Categorical Data Analysis classes: wide date: '2024-06-03' -excerpt: Learn the key differences between the G-Test and Chi-Square Test for analyzing - categorical data, and discover their applications in fields like genetics, market - research, and large datasets. +excerpt: Learn the key differences between the G-Test and Chi-Square Test for analyzing categorical data, and discover their applications in fields like genetics, market research, and large datasets. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -22,14 +20,10 @@ keywords: - Genetic studies - Market research - Large datasets -seo_description: Explore the differences between the G-Test and Chi-Square Test, two - methods for analyzing categorical data, with use cases in genetic studies, market - research, and large datasets. +seo_description: Explore the differences between the G-Test and Chi-Square Test, two methods for analyzing categorical data, with use cases in genetic studies, market research, and large datasets. seo_title: 'G-Test vs. Chi-Square Test: A Comparison for Categorical Data Analysis' seo_type: article -summary: The G-Test and Chi-Square Test are two widely used statistical methods for - analyzing categorical data. This article compares their formulas, assumptions, advantages, - and applications in fields like genetic studies, market research, and large datasets. +summary: The G-Test and Chi-Square Test are two widely used statistical methods for analyzing categorical data. This article compares their formulas, assumptions, advantages, and applications in fields like genetic studies, market research, and large datasets. tags: - G-test - Chi-square test diff --git a/_posts/2024-06-04-poisson_distribution.md b/_posts/2024-06-04-poisson_distribution.md index 90830311..52363071 100644 --- a/_posts/2024-06-04-poisson_distribution.md +++ b/_posts/2024-06-04-poisson_distribution.md @@ -29,6 +29,7 @@ tags: - Statistical testing - R - R +- r title: Modeling Count Events with Poisson Distribution in R --- diff --git a/_posts/2024-06-05-data_science_in_health_tech.md b/_posts/2024-06-05-data_science_in_health_tech.md index 6ca792fa..13d13f07 100644 --- a/_posts/2024-06-05-data_science_in_health_tech.md +++ b/_posts/2024-06-05-data_science_in_health_tech.md @@ -26,14 +26,10 @@ keywords: - Machine learning for health - Healthcare operations improvement - Patient outcomes and ai -seo_description: Discover how data science is revolutionizing healthcare technology - through predictive analytics, machine learning, personalized medicine, and real-time - monitoring to improve patient care and operational efficiency. +seo_description: Discover how data science is revolutionizing healthcare technology through predictive analytics, machine learning, personalized medicine, and real-time monitoring to improve patient care and operational efficiency. seo_title: The Advantages of Data Science in Healthcare Technology seo_type: article -summary: This article explores how data science is transforming healthcare technology, - focusing on predictive analytics, early diagnosis, personalized medicine, and improving - patient outcomes through machine learning and real-time monitoring. +summary: This article explores how data science is transforming healthcare technology, focusing on predictive analytics, early diagnosis, personalized medicine, and improving patient outcomes through machine learning and real-time monitoring. tags: - Data science - Health tech diff --git a/_posts/2024-06-05-sensor_activations_models.md b/_posts/2024-06-05-sensor_activations_models.md index 01457f39..629633da 100644 --- a/_posts/2024-06-05-sensor_activations_models.md +++ b/_posts/2024-06-05-sensor_activations_models.md @@ -22,14 +22,11 @@ keywords: - Residual analysis - Python programming for data analysis - Python -seo_description: Learn how to model sensor activations with the Poisson distribution - in Python. This tutorial covers data preparation, residual analysis, goodness-of-fit, - and cross-validation for accurate predictions. +- python +seo_description: Learn how to model sensor activations with the Poisson distribution in Python. This tutorial covers data preparation, residual analysis, goodness-of-fit, and cross-validation for accurate predictions. seo_title: Modeling Sensor Activations Using Poisson Distribution in Python seo_type: article -summary: This tutorial explores how to model sensor activations using the Poisson - distribution in Python, covering data preparation, model evaluation, residual analysis, - and cross-validation techniques. +summary: This tutorial explores how to model sensor activations using the Poisson distribution in Python, covering data preparation, model evaluation, residual analysis, and cross-validation techniques. tags: - Poisson distribution - Count data @@ -47,6 +44,7 @@ tags: - Python programming - Educational tutorial - Python +- python title: Modeling Sensor Activations with Poisson Distribution in Python --- diff --git a/_posts/2024-06-06-wine_sensory_evaluation.md b/_posts/2024-06-06-wine_sensory_evaluation.md index 110c3de9..b1e9dd22 100644 --- a/_posts/2024-06-06-wine_sensory_evaluation.md +++ b/_posts/2024-06-06-wine_sensory_evaluation.md @@ -29,8 +29,7 @@ tags: - Anova - Regression analysis - Wine quality -title: 'Wine Sensory Evaluation: From Sensory Lexicons and Emotions to Data Statistical - Analysis Techniques' +title: 'Wine Sensory Evaluation: From Sensory Lexicons and Emotions to Data Statistical Analysis Techniques' --- ## Abstract diff --git a/_posts/2024-06-07-z-score.md b/_posts/2024-06-07-z-score.md index eee178ce..07c5e46d 100644 --- a/_posts/2024-06-07-z-score.md +++ b/_posts/2024-06-07-z-score.md @@ -25,14 +25,11 @@ keywords: - Data comparison techniques - R - R -seo_description: Learn the basics of Z-Scores for standardizing data, detecting outliers, - and comparing data points across datasets. This guide offers practical insights - and examples using R programming. +- r +seo_description: Learn the basics of Z-Scores for standardizing data, detecting outliers, and comparing data points across datasets. This guide offers practical insights and examples using R programming. seo_title: 'Data Analysis with Z-Scores: A Quick Guide to Mastering Standard Scores' seo_type: article -summary: This tutorial provides an introduction to Z-Scores, explaining their role - in standardizing data, detecting outliers, and comparing data points across different - datasets, with examples in R programming. +summary: This tutorial provides an introduction to Z-Scores, explaining their role in standardizing data, detecting outliers, and comparing data points across different datasets, with examples in R programming. tags: - Z-score - Standard score @@ -46,6 +43,7 @@ tags: - Normal distribution - R - R +- r title: 'Data Analysis Skills with Z-Scores: A Quick Guide' --- diff --git a/_posts/2024-06-11-survival_analysis.md b/_posts/2024-06-11-survival_analysis.md index d2a11cd7..fa5a8147 100644 --- a/_posts/2024-06-11-survival_analysis.md +++ b/_posts/2024-06-11-survival_analysis.md @@ -28,6 +28,7 @@ tags: - Curve fitting - Medical statistics - Python +- python title: 'Estimating Survival Functions: Parametric and Non-Parametric Approaches' --- diff --git a/_posts/2024-06-13-Stepwise_regression.md b/_posts/2024-06-13-Stepwise_regression.md index 417afd32..f703efca 100644 --- a/_posts/2024-06-13-Stepwise_regression.md +++ b/_posts/2024-06-13-Stepwise_regression.md @@ -27,6 +27,9 @@ tags: - Julia - Statistics - Data science +- python +- r +- julia title: 'Stepwise Regression: Methodology, Applications, and Concerns' --- diff --git a/_posts/2024-06-14-matthew_correlation.md b/_posts/2024-06-14-matthew_correlation.md index 5d96d177..22e19a90 100644 --- a/_posts/2024-06-14-matthew_correlation.md +++ b/_posts/2024-06-14-matthew_correlation.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-06-14' -excerpt: Dive deep into Matthew's Correlation Coefficient (MCC), a powerful metric - for evaluating binary classification models, especially in imbalanced datasets. +excerpt: Dive deep into Matthew's Correlation Coefficient (MCC), a powerful metric for evaluating binary classification models, especially in imbalanced datasets. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_7.jpg @@ -15,7 +14,7 @@ header: twitter_image: /assets/images/data_science_7.jpg keywords: - Mcc -- "Matthew\u2019s correlation coefficient" +- Matthew’s correlation coefficient - Binary classification - Confusion matrix - Model evaluation @@ -32,15 +31,15 @@ keywords: - Fortran - Sh - C -seo_description: "Learn about Matthew\u2019s Correlation Coefficient (MCC), an essential\ - \ metric for evaluating binary classification models, particularly in imbalanced\ - \ datasets, and how it improves upon traditional metrics." -seo_title: "Matthew\u2019s Correlation Coefficient (MCC): A Guide to Binary Classification" +- python +- fortran +- sh +- c +seo_description: Learn about Matthew’s Correlation Coefficient (MCC), an essential metric for evaluating binary classification models, particularly in imbalanced datasets, and how it improves upon traditional metrics. +seo_title: 'Matthew’s Correlation Coefficient (MCC): A Guide to Binary Classification' seo_type: article subtitle: Understanding and Applying MCC in Binary Classification -summary: "This article provides a comprehensive explanation of Matthew\u2019s Correlation\ - \ Coefficient (MCC), its importance in binary classification, and how it compares\ - \ to other performance metrics like accuracy, precision, and recall." +summary: This article provides a comprehensive explanation of Matthew’s Correlation Coefficient (MCC), its importance in binary classification, and how it compares to other performance metrics like accuracy, precision, and recall. tags: - Mcc - Evaluation metrics @@ -59,7 +58,11 @@ tags: - Fortran - Sh - C -title: "Matthew\u2019s Correlation Coefficient (MCC): A Detailed Explanation" +- python +- fortran +- sh +- c +title: 'Matthew’s Correlation Coefficient (MCC): A Detailed Explanation' --- ## Introduction diff --git a/_posts/2024-06-15-EMI_RSSI_SIGNAL.md b/_posts/2024-06-15-EMI_RSSI_SIGNAL.md index 8a26643b..4c0bd379 100644 --- a/_posts/2024-06-15-EMI_RSSI_SIGNAL.md +++ b/_posts/2024-06-15-EMI_RSSI_SIGNAL.md @@ -26,8 +26,7 @@ tags: - Frequency selection - Data quality - Network performance -title: 'Impact of Electromagnetic Interference on RSSI Signal: Detailed Insights and - Implications' +title: 'Impact of Electromagnetic Interference on RSSI Signal: Detailed Insights and Implications' --- Electromagnetic interference (EMI), also known as electrical magnetic distortion, is a phenomenon that can significantly impact the performance of wireless communication systems. One of the key metrics affected by EMI is the Received Signal Strength Indicator (RSSI), which measures the power level of the received signal. diff --git a/_posts/2024-06-29-GLM.md b/_posts/2024-06-29-GLM.md index 024182c4..ec6729bc 100644 --- a/_posts/2024-06-29-GLM.md +++ b/_posts/2024-06-29-GLM.md @@ -27,6 +27,8 @@ tags: - Statistical analysis - Bash - Python +- bash +- python title: Statistical Analysis with Generalized Linear Models --- diff --git a/_posts/2024-06-30-RSSI_body_effects.md b/_posts/2024-06-30-RSSI_body_effects.md index 78cf0b24..2939b3f6 100644 --- a/_posts/2024-06-30-RSSI_body_effects.md +++ b/_posts/2024-06-30-RSSI_body_effects.md @@ -25,14 +25,11 @@ keywords: - Antenna design adjustments - Python - Python -seo_description: Explore how the human body affects RSSI in wireless communication. - Learn about absorption, reflection, shadowing, and practical approaches to mitigate - signal quality issues. +- python +seo_description: Explore how the human body affects RSSI in wireless communication. Learn about absorption, reflection, shadowing, and practical approaches to mitigate signal quality issues. seo_title: 'How the Human Body Affects RSSI: Analysis and Practical Solutions' seo_type: article -summary: This article provides a comprehensive analysis of how the human body impacts - RSSI, covering absorption, reflection, shadowing, and proximity effects, and offering - practical approaches to mitigate signal interference. +summary: This article provides a comprehensive analysis of how the human body impacts RSSI, covering absorption, reflection, shadowing, and proximity effects, and offering practical approaches to mitigate signal interference. tags: - Rssi - Absorption @@ -46,6 +43,7 @@ tags: - Signal quality - Python - Python +- python title: 'How the Human Body Affects RSSI: Detailed Analysis and Practical Approaches' --- diff --git a/_posts/2024-06-30-RSSI_humanbody.md b/_posts/2024-06-30-RSSI_humanbody.md index d14ad466..e0224dc4 100644 --- a/_posts/2024-06-30-RSSI_humanbody.md +++ b/_posts/2024-06-30-RSSI_humanbody.md @@ -4,8 +4,7 @@ categories: - Signal Processing classes: wide date: '2024-06-30' -excerpt: Explore the impact of human presence on RSSI and the challenges it introduces, - along with effective mitigation strategies in wireless communication systems. +excerpt: Explore the impact of human presence on RSSI and the challenges it introduces, along with effective mitigation strategies in wireless communication systems. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -22,14 +21,10 @@ keywords: - Shadowing - Interference - Beamforming -seo_description: Discover how the presence of a human body impacts RSSI in wireless - networks and explore strategies for overcoming challenges like signal attenuation, - interference, and multipath effects. +seo_description: Discover how the presence of a human body impacts RSSI in wireless networks and explore strategies for overcoming challenges like signal attenuation, interference, and multipath effects. seo_title: 'Effects of a Human Body on RSSI: Challenges and Mitigations' seo_type: article -summary: This article examines how human bodies affect Received Signal Strength Indicator - (RSSI), the resulting challenges like signal attenuation and interference, and key - techniques for mitigating these effects. +summary: This article examines how human bodies affect Received Signal Strength Indicator (RSSI), the resulting challenges like signal attenuation and interference, and key techniques for mitigating these effects. tags: - Rssi - Signal attenuation diff --git a/_posts/2024-07-02-monitoring_drift.md b/_posts/2024-07-02-monitoring_drift.md index 336fa597..3754d6f3 100644 --- a/_posts/2024-07-02-monitoring_drift.md +++ b/_posts/2024-07-02-monitoring_drift.md @@ -28,15 +28,11 @@ keywords: - Technology - Python - Python -seo_description: Explore advanced methods for machine learning monitoring by moving - beyond univariate data drift detection. Learn about direct loss estimation, detecting - outliers, and addressing alarm fatigue in production AI systems. +- python +seo_description: Explore advanced methods for machine learning monitoring by moving beyond univariate data drift detection. Learn about direct loss estimation, detecting outliers, and addressing alarm fatigue in production AI systems. seo_title: 'Machine Learning Monitoring: Moving Beyond Univariate Data Drift Detection' seo_type: article -summary: A deep dive into advanced machine learning monitoring techniques that extend - beyond traditional univariate data drift detection. This article covers methods - such as direct loss estimation, outlier detection, and best practices for addressing - alarm fatigue in AI systems deployed in production. +summary: A deep dive into advanced machine learning monitoring techniques that extend beyond traditional univariate data drift detection. This article covers methods such as direct loss estimation, outlier detection, and best practices for addressing alarm fatigue in AI systems deployed in production. tags: - Data drift - Direct loss estimation @@ -53,6 +49,7 @@ tags: - Technology - Python - Python +- python title: 'Machine Learning Monitoring: Moving Beyond Univariate Data Drift Detection' --- diff --git a/_posts/2024-07-03-ancova.md b/_posts/2024-07-03-ancova.md index 07d08126..ca125585 100644 --- a/_posts/2024-07-03-ancova.md +++ b/_posts/2024-07-03-ancova.md @@ -26,6 +26,8 @@ tags: - Generalized estimating equations - R - Unknown +- r +- unknown title: Advanced Non-Parametric ANCOVA and Robust Alternatives --- diff --git a/_posts/2024-07-04-Logram_test.md b/_posts/2024-07-04-Logram_test.md index ea0216f1..3bf0c156 100644 --- a/_posts/2024-07-04-Logram_test.md +++ b/_posts/2024-07-04-Logram_test.md @@ -27,6 +27,8 @@ tags: - Hypothesis testing - Python - R +- python +- r title: Understanding the Logrank Test in Survival Analysis --- diff --git a/_posts/2024-07-05-savitzky_golay.md b/_posts/2024-07-05-savitzky_golay.md index 855818fc..b5c6a4e9 100644 --- a/_posts/2024-07-05-savitzky_golay.md +++ b/_posts/2024-07-05-savitzky_golay.md @@ -27,14 +27,12 @@ keywords: - Unknown - Python - Unknown -seo_description: Learn about smoothing time series data using Moving Averages and - Savitzky-Golay filters. Explore their differences, benefits, and Python implementations - for signal and data processing. +- python +- unknown +seo_description: Learn about smoothing time series data using Moving Averages and Savitzky-Golay filters. Explore their differences, benefits, and Python implementations for signal and data processing. seo_title: 'Time Series Smoothing: Moving Averages vs. Savitzky-Golay Filters' seo_type: article -summary: 'This article compares two popular techniques for smoothing time series data: - Moving Averages and Savitzky-Golay filters, focusing on their applications, benefits, - and implementation in Python.' +summary: 'This article compares two popular techniques for smoothing time series data: Moving Averages and Savitzky-Golay filters, focusing on their applications, benefits, and implementation in Python.' tags: - Time series - Data smoothing @@ -48,6 +46,8 @@ tags: - Unknown - Python - Unknown +- python +- unknown title: 'Smoothing Time Series Data: Moving Averages vs. Savitzky-Golay Filters' --- diff --git a/_posts/2024-07-07-logistic-model.md b/_posts/2024-07-07-logistic-model.md index 3908830f..1ba17583 100644 --- a/_posts/2024-07-07-logistic-model.md +++ b/_posts/2024-07-07-logistic-model.md @@ -27,9 +27,7 @@ keywords: - Machine learning algorithms - Classification models - Predictive modeling -seo_description: A comprehensive guide to Logistic Regression, covering binary classification, - logit models, probability, maximum-likelihood estimation, odds ratios, and the contributions - of Joseph Berkson. Explore its use in machine learning and predictive modeling. +seo_description: A comprehensive guide to Logistic Regression, covering binary classification, logit models, probability, maximum-likelihood estimation, odds ratios, and the contributions of Joseph Berkson. Explore its use in machine learning and predictive modeling. seo_title: 'The Logistic Model: Explained' seo_type: article tags: diff --git a/_posts/2024-07-08-PSOD.md b/_posts/2024-07-08-PSOD.md index 1697537f..a08733eb 100644 --- a/_posts/2024-07-08-PSOD.md +++ b/_posts/2024-07-08-PSOD.md @@ -28,6 +28,7 @@ tags: - Pseudo-labeling - Iterative refinement - Python +- python title: Pseudo-Supervised Outlier Detection --- diff --git a/_posts/2024-07-09-error_bars.md b/_posts/2024-07-09-error_bars.md index 20f3cde4..0b606b9b 100644 --- a/_posts/2024-07-09-error_bars.md +++ b/_posts/2024-07-09-error_bars.md @@ -25,14 +25,10 @@ keywords: - Statistical reporting - Scientific analysis - Error representation in research -seo_description: Learn how error bars represent variability, standard deviation, standard - error, and confidence intervals in scientific research, improving the accuracy and - clarity of reporting findings. +seo_description: Learn how error bars represent variability, standard deviation, standard error, and confidence intervals in scientific research, improving the accuracy and clarity of reporting findings. seo_title: 'Understanding Error Bars: A Guide to Scientific Reporting' seo_type: article -summary: This article explores the significance of error bars in scientific reporting, - focusing on their use in representing variability, standard deviation, standard - error, and confidence intervals in research findings. +summary: This article explores the significance of error bars in scientific reporting, focusing on their use in representing variability, standard deviation, standard error, and confidence intervals in research findings. tags: - Research paper writing - Academic writing tips diff --git a/_posts/2024-07-10-prob_distributions_clinical.md b/_posts/2024-07-10-prob_distributions_clinical.md index f7852d1b..8140fe10 100644 --- a/_posts/2024-07-10-prob_distributions_clinical.md +++ b/_posts/2024-07-10-prob_distributions_clinical.md @@ -19,13 +19,10 @@ keywords: - Binomial distribution - Statistical analysis in healthcare - Trial outcome analysis -seo_description: Learn about common probability distributions used in clinical trials, - including their roles in hypothesis testing and statistical analysis of healthcare - data. +seo_description: Learn about common probability distributions used in clinical trials, including their roles in hypothesis testing and statistical analysis of healthcare data. seo_title: Common Probability Distributions in Clinical Trials seo_type: article -summary: This article explores key probability distributions used in clinical trials, - focusing on their applications in hypothesis testing and outcome analysis. +summary: This article explores key probability distributions used in clinical trials, focusing on their applications in hypothesis testing and outcome analysis. tags: - Probability distributions - Clinical trials diff --git a/_posts/2024-07-11-pre_commit.md b/_posts/2024-07-11-pre_commit.md index 10aa7d4b..13082c54 100644 --- a/_posts/2024-07-11-pre_commit.md +++ b/_posts/2024-07-11-pre_commit.md @@ -19,6 +19,8 @@ tags: - Devops - Bash - Yaml +- bash +- yaml title: Streamlining Your Workflow with Pre-commit Hooks in Python Projects --- diff --git a/_posts/2024-07-13-CLT.md b/_posts/2024-07-13-CLT.md index c3bdc29c..851d250c 100644 --- a/_posts/2024-07-13-CLT.md +++ b/_posts/2024-07-13-CLT.md @@ -15,12 +15,13 @@ header: seo_type: article tags: - Central limit theorem -- "Lindeberg\u2013l\xE9vy clt" +- Lindeberg–lévy clt - Lyapunov clt -- "Lindeberg\u2013feller clt" +- Lindeberg–feller clt - Orey's clt - Prokhorov's theorem - Python +- python title: 'Central Limit Theorems: A Comprehensive Overview' --- diff --git a/_posts/2024-07-14-confidence-intervales.md b/_posts/2024-07-14-confidence-intervales.md index e0807f72..03d788f3 100644 --- a/_posts/2024-07-14-confidence-intervales.md +++ b/_posts/2024-07-14-confidence-intervales.md @@ -18,8 +18,7 @@ tags: - Linear regression - Confidence interval - Prediction interval -title: 'Understanding Uncertainty in Statistical Estimates: Confidence and Prediction - Intervals' +title: 'Understanding Uncertainty in Statistical Estimates: Confidence and Prediction Intervals' --- Statistical estimates always have some uncertainty. Consider a simple example of modeling house prices based solely on their area using linear regression. A prediction from this model wouldn’t reveal the exact value of a house based on its area, because different houses of the same size can have different prices. Instead, the model predicts the mean value related to the outcome for a particular input. diff --git a/_posts/2024-07-14-copulas.md b/_posts/2024-07-14-copulas.md index 2d09a4ca..50670ac5 100644 --- a/_posts/2024-07-14-copulas.md +++ b/_posts/2024-07-14-copulas.md @@ -19,6 +19,7 @@ tags: - Garch - Financial models - Python +- python title: Copula, GARCH, and Other Financial Models --- diff --git a/_posts/2024-07-15-outlier_detection_doping.md b/_posts/2024-07-15-outlier_detection_doping.md index 72dceebf..a8815118 100644 --- a/_posts/2024-07-15-outlier_detection_doping.md +++ b/_posts/2024-07-15-outlier_detection_doping.md @@ -22,20 +22,18 @@ keywords: - Robust data models - Python - Python -seo_description: Learn how to test and evaluate outlier detection models using data - doping techniques. Understand the impact of doping on model performance and outlier - identification. +- python +seo_description: Learn how to test and evaluate outlier detection models using data doping techniques. Understand the impact of doping on model performance and outlier identification. seo_title: Evaluating Outlier Detectors with Data Doping Techniques seo_type: article -summary: This article explores techniques for testing and evaluating outlier detection - models using data doping, highlighting key methodologies and their impact on model - performance. +summary: This article explores techniques for testing and evaluating outlier detection models using data doping, highlighting key methodologies and their impact on model performance. tags: - Outlier detection - Data doping - Model evaluation - Python - Python +- python title: Testing and Evaluating Outlier Detectors Using Doping --- diff --git a/_posts/2024-07-16-Einstein.md b/_posts/2024-07-16-Einstein.md index 33b02d7d..a1da0d1a 100644 --- a/_posts/2024-07-16-Einstein.md +++ b/_posts/2024-07-16-Einstein.md @@ -20,14 +20,10 @@ keywords: - Software development best practices - Scientific research methods - Applying simplicity in technology -seo_description: Explore how Einstein's principle of simplicity influences scientific - research, data analysis, communication, and software development, enhancing clarity - and efficiency across disciplines. +seo_description: Explore how Einstein's principle of simplicity influences scientific research, data analysis, communication, and software development, enhancing clarity and efficiency across disciplines. seo_title: Applying Einstein's Principle of Simplicity in Science, Data, and Software seo_type: article -summary: This article explores how Einstein's principle of simplicity can be applied - across various fields, including scientific research, data analysis, effective communication, - and software development. +summary: This article explores how Einstein's principle of simplicity can be applied across various fields, including scientific research, data analysis, effective communication, and software development. tags: - Einstein - Simplicity diff --git a/_posts/2024-07-19-clt_revisited.md b/_posts/2024-07-19-clt_revisited.md index 52bd14d7..f17eb999 100644 --- a/_posts/2024-07-19-clt_revisited.md +++ b/_posts/2024-07-19-clt_revisited.md @@ -6,9 +6,7 @@ categories: - Mathematical Analysis classes: wide date: '2024-07-19' -excerpt: This article rigorously explores the Central Limit Theorem for m-dependent - random variables under sub-linear expectations, presenting new inequalities, proof - outlines, and implications in modeling dependent sequences. +excerpt: This article rigorously explores the Central Limit Theorem for m-dependent random variables under sub-linear expectations, presenting new inequalities, proof outlines, and implications in modeling dependent sequences. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_8.jpg @@ -20,23 +18,17 @@ keywords: - Central limit theorem - M-dependence - Sub-linear expectations -- "Rosenthal\u2019s inequality" +- Rosenthal’s inequality - Truncated variables -seo_description: A detailed study on the extension of the Central Limit Theorem for - m-dependent random variables under sub-linear expectations, focusing on Rosenthal's - inequality and handling truncated variables. +seo_description: A detailed study on the extension of the Central Limit Theorem for m-dependent random variables under sub-linear expectations, focusing on Rosenthal's inequality and handling truncated variables. seo_title: Central Limit Theorem for m-dependent Random Variables seo_type: article -summary: This article extends the classical Central Limit Theorem (CLT) to m-dependent - random variables within the sub-linear expectation framework. It incorporates Rosenthal's - inequality for m-dependent sequences, examines truncated conditions, and discusses - the broader implications for real-world systems characterized by uncertainty and - dependencies. +summary: This article extends the classical Central Limit Theorem (CLT) to m-dependent random variables within the sub-linear expectation framework. It incorporates Rosenthal's inequality for m-dependent sequences, examines truncated conditions, and discusses the broader implications for real-world systems characterized by uncertainty and dependencies. tags: - Central limit theorem - M-dependence - Sub-linear expectations -- "Rosenthal\u2019s inequality" +- Rosenthal’s inequality title: Central Limit Theorem for m-dependent Random Variables Under Sub-linear Expectations --- diff --git a/_posts/2024-07-20-FPOF.md b/_posts/2024-07-20-FPOF.md index 2cd28af8..340c916e 100644 --- a/_posts/2024-07-20-FPOF.md +++ b/_posts/2024-07-20-FPOF.md @@ -20,6 +20,7 @@ tags: - Unsupervised learning - Data analysis - Python +- python title: 'Frequent Patterns Outlier Factor ' --- diff --git a/_posts/2024-07-20-sequential_change.md b/_posts/2024-07-20-sequential_change.md index d422968c..c339510b 100644 --- a/_posts/2024-07-20-sequential_change.md +++ b/_posts/2024-07-20-sequential_change.md @@ -19,6 +19,7 @@ tags: - Structural changes - Real-time processing - Python +- python title: Sequential Detection of Switches in Models with Changing Structures --- diff --git a/_posts/2024-07-21-iknn.md b/_posts/2024-07-21-iknn.md index 72131c05..391d33ca 100644 --- a/_posts/2024-07-21-iknn.md +++ b/_posts/2024-07-21-iknn.md @@ -18,6 +18,7 @@ tags: - Iknn - Python - Python +- python title: 'Introducing ikNN: An Interpretable k Nearest Neighbors Model' --- diff --git a/_posts/2024-07-31-Custom_libraries.md b/_posts/2024-07-31-Custom_libraries.md index 5962694f..f053d204 100644 --- a/_posts/2024-07-31-Custom_libraries.md +++ b/_posts/2024-07-31-Custom_libraries.md @@ -21,6 +21,7 @@ tags: - Software development - Automation - Python +- python title: Building Custom Python Libraries for Your Industry Needs --- diff --git a/_posts/2024-08-01-Data_leakeage.md b/_posts/2024-08-01-Data_leakeage.md index 8a6e214f..48a6b367 100644 --- a/_posts/2024-08-01-Data_leakeage.md +++ b/_posts/2024-08-01-Data_leakeage.md @@ -17,6 +17,7 @@ tags: - Data science - Model integrity - Python +- python title: 'Understanding Data Leakage in Machine Learning: Causes, Types, and Prevention' --- diff --git a/_posts/2024-08-03-feature_engineering.md b/_posts/2024-08-03-feature_engineering.md index 44bc5c4c..5035ae30 100644 --- a/_posts/2024-08-03-feature_engineering.md +++ b/_posts/2024-08-03-feature_engineering.md @@ -5,9 +5,7 @@ categories: - Data Science classes: wide date: '2024-08-03' -excerpt: Discover the importance of feature engineering in enhancing machine learning - models. Learn essential techniques for transforming raw data into valuable inputs - that drive better predictive performance. +excerpt: Discover the importance of feature engineering in enhancing machine learning models. Learn essential techniques for transforming raw data into valuable inputs that drive better predictive performance. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_1.jpg @@ -24,13 +22,11 @@ keywords: - Predictive analytics - Python - Python -seo_description: Explore powerful feature engineering techniques that boost the performance - of machine learning models by improving data preprocessing and feature selection. +- python +seo_description: Explore powerful feature engineering techniques that boost the performance of machine learning models by improving data preprocessing and feature selection. seo_title: Feature Engineering for Better Machine Learning Models seo_type: article -summary: This article delves into various feature engineering techniques essential - for improving machine learning model performance. It covers data preprocessing, - feature selection, transformation methods, and tips to enhance predictive accuracy. +summary: This article delves into various feature engineering techniques essential for improving machine learning model performance. It covers data preprocessing, feature selection, transformation methods, and tips to enhance predictive accuracy. tags: - Feature engineering - Data preprocessing @@ -39,6 +35,7 @@ tags: - Model performance - Python - Python +- python title: Feature Engineering Techniques for Improved Machine Learning --- diff --git a/_posts/2024-08-15-structural_equations.md b/_posts/2024-08-15-structural_equations.md index 83db72fa..ce985cb4 100644 --- a/_posts/2024-08-15-structural_equations.md +++ b/_posts/2024-08-15-structural_equations.md @@ -5,9 +5,7 @@ categories: - Research Methods classes: wide date: '2024-08-15' -excerpt: Learn the fundamentals of Structural Equation Modeling (SEM) with latent - variables. This guide covers measurement models, path analysis, factor loadings, - and more for researchers and statisticians. +excerpt: Learn the fundamentals of Structural Equation Modeling (SEM) with latent variables. This guide covers measurement models, path analysis, factor loadings, and more for researchers and statisticians. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_6.jpg @@ -24,14 +22,10 @@ keywords: - Variance-covariance matrix - Measurement models - Exogenous and endogenous variables -seo_description: Explore a detailed guide on Structural Equation Modeling (SEM) with - latent variables, including path analysis, measurement models, and techniques for - handling exogenous and endogenous variables. +seo_description: Explore a detailed guide on Structural Equation Modeling (SEM) with latent variables, including path analysis, measurement models, and techniques for handling exogenous and endogenous variables. seo_title: Guide to Structural Equation Modeling with Latent Variables seo_type: article -summary: This comprehensive guide explains the key concepts and techniques of Structural - Equation Modeling (SEM) with latent variables. It includes path analysis, factor - loadings, variance-covariance matrices, and handling endogenous and exogenous variables. +summary: This comprehensive guide explains the key concepts and techniques of Structural Equation Modeling (SEM) with latent variables. It includes path analysis, factor loadings, variance-covariance matrices, and handling endogenous and exogenous variables. tags: - Structural equation modeling (sem) - Latent variables diff --git a/_posts/2024-08-16-utility_functions_python.md b/_posts/2024-08-16-utility_functions_python.md index a8b145e9..79cb8be7 100644 --- a/_posts/2024-08-16-utility_functions_python.md +++ b/_posts/2024-08-16-utility_functions_python.md @@ -6,9 +6,7 @@ categories: - Software Development classes: wide date: '2024-08-16' -excerpt: Learn how to design and implement utility classes in Python. This guide covers - best practices, real-world examples, and tips for building reusable, efficient code - using object-oriented programming. +excerpt: Learn how to design and implement utility classes in Python. This guide covers best practices, real-world examples, and tips for building reusable, efficient code using object-oriented programming. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_7.jpg @@ -24,14 +22,11 @@ keywords: - Software development - Design patterns - Python -seo_description: Explore the design and implementation of Python utility classes. - This article provides examples, best practices, and insights for creating reusable - components using object-oriented programming. +- python +seo_description: Explore the design and implementation of Python utility classes. This article provides examples, best practices, and insights for creating reusable components using object-oriented programming. seo_title: 'Python Utility Classes: Design and Implementation Guide' seo_type: article -summary: This article provides a deep dive into Python utility classes, discussing - their design, best practices, and implementation. It covers object-oriented programming - principles and shows how to build reusable and efficient utility classes in Python. +summary: This article provides a deep dive into Python utility classes, discussing their design, best practices, and implementation. It covers object-oriented programming principles and shows how to build reusable and efficient utility classes in Python. tags: - Python - Utility classes @@ -39,6 +34,7 @@ tags: - Code reusability - Software design patterns - Python +- python title: 'Python Utility Classes: Best Practices and Examples' --- diff --git a/_posts/2024-08-19-pre_comit_tutorial.md b/_posts/2024-08-19-pre_comit_tutorial.md index eba2da7b..2fd68bf7 100644 --- a/_posts/2024-08-19-pre_comit_tutorial.md +++ b/_posts/2024-08-19-pre_comit_tutorial.md @@ -6,9 +6,7 @@ categories: - Version Control classes: wide date: '2024-08-19' -excerpt: Learn how to use pre-commit tools in Python to enforce code quality and consistency - before committing changes. This guide covers the setup, configuration, and best - practices for using Git hooks to streamline your workflow. +excerpt: Learn how to use pre-commit tools in Python to enforce code quality and consistency before committing changes. This guide covers the setup, configuration, and best practices for using Git hooks to streamline your workflow. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_1.jpg @@ -25,15 +23,12 @@ keywords: - Python development workflow - Bash - Yaml -seo_description: Explore pre-commit tools in Python for ensuring code quality and - managing Git hooks. Learn how to integrate automated checks into your development - process to improve code consistency. +- bash +- yaml +seo_description: Explore pre-commit tools in Python for ensuring code quality and managing Git hooks. Learn how to integrate automated checks into your development process to improve code consistency. seo_title: 'Pre-Commit Tools in Python: Best Practices and Guide' seo_type: article -summary: This guide provides an in-depth overview of pre-commit tools in Python, covering - how to set up and configure them to improve code quality and automate Git hooks. - It includes best practices for using pre-commit to ensure consistency and streamline - the development process. +summary: This guide provides an in-depth overview of pre-commit tools in Python, covering how to set up and configure them to improve code quality and automate Git hooks. It includes best practices for using pre-commit to ensure consistency and streamline the development process. tags: - Python - Pre-commit @@ -43,6 +38,8 @@ tags: - Automation - Bash - Yaml +- bash +- yaml title: A Comprehensive Guide to Pre-Commit Tools in Python --- diff --git a/_posts/2024-08-24-circular_economy.md b/_posts/2024-08-24-circular_economy.md index 00b28fcd..bf249c87 100644 --- a/_posts/2024-08-24-circular_economy.md +++ b/_posts/2024-08-24-circular_economy.md @@ -6,9 +6,7 @@ categories: - Circular Economy classes: wide date: '2024-08-24' -excerpt: Explore how Python and network analysis can be used to implement and optimize - circular economy models. Learn how systems thinking and data science tools can drive - sustainability and resource efficiency. +excerpt: Explore how Python and network analysis can be used to implement and optimize circular economy models. Learn how systems thinking and data science tools can drive sustainability and resource efficiency. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_7.jpg @@ -25,14 +23,11 @@ keywords: - Resource efficiency - Python - Python -seo_description: Learn to implement circular economy models using Python and network - analysis techniques. This guide covers how data science and systems thinking can - promote sustainability and resource management. +- python +seo_description: Learn to implement circular economy models using Python and network analysis techniques. This guide covers how data science and systems thinking can promote sustainability and resource management. seo_title: Circular Economy Models with Python and Network Analysis seo_type: article -summary: This article explores the implementation of circular economy models using - Python and network analysis. It focuses on how data science and systems thinking - can be applied to improve resource efficiency, sustainability, and waste reduction. +summary: This article explores the implementation of circular economy models using Python and network analysis. It focuses on how data science and systems thinking can be applied to improve resource efficiency, sustainability, and waste reduction. tags: - Python - Network analysis @@ -42,6 +37,7 @@ tags: - Resource efficiency - Python - Python +- python title: Implementing Circular Economy Models with Python and Network Analysis --- diff --git a/_posts/2024-08-24-kruskal_wallis.md b/_posts/2024-08-24-kruskal_wallis.md index faeba976..eff4e90f 100644 --- a/_posts/2024-08-24-kruskal_wallis.md +++ b/_posts/2024-08-24-kruskal_wallis.md @@ -5,9 +5,7 @@ categories: - Data Analysis classes: wide date: '2024-08-24' -excerpt: Discover the Kruskal-Wallis Test, a powerful non-parametric statistical method - used for comparing multiple groups. Learn when and how to apply it in data analysis - where assumptions of normality don't hold. +excerpt: Discover the Kruskal-Wallis Test, a powerful non-parametric statistical method used for comparing multiple groups. Learn when and how to apply it in data analysis where assumptions of normality don't hold. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_7.jpg @@ -25,15 +23,12 @@ keywords: - Python - R - Python -seo_description: Explore the Kruskal-Wallis Test, a non-parametric alternative to - ANOVA for comparing independent samples. Understand its applications, assumptions, - and how to interpret results in data analysis. +- r +- python +seo_description: Explore the Kruskal-Wallis Test, a non-parametric alternative to ANOVA for comparing independent samples. Understand its applications, assumptions, and how to interpret results in data analysis. seo_title: 'Kruskal-Wallis Test: Guide to Non-Parametric Statistical Analysis' seo_type: article -summary: This comprehensive guide explains the Kruskal-Wallis Test, a non-parametric - statistical method ideal for comparing multiple independent samples without assuming - normal distribution. It discusses when to use the test, its assumptions, and how - to interpret the results in data analysis. +summary: This comprehensive guide explains the Kruskal-Wallis Test, a non-parametric statistical method ideal for comparing multiple independent samples without assuming normal distribution. It discusses when to use the test, its assumptions, and how to interpret the results in data analysis. tags: - Kruskal-wallis test - Non-parametric methods @@ -44,6 +39,8 @@ tags: - Python - R - Python +- r +- python title: 'The Kruskal-Wallis Test: A Comprehensive Guide to Non-Parametric Analysis' --- diff --git a/_posts/2024-08-25-Vehicle_Routing_Problem.md b/_posts/2024-08-25-Vehicle_Routing_Problem.md index 9bf50a89..3e845425 100644 --- a/_posts/2024-08-25-Vehicle_Routing_Problem.md +++ b/_posts/2024-08-25-Vehicle_Routing_Problem.md @@ -6,9 +6,7 @@ categories: - Logistics classes: wide date: '2024-08-25' -excerpt: Learn how to solve the Vehicle Routing Problem (VRP) using Python and optimization - algorithms. This guide covers strategies for efficient transportation and logistics - solutions. +excerpt: Learn how to solve the Vehicle Routing Problem (VRP) using Python and optimization algorithms. This guide covers strategies for efficient transportation and logistics solutions. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_8.jpg @@ -27,15 +25,12 @@ keywords: - Python - Bash - Python -seo_description: Explore how to implement solutions for the Vehicle Routing Problem - (VRP) using Python. This article covers optimization techniques and algorithms for - transportation and logistics management. +- bash +- python +seo_description: Explore how to implement solutions for the Vehicle Routing Problem (VRP) using Python. This article covers optimization techniques and algorithms for transportation and logistics management. seo_title: 'Vehicle Routing Problem Solutions with Python: Optimization Guide' seo_type: article -summary: This comprehensive guide explains how to solve the Vehicle Routing Problem - (VRP) using Python. It covers key optimization algorithms and their applications - in transportation, logistics, and supply chain management to improve operational - efficiency. +summary: This comprehensive guide explains how to solve the Vehicle Routing Problem (VRP) using Python. It covers key optimization algorithms and their applications in transportation, logistics, and supply chain management to improve operational efficiency. tags: - Vehicle routing problem - Python @@ -47,6 +42,8 @@ tags: - Python - Bash - Python +- bash +- python title: Implementing Vehicle Routing Problem Solutions with Python --- diff --git a/_posts/2024-08-26-energie.md b/_posts/2024-08-26-energie.md index fe56efde..ead8941f 100644 --- a/_posts/2024-08-26-energie.md +++ b/_posts/2024-08-26-energie.md @@ -4,9 +4,7 @@ categories: - Energy Management classes: wide date: '2024-08-26' -excerpt: Explore energy optimization strategies for production facilities to reduce - costs and improve efficiency. This model incorporates cogeneration plants, machine - flexibility, and operational adjustments for maximum savings. +excerpt: Explore energy optimization strategies for production facilities to reduce costs and improve efficiency. This model incorporates cogeneration plants, machine flexibility, and operational adjustments for maximum savings. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_4.jpg @@ -23,15 +21,11 @@ keywords: - Operational flexibility - Python - Python -seo_description: Learn how to implement energy optimization models in production facilities, - focusing on reducing energy costs, improving efficiency, and leveraging optimization - algorithms for operational flexibility. +- python +seo_description: Learn how to implement energy optimization models in production facilities, focusing on reducing energy costs, improving efficiency, and leveraging optimization algorithms for operational flexibility. seo_title: 'Energy Optimization in Production Facilities: Cost-Saving Models' seo_type: article -summary: This article provides an in-depth look at energy optimization models designed - for production facilities. It covers key strategies such as cogeneration plants, - machine flexibility, and optimization algorithms to reduce energy costs and enhance - production efficiency. +summary: This article provides an in-depth look at energy optimization models designed for production facilities. It covers key strategies such as cogeneration plants, machine flexibility, and optimization algorithms to reduce energy costs and enhance production efficiency. tags: - Energy optimization - Production facility @@ -45,6 +39,7 @@ tags: - Production efficiency - Python - Python +- python title: 'Energy Optimization for a Production Facility: A Model for Cost Savings' --- diff --git a/_posts/2024-08-27-coeeficient_variation.md b/_posts/2024-08-27-coeeficient_variation.md index 1aec7734..3018d6d4 100644 --- a/_posts/2024-08-27-coeeficient_variation.md +++ b/_posts/2024-08-27-coeeficient_variation.md @@ -5,9 +5,7 @@ categories: - Data Analysis classes: wide date: '2024-08-27' -excerpt: Learn how to calculate and interpret the Coefficient of Variation (CV), a - crucial statistical measure of relative variability. This guide explores its applications - and limitations in various data analysis contexts. +excerpt: Learn how to calculate and interpret the Coefficient of Variation (CV), a crucial statistical measure of relative variability. This guide explores its applications and limitations in various data analysis contexts. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_2.jpg @@ -24,15 +22,11 @@ keywords: - Interpreting data variability - Rust - Rust -seo_description: Explore the Coefficient of Variation (CV) as a statistical tool for - assessing variability. Understand its advantages and limitations in data interpretation - and analysis. +- rust +seo_description: Explore the Coefficient of Variation (CV) as a statistical tool for assessing variability. Understand its advantages and limitations in data interpretation and analysis. seo_title: 'Coefficient of Variation: A Guide to Applications and Limitations' seo_type: article -summary: This article explains the Coefficient of Variation (CV), a statistical measure - used to compare variability across datasets. It discusses its applications in fields - like economics, biology, and finance, as well as its limitations when interpreting - data with different units or scales. +summary: This article explains the Coefficient of Variation (CV), a statistical measure used to compare variability across datasets. It discusses its applications in fields like economics, biology, and finance, as well as its limitations when interpreting data with different units or scales. tags: - Coefficient of variation - Statistical measures @@ -41,6 +35,7 @@ tags: - Relative standard deviation - Rust - Rust +- rust title: 'Understanding the Coefficient of Variation: Applications and Limitations' --- diff --git a/_posts/2024-08-28-mathematics.md b/_posts/2024-08-28-mathematics.md index 459b7196..8784e26c 100644 --- a/_posts/2024-08-28-mathematics.md +++ b/_posts/2024-08-28-mathematics.md @@ -7,9 +7,7 @@ categories: - Society classes: wide date: '2024-08-28' -excerpt: Explore how mathematics shapes modern society across fields like technology, - education, and problem-solving. This article delves into the often overlooked impact - of mathematics on innovation and societal progress. +excerpt: Explore how mathematics shapes modern society across fields like technology, education, and problem-solving. This article delves into the often overlooked impact of mathematics on innovation and societal progress. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_2.jpg @@ -24,15 +22,10 @@ keywords: - Technology and math - Societal impact of mathematics - Mathematical thinking -seo_description: Discover the critical role mathematics plays in modern society, from - technological advancements to its foundational importance in education. Learn how - math drives innovation and impacts societal development. +seo_description: Discover the critical role mathematics plays in modern society, from technological advancements to its foundational importance in education. Learn how math drives innovation and impacts societal development. seo_title: 'The Power of Mathematics in Modern Society: Technology and Education' seo_type: article -summary: This article highlights the undervalued role of mathematics in modern society, - focusing on its contributions to technology, education, and societal progress. It - discusses how mathematical thinking underpins innovation, problem-solving, and advancements - across various industries. +summary: This article highlights the undervalued role of mathematics in modern society, focusing on its contributions to technology, education, and societal progress. It discusses how mathematical thinking underpins innovation, problem-solving, and advancements across various industries. tags: - Mathematics - Technology diff --git a/_posts/2024-08-31-PAPE.md b/_posts/2024-08-31-PAPE.md index ea018475..3a30ca6d 100644 --- a/_posts/2024-08-31-PAPE.md +++ b/_posts/2024-08-31-PAPE.md @@ -6,9 +6,7 @@ categories: - Model Performance classes: wide date: '2024-08-31' -excerpt: Explore adaptive performance estimation techniques in machine learning, including - methods like CBPE and PAPE. Learn how these approaches help monitor model performance - and detect issues like data drift and covariate shift. +excerpt: Explore adaptive performance estimation techniques in machine learning, including methods like CBPE and PAPE. Learn how these approaches help monitor model performance and detect issues like data drift and covariate shift. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_9.jpg @@ -24,15 +22,10 @@ keywords: - Data drift detection - Covariate shift management - Model performance tracking -seo_description: Learn about adaptive performance estimation in machine learning with - a focus on methods like CBPE and PAPE. Understand how to manage performance monitoring, - data drift, and covariate shift for better model outcomes. +seo_description: Learn about adaptive performance estimation in machine learning with a focus on methods like CBPE and PAPE. Understand how to manage performance monitoring, data drift, and covariate shift for better model outcomes. seo_title: 'Adaptive Machine Learning Performance Estimation: CBPE and PAPE' seo_type: article -summary: This article dives into adaptive performance estimation techniques in machine - learning, comparing methods such as Confidence-Based Performance Estimation (CBPE) - and Predictive Adaptive Performance Estimation (PAPE). It covers their roles in - detecting data drift, covariate shift, and maintaining optimal model performance. +summary: This article dives into adaptive performance estimation techniques in machine learning, comparing methods such as Confidence-Based Performance Estimation (CBPE) and Predictive Adaptive Performance Estimation (PAPE). It covers their roles in detecting data drift, covariate shift, and maintaining optimal model performance. tags: - Machine learning - Performance monitoring diff --git a/_posts/2024-08-31-pedestrian_movement.md b/_posts/2024-08-31-pedestrian_movement.md index eeadb0d7..fc9873ed 100644 --- a/_posts/2024-08-31-pedestrian_movement.md +++ b/_posts/2024-08-31-pedestrian_movement.md @@ -5,9 +5,7 @@ categories: - Simulation Models classes: wide date: '2024-08-31' -excerpt: Explore the simulation of pedestrian evacuation in environments impacted - by smoke. This guide covers key models such as the Social Force Model and Advection-Diffusion - Equation to assess evacuation efficiency under smoke propagation conditions. +excerpt: Explore the simulation of pedestrian evacuation in environments impacted by smoke. This guide covers key models such as the Social Force Model and Advection-Diffusion Equation to assess evacuation efficiency under smoke propagation conditions. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_6.jpg @@ -25,15 +23,13 @@ keywords: - Bash - Python - Fortran -seo_description: Learn how to simulate pedestrian evacuation in smoke-affected environments - using the Social Force Model and Advection-Diffusion Equation. Explore numerical - methods to optimize emergency preparedness strategies. +- bash +- python +- fortran +seo_description: Learn how to simulate pedestrian evacuation in smoke-affected environments using the Social Force Model and Advection-Diffusion Equation. Explore numerical methods to optimize emergency preparedness strategies. seo_title: Pedestrian Evacuation Simulation in Smoke-Affected Environments seo_type: article -summary: This article examines simulation models for pedestrian evacuation in smoke-affected - environments. It focuses on the Social Force Model, smoke propagation dynamics through - the Advection-Diffusion Equation, and numerical methods for optimizing evacuation - strategies during emergencies. +summary: This article examines simulation models for pedestrian evacuation in smoke-affected environments. It focuses on the Social Force Model, smoke propagation dynamics through the Advection-Diffusion Equation, and numerical methods for optimizing evacuation strategies during emergencies. tags: - Pedestrian evacuation - Smoke propagation @@ -44,6 +40,9 @@ tags: - Bash - Python - Fortran +- bash +- python +- fortran title: Simulating Pedestrian Evacuation in Smoke-Affected Environments --- diff --git a/_posts/2024-09-01-graph_theory.md b/_posts/2024-09-01-graph_theory.md index 4b897057..fbdc54e0 100644 --- a/_posts/2024-09-01-graph_theory.md +++ b/_posts/2024-09-01-graph_theory.md @@ -5,9 +5,7 @@ categories: - Supply Chain Management classes: wide date: '2024-09-01' -excerpt: Explore how graph theory is applied to optimize production systems and supply - chains. Learn how network optimization and resource allocation techniques improve - efficiency and streamline operations. +excerpt: Explore how graph theory is applied to optimize production systems and supply chains. Learn how network optimization and resource allocation techniques improve efficiency and streamline operations. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_4.jpg @@ -22,15 +20,10 @@ keywords: - Supply chain management - Optimization strategies - Production systems efficiency -seo_description: Discover the role of graph theory in optimizing production systems - and supply chains. This article covers network optimization, resource allocation, - and key strategies for improving operational efficiency. +seo_description: Discover the role of graph theory in optimizing production systems and supply chains. This article covers network optimization, resource allocation, and key strategies for improving operational efficiency. seo_title: Graph Theory in Production Systems and Supply Chain Optimization seo_type: article -summary: This article examines the practical applications of graph theory in optimizing - production systems and supply chains. It focuses on network optimization and resource - allocation techniques that enhance operational efficiency and decision-making in - supply chain management. +summary: This article examines the practical applications of graph theory in optimizing production systems and supply chains. It focuses on network optimization and resource allocation techniques that enhance operational efficiency and decision-making in supply chain management. tags: - Graph theory - Network optimization diff --git a/_posts/2024-09-01-math_and_music.md b/_posts/2024-09-01-math_and_music.md index e2b92f67..37dcc3af 100644 --- a/_posts/2024-09-01-math_and_music.md +++ b/_posts/2024-09-01-math_and_music.md @@ -6,9 +6,7 @@ categories: - Technology classes: wide date: '2024-09-01' -excerpt: Discover how mathematics influences electronic music creation through sound - synthesis, rhythm, and algorithmic composition. Explore the role of numbers in shaping - digital signal processing and generative music. +excerpt: Discover how mathematics influences electronic music creation through sound synthesis, rhythm, and algorithmic composition. Explore the role of numbers in shaping digital signal processing and generative music. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_6.jpg @@ -23,15 +21,10 @@ keywords: - Digital signal processing - Generative music - Rhythm and numbers -seo_description: Explore how mathematics drives electronic music production, from - sound synthesis to algorithmic composition. Learn how numbers shape rhythm, signal - processing, and generative music. +seo_description: Explore how mathematics drives electronic music production, from sound synthesis to algorithmic composition. Learn how numbers shape rhythm, signal processing, and generative music. seo_title: 'The Role of Mathematics in Electronic Music: Sound, Rhythm, and Composition' seo_type: article -summary: This article explores the intersection of mathematics and electronic music, - highlighting how algorithms and mathematical principles influence sound synthesis, - rhythm, and generative music creation. It delves into the technical aspects of digital - signal processing and algorithmic composition in music technology. +summary: This article explores the intersection of mathematics and electronic music, highlighting how algorithms and mathematical principles influence sound synthesis, rhythm, and generative music creation. It delves into the technical aspects of digital signal processing and algorithmic composition in music technology. tags: - Sound synthesis - Algorithmic composition diff --git a/_posts/2024-09-03-climate_change.md b/_posts/2024-09-03-climate_change.md index f654dedb..eb1a0348 100644 --- a/_posts/2024-09-03-climate_change.md +++ b/_posts/2024-09-03-climate_change.md @@ -7,8 +7,7 @@ categories: - Technology classes: wide date: '2024-09-03' -excerpt: Discover how data science is transforming the fight against climate change - with new methods for understanding and reducing global warming impacts. +excerpt: Discover how data science is transforming the fight against climate change with new methods for understanding and reducing global warming impacts. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_4.jpg @@ -25,19 +24,10 @@ keywords: - Machine learning in climate change - Environmental science - Policy-making -seo_description: Explore how data science is driving innovation in climate modeling, - risk assessment, and policy-making to mitigate global warming. Learn about the latest - applications of machine learning and data analysis in tackling the climate crisis. +seo_description: Explore how data science is driving innovation in climate modeling, risk assessment, and policy-making to mitigate global warming. Learn about the latest applications of machine learning and data analysis in tackling the climate crisis. seo_title: 'Data Science and Climate Change: Solutions for Global Warming' seo_type: article -summary: As the climate crisis intensifies, data science has emerged as a key player - in understanding and mitigating global warming. This article delves into how cutting-edge - techniques such as climate modeling, machine learning, and data analysis are transforming - our ability to assess climate risks and inform policy decisions. From renewable - energy forecasting to advanced risk assessment strategies, data science is providing - powerful tools to combat climate change. Explore the innovative ways in which technology - is shaping the future of environmental science and policy-making, helping us tackle - one of the greatest challenges of our time. +summary: As the climate crisis intensifies, data science has emerged as a key player in understanding and mitigating global warming. This article delves into how cutting-edge techniques such as climate modeling, machine learning, and data analysis are transforming our ability to assess climate risks and inform policy decisions. From renewable energy forecasting to advanced risk assessment strategies, data science is providing powerful tools to combat climate change. Explore the innovative ways in which technology is shaping the future of environmental science and policy-making, helping us tackle one of the greatest challenges of our time. tags: - Climate modeling - Data analysis @@ -45,8 +35,7 @@ tags: - Risk assessment - Policy-making - Machine learning -title: 'Data Science and the Climate Crisis: Innovative Approaches to Understanding - and Mitigating Global Warming' +title: 'Data Science and the Climate Crisis: Innovative Approaches to Understanding and Mitigating Global Warming' --- ## Introduction diff --git a/_posts/2024-09-03-fundamentals_matter.md b/_posts/2024-09-03-fundamentals_matter.md index 74d956a4..65fecaa2 100644 --- a/_posts/2024-09-03-fundamentals_matter.md +++ b/_posts/2024-09-03-fundamentals_matter.md @@ -5,8 +5,7 @@ categories: - Technology classes: wide date: '2024-09-03' -excerpt: Learn why a deep understanding of machine learning fundamentals is more valuable - than expertise in specific tools and frameworks. +excerpt: Learn why a deep understanding of machine learning fundamentals is more valuable than expertise in specific tools and frameworks. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_5.jpg @@ -23,20 +22,10 @@ keywords: - Machine learning success - Adapting to new tools - Technology in machine learning -seo_description: Understand why mastering the fundamentals of machine learning is - more critical than focusing on specific tools. Explore key principles that drive - successful machine learning projects. +seo_description: Understand why mastering the fundamentals of machine learning is more critical than focusing on specific tools. Explore key principles that drive successful machine learning projects. seo_title: 'Machine Learning Fundamentals vs Tools: What Matters Most' seo_type: article -summary: "Machine learning has become one of the most influential fields in technology\ - \ today, with new tools and frameworks constantly emerging. However, despite the\ - \ rapid development of sophisticated software, it's the foundational principles\ - \ of machine learning that ultimately determine success. In this article, we explore\ - \ why a strong grasp of the fundamentals\u2014such as algorithms, data preprocessing,\ - \ and model evaluation\u2014matters more than expertise in any specific tool. By\ - \ understanding these core concepts, data scientists and engineers can adapt to\ - \ new tools and technologies more effectively, leading to better outcomes in their\ - \ machine learning projects." +summary: Machine learning has become one of the most influential fields in technology today, with new tools and frameworks constantly emerging. However, despite the rapid development of sophisticated software, it's the foundational principles of machine learning that ultimately determine success. In this article, we explore why a strong grasp of the fundamentals—such as algorithms, data preprocessing, and model evaluation—matters more than expertise in any specific tool. By understanding these core concepts, data scientists and engineers can adapt to new tools and technologies more effectively, leading to better outcomes in their machine learning projects. tags: - Machine learning - Fundamentals diff --git a/_posts/2024-09-04-outlier_detection.md b/_posts/2024-09-04-outlier_detection.md index 92e317b3..85abd52b 100644 --- a/_posts/2024-09-04-outlier_detection.md +++ b/_posts/2024-09-04-outlier_detection.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-09-04' -excerpt: Explore the intricacies of outlier detection using distance metrics and metric - learning techniques. This article delves into methods such as Random Forests and - distance metric learning to improve outlier detection accuracy. +excerpt: Explore the intricacies of outlier detection using distance metrics and metric learning techniques. This article delves into methods such as Random Forests and distance metric learning to improve outlier detection accuracy. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_4.jpg @@ -23,15 +21,11 @@ keywords: - Anomaly detection methods - Machine learning outlier techniques - Python -seo_description: Learn about outlier detection techniques in machine learning, focusing - on distance metrics and metric learning. Discover how these methods enhance the - accuracy of detecting anomalies and outliers. +- python +seo_description: Learn about outlier detection techniques in machine learning, focusing on distance metrics and metric learning. Discover how these methods enhance the accuracy of detecting anomalies and outliers. seo_title: 'Outlier Detection in Machine Learning: Exploring Distance Metric Learning' seo_type: article -summary: This comprehensive guide explores outlier detection using distance metrics - and metric learning techniques. It highlights the role of algorithms such as Random - Forests and distance metric learning in identifying anomalies and improving detection - accuracy in machine learning models. +summary: This comprehensive guide explores outlier detection using distance metrics and metric learning techniques. It highlights the role of algorithms such as Random Forests and distance metric learning in identifying anomalies and improving detection accuracy in machine learning models. tags: - Outlier detection - Distance metrics @@ -39,6 +33,7 @@ tags: - Distance metric learning - Anomaly detection - Python +- python title: 'Understanding Outlier Detection: A Deep Dive into Distance Metric Learning' --- diff --git a/_posts/2024-09-05-detecting_drift.md b/_posts/2024-09-05-detecting_drift.md index 0118f4bc..0b7f4b28 100644 --- a/_posts/2024-09-05-detecting_drift.md +++ b/_posts/2024-09-05-detecting_drift.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-09-05' -excerpt: Explore the challenges of using traditional hypothesis testing for detecting - data drift in machine learning models and learn how Bayesian probability offers - a more robust alternative for monitoring data shifts. +excerpt: Explore the challenges of using traditional hypothesis testing for detecting data drift in machine learning models and learn how Bayesian probability offers a more robust alternative for monitoring data shifts. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_9.jpg @@ -22,23 +20,17 @@ keywords: - Data monitoring in machine learning - Bayesian methods in data science - Model adaptation and data drift -seo_description: Understand why hypothesis testing may fall short for detecting data - drift and explore how Bayesian probability provides a better framework for monitoring - and adapting to data shifts in machine learning models. +seo_description: Understand why hypothesis testing may fall short for detecting data drift and explore how Bayesian probability provides a better framework for monitoring and adapting to data shifts in machine learning models. seo_title: 'Data Drift Detection: Limitations of Hypothesis Testing and Bayesian Alternatives' seo_type: article -summary: This article explores the limitations of using hypothesis testing to detect - data drift in machine learning models. It introduces Bayesian probability as an - alternative approach, offering a more flexible and adaptive method for monitoring - data shifts and maintaining model performance. +summary: This article explores the limitations of using hypothesis testing to detect data drift in machine learning models. It introduces Bayesian probability as an alternative approach, offering a more flexible and adaptive method for monitoring data shifts and maintaining model performance. tags: - Data drift - Hypothesis testing - Bayesian probability - Data monitoring - Model adaptation -title: 'The Limitations of Hypothesis Testing for Detecting Data Drift: A Bayesian - Alternative' +title: 'The Limitations of Hypothesis Testing for Detecting Data Drift: A Bayesian Alternative' --- With statistics at the heart of data science, hypothesis testing is a logical first step for detecting data drift. The fundamental idea behind hypothesis testing is straightforward: define a null hypothesis that assumes no drift in the data, then use the p-value to determine whether this hypothesis should be rejected. However, when applied to detecting data drift in production environments, traditional hypothesis testing can be unreliable and potentially misleading. This article explores the limitations of hypothesis testing for this purpose and suggests Bayesian probability as a more effective alternative. diff --git a/_posts/2024-09-05-real_time_data_streaming.md b/_posts/2024-09-05-real_time_data_streaming.md index 98ea283b..27d2cc6e 100644 --- a/_posts/2024-09-05-real_time_data_streaming.md +++ b/_posts/2024-09-05-real_time_data_streaming.md @@ -5,9 +5,7 @@ categories: - Real-time Processing classes: wide date: '2024-09-05' -excerpt: Learn how to implement real-time data streaming using Python and Apache Kafka. - This guide covers key concepts, setup, and best practices for managing data streams - in real-time processing pipelines. +excerpt: Learn how to implement real-time data streaming using Python and Apache Kafka. This guide covers key concepts, setup, and best practices for managing data streams in real-time processing pipelines. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_6.jpg @@ -26,15 +24,12 @@ keywords: - Python - Bash - Python -seo_description: Explore real-time data streaming using Python and Apache Kafka. This - article explains the setup, core concepts, and best practices for creating efficient - real-time data processing pipelines. +- bash +- python +seo_description: Explore real-time data streaming using Python and Apache Kafka. This article explains the setup, core concepts, and best practices for creating efficient real-time data processing pipelines. seo_title: Real-time Data Streaming with Python and Apache Kafka seo_type: article -summary: This article provides a comprehensive guide to implementing real-time data - streaming using Python and Apache Kafka. It explains how to set up Kafka, stream - data efficiently, and manage real-time data pipelines in Python, with a focus on - best practices for data engineering. +summary: This article provides a comprehensive guide to implementing real-time data streaming using Python and Apache Kafka. It explains how to set up Kafka, stream data efficiently, and manage real-time data pipelines in Python, with a focus on best practices for data engineering. tags: - Apache kafka - Python @@ -45,6 +40,8 @@ tags: - Python - Bash - Python +- bash +- python title: Real-time Data Streaming using Python and Kafka --- diff --git a/_posts/2024-09-06-covariate_shift.md b/_posts/2024-09-06-covariate_shift.md index 79596ff0..ef3b1726 100644 --- a/_posts/2024-09-06-covariate_shift.md +++ b/_posts/2024-09-06-covariate_shift.md @@ -3,9 +3,7 @@ author_profile: false categories: - Machine Learning date: '2024-09-06' -excerpt: Learn how to manage covariate shifts in machine learning models through effective - model monitoring, feature engineering, and adaptation strategies to maintain model - accuracy and performance. +excerpt: Learn how to manage covariate shifts in machine learning models through effective model monitoring, feature engineering, and adaptation strategies to maintain model accuracy and performance. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_8.jpg @@ -20,15 +18,10 @@ keywords: - Model adaptation strategies - Managing data drift in machine learning - Maintaining model accuracy -seo_description: Explore techniques for managing covariate shifts in machine learning, - including model monitoring, feature engineering, and model adaptation. Learn how - to mitigate data drift and maintain model performance. +seo_description: Explore techniques for managing covariate shifts in machine learning, including model monitoring, feature engineering, and model adaptation. Learn how to mitigate data drift and maintain model performance. seo_title: 'Managing Covariate Shifts in Machine Learning: Strategies for Model Adaptation' seo_type: article -summary: This article covers strategies for managing covariate shifts in machine learning - models. It explains how to monitor models, adapt to changing data distributions, - and implement feature engineering to address data drift and ensure continued model - performance. +summary: This article covers strategies for managing covariate shifts in machine learning models. It explains how to monitor models, adapt to changing data distributions, and implement feature engineering to address data drift and ensure continued model performance. tags: - Covariate shift - Model monitoring diff --git a/_posts/2024-09-06-normality.md b/_posts/2024-09-06-normality.md index 606b048d..589a299a 100644 --- a/_posts/2024-09-06-normality.md +++ b/_posts/2024-09-06-normality.md @@ -6,9 +6,7 @@ categories: - Machine Learning classes: wide date: '2024-09-06' -excerpt: Explore the complexity of real-world data distributions beyond the normal - distribution. Learn about log-normal distributions, heavy-tailed phenomena, and - how the Central Limit Theorem and Extreme Value Theory influence data analysis. +excerpt: Explore the complexity of real-world data distributions beyond the normal distribution. Learn about log-normal distributions, heavy-tailed phenomena, and how the Central Limit Theorem and Extreme Value Theory influence data analysis. header: image: /assets/images/data_science_1.jpg og_image: /assets/images/data_science_9.jpg @@ -23,15 +21,10 @@ keywords: - Central limit theorem applications - Extreme value theory - Statistical analysis beyond normality -seo_description: Discover the intricacies of real-world data distributions, including - heavy-tailed distributions, the Central Limit Theorem, and Extreme Value Theory. - Learn how these concepts affect statistical analysis and machine learning. +seo_description: Discover the intricacies of real-world data distributions, including heavy-tailed distributions, the Central Limit Theorem, and Extreme Value Theory. Learn how these concepts affect statistical analysis and machine learning. seo_title: 'Beyond Normal Distributions: Exploring Real-World Data Complexity' seo_type: article -summary: This article delves into the complexity of real-world data distributions, - moving beyond the assumptions of normality. It covers the importance of log-normal - and heavy-tailed distributions, the Central Limit Theorem, and the application of - Extreme Value Theory in data analysis. +summary: This article delves into the complexity of real-world data distributions, moving beyond the assumptions of normality. It covers the importance of log-normal and heavy-tailed distributions, the Central Limit Theorem, and the application of Extreme Value Theory in data analysis. tags: - Normal distribution - Central limit theorem diff --git a/_posts/2024-09-06-sequential_detection_switches.md b/_posts/2024-09-06-sequential_detection_switches.md index 780516c9..a36b9542 100644 --- a/_posts/2024-09-06-sequential_detection_switches.md +++ b/_posts/2024-09-06-sequential_detection_switches.md @@ -6,9 +6,7 @@ categories: - Data Analysis classes: wide date: '2024-09-06' -excerpt: Learn about sequential detection techniques for identifying switches in models - with changing structures. Explore methods for detecting structural changes in time-series - data and dynamic systems. +excerpt: Learn about sequential detection techniques for identifying switches in models with changing structures. Explore methods for detecting structural changes in time-series data and dynamic systems. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_4.jpg @@ -23,15 +21,10 @@ keywords: - Time-series analysis - Dynamic systems modeling - Model structure shifts -seo_description: Discover sequential detection methods for identifying structural - changes in models. Learn how to apply change-point detection and sequential analysis - in dynamic systems and time-series data. +seo_description: Discover sequential detection methods for identifying structural changes in models. Learn how to apply change-point detection and sequential analysis in dynamic systems and time-series data. seo_title: 'Sequential Detection of Structural Changes in Models: Techniques and Methods' seo_type: article -summary: This article explores sequential detection techniques used for identifying - switches in models with changing structures. It focuses on methods like change-point - detection and sequential analysis, particularly in time-series data and dynamic - systems. +summary: This article explores sequential detection techniques used for identifying switches in models with changing structures. It focuses on methods like change-point detection and sequential analysis, particularly in time-series data and dynamic systems. tags: - Change-point detection - Sequential analysis diff --git a/_posts/2024-09-07-energie_efficiency.md b/_posts/2024-09-07-energie_efficiency.md index 5082d2bd..e74a0e90 100644 --- a/_posts/2024-09-07-energie_efficiency.md +++ b/_posts/2024-09-07-energie_efficiency.md @@ -6,9 +6,7 @@ categories: - Sustainability classes: wide date: '2024-09-07' -excerpt: Explore how Python and machine learning can be applied to analyze and improve - building energy efficiency. Learn key techniques for assessing sustainability, optimizing - energy usage, and reducing carbon footprints. +excerpt: Explore how Python and machine learning can be applied to analyze and improve building energy efficiency. Learn key techniques for assessing sustainability, optimizing energy usage, and reducing carbon footprints. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_6.jpg @@ -24,15 +22,11 @@ keywords: - Sustainable building practices - Carbon footprint reduction - Python -seo_description: Learn how to apply machine learning techniques and Python to building - energy efficiency analysis. This article focuses on optimizing energy usage, sustainability, - and reducing environmental impact. +- python +seo_description: Learn how to apply machine learning techniques and Python to building energy efficiency analysis. This article focuses on optimizing energy usage, sustainability, and reducing environmental impact. seo_title: Building Energy Efficiency Analysis with Python and Machine Learning seo_type: article -summary: This article covers the application of Python and machine learning to analyze - building energy efficiency. It explores techniques for optimizing energy consumption, - improving sustainability, and reducing carbon footprints, helping to create more - energy-efficient structures. +summary: This article covers the application of Python and machine learning to analyze building energy efficiency. It explores techniques for optimizing energy consumption, improving sustainability, and reducing carbon footprints, helping to create more energy-efficient structures. tags: - Energy efficiency - Python @@ -40,6 +34,7 @@ tags: - Building analysis - Sustainability - Python +- python title: Building Energy Efficiency Analysis with Python and Machine Learning --- diff --git a/_posts/2024-09-08-nonparametric_tests.md b/_posts/2024-09-08-nonparametric_tests.md index 77da321e..2cf00e24 100644 --- a/_posts/2024-09-08-nonparametric_tests.md +++ b/_posts/2024-09-08-nonparametric_tests.md @@ -4,9 +4,7 @@ categories: - Statistics classes: wide date: '2024-09-08' -excerpt: Explore the full potential of nonparametric tests, going beyond the Mann-Whitney - Test. Learn how techniques like quantile regression and other nonparametric methods - offer robust alternatives in statistical analysis. +excerpt: Explore the full potential of nonparametric tests, going beyond the Mann-Whitney Test. Learn how techniques like quantile regression and other nonparametric methods offer robust alternatives in statistical analysis. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_3.jpg @@ -24,15 +22,14 @@ keywords: - Bash - Ruby - Python -seo_description: Discover the real power of nonparametric tests, moving beyond Mann-Whitney - to explore quantile regression and other robust statistical techniques for data - analysis without distributional assumptions. +- r +- bash +- ruby +- python +seo_description: Discover the real power of nonparametric tests, moving beyond Mann-Whitney to explore quantile regression and other robust statistical techniques for data analysis without distributional assumptions. seo_title: 'Nonparametric Tests Beyond Mann-Whitney: Unlocking Statistical Power' seo_type: article -summary: This article explores the broader landscape of nonparametric tests, focusing - on methods that go beyond the Mann-Whitney Test. It covers powerful techniques like - quantile regression and highlights how these approaches are used for robust statistical - analysis without strict distributional assumptions. +summary: This article explores the broader landscape of nonparametric tests, focusing on methods that go beyond the Mann-Whitney Test. It covers powerful techniques like quantile regression and highlights how these approaches are used for robust statistical analysis without strict distributional assumptions. tags: - Nonparametric tests - Quantile regression @@ -42,6 +39,10 @@ tags: - Bash - Ruby - Python +- r +- bash +- ruby +- python title: 'The Real Power of Nonparametric Tests: Beyond Mann-Whitney' --- diff --git a/_posts/2024-09-09-kmeans.md b/_posts/2024-09-09-kmeans.md index 6f95305a..8d64503a 100644 --- a/_posts/2024-09-09-kmeans.md +++ b/_posts/2024-09-09-kmeans.md @@ -5,9 +5,7 @@ categories: - Data Science classes: wide date: '2024-09-09' -excerpt: KMeans is widely used, but it's not always the best clustering algorithm - for your data. Explore alternative methods like Gaussian Mixture Models and other - clustering techniques to improve your machine learning results. +excerpt: KMeans is widely used, but it's not always the best clustering algorithm for your data. Explore alternative methods like Gaussian Mixture Models and other clustering techniques to improve your machine learning results. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_7.jpg @@ -22,15 +20,10 @@ keywords: - Unsupervised learning techniques - Better clustering methods - Machine learning clustering -seo_description: Learn why KMeans may not always be the best choice for clustering. - Explore alternatives like Gaussian Mixture Models and other advanced algorithms - for better results in unsupervised learning. +seo_description: Learn why KMeans may not always be the best choice for clustering. Explore alternatives like Gaussian Mixture Models and other advanced algorithms for better results in unsupervised learning. seo_title: 'Alternatives to KMeans: Exploring Clustering Algorithms in Machine Learning' seo_type: article -summary: This article discusses the limitations of KMeans as a clustering algorithm - and introduces alternatives such as Gaussian Mixture Models and other clustering - techniques. It provides insights into when to move beyond KMeans for better performance - in unsupervised learning tasks. +summary: This article discusses the limitations of KMeans as a clustering algorithm and introduces alternatives such as Gaussian Mixture Models and other clustering techniques. It provides insights into when to move beyond KMeans for better performance in unsupervised learning tasks. tags: - Kmeans - Clustering algorithms diff --git a/_posts/2024-09-10-wilcoxon.md b/_posts/2024-09-10-wilcoxon.md index d0089f9b..deaea8a1 100644 --- a/_posts/2024-09-10-wilcoxon.md +++ b/_posts/2024-09-10-wilcoxon.md @@ -5,9 +5,7 @@ categories: - Data Analysis classes: wide date: '2024-09-10' -excerpt: Learn about the Wilcoxon Signed-Rank Test, a robust non-parametric method - for comparing paired samples, especially useful when data is skewed or contains - outliers. +excerpt: Learn about the Wilcoxon Signed-Rank Test, a robust non-parametric method for comparing paired samples, especially useful when data is skewed or contains outliers. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -24,15 +22,12 @@ keywords: - Statistical analysis for outliers - R - Python -seo_description: Explore the Wilcoxon Signed-Rank Test, a non-parametric alternative - to the paired t-test, suitable for skewed data, outliers, and small sample sizes - in statistical analysis. +- r +- python +seo_description: Explore the Wilcoxon Signed-Rank Test, a non-parametric alternative to the paired t-test, suitable for skewed data, outliers, and small sample sizes in statistical analysis. seo_title: 'Wilcoxon Signed-Rank Test: Non-Parametric Alternative to Paired T-Test' seo_type: article -summary: This article explores the Wilcoxon Signed-Rank Test, a non-parametric alternative - to the paired t-test. It explains how this test is ideal for analyzing paired data - when assumptions of normality are violated, such as with skewed data, outliers, - or small sample sizes. +summary: This article explores the Wilcoxon Signed-Rank Test, a non-parametric alternative to the paired t-test. It explains how this test is ideal for analyzing paired data when assumptions of normality are violated, such as with skewed data, outliers, or small sample sizes. tags: - Wilcoxon signed-rank test - Non-parametric tests @@ -41,8 +36,9 @@ tags: - Robust statistical methods - R - Python -title: 'Understanding the Wilcoxon Signed-Rank Test: A Non-Parametric Alternative - to the Paired T-Test' +- r +- python +title: 'Understanding the Wilcoxon Signed-Rank Test: A Non-Parametric Alternative to the Paired T-Test' --- ## The Wilcoxon Signed-Rank Test: An Overview diff --git a/_posts/2024-09-11-cross_validation.md b/_posts/2024-09-11-cross_validation.md index 1233c5f0..db8b2e52 100644 --- a/_posts/2024-09-11-cross_validation.md +++ b/_posts/2024-09-11-cross_validation.md @@ -5,9 +5,7 @@ categories: - Data Science classes: wide date: '2024-09-11' -excerpt: An exploration of cross-validation techniques in machine learning, focusing - on methods to evaluate and enhance model performance while mitigating overfitting - risks. +excerpt: An exploration of cross-validation techniques in machine learning, focusing on methods to evaluate and enhance model performance while mitigating overfitting risks. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_8.jpg @@ -22,15 +20,10 @@ keywords: - Preventing overfitting - Machine learning model validation - Data science methodologies -seo_description: Explore various cross-validation techniques in machine learning, - their importance, and how they help ensure robust model performance by minimizing - overfitting. +seo_description: Explore various cross-validation techniques in machine learning, their importance, and how they help ensure robust model performance by minimizing overfitting. seo_title: Cross-Validation Techniques for Robust Machine Learning Models seo_type: article -summary: Cross-validation is a critical technique in machine learning for assessing - model performance and preventing overfitting. This article covers key cross-validation - methods, including k-fold, stratified, and leave-one-out cross-validation, and discusses - their role in building reliable and generalizable machine learning models. +summary: Cross-validation is a critical technique in machine learning for assessing model performance and preventing overfitting. This article covers key cross-validation methods, including k-fold, stratified, and leave-one-out cross-validation, and discusses their role in building reliable and generalizable machine learning models. tags: - Cross-validation - Model evaluation diff --git a/_posts/2024-09-12-importance_sampling.md b/_posts/2024-09-12-importance_sampling.md index 081fe334..8f0a5e07 100644 --- a/_posts/2024-09-12-importance_sampling.md +++ b/_posts/2024-09-12-importance_sampling.md @@ -5,9 +5,7 @@ categories: - Risk Management classes: wide date: '2024-09-12' -excerpt: Importance Sampling offers an efficient alternative to traditional Monte - Carlo simulations for portfolio credit risk estimation by focusing on rare, significant - loss events. +excerpt: Importance Sampling offers an efficient alternative to traditional Monte Carlo simulations for portfolio credit risk estimation by focusing on rare, significant loss events. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_1.jpg @@ -27,15 +25,14 @@ keywords: - R - Ruby - Rust -seo_description: Learn how Importance Sampling enhances Monte Carlo simulations in - estimating portfolio credit risk, especially in the context of copula models and - rare events. +- python +- r +- ruby +- rust +seo_description: Learn how Importance Sampling enhances Monte Carlo simulations in estimating portfolio credit risk, especially in the context of copula models and rare events. seo_title: Importance Sampling for Portfolio Credit Risk seo_type: article -summary: Importance Sampling is an advanced technique used to improve the efficiency - of Monte Carlo simulations in estimating portfolio credit risk. By focusing computational - resources on rare but impactful loss events, it enhances the accuracy of risk predictions, - particularly when working with complex copula models. +summary: Importance Sampling is an advanced technique used to improve the efficiency of Monte Carlo simulations in estimating portfolio credit risk. By focusing computational resources on rare but impactful loss events, it enhances the accuracy of risk predictions, particularly when working with complex copula models. tags: - Importance sampling - Monte carlo simulation @@ -46,6 +43,10 @@ tags: - R - Ruby - Rust +- python +- r +- ruby +- rust title: Importance Sampling for Portfolio Credit Risk --- diff --git a/_posts/2024-09-13-multi_colinearity.md b/_posts/2024-09-13-multi_colinearity.md index 3e8737bd..889f3f86 100644 --- a/_posts/2024-09-13-multi_colinearity.md +++ b/_posts/2024-09-13-multi_colinearity.md @@ -4,8 +4,7 @@ categories: - Statistics classes: wide date: '2024-09-13' -excerpt: Multicollinearity is a common issue in regression analysis. Learn about its - implications, misconceptions, and techniques to manage it in statistical modeling. +excerpt: Multicollinearity is a common issue in regression analysis. Learn about its implications, misconceptions, and techniques to manage it in statistical modeling. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_2.jpg @@ -21,16 +20,10 @@ keywords: - Ridge regression - Statistical modeling - Regression diagnostics -seo_description: An in-depth exploration of multicollinearity in regression analysis, - its consequences, common misconceptions, identification techniques, and methods - to address it. +seo_description: An in-depth exploration of multicollinearity in regression analysis, its consequences, common misconceptions, identification techniques, and methods to address it. seo_title: Understanding Multicollinearity in Regression Models seo_type: article -summary: Multicollinearity occurs when independent variables in a regression model - are highly correlated, leading to unreliable coefficient estimates. This article - explores the causes and consequences of multicollinearity, clarifies misconceptions, - and discusses various techniques, such as variance inflation factor (VIF) and ridge - regression, to detect and mitigate its effects. +summary: Multicollinearity occurs when independent variables in a regression model are highly correlated, leading to unreliable coefficient estimates. This article explores the causes and consequences of multicollinearity, clarifies misconceptions, and discusses various techniques, such as variance inflation factor (VIF) and ridge regression, to detect and mitigate its effects. tags: - Multicollinearity - Regression analysis diff --git a/_posts/2024-09-14-ML_supply_chain.md b/_posts/2024-09-14-ML_supply_chain.md index 2bc30e03..5c08dc00 100644 --- a/_posts/2024-09-14-ML_supply_chain.md +++ b/_posts/2024-09-14-ML_supply_chain.md @@ -6,9 +6,7 @@ categories: - Operations Management classes: wide date: '2024-09-14' -excerpt: Learn how machine learning optimizes supply chain operations by enhancing - demand forecasting, inventory management, logistics, and more, driving efficiency - and business value. +excerpt: Learn how machine learning optimizes supply chain operations by enhancing demand forecasting, inventory management, logistics, and more, driving efficiency and business value. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_7.jpg @@ -24,16 +22,10 @@ keywords: - Logistics optimization - Operations management - Predictive analytics in supply chain -seo_description: Explore how machine learning can optimize supply chain operations, - enhance efficiency, and drive business value through demand forecasting, inventory - management, and logistics. +seo_description: Explore how machine learning can optimize supply chain operations, enhance efficiency, and drive business value through demand forecasting, inventory management, and logistics. seo_title: 'Machine Learning in Supply Chain: Optimization and Efficiency' seo_type: article -summary: Machine learning is revolutionizing supply chain management by optimizing - key processes such as demand forecasting, inventory management, and logistics. This - article explores how machine learning models improve operational efficiency, reduce - costs, and drive business value through data-driven decision-making in supply chain - operations. +summary: Machine learning is revolutionizing supply chain management by optimizing key processes such as demand forecasting, inventory management, and logistics. This article explores how machine learning models improve operational efficiency, reduce costs, and drive business value through data-driven decision-making in supply chain operations. tags: - Supply chain - Machine learning diff --git a/_posts/2024-09-15-forest_fiers.md b/_posts/2024-09-15-forest_fiers.md index bb99f49f..5d07356e 100644 --- a/_posts/2024-09-15-forest_fiers.md +++ b/_posts/2024-09-15-forest_fiers.md @@ -6,9 +6,7 @@ categories: - Disaster Management classes: wide date: '2024-09-15' -excerpt: This article delves into the role of machine learning in managing forest - fires in Portugal, offering a detailed analysis of early detection, risk assessment, - and strategic response, with a focus on the challenges posed by eucalyptus forests. +excerpt: This article delves into the role of machine learning in managing forest fires in Portugal, offering a detailed analysis of early detection, risk assessment, and strategic response, with a focus on the challenges posed by eucalyptus forests. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_7.jpg @@ -24,16 +22,10 @@ keywords: - Environmental protection - Disaster management - Forest fire detection in portugal -seo_description: Explore how machine learning enhances forest fire management in Portugal, - addressing early detection, risk assessment, and the impact of eucalyptus plantations. -seo_title: 'Machine Learning and Forest Fires: Insights from Portugal''s Wildfire - Management' +seo_description: Explore how machine learning enhances forest fire management in Portugal, addressing early detection, risk assessment, and the impact of eucalyptus plantations. +seo_title: 'Machine Learning and Forest Fires: Insights from Portugal''s Wildfire Management' seo_type: article -summary: Machine learning plays a vital role in improving forest fire management in - Portugal by enhancing early detection, risk assessment, and response strategies. - This article explores the challenges specific to Portugal, particularly the prevalence - of eucalyptus forests, and how data-driven approaches are transforming fire prevention - and control efforts. +summary: Machine learning plays a vital role in improving forest fire management in Portugal by enhancing early detection, risk assessment, and response strategies. This article explores the challenges specific to Portugal, particularly the prevalence of eucalyptus forests, and how data-driven approaches are transforming fire prevention and control efforts. tags: - Forest fires - Machine learning diff --git a/_posts/2024-09-16-ML_and_forest_fires.md b/_posts/2024-09-16-ML_and_forest_fires.md index 23ad0ddb..c3ffcaf8 100644 --- a/_posts/2024-09-16-ML_and_forest_fires.md +++ b/_posts/2024-09-16-ML_and_forest_fires.md @@ -6,9 +6,7 @@ categories: - Disaster Management classes: wide date: '2024-09-16' -excerpt: Machine learning is revolutionizing forest fire management through advanced - models, real-time data integration, and emerging technologies like IoT and blockchain, - offering a holistic and adaptive strategy for combating forest fires. +excerpt: Machine learning is revolutionizing forest fire management through advanced models, real-time data integration, and emerging technologies like IoT and blockchain, offering a holistic and adaptive strategy for combating forest fires. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_6.jpg @@ -27,11 +25,8 @@ keywords: - Ai for disaster response - Blockchain for environmental monitoring - Sustainable technologies for fire management -seo_description: Explore advanced machine learning applications for forest fire management, - including deep learning, big data integration, IoT, and ethical considerations for - a holistic approach. -seo_title: 'Machine Learning in Forest Fire Management: Advanced Applications and - Holistic Strategies' +seo_description: Explore advanced machine learning applications for forest fire management, including deep learning, big data integration, IoT, and ethical considerations for a holistic approach. +seo_title: 'Machine Learning in Forest Fire Management: Advanced Applications and Holistic Strategies' seo_type: article tags: - Forest fires diff --git a/_posts/2024-09-17-feature_engenniring.md b/_posts/2024-09-17-feature_engenniring.md index f78d488a..97ef684d 100644 --- a/_posts/2024-09-17-feature_engenniring.md +++ b/_posts/2024-09-17-feature_engenniring.md @@ -5,9 +5,7 @@ categories: - Data Science classes: wide date: '2024-09-17' -excerpt: Feature engineering is crucial in machine learning, but it's easy to make - mistakes that lead to inaccurate models. This article highlights five common pitfalls - and provides strategies to avoid them. +excerpt: Feature engineering is crucial in machine learning, but it's easy to make mistakes that lead to inaccurate models. This article highlights five common pitfalls and provides strategies to avoid them. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_2.jpg @@ -27,9 +25,8 @@ keywords: - Robust feature engineering - Data cleaning for machine learning - Python -seo_description: Explore five common mistakes in feature engineering, including data - leakage and over-engineering, and learn how to avoid them for more robust machine - learning models. +- python +seo_description: Explore five common mistakes in feature engineering, including data leakage and over-engineering, and learn how to avoid them for more robust machine learning models. seo_title: Avoiding 5 Common Feature Engineering Mistakes in Machine Learning seo_type: article tags: @@ -37,6 +34,7 @@ tags: - Data preprocessing - Machine learning - Python +- python title: 5 Common Mistakes in Feature Engineering and How to Avoid Them --- diff --git a/_posts/2024-09-17-ml_healthcare.md b/_posts/2024-09-17-ml_healthcare.md index ff3d5263..59f320e4 100644 --- a/_posts/2024-09-17-ml_healthcare.md +++ b/_posts/2024-09-17-ml_healthcare.md @@ -6,9 +6,7 @@ categories: - Data Analytics classes: wide date: '2024-09-17' -excerpt: Discover how machine learning is revolutionizing healthcare analytics, from - predictive patient outcomes to personalized medicine, and the challenges faced in - integrating ML into healthcare. +excerpt: Discover how machine learning is revolutionizing healthcare analytics, from predictive patient outcomes to personalized medicine, and the challenges faced in integrating ML into healthcare. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_3.jpg @@ -27,20 +25,10 @@ keywords: - Clinical implementation challenges - Predictive patient outcomes - Real-time medical data analysis -seo_description: Explore the impact of machine learning on healthcare analytics, including - advancements in predictive patient outcomes, personalized medicine, medical imaging, - and the challenges of integrating ML into healthcare systems. -seo_title: How Machine Learning is Revolutionizing Healthcare Analytics for Improved - Patient Care +seo_description: Explore the impact of machine learning on healthcare analytics, including advancements in predictive patient outcomes, personalized medicine, medical imaging, and the challenges of integrating ML into healthcare systems. +seo_title: How Machine Learning is Revolutionizing Healthcare Analytics for Improved Patient Care seo_type: article -summary: Machine learning is reshaping healthcare analytics by enabling advanced predictive - models, personalized treatment plans, and real-time analysis of medical data. This - article highlights how ML is being applied in critical areas such as predictive - patient outcomes, medical imaging, and personalized medicine. It also explores the - challenges of integrating machine learning into healthcare systems, including data - privacy concerns, interpretability issues, and the complexity of clinical implementation. - With its potential to enhance patient care and optimize resource allocation, machine - learning is poised to revolutionize the healthcare industry. +summary: Machine learning is reshaping healthcare analytics by enabling advanced predictive models, personalized treatment plans, and real-time analysis of medical data. This article highlights how ML is being applied in critical areas such as predictive patient outcomes, medical imaging, and personalized medicine. It also explores the challenges of integrating machine learning into healthcare systems, including data privacy concerns, interpretability issues, and the complexity of clinical implementation. With its potential to enhance patient care and optimize resource allocation, machine learning is poised to revolutionize the healthcare industry. tags: - Healthcare analytics - Machine learning diff --git a/_posts/2024-09-18-baysean_statistics.md b/_posts/2024-09-18-baysean_statistics.md index f7de089a..d0358659 100644 --- a/_posts/2024-09-18-baysean_statistics.md +++ b/_posts/2024-09-18-baysean_statistics.md @@ -5,9 +5,7 @@ categories: - Statistics classes: wide date: '2024-09-18' -excerpt: Unlock the power of Bayesian statistics in machine learning through probabilistic - reasoning, offering insights into model uncertainty, predictive distributions, and - real-world applications. +excerpt: Unlock the power of Bayesian statistics in machine learning through probabilistic reasoning, offering insights into model uncertainty, predictive distributions, and real-world applications. header: image: /assets/images/bayes_stats_1.png og_image: /assets/images/data_science_7.jpg @@ -26,20 +24,10 @@ keywords: - Probabilistic programming - Bayesian networks - Uncertainty quantification -seo_description: Explore Bayesian statistics in machine learning, highlighting probabilistic - reasoning, uncertainty quantification, and practical applications across various - domains. +seo_description: Explore Bayesian statistics in machine learning, highlighting probabilistic reasoning, uncertainty quantification, and practical applications across various domains. seo_title: Demystifying Bayesian Statistics in Machine Learning seo_type: article -summary: Bayesian statistics provides a powerful framework for dealing with uncertainty - in machine learning models, making it essential for building robust predictive systems. - This article explores the principles of Bayesian inference, probabilistic reasoning, - and how these concepts apply to machine learning. It delves into practical tools - such as Markov Chain Monte Carlo (MCMC) methods and probabilistic programming, demonstrating - how Bayesian approaches enhance model interpretability and predictive accuracy. - Whether it's for uncertainty quantification or developing Bayesian networks, this - guide offers valuable insights into the real-world applications of Bayesian statistics - in AI. +summary: Bayesian statistics provides a powerful framework for dealing with uncertainty in machine learning models, making it essential for building robust predictive systems. This article explores the principles of Bayesian inference, probabilistic reasoning, and how these concepts apply to machine learning. It delves into practical tools such as Markov Chain Monte Carlo (MCMC) methods and probabilistic programming, demonstrating how Bayesian approaches enhance model interpretability and predictive accuracy. Whether it's for uncertainty quantification or developing Bayesian networks, this guide offers valuable insights into the real-world applications of Bayesian statistics in AI. tags: - Bayesian statistics - Probabilistic reasoning diff --git a/_posts/2024-09-19-build_ds_team.md b/_posts/2024-09-19-build_ds_team.md index 992288c6..aaef6cbb 100644 --- a/_posts/2024-09-19-build_ds_team.md +++ b/_posts/2024-09-19-build_ds_team.md @@ -6,9 +6,7 @@ categories: - Organizational Behavior classes: wide date: '2024-09-19' -excerpt: Discover the implications of assigning different job titles in data science - teams, examining how uniform or specialized titles affect team unity, role clarity, - and individual motivation. +excerpt: Discover the implications of assigning different job titles in data science teams, examining how uniform or specialized titles affect team unity, role clarity, and individual motivation. header: image: /assets/images/data_team.png og_image: /assets/images/data_science_5.jpg @@ -27,21 +25,11 @@ keywords: - Career development - Employee motivation - Team management -seo_description: Explore the pros and cons of assigning uniform versus specialized - job titles in data science teams. Learn how job titles impact team dynamics, collaboration, - and organizational success. -seo_title: 'Uniform vs. Specialized Job Titles in Data Science Teams: Impact and Best - Practices' +seo_description: Explore the pros and cons of assigning uniform versus specialized job titles in data science teams. Learn how job titles impact team dynamics, collaboration, and organizational success. +seo_title: 'Uniform vs. Specialized Job Titles in Data Science Teams: Impact and Best Practices' seo_type: article -subtitle: Exploring the Impact of Uniform vs. Specialized Job Titles in Data Science - Teams -summary: This article explores the debate on whether data science teams should assign - uniform or specialized job titles to team members such as software engineers and - machine learning researchers. It examines the arguments for and against both approaches, - considering factors like team unity, role clarity, individual motivation, and organizational - culture. By analyzing the impact of job titles on team dynamics and performance, - the article provides recommendations to help organizations make informed decisions - that align with their strategic goals and foster a productive work environment. +subtitle: Exploring the Impact of Uniform vs. Specialized Job Titles in Data Science Teams +summary: This article explores the debate on whether data science teams should assign uniform or specialized job titles to team members such as software engineers and machine learning researchers. It examines the arguments for and against both approaches, considering factors like team unity, role clarity, individual motivation, and organizational culture. By analyzing the impact of job titles on team dynamics and performance, the article provides recommendations to help organizations make informed decisions that align with their strategic goals and foster a productive work environment. tags: - Data science teams - Job titles @@ -53,8 +41,7 @@ tags: - Human resources - Career development - Employee motivation -title: 'The Great Title Debate: Should Data Science Teams Assign Different Job Titles - to Specialized Roles?' +title: 'The Great Title Debate: Should Data Science Teams Assign Different Job Titles to Specialized Roles?' toc: false --- diff --git a/_posts/2024-09-20-model_customer_behaviour.md b/_posts/2024-09-20-model_customer_behaviour.md index 100a4adc..ff7d2dd7 100644 --- a/_posts/2024-09-20-model_customer_behaviour.md +++ b/_posts/2024-09-20-model_customer_behaviour.md @@ -5,8 +5,7 @@ categories: - Data Science classes: wide date: '2024-09-20' -excerpt: Understand how Markov chains can be used to model customer behavior in cloud - services, enabling predictions of usage patterns and helping optimize service offerings. +excerpt: Understand how Markov chains can be used to model customer behavior in cloud services, enabling predictions of usage patterns and helping optimize service offerings. header: image: /assets/images/consumer_behaviour.jpeg og_image: /assets/images/data_science_1.jpg @@ -26,19 +25,12 @@ keywords: - Customer behavior prediction - Data-driven decision-making - Python -seo_description: Explore how Markov chains can model and predict customer behavior - in cloud services. Learn how this statistical method enhances data-driven decision-making - and customer retention strategies. +- python +seo_description: Explore how Markov chains can model and predict customer behavior in cloud services. Learn how this statistical method enhances data-driven decision-making and customer retention strategies. seo_title: 'Deciphering Cloud Customer Behavior: A Deep Dive into Markov Chain Modeling' seo_type: article subtitle: A Deep Dive into Markov Chain Modeling -summary: This article explores how Markov chains can be used to model customer behavior - in cloud services, providing actionable insights into usage patterns, customer churn, - and service optimization. By leveraging this powerful statistical method, cloud - service providers can make data-driven decisions to enhance customer engagement, - predict future usage trends, and increase retention rates. Through code examples - and practical applications, readers are introduced to the mechanics of Markov chains - and their potential impact on cloud-based services. +summary: This article explores how Markov chains can be used to model customer behavior in cloud services, providing actionable insights into usage patterns, customer churn, and service optimization. By leveraging this powerful statistical method, cloud service providers can make data-driven decisions to enhance customer engagement, predict future usage trends, and increase retention rates. Through code examples and practical applications, readers are introduced to the mechanics of Markov chains and their potential impact on cloud-based services. tags: - Cloud computing - Customer behavior @@ -46,6 +38,7 @@ tags: - Data analysis - Predictive modeling - Python +- python title: Deciphering Cloud Customer Behavior toc: false toc_label: The Complexity of Real-World Data Distributions diff --git a/_posts/2024-09-21-data_design.md b/_posts/2024-09-21-data_design.md index 3e3e60ea..34873500 100644 --- a/_posts/2024-09-21-data_design.md +++ b/_posts/2024-09-21-data_design.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2024-09-21' -excerpt: This article explores the often-overlooked importance of data quality in - the data industry and emphasizes the urgent need for defined roles in data design, - collection, and quality assurance. +excerpt: This article explores the often-overlooked importance of data quality in the data industry and emphasizes the urgent need for defined roles in data design, collection, and quality assurance. header: image: /assets/images/what-is-data-quality.jpg og_image: /assets/images/data_science_9.jpg @@ -25,21 +23,11 @@ keywords: - Importance of data roles - Data validation - Data governance -seo_description: Explore the vital importance of data quality, the need for defined - roles in data design and collection, and how data quality impacts data science and - engineering. +seo_description: Explore the vital importance of data quality, the need for defined roles in data design and collection, and how data quality impacts data science and engineering. seo_title: The Critical Role of Data Quality in the Data Industry seo_type: article -subtitle: The Importance of Data Design, Quality Assurance, and the Urgent Need for - Defined Roles in the Data Industry -summary: Data quality is a crucial, yet often overlooked, aspect of data science and - engineering. Without proper attention to data design, collection, and validation, - even the most sophisticated analyses can be flawed. This article highlights the - importance of establishing clear roles in data quality assurance and governance, - ensuring that organizations can confidently rely on the data they use for decision-making. - From defining data collection standards to ensuring ongoing data validation, this - guide covers key strategies for maintaining high-quality data across the lifecycle - of any data-driven project. +subtitle: The Importance of Data Design, Quality Assurance, and the Urgent Need for Defined Roles in the Data Industry +summary: Data quality is a crucial, yet often overlooked, aspect of data science and engineering. Without proper attention to data design, collection, and validation, even the most sophisticated analyses can be flawed. This article highlights the importance of establishing clear roles in data quality assurance and governance, ensuring that organizations can confidently rely on the data they use for decision-making. From defining data collection standards to ensuring ongoing data validation, this guide covers key strategies for maintaining high-quality data across the lifecycle of any data-driven project. tags: - Data science - Data engineering diff --git a/_posts/2024-09-21-data_drift_example.md b/_posts/2024-09-21-data_drift_example.md index 14eb0b99..c5af7fa5 100644 --- a/_posts/2024-09-21-data_drift_example.md +++ b/_posts/2024-09-21-data_drift_example.md @@ -4,8 +4,7 @@ categories: - Data Science classes: wide date: '2024-09-21' -excerpt: A comprehensive exploration of data drift in credit risk models, examining - practical methods to identify and address drift using multivariate techniques. +excerpt: A comprehensive exploration of data drift in credit risk models, examining practical methods to identify and address drift using multivariate techniques. header: image: /assets/images/data_drift.png og_image: /assets/images/data_science_1.jpg @@ -24,17 +23,10 @@ keywords: - Detecting data drift - Credit risk assessment - Adapting models to data changes -seo_description: Explore a practical approach to solving data drift in credit risk - models, focusing on multivariate analysis and its impact on model performance. +seo_description: Explore a practical approach to solving data drift in credit risk models, focusing on multivariate analysis and its impact on model performance. seo_title: 'Addressing Data Drift in Credit Risk Models: A Case Study' seo_type: article -summary: Data drift can significantly affect the accuracy of credit risk models, making - early detection and correction essential for maintaining model reliability. This - article delves into practical approaches for identifying and addressing data drift, - particularly through multivariate analysis. By examining the impact of data drift - on model performance, the guide provides actionable strategies for maintaining the - robustness of credit risk models, ensuring they remain effective over time despite - changes in underlying data distributions. +summary: Data drift can significantly affect the accuracy of credit risk models, making early detection and correction essential for maintaining model reliability. This article delves into practical approaches for identifying and addressing data drift, particularly through multivariate analysis. By examining the impact of data drift on model performance, the guide provides actionable strategies for maintaining the robustness of credit risk models, ensuring they remain effective over time despite changes in underlying data distributions. tags: - Credit risk modeling - Data drift diff --git a/_posts/2024-09-22-randomized_inference.md b/_posts/2024-09-22-randomized_inference.md index 91edbd80..f1946d45 100644 --- a/_posts/2024-09-22-randomized_inference.md +++ b/_posts/2024-09-22-randomized_inference.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-09-22' -excerpt: COPOD is a popular anomaly detection model, but how well does it perform - in practice? This article discusses critical validation issues in third-party models - and lessons learned from COPOD. +excerpt: COPOD is a popular anomaly detection model, but how well does it perform in practice? This article discusses critical validation issues in third-party models and lessons learned from COPOD. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_1.jpg @@ -15,24 +13,17 @@ header: show_overlay_excerpt: false teaser: /assets/images/data_science_4.jpg twitter_image: /assets/images/data_science_1.jpg -seo_description: Learn the importance of validating anomaly detection models like - COPOD. Explore the pitfalls of assuming variable independence in high-dimensional - data. +seo_description: Learn the importance of validating anomaly detection models like COPOD. Explore the pitfalls of assuming variable independence in high-dimensional data. seo_title: 'COPOD Model Validation: Lessons for Anomaly Detection' seo_type: article -summary: Anomaly detection models like COPOD are widely used, but proper validation - is essential to ensure their reliability, especially in high-dimensional datasets. - This article explores the challenges of validating third-party models, focusing - on common pitfalls such as the assumption of variable independence. By examining - the performance of COPOD in real-world scenarios, this guide offers insights into - best practices for model validation, helping data scientists avoid common mistakes - and improve the robustness of their anomaly detection techniques. +summary: Anomaly detection models like COPOD are widely used, but proper validation is essential to ensure their reliability, especially in high-dimensional datasets. This article explores the challenges of validating third-party models, focusing on common pitfalls such as the assumption of variable independence. By examining the performance of COPOD in real-world scenarios, this guide offers insights into best practices for model validation, helping data scientists avoid common mistakes and improve the robustness of their anomaly detection techniques. tags: - Anomaly detection - Model validation - Copod - Python - Python +- python title: 'Validating Anomaly Detection Models: Lessons from COPOD' --- diff --git a/_posts/2024-09-23-improving_decision_trees.md b/_posts/2024-09-23-improving_decision_trees.md index 092c92c1..b1996f0d 100644 --- a/_posts/2024-09-23-improving_decision_trees.md +++ b/_posts/2024-09-23-improving_decision_trees.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-09-23' -excerpt: A deep dive into using Genetic Algorithms to create more accurate, interpretable - decision trees for classification tasks. +excerpt: A deep dive into using Genetic Algorithms to create more accurate, interpretable decision trees for classification tasks. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_4.jpg @@ -21,14 +20,11 @@ keywords: - Classification - Python - Python -seo_description: Explore how Genetic Algorithms can significantly improve the performance - of decision trees in machine learning, yielding interpretable models with higher - accuracy and the same size as standard trees. +- python +seo_description: Explore how Genetic Algorithms can significantly improve the performance of decision trees in machine learning, yielding interpretable models with higher accuracy and the same size as standard trees. seo_title: Enhancing Decision Trees Using Genetic Algorithms for Better Performance seo_type: article -summary: This article explains how to enhance decision tree performance using Genetic - Algorithms. The approach allows for small, interpretable trees that outperform those - created with standard greedy methods. +summary: This article explains how to enhance decision tree performance using Genetic Algorithms. The approach allows for small, interpretable trees that outperform those created with standard greedy methods. tags: - Decision trees - Genetic algorithms @@ -36,6 +32,7 @@ tags: - Classification models - Python - Python +- python title: Improving Decision Tree Performance with Genetic Algorithms --- diff --git a/_posts/2024-09-24-sample_size_clinical.md b/_posts/2024-09-24-sample_size_clinical.md index 413e2e42..fd165c1a 100644 --- a/_posts/2024-09-24-sample_size_clinical.md +++ b/_posts/2024-09-24-sample_size_clinical.md @@ -5,9 +5,7 @@ categories: - Biostatistics classes: wide date: '2024-09-24' -excerpt: A complete guide to writing the sample size justification section for your - clinical trial protocol, covering key statistical concepts like power, error thresholds, - and outcome assumptions. +excerpt: A complete guide to writing the sample size justification section for your clinical trial protocol, covering key statistical concepts like power, error thresholds, and outcome assumptions. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_3.jpg @@ -21,18 +19,10 @@ keywords: - Statistical power - Type 1 and type 2 errors - Biostatistics in clinical research -seo_description: Learn how to write a comprehensive sample size justification in your - clinical protocol, ensuring adequate power and statistical rigor in your trial design. +seo_description: Learn how to write a comprehensive sample size justification in your clinical protocol, ensuring adequate power and statistical rigor in your trial design. seo_title: Writing a Proper Sample Size Justification for Clinical Protocols seo_type: article -summary: Proper sample size justification is a critical component of clinical trial - design, ensuring that the study has enough statistical power to detect meaningful - outcomes. This guide walks you through the process of writing a thorough sample - size justification for clinical protocols, covering essential biostatistical concepts - such as power analysis, Type I and Type II errors, and outcome assumptions. By understanding - these principles, researchers can design more robust trials that meet regulatory - standards while minimizing the risk of invalid results due to inadequate sample - sizes. +summary: Proper sample size justification is a critical component of clinical trial design, ensuring that the study has enough statistical power to detect meaningful outcomes. This guide walks you through the process of writing a thorough sample size justification for clinical protocols, covering essential biostatistical concepts such as power analysis, Type I and Type II errors, and outcome assumptions. By understanding these principles, researchers can design more robust trials that meet regulatory standards while minimizing the risk of invalid results due to inadequate sample sizes. tags: - Sample size justification - Clinical protocol diff --git a/_posts/2024-09-25-simuled_anneling.md b/_posts/2024-09-25-simuled_anneling.md index ae450f4b..8018aba7 100644 --- a/_posts/2024-09-25-simuled_anneling.md +++ b/_posts/2024-09-25-simuled_anneling.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-09-25' -excerpt: Discover how simulated annealing, inspired by metallurgy, offers a powerful - optimization method for machine learning models, especially when dealing with complex - and non-convex loss functions. +excerpt: Discover how simulated annealing, inspired by metallurgy, offers a powerful optimization method for machine learning models, especially when dealing with complex and non-convex loss functions. header: image: /assets/images/machine_learning/machine_learning.jpg og_image: /assets/images/data_science_1.jpg @@ -22,15 +20,11 @@ keywords: - Global optimization - Non-convex loss functions - Python -seo_description: Explore how simulated annealing, a probabilistic technique, can optimize - machine learning models by navigating complex loss functions and improving model - performance. +- python +seo_description: Explore how simulated annealing, a probabilistic technique, can optimize machine learning models by navigating complex loss functions and improving model performance. seo_title: Optimizing Machine Learning Models with Simulated Annealing seo_type: article -summary: Simulated annealing is a probabilistic optimization technique inspired by - metallurgy. This method is especially useful for optimizing machine learning models - with complex, non-convex loss functions, allowing them to escape local minima and - find global solutions. +summary: Simulated annealing is a probabilistic optimization technique inspired by metallurgy. This method is especially useful for optimizing machine learning models with complex, non-convex loss functions, allowing them to escape local minima and find global solutions. tags: - Optimization - Simulated annealing @@ -39,6 +33,7 @@ tags: - Machine learning models - Non-convex optimization - Python +- python title: Optimizing Machine Learning Models using Simulated Annealing --- diff --git a/_posts/2024-09-27-entropy_data_science.md b/_posts/2024-09-27-entropy_data_science.md index 8bc52e5a..abfecf24 100644 --- a/_posts/2024-09-27-entropy_data_science.md +++ b/_posts/2024-09-27-entropy_data_science.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-09-27' -excerpt: Explore the deep connection between entropy, data science, and machine learning. - Understand how entropy drives decision trees, uncertainty measures, feature selection, - and information theory in modern AI. +excerpt: Explore the deep connection between entropy, data science, and machine learning. Understand how entropy drives decision trees, uncertainty measures, feature selection, and information theory in modern AI. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_5.jpg @@ -27,15 +25,11 @@ keywords: - Data science - Machine learning - Python -seo_description: An in-depth exploration of how entropy plays a critical role in data - science and machine learning, including decision trees, uncertainty quantification, - and information theory. +- python +seo_description: An in-depth exploration of how entropy plays a critical role in data science and machine learning, including decision trees, uncertainty quantification, and information theory. seo_title: 'Entropy in Data Science and Machine Learning: Comprehensive Analysis' seo_type: article -summary: "This article explores how entropy, a concept from information theory, is\ - \ used in data science and machine learning. It delves into entropy\u2019s role\ - \ in decision trees, classification, clustering, anomaly detection, and reinforcement\ - \ learning." +summary: This article explores how entropy, a concept from information theory, is used in data science and machine learning. It delves into entropy’s role in decision trees, classification, clustering, anomaly detection, and reinforcement learning. tags: - Entropy - Information theory @@ -44,6 +38,7 @@ tags: - Decision trees - Probability - Python +- python title: 'Entropy in Data Science and Machine Learning: A Deep Dive' --- diff --git a/_posts/2024-09-28-roc.auc.md b/_posts/2024-09-28-roc.auc.md index 09aff345..f1467820 100644 --- a/_posts/2024-09-28-roc.auc.md +++ b/_posts/2024-09-28-roc.auc.md @@ -4,8 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-09-28' -excerpt: Explore the differences between ROC AUC and Precision-Recall AUC in machine - learning and learn when to use each metric for classification tasks. +excerpt: Explore the differences between ROC AUC and Precision-Recall AUC in machine learning and learn when to use each metric for classification tasks. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_7.jpg @@ -19,8 +18,7 @@ keywords: - Machine learning metrics - Classification evaluation - Imbalanced datasets -seo_description: A deep dive into ROC AUC and Precision-Recall AUC, focusing on their - differences, strengths, and the best scenarios to use each metric in machine learning. +seo_description: A deep dive into ROC AUC and Precision-Recall AUC, focusing on their differences, strengths, and the best scenarios to use each metric in machine learning. seo_title: ROC AUC vs Precision-Recall AUC in Machine Learning seo_type: article tags: @@ -28,8 +26,7 @@ tags: - Roc auc - Precision-recall auc - Model performance -title: Understanding the Differences Between ROC AUC and Precision-Recall AUC in Machine - Learning +title: Understanding the Differences Between ROC AUC and Precision-Recall AUC in Machine Learning toc: false --- diff --git a/_posts/2024-09-29-business_intelligence_machine_learning.md b/_posts/2024-09-29-business_intelligence_machine_learning.md index 18bad54b..26eb39c7 100644 --- a/_posts/2024-09-29-business_intelligence_machine_learning.md +++ b/_posts/2024-09-29-business_intelligence_machine_learning.md @@ -4,8 +4,7 @@ categories: - Business Intelligence classes: wide date: '2024-09-29' -excerpt: The fusion of Business Intelligence and Machine Learning offers a pathway - from historical analysis to predictive and prescriptive decision-making. +excerpt: The fusion of Business Intelligence and Machine Learning offers a pathway from historical analysis to predictive and prescriptive decision-making. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -18,16 +17,10 @@ keywords: - Machine learning - Data-driven decision making - Predictive analytics -seo_description: Exploring the fusion of Business Intelligence and Machine Learning, - this article discusses how their integration enhances real-time decision-making, - forecasting, and customer behavior analysis. +seo_description: Exploring the fusion of Business Intelligence and Machine Learning, this article discusses how their integration enhances real-time decision-making, forecasting, and customer behavior analysis. seo_title: 'Bridging Business Intelligence and Machine Learning: A Strategic Approach' seo_type: article -summary: This article examines the integration of Business Intelligence and Machine - Learning, focusing on how this fusion enables businesses to transition from retrospective - analysis to predictive and prescriptive decision-making. Key applications, such - as forecasting, customer behavior analysis, and resource optimization, are discussed, - along with practical examples from leading companies. +summary: This article examines the integration of Business Intelligence and Machine Learning, focusing on how this fusion enables businesses to transition from retrospective analysis to predictive and prescriptive decision-making. Key applications, such as forecasting, customer behavior analysis, and resource optimization, are discussed, along with practical examples from leading companies. tags: - Bi - Ml diff --git a/_posts/2024-09-29-causal_inference.md b/_posts/2024-09-29-causal_inference.md index 5b91ca2f..3ec84945 100644 --- a/_posts/2024-09-29-causal_inference.md +++ b/_posts/2024-09-29-causal_inference.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-09-29' -excerpt: Monotonic constraints are crucial for building reliable and interpretable - machine learning models. Discover how they are applied in causal ML and business - decisions. +excerpt: Monotonic constraints are crucial for building reliable and interpretable machine learning models. Discover how they are applied in causal ML and business decisions. header: image: /assets/images/Causal-Inference-Hero.png og_image: /assets/images/data_science_2.jpg @@ -23,24 +21,18 @@ keywords: - Business analytics - Python - Python -seo_description: Learn how monotonic constraints improve predictions in causal machine - learning and real-world applications like real estate, healthcare, and marketing. +- python +seo_description: Learn how monotonic constraints improve predictions in causal machine learning and real-world applications like real estate, healthcare, and marketing. seo_title: Causal Machine Learning with Monotonic Constraints seo_type: article -summary: Monotonic constraints play a vital role in enhancing the reliability and - interpretability of machine learning models, particularly in causal inference and - decision-making applications. This article explores how monotonic constraints are - implemented in techniques like decision trees and gradient boosting, ensuring that - models behave predictably in response to input changes. With real-world applications - in fields such as real estate, healthcare, and marketing, these constraints help - businesses make more accurate and actionable predictions while maintaining transparency - in their machine learning models. +summary: Monotonic constraints play a vital role in enhancing the reliability and interpretability of machine learning models, particularly in causal inference and decision-making applications. This article explores how monotonic constraints are implemented in techniques like decision trees and gradient boosting, ensuring that models behave predictably in response to input changes. With real-world applications in fields such as real estate, healthcare, and marketing, these constraints help businesses make more accurate and actionable predictions while maintaining transparency in their machine learning models. tags: - Causal ml - Monotonic constraints - Business applications - Python - Python +- python title: 'Causal Insights in Machine Learning: Monotonic Constraints for Better Predictions' --- diff --git a/_posts/2024-09-30-ds_projects.md b/_posts/2024-09-30-ds_projects.md index 572be59e..ab6f227d 100644 --- a/_posts/2024-09-30-ds_projects.md +++ b/_posts/2024-09-30-ds_projects.md @@ -5,8 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-09-30' -excerpt: This checklist helps Data Science professionals ensure thorough validation - of their projects before declaring success and deploying models. +excerpt: This checklist helps Data Science professionals ensure thorough validation of their projects before declaring success and deploying models. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_2.jpg @@ -19,8 +18,7 @@ keywords: - Model deployment - Research validation - Best practices -seo_description: A detailed checklist for Data Science professionals to validate research - and model integrity before deployment. +seo_description: A detailed checklist for Data Science professionals to validate research and model integrity before deployment. seo_title: 'Data Science Project Checklist: Ensure Success Before Deployment' seo_type: article tags: diff --git a/_posts/2024-09-30-exploratory_data_analysis_techniques_pandas.md b/_posts/2024-09-30-exploratory_data_analysis_techniques_pandas.md index d9708cb6..4f2e6f17 100644 --- a/_posts/2024-09-30-exploratory_data_analysis_techniques_pandas.md +++ b/_posts/2024-09-30-exploratory_data_analysis_techniques_pandas.md @@ -4,9 +4,7 @@ categories: - Data Science classes: wide date: '2024-09-30' -excerpt: Explore how to perform effective Exploratory Data Analysis (EDA) using Pandas, - a powerful Python library. Learn data loading, cleaning, visualization, and advanced - EDA techniques. +excerpt: Explore how to perform effective Exploratory Data Analysis (EDA) using Pandas, a powerful Python library. Learn data loading, cleaning, visualization, and advanced EDA techniques. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_5.jpg @@ -20,20 +18,18 @@ keywords: - Data science pandas - Python - Python -seo_description: A detailed guide on performing Exploratory Data Analysis (EDA) using - the Pandas library in Python, covering data loading, cleaning, visualization, and - advanced techniques. -seo_title: 'Exploratory Data Analysis (EDA) Techniques with Pandas: A Comprehensive - Guide' +- python +seo_description: A detailed guide on performing Exploratory Data Analysis (EDA) using the Pandas library in Python, covering data loading, cleaning, visualization, and advanced techniques. +seo_title: 'Exploratory Data Analysis (EDA) Techniques with Pandas: A Comprehensive Guide' seo_type: article -summary: A comprehensive guide on Exploratory Data Analysis (EDA) using Pandas, covering - essential techniques for understanding, cleaning, and analyzing datasets in Python. +summary: A comprehensive guide on Exploratory Data Analysis (EDA) using Pandas, covering essential techniques for understanding, cleaning, and analyzing datasets in Python. tags: - Python - Pandas - Eda - Python - Python +- python title: Exploratory Data Analysis (EDA) Techniques with Pandas --- diff --git a/_posts/2024-10-01-automated_prompt_engineering.md b/_posts/2024-10-01-automated_prompt_engineering.md index cc6bbd89..b066f020 100644 --- a/_posts/2024-10-01-automated_prompt_engineering.md +++ b/_posts/2024-10-01-automated_prompt_engineering.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-10-01' -excerpt: Explore Automated Prompt Engineering (APE), a powerful method to automate - and optimize prompts for Large Language Models, enhancing their task performance - and efficiency. +excerpt: Explore Automated Prompt Engineering (APE), a powerful method to automate and optimize prompts for Large Language Models, enhancing their task performance and efficiency. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_7.jpg @@ -23,14 +21,11 @@ keywords: - Random prompt optimization - Python - Python -seo_description: An in-depth exploration of Automated Prompt Engineering (APE), its - strategies, and how it automates the process of generating and refining prompts - for improving Large Language Models. +- python +seo_description: An in-depth exploration of Automated Prompt Engineering (APE), its strategies, and how it automates the process of generating and refining prompts for improving Large Language Models. seo_title: 'Automated Prompt Engineering (APE): Optimizing LLMs' seo_type: article -summary: This article delves into Automated Prompt Engineering (APE), explaining how - it automates and optimizes the prompt generation process to enhance the performance - of Large Language Models. +summary: This article delves into Automated Prompt Engineering (APE), explaining how it automates and optimizes the prompt generation process to enhance the performance of Large Language Models. tags: - Automated prompt engineering - Hyperparameter optimization @@ -38,8 +33,8 @@ tags: - Large language models - Python - Python -title: 'Automated Prompt Engineering (APE): Optimizing Large Language Models through - Automation' +- python +title: 'Automated Prompt Engineering (APE): Optimizing Large Language Models through Automation' toc: false toc_icon: robot toc_label: Automated Prompt Engineering Overview diff --git a/_posts/2024-10-01-edge_machine_learning.md b/_posts/2024-10-01-edge_machine_learning.md index c6b61551..87175916 100644 --- a/_posts/2024-10-01-edge_machine_learning.md +++ b/_posts/2024-10-01-edge_machine_learning.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-10-01' -excerpt: This article dives into the implementation of continuous machine learning - deployment on edge devices, using MLOps and IoT management tools for a real-world - agriculture use case. +excerpt: This article dives into the implementation of continuous machine learning deployment on edge devices, using MLOps and IoT management tools for a real-world agriculture use case. header: image: /assets/images/Edge-Computing.png og_image: /assets/images/data_science_2.jpg @@ -26,18 +24,13 @@ keywords: - Ai for agriculture - Machine learning pipelines for edge devices - Yaml +- yaml math: true -seo_description: Explore how to implement continuous machine learning deployment on - edge devices using MLOps platforms, focusing on a real-world example of a smart - agriculture system. +seo_description: Explore how to implement continuous machine learning deployment on edge devices using MLOps platforms, focusing on a real-world example of a smart agriculture system. seo_title: 'Continuous Machine Learning Deployment for Edge Devices: A Practical Approach' seo_type: article social_image: /assets/images/edge-devices.png -summary: This article explores how to implement continuous machine learning deployment - on edge devices using MLOps and IoT management platforms. Focusing on a real-world - smart agriculture use case, it highlights the benefits of edge inference for real-time - processing, lower latency, and improved decision-making in environments with limited - connectivity. +summary: This article explores how to implement continuous machine learning deployment on edge devices using MLOps and IoT management platforms. Focusing on a real-world smart agriculture use case, it highlights the benefits of edge inference for real-time processing, lower latency, and improved decision-making in environments with limited connectivity. tags: - Mlops - Edge ai @@ -45,6 +38,7 @@ tags: - Smart devices - Iot - Yaml +- yaml title: Implementing Continuous Machine Learning Deployment on Edge Devices --- diff --git a/_posts/2024-10-02-building_data_driven_business_strategy.md b/_posts/2024-10-02-building_data_driven_business_strategy.md index f0d3e215..679dee7d 100644 --- a/_posts/2024-10-02-building_data_driven_business_strategy.md +++ b/_posts/2024-10-02-building_data_driven_business_strategy.md @@ -4,8 +4,7 @@ categories: - Business Intelligence classes: wide date: '2024-10-02' -excerpt: A data-driven business strategy integrates Business Intelligence and Data - Science to drive informed decisions, optimize resources, and stay competitive. +excerpt: A data-driven business strategy integrates Business Intelligence and Data Science to drive informed decisions, optimize resources, and stay competitive. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_9.jpg @@ -18,24 +17,16 @@ keywords: - Data science - Data-driven strategy - Predictive analytics -seo_description: This article explores how organizations can build a data-driven business - strategy by blending Business Intelligence (BI) and Data Science (DS) to enhance - decision-making and competitiveness. -seo_title: 'Building a Data-Driven Business Strategy: The Role of Business Intelligence - and Data Science' +seo_description: This article explores how organizations can build a data-driven business strategy by blending Business Intelligence (BI) and Data Science (DS) to enhance decision-making and competitiveness. +seo_title: 'Building a Data-Driven Business Strategy: The Role of Business Intelligence and Data Science' seo_type: article -summary: Discover how Business Intelligence and Data Science can work together to - build a data-driven business strategy, from leveraging historical data for insights - to using predictive analytics for forward-looking decisions. Learn from case studies - of companies like Walmart, Uber, and Netflix, and explore the necessary infrastructure - to support a data-driven organization. +summary: Discover how Business Intelligence and Data Science can work together to build a data-driven business strategy, from leveraging historical data for insights to using predictive analytics for forward-looking decisions. Learn from case studies of companies like Walmart, Uber, and Netflix, and explore the necessary infrastructure to support a data-driven organization. tags: - Bi - Data science - Predictive analytics - Data strategy -title: 'Building a Data-Driven Business Strategy: The Role of Business Intelligence - and Data Science' +title: 'Building a Data-Driven Business Strategy: The Role of Business Intelligence and Data Science' --- In today’s rapidly evolving business, data has become the lifeblood of organizations. Businesses, regardless of their size or industry, generate enormous volumes of data daily, and the ability to extract actionable insights from this data is pivotal for maintaining competitiveness. Consequently, the need for a data-driven strategy—one that leverages both Business Intelligence (BI) and Data Science (DS)—has never been more critical. diff --git a/_posts/2024-10-02-entropy.md b/_posts/2024-10-02-entropy.md index b7f726d0..00237a38 100644 --- a/_posts/2024-10-02-entropy.md +++ b/_posts/2024-10-02-entropy.md @@ -5,8 +5,7 @@ categories: - Information Theory classes: wide date: '2024-10-02' -excerpt: Explore entropy's role in thermodynamics, information theory, and quantum - mechanics, and its broader implications in physics and beyond. +excerpt: Explore entropy's role in thermodynamics, information theory, and quantum mechanics, and its broader implications in physics and beyond. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_1.jpg @@ -14,9 +13,7 @@ header: show_overlay_excerpt: false teaser: /assets/images/data_science_5.jpg twitter_image: /assets/images/data_science_1.jpg -seo_description: An in-depth exploration of entropy in thermodynamics, statistical - mechanics, and information theory, from classical formulations to quantum mechanics - applications. +seo_description: An in-depth exploration of entropy in thermodynamics, statistical mechanics, and information theory, from classical formulations to quantum mechanics applications. seo_title: 'Entropy and Information Theory: A Comprehensive Analysis' seo_type: article tags: diff --git a/_posts/2024-10-03-differentiating_machine_learning_engineering_mlops.md b/_posts/2024-10-03-differentiating_machine_learning_engineering_mlops.md index 56df91f1..aeaecb14 100644 --- a/_posts/2024-10-03-differentiating_machine_learning_engineering_mlops.md +++ b/_posts/2024-10-03-differentiating_machine_learning_engineering_mlops.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-10-03' -excerpt: This article explores the fine line between Machine Learning Engineering - (MLE) and MLOps roles, delving into their shared responsibilities, unique contributions, - and how these roles integrate in small to large teams. +excerpt: This article explores the fine line between Machine Learning Engineering (MLE) and MLOps roles, delving into their shared responsibilities, unique contributions, and how these roles integrate in small to large teams. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_5.jpg @@ -20,22 +18,16 @@ keywords: - Ai infrastructure - Model deployment - Ml pipelines -seo_description: An in-depth exploration of the roles of Machine Learning Engineers - (MLE) and MLOps engineers, their overlaps, and distinctions in modern ML pipelines. -seo_title: 'Differentiating Machine Learning Engineering and MLOps: Key Responsibilities - and Overlaps' +seo_description: An in-depth exploration of the roles of Machine Learning Engineers (MLE) and MLOps engineers, their overlaps, and distinctions in modern ML pipelines. +seo_title: 'Differentiating Machine Learning Engineering and MLOps: Key Responsibilities and Overlaps' seo_type: article -summary: Machine Learning Engineering (MLE) and MLOps are two interconnected yet distinct - roles in the AI landscape. This article delves into the responsibilities and challenges - of both roles, highlighting where they overlap and where they diverge, especially - in real-world machine learning projects. +summary: Machine Learning Engineering (MLE) and MLOps are two interconnected yet distinct roles in the AI landscape. This article delves into the responsibilities and challenges of both roles, highlighting where they overlap and where they diverge, especially in real-world machine learning projects. tags: - Machine learning engineering - Mlops - Ml infrastructure - Model deployment -title: 'Differentiating Machine Learning Engineering and MLOps: A Fine Line Between - Two Critical Roles' +title: 'Differentiating Machine Learning Engineering and MLOps: A Fine Line Between Two Critical Roles' --- The emergence of artificial intelligence and machine learning (ML) as cornerstones of modern technology has introduced several specialized roles that drive the development and deployment of intelligent systems. Among these, two crucial roles stand out: Machine Learning Engineer (MLE) and MLOps Engineer. While these roles are integral to delivering machine learning models from research to production, the fine line between their responsibilities has blurred, particularly in smaller teams. diff --git a/_posts/2024-10-04-guide_arima_time_series_modeling.md b/_posts/2024-10-04-guide_arima_time_series_modeling.md index e940f151..0fb7e503 100644 --- a/_posts/2024-10-04-guide_arima_time_series_modeling.md +++ b/_posts/2024-10-04-guide_arima_time_series_modeling.md @@ -4,9 +4,7 @@ categories: - Time Series Analysis classes: wide date: '2024-10-04' -excerpt: A detailed exploration of the ARIMA model for time series forecasting. Understand - its components, parameter identification techniques, and comparison with ARIMAX, - SARIMA, and ARMA. +excerpt: A detailed exploration of the ARIMA model for time series forecasting. Understand its components, parameter identification techniques, and comparison with ARIMAX, SARIMA, and ARMA. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_4.jpg @@ -22,15 +20,12 @@ keywords: - Arma - Python - R -seo_description: Learn the fundamentals of ARIMA (AutoRegressive Integrated Moving - Average) modeling, including components, parameter identification, validation, and - practical applications. +- python +- r +seo_description: Learn the fundamentals of ARIMA (AutoRegressive Integrated Moving Average) modeling, including components, parameter identification, validation, and practical applications. seo_title: ARIMA Time Series Modeling Explained seo_type: article -summary: This guide delves into the AutoRegressive Integrated Moving Average (ARIMA) - model, a powerful tool for time series forecasting. It covers the essential components, - how to identify model parameters, validation techniques, and how ARIMA compares - with other time series models like ARIMAX, SARIMA, and ARMA. +summary: This guide delves into the AutoRegressive Integrated Moving Average (ARIMA) model, a powerful tool for time series forecasting. It covers the essential components, how to identify model parameters, validation techniques, and how ARIMA compares with other time series models like ARIMAX, SARIMA, and ARMA. tags: - Arima - Time series modeling @@ -38,6 +33,8 @@ tags: - Data science - Python - R +- python +- r title: A Comprehensive Guide to ARIMA Time Series Modeling --- diff --git a/_posts/2024-10-05-simple_distribution.md b/_posts/2024-10-05-simple_distribution.md index a53d83bb..bf387dfd 100644 --- a/_posts/2024-10-05-simple_distribution.md +++ b/_posts/2024-10-05-simple_distribution.md @@ -5,8 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-10-05' -excerpt: An in-depth review of the role of simple distributional properties, like - mean and standard deviation, in time-series classification as a baseline approach. +excerpt: An in-depth review of the role of simple distributional properties, like mean and standard deviation, in time-series classification as a baseline approach. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_5.jpg @@ -18,20 +17,15 @@ keywords: - Time-series classification - Simple distributional properties - Deep learning -seo_description: Explore the effectiveness of using simple distributional properties - as a baseline for time-series classification, compared to complex deep learning - models. +seo_description: Explore the effectiveness of using simple distributional properties as a baseline for time-series classification, compared to complex deep learning models. seo_title: Comprehensive Review of Distributional Properties in Time-Series Classification seo_type: article -summary: This article reviews time-series classification techniques, highlighting - the importance of simple distributional properties such as mean and standard deviation - as a baseline. +summary: This article reviews time-series classification techniques, highlighting the importance of simple distributional properties such as mean and standard deviation as a baseline. tags: - Time-series classification - Distributional properties - Deep learning -title: A Comprehensive Review of Simple Distributional Properties as a Baseline for - Time-Series Classification +title: A Comprehensive Review of Simple Distributional Properties as a Baseline for Time-Series Classification --- ## 1. Overview of Time-Series Classification diff --git a/_posts/2024-10-06-evaluating_distributions.md b/_posts/2024-10-06-evaluating_distributions.md index f7656352..0af07b3c 100644 --- a/_posts/2024-10-06-evaluating_distributions.md +++ b/_posts/2024-10-06-evaluating_distributions.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-10-06' -excerpt: A comprehensive review of simple distributional properties such as mean and - standard deviation as a strong baseline for time-series classification in standardized - benchmarks. +excerpt: A comprehensive review of simple distributional properties such as mean and standard deviation as a strong baseline for time-series classification in standardized benchmarks. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_1.jpg @@ -21,20 +19,15 @@ keywords: - Distributional properties - Machine learning - Benchmarking -seo_description: Explore the performance of simple distributional properties in time-series - classification benchmarks using the UEA/UCR repository, and the relevance of these - models in complex tasks. +seo_description: Explore the performance of simple distributional properties in time-series classification benchmarks using the UEA/UCR repository, and the relevance of these models in complex tasks. seo_title: Simple Distributional Properties for Time-Series Classification Benchmarks seo_type: article -summary: This article discusses the use of simple distributional properties as a baseline - for time-series classification, focusing on benchmarks from the UEA/UCR repository - and comparing simple and complex models. +summary: This article discusses the use of simple distributional properties as a baseline for time-series classification, focusing on benchmarks from the UEA/UCR repository and comparing simple and complex models. tags: - Time-series classification - Uea/ucr repository - Simple models -title: Evaluating Simple Distributional Properties for Time-Series Classification - Benchmarks +title: Evaluating Simple Distributional Properties for Time-Series Classification Benchmarks --- ## The UEA/UCR Time-Series Classification Repository diff --git a/_posts/2024-10-07-extending_simple_model.md b/_posts/2024-10-07-extending_simple_model.md index 74f28c55..146fa416 100644 --- a/_posts/2024-10-07-extending_simple_model.md +++ b/_posts/2024-10-07-extending_simple_model.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-10-07' -excerpt: Explore how simple distributional models for time-series classification can - be extended with additional feature sets like catch22 to improve performance without - sacrificing interpretability. +excerpt: Explore how simple distributional models for time-series classification can be extended with additional feature sets like catch22 to improve performance without sacrificing interpretability. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_3.jpg @@ -20,14 +18,10 @@ keywords: - Catch22 - Simple models - Feature engineering -seo_description: A review of how simple time-series classification models can be extended - using feature sets like catch22 and the practical implications of balancing complexity - and interpretability. +seo_description: A review of how simple time-series classification models can be extended using feature sets like catch22 and the practical implications of balancing complexity and interpretability. seo_title: 'Extending Simple Models: Adding Catch22 for Time-Series Classification' seo_type: article -summary: This article discusses when and how to extend simple time-series classification - models by introducing additional features, such as catch22, and the practical implications - of using these models in various domains. +summary: This article discusses when and how to extend simple time-series classification models by introducing additional features, such as catch22, and the practical implications of using these models in various domains. tags: - Time-series classification - Catch22 diff --git a/_posts/2024-10-08-implementing_time_series.md b/_posts/2024-10-08-implementing_time_series.md index 361f1093..95ead6d0 100644 --- a/_posts/2024-10-08-implementing_time_series.md +++ b/_posts/2024-10-08-implementing_time_series.md @@ -5,9 +5,7 @@ categories: - Machine Learning classes: wide date: '2024-10-08' -excerpt: Explore time-series classification in Python with step-by-step examples using - simple models, the catch22 feature set, and UEA/UCR repository benchmarking with - statistical tests. +excerpt: Explore time-series classification in Python with step-by-step examples using simple models, the catch22 feature set, and UEA/UCR repository benchmarking with statistical tests. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -21,22 +19,19 @@ keywords: - Python - Uea/ucr - Python -seo_description: Learn how to implement time-series classification in Python using - simple models, catch22 features, and benchmarking with statistical tests using UEA/UCR - datasets. +- python +seo_description: Learn how to implement time-series classification in Python using simple models, catch22 features, and benchmarking with statistical tests using UEA/UCR datasets. seo_title: 'Python Code for Time-Series Classification: Simple Models to Catch22' seo_type: article -summary: This article provides Python code for time-series classification, covering - simple models, catch22 features, and benchmarking with UEA/UCR repository datasets - and statistical significance testing. +summary: This article provides Python code for time-series classification, covering simple models, catch22 features, and benchmarking with UEA/UCR repository datasets and statistical significance testing. tags: - Python - Time-series classification - Catch22 - Uea/ucr - Python -title: 'Implementing Time-Series Classification: From Simple Models to Advanced Feature - Sets' +- python +title: 'Implementing Time-Series Classification: From Simple Models to Advanced Feature Sets' --- --- diff --git a/_posts/2024-10-09-magnitude_matter_machine_learning.md b/_posts/2024-10-09-magnitude_matter_machine_learning.md index 435eefe4..4fd05718 100644 --- a/_posts/2024-10-09-magnitude_matter_machine_learning.md +++ b/_posts/2024-10-09-magnitude_matter_machine_learning.md @@ -4,10 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-10-09' -excerpt: The magnitude of variables in machine learning models can have significant - impacts, particularly on linear regression, neural networks, and models using distance - metrics. This article explores why feature scaling is crucial and which models are - sensitive to variable magnitude. +excerpt: The magnitude of variables in machine learning models can have significant impacts, particularly on linear regression, neural networks, and models using distance metrics. This article explores why feature scaling is crucial and which models are sensitive to variable magnitude. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_3.jpg @@ -23,14 +20,11 @@ keywords: - Neural networks - Support vector machines - Python -seo_description: An in-depth discussion on the importance of variable magnitude in - machine learning models, its impact on regression coefficients, and how feature - scaling improves model performance. +- python +seo_description: An in-depth discussion on the importance of variable magnitude in machine learning models, its impact on regression coefficients, and how feature scaling improves model performance. seo_title: Does the Magnitude of the Variable Matter in Machine Learning Models? seo_type: article -summary: This article discusses the importance of variable magnitude in machine learning - models, how feature scaling enhances model performance, and the distinctions between - models that are sensitive to the scale of variables and those that are not. +summary: This article discusses the importance of variable magnitude in machine learning models, how feature scaling enhances model performance, and the distinctions between models that are sensitive to the scale of variables and those that are not. tags: - Feature scaling - Linear regression @@ -40,6 +34,7 @@ tags: - Pca - Random forests - Python +- python title: Does the Magnitude of the Variable Matter in Machine Learning? --- diff --git a/_posts/2024-10-10-understanding_data_drift_what_why_matters_machine_learning.md b/_posts/2024-10-10-understanding_data_drift_what_why_matters_machine_learning.md index 95ff719e..21cd7493 100644 --- a/_posts/2024-10-10-understanding_data_drift_what_why_matters_machine_learning.md +++ b/_posts/2024-10-10-understanding_data_drift_what_why_matters_machine_learning.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-10-10' -excerpt: Data drift can significantly affect the performance of machine learning models - over time. Learn about different types of drift and how they impact model predictions - in dynamic environments. +excerpt: Data drift can significantly affect the performance of machine learning models over time. Learn about different types of drift and how they impact model predictions in dynamic environments. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -20,15 +18,10 @@ keywords: - Covariate drift - Concept drift - Label drift -seo_description: This article explores data drift in machine learning, its types, - and how changes in input data can affect model performance. It covers covariate, - label, and concept drift, with real-world examples from finance and healthcare. +seo_description: This article explores data drift in machine learning, its types, and how changes in input data can affect model performance. It covers covariate, label, and concept drift, with real-world examples from finance and healthcare. seo_title: 'Understanding Data Drift in Machine Learning: Types and Impact' seo_type: article -summary: This article explains the concept of data drift, focusing on how changes - in data distribution affect machine learning model performance. We discuss the different - types of data drift, such as covariate, label, and concept drift, providing examples - from industries like finance and healthcare. +summary: This article explains the concept of data drift, focusing on how changes in data distribution affect machine learning model performance. We discuss the different types of data drift, such as covariate, label, and concept drift, providing examples from industries like finance and healthcare. tags: - Data drift - Machine learning models diff --git a/_posts/2024-10-11-model_drift_why_even_best_machine_learning_models_fail_over_time.md b/_posts/2024-10-11-model_drift_why_even_best_machine_learning_models_fail_over_time.md index 5b136f8c..8bc21c92 100644 --- a/_posts/2024-10-11-model_drift_why_even_best_machine_learning_models_fail_over_time.md +++ b/_posts/2024-10-11-model_drift_why_even_best_machine_learning_models_fail_over_time.md @@ -4,9 +4,7 @@ categories: - Machine Learning classes: wide date: '2024-10-11' -excerpt: Even the best machine learning models experience performance degradation - over time due to model drift. Learn about the causes of model drift and how it affects - production systems. +excerpt: Even the best machine learning models experience performance degradation over time due to model drift. Learn about the causes of model drift and how it affects production systems. header: image: /assets/images/data_science_3.jpg og_image: /assets/images/data_science_3.jpg @@ -20,15 +18,10 @@ keywords: - Data drift - Model degradation - Ai in production -seo_description: This article explores the concept of model drift and how changes - in data or target variables degrade the accuracy of machine learning models over - time, with case studies from real-world applications. +seo_description: This article explores the concept of model drift and how changes in data or target variables degrade the accuracy of machine learning models over time, with case studies from real-world applications. seo_title: 'Why Machine Learning Models Fail Over Time: Understanding Model Drift' seo_type: article -summary: This article examines model drift, focusing on how data drift, changes in - underlying patterns, and new unseen data can degrade machine learning model accuracy - over time. We explore the causes of model drift and provide case studies from industries - like finance and healthcare. +summary: This article examines model drift, focusing on how data drift, changes in underlying patterns, and new unseen data can degrade machine learning model accuracy over time. We explore the causes of model drift and provide case studies from industries like finance and healthcare. tags: - Model drift - Data drift diff --git a/_posts/2024-10-12-how_data_science_reshaping_business_strategy_age_machine_learning.md b/_posts/2024-10-12-how_data_science_reshaping_business_strategy_age_machine_learning.md index a0c6eb0c..770aab0b 100644 --- a/_posts/2024-10-12-how_data_science_reshaping_business_strategy_age_machine_learning.md +++ b/_posts/2024-10-12-how_data_science_reshaping_business_strategy_age_machine_learning.md @@ -6,10 +6,7 @@ categories: - Business Strategy classes: wide date: '2024-10-12' -excerpt: Data-driven decision-making, powered by data science and machine learning, - is becoming central to business strategy. Learn how companies are integrating data - science into strategic planning to improve outcomes in customer segmentation, churn - prediction, and recommendation systems. +excerpt: Data-driven decision-making, powered by data science and machine learning, is becoming central to business strategy. Learn how companies are integrating data science into strategic planning to improve outcomes in customer segmentation, churn prediction, and recommendation systems. header: image: /assets/images/data_science_9.jpg og_image: /assets/images/data_science_9.jpg @@ -24,16 +21,10 @@ keywords: - Customer segmentation - Churn prediction - Recommendation systems -seo_description: This article explores how data science and machine learning are reshaping - business strategy, focusing on key use cases like customer segmentation, churn prediction, - and recommendation systems. +seo_description: This article explores how data science and machine learning are reshaping business strategy, focusing on key use cases like customer segmentation, churn prediction, and recommendation systems. seo_title: How Data Science is Transforming Business Strategy with Machine Learning seo_type: article -summary: This article examines how data science and machine learning are transforming - business strategy, highlighting key use cases such as customer segmentation, churn - prediction, and recommendation systems. It compares traditional decision-making - approaches with data-driven methods and discusses the benefits of integrating data - science into strategic planning. +summary: This article examines how data science and machine learning are transforming business strategy, highlighting key use cases such as customer segmentation, churn prediction, and recommendation systems. It compares traditional decision-making approaches with data-driven methods and discusses the benefits of integrating data science into strategic planning. tags: - Data science - Machine learning diff --git a/_posts/2024-10-29-exponential_smoothing_methods_time_series_forecasting.md b/_posts/2024-10-29-exponential_smoothing_methods_time_series_forecasting.md index 76d925e6..99f7d351 100644 --- a/_posts/2024-10-29-exponential_smoothing_methods_time_series_forecasting.md +++ b/_posts/2024-10-29-exponential_smoothing_methods_time_series_forecasting.md @@ -4,10 +4,7 @@ categories: - Time Series Analysis classes: wide date: '2024-10-29' -excerpt: This detailed guide covers exponential smoothing methods for time series - forecasting, including simple, double, and triple exponential smoothing (ETS). Learn - how these methods work, how they compare to ARIMA, and practical applications in - retail, finance, and inventory management. +excerpt: This detailed guide covers exponential smoothing methods for time series forecasting, including simple, double, and triple exponential smoothing (ETS). Learn how these methods work, how they compare to ARIMA, and practical applications in retail, finance, and inventory management. header: image: /assets/images/data_science_2.jpg og_image: /assets/images/data_science_2.jpg @@ -24,16 +21,12 @@ keywords: - Inventory management - Python - R -seo_description: Explore simple, double, and triple exponential smoothing methods - (ETS) for time series forecasting. Learn how these methods compare to ARIMA models - and their applications in retail, finance, and inventory management. -seo_title: A Comprehensive Guide to Exponential Smoothing Methods for Time Series - Forecasting +- python +- r +seo_description: Explore simple, double, and triple exponential smoothing methods (ETS) for time series forecasting. Learn how these methods compare to ARIMA models and their applications in retail, finance, and inventory management. +seo_title: A Comprehensive Guide to Exponential Smoothing Methods for Time Series Forecasting seo_type: article -summary: Explore the different types of exponential smoothing methods, how they work, - and their practical applications in time series forecasting. This article compares - ETS methods with ARIMA models and includes use cases in retail, inventory management, - and finance. +summary: Explore the different types of exponential smoothing methods, how they work, and their practical applications in time series forecasting. This article compares ETS methods with ARIMA models and includes use cases in retail, inventory management, and finance. tags: - Exponential smoothing - Ets @@ -42,6 +35,8 @@ tags: - Data science - Python - R +- python +- r title: Introduction to Exponential Smoothing Methods for Time Series Forecasting --- diff --git a/_posts/2024-10-30-introduction_seasonal_decomposition_time_series.md b/_posts/2024-10-30-introduction_seasonal_decomposition_time_series.md index 5c931310..2d9d7e1d 100644 --- a/_posts/2024-10-30-introduction_seasonal_decomposition_time_series.md +++ b/_posts/2024-10-30-introduction_seasonal_decomposition_time_series.md @@ -4,9 +4,7 @@ categories: - Time Series Analysis classes: wide date: '2024-10-30' -excerpt: This article provides an in-depth look at STL and X-13-SEATS, two powerful - methods for decomposing time series into trend, seasonal, and residual components. - Learn how these methods help model seasonality in time series forecasting. +excerpt: This article provides an in-depth look at STL and X-13-SEATS, two powerful methods for decomposing time series into trend, seasonal, and residual components. Learn how these methods help model seasonality in time series forecasting. header: image: /assets/images/data_science_8.jpg og_image: /assets/images/data_science_8.jpg @@ -23,15 +21,12 @@ keywords: - Python - Python - R -seo_description: Learn how Seasonal-Trend decomposition using LOESS (STL) and X-13-SEATS - methods help model seasonality in time series data, with practical examples in R - and Python. +- python +- r +seo_description: Learn how Seasonal-Trend decomposition using LOESS (STL) and X-13-SEATS methods help model seasonality in time series data, with practical examples in R and Python. seo_title: STL and X-13 Methods for Time Series Decomposition seo_type: article -summary: Explore STL (Seasonal-Trend decomposition using LOESS) and X-13-SEATS, two - prominent methods for time series decomposition, and their importance in modeling - seasonality. The article includes practical examples and code implementation in - both R and Python. +summary: Explore STL (Seasonal-Trend decomposition using LOESS) and X-13-SEATS, two prominent methods for time series decomposition, and their importance in modeling seasonality. The article includes practical examples and code implementation in both R and Python. tags: - Seasonal decomposition - Time series @@ -40,6 +35,8 @@ tags: - Forecasting - Python - R +- python +- r title: 'Introduction to Seasonal Decomposition of Time Series: STL and X-13 Methods' --- diff --git a/_posts/2024-10-31-machine_learning_fall_prediction.md b/_posts/2024-10-31-machine_learning_fall_prediction.md index d9e81fb6..f2715aad 100644 --- a/_posts/2024-10-31-machine_learning_fall_prediction.md +++ b/_posts/2024-10-31-machine_learning_fall_prediction.md @@ -4,9 +4,7 @@ categories: - HealthTech classes: wide date: '2024-10-31' -excerpt: Machine learning is revolutionizing fall prevention in elderly care by predicting - the likelihood of falls through wearable sensor data, mobility analysis, and health - history insights. +excerpt: Machine learning is revolutionizing fall prevention in elderly care by predicting the likelihood of falls through wearable sensor data, mobility analysis, and health history insights. header: image: /assets/images/data_science_7.jpg og_image: /assets/images/data_science_5.jpg @@ -20,14 +18,10 @@ keywords: - Wearable technology - Elderly care - Health monitoring -seo_description: Learn how machine learning models are used to predict and prevent - falls among the elderly by analyzing sensor data, wearables, and health history. +seo_description: Learn how machine learning models are used to predict and prevent falls among the elderly by analyzing sensor data, wearables, and health history. seo_title: Machine Learning for Fall Prevention in the Elderly seo_type: article -summary: Falls among the elderly are a significant public health concern. Machine - learning can help predict and prevent falls by analyzing data from wearables, sensors, - and other health records, offering timely interventions that can improve quality - of life. +summary: Falls among the elderly are a significant public health concern. Machine learning can help predict and prevent falls by analyzing data from wearables, sensors, and other health records, offering timely interventions that can improve quality of life. tags: - Machine learning - Healthcare diff --git a/_posts/2024-11-01-data_driven_elderly_care.md b/_posts/2024-11-01-data_driven_elderly_care.md index df58a044..27b3ec8c 100644 --- a/_posts/2024-11-01-data_driven_elderly_care.md +++ b/_posts/2024-11-01-data_driven_elderly_care.md @@ -4,9 +4,7 @@ categories: - HealthTech classes: wide date: '2024-11-01' -excerpt: Data science is revolutionizing chronic disease management among the elderly - by leveraging predictive analytics to monitor disease progression, manage medications, - and create personalized treatment plans. +excerpt: Data science is revolutionizing chronic disease management among the elderly by leveraging predictive analytics to monitor disease progression, manage medications, and create personalized treatment plans. header: image: /assets/images/data_science_6.jpg og_image: /assets/images/data_science_7.jpg @@ -20,15 +18,10 @@ keywords: - Elderly care - Data-driven healthcare - Personalized medicine -seo_description: Discover how data-driven approaches, powered by predictive analytics, - help manage chronic diseases like diabetes, arthritis, and cardiovascular conditions - in elderly populations. +seo_description: Discover how data-driven approaches, powered by predictive analytics, help manage chronic diseases like diabetes, arthritis, and cardiovascular conditions in elderly populations. seo_title: Data Science for Managing Chronic Diseases in the Elderly seo_type: article -summary: Data-driven approaches are improving the management of chronic diseases in - elderly populations by harnessing the power of predictive analytics. These methods - allow healthcare providers to monitor disease progression, optimize medication regimens, - and tailor treatment plans based on real-time individual health data. +summary: Data-driven approaches are improving the management of chronic diseases in elderly populations by harnessing the power of predictive analytics. These methods allow healthcare providers to monitor disease progression, optimize medication regimens, and tailor treatment plans based on real-time individual health data. tags: - Chronic disease management - Predictive analytics diff --git a/_posts/2024-11-30-outliers.md b/_posts/2024-11-30-outliers.md index b2d0d436..c439efae 100644 --- a/_posts/2024-11-30-outliers.md +++ b/_posts/2024-11-30-outliers.md @@ -15,8 +15,7 @@ header: teaser: /assets/images/data_science_5.jpg twitter_image: /assets/images/data_science_8.jpg seo_type: article -subtitle: Understanding and Managing Data Points that Deviate Significantly from the - Norm +subtitle: Understanding and Managing Data Points that Deviate Significantly from the Norm tags: - Outliers - Robust statistics diff --git a/_posts/2024-12-01-remote_monitoring_elderly_care.md b/_posts/2024-12-01-remote_monitoring_elderly_care.md index 7a421773..e6982ee0 100644 --- a/_posts/2024-12-01-remote_monitoring_elderly_care.md +++ b/_posts/2024-12-01-remote_monitoring_elderly_care.md @@ -4,9 +4,7 @@ categories: - HealthTech classes: wide date: '2024-12-01' -excerpt: The integration of IoT and big data is revolutionizing elderly care by enabling - remote monitoring systems that track vital signs, detect emergencies, and ensure - quick responses to health risks. +excerpt: The integration of IoT and big data is revolutionizing elderly care by enabling remote monitoring systems that track vital signs, detect emergencies, and ensure quick responses to health risks. header: image: /assets/images/data_science_5.jpg og_image: /assets/images/data_science_4.jpg @@ -20,23 +18,17 @@ keywords: - Elderly care - Health emergencies - Smart homes -seo_description: Explore how IoT-enabled devices, wearables, and health monitors are - using big data to remotely monitor elderly individuals and detect health emergencies - in real time. +seo_description: Explore how IoT-enabled devices, wearables, and health monitors are using big data to remotely monitor elderly individuals and detect health emergencies in real time. seo_title: IoT and Big Data in Remote Monitoring for Elderly Care seo_type: article -summary: IoT-enabled devices and big data are transforming elderly care by enabling - real-time remote monitoring. From wearable devices to smart home systems, these - technologies offer continuous health tracking and quick responses to emergencies - like heart attacks, strokes, or falls, ensuring that seniors remain safe and healthy. +summary: IoT-enabled devices and big data are transforming elderly care by enabling real-time remote monitoring. From wearable devices to smart home systems, these technologies offer continuous health tracking and quick responses to emergencies like heart attacks, strokes, or falls, ensuring that seniors remain safe and healthy. tags: - Elderly care - Iot - Big data - Remote monitoring - Health monitoring -title: 'Remote Monitoring and Elderly Care: How IoT and Big Data are Keeping Seniors - Safe' +title: 'Remote Monitoring and Elderly Care: How IoT and Big Data are Keeping Seniors Safe' --- ## Introduction diff --git a/_posts/2024-12-30-predicting_hospital_readmissions.md b/_posts/2024-12-30-predicting_hospital_readmissions.md index f9d17987..68e2b4c7 100644 --- a/_posts/2024-12-30-predicting_hospital_readmissions.md +++ b/_posts/2024-12-30-predicting_hospital_readmissions.md @@ -4,9 +4,7 @@ categories: - HealthTech classes: wide date: '2024-12-30' -excerpt: Machine learning models are revolutionizing post-hospitalization care by - predicting hospital readmissions in elderly patients, helping healthcare providers - optimize treatment and reduce complications. +excerpt: Machine learning models are revolutionizing post-hospitalization care by predicting hospital readmissions in elderly patients, helping healthcare providers optimize treatment and reduce complications. header: image: /assets/images/data_science_4.jpg og_image: /assets/images/data_science_9.jpg @@ -20,15 +18,10 @@ keywords: - Elderly patients - Post-hospital care - Predictive analytics -seo_description: Explore how machine learning models can predict hospital readmissions - among elderly patients by analyzing post-discharge data, treatment adherence, and - health conditions. +seo_description: Explore how machine learning models can predict hospital readmissions among elderly patients by analyzing post-discharge data, treatment adherence, and health conditions. seo_title: Machine Learning for Predicting Hospital Readmissions in Elderly Patients seo_type: article -summary: Hospital readmissions among elderly patients are a significant healthcare - challenge. This article examines how machine learning algorithms are being used - to predict readmission risks by analyzing post-discharge data, health records, and - treatment adherence, enabling optimized care and timely interventions. +summary: Hospital readmissions among elderly patients are a significant healthcare challenge. This article examines how machine learning algorithms are being used to predict readmission risks by analyzing post-discharge data, health records, and treatment adherence, enabling optimized care and timely interventions. tags: - Hospital readmissions - Predictive analytics From 3f97d6b20b1b8b96b54b52530b06d6feae72cf6d Mon Sep 17 00:00:00 2001 From: Diogo Ribeiro Date: Sun, 13 Oct 2024 23:33:06 +0100 Subject: [PATCH 3/7] feat: new article --- .../-_ideas/2030-01-01-new_articles_topics.md | 3 - ...ole_data_science_predictive_maintenance.md | 249 ++++++++++++++++++ 2 files changed, 249 insertions(+), 3 deletions(-) create mode 100644 _posts/2020-01-06-role_data_science_predictive_maintenance.md diff --git a/_posts/-_ideas/2030-01-01-new_articles_topics.md b/_posts/-_ideas/2030-01-01-new_articles_topics.md index 4a0b9a55..84633cd5 100644 --- a/_posts/-_ideas/2030-01-01-new_articles_topics.md +++ b/_posts/-_ideas/2030-01-01-new_articles_topics.md @@ -19,9 +19,6 @@ There are several interesting article topics you can explore under the umbrella - **Overview**: Explain what predictive maintenance (PdM) is and how it differs from preventive and reactive maintenance. - **Focus**: Basic techniques and traditional approaches to predictive maintenance, including time-based and condition-based maintenance strategies. -### 2. The Role of Data Science in Predictive Maintenance - - **Overview**: Explore how data science methods, such as statistical analysis and predictive modeling, enable organizations to forecast failures and optimize maintenance schedules. - - **Focus**: Key data science techniques used in PdM, such as regression, anomaly detection, and clustering. ### 3. How Big Data is Transforming Predictive Maintenance - **Overview**: Discuss how the vast amounts of data generated by IoT sensors, machinery, and operational systems contribute to more accurate predictions and better decision-making. diff --git a/_posts/2020-01-06-role_data_science_predictive_maintenance.md b/_posts/2020-01-06-role_data_science_predictive_maintenance.md new file mode 100644 index 00000000..8a0e9b07 --- /dev/null +++ b/_posts/2020-01-06-role_data_science_predictive_maintenance.md @@ -0,0 +1,249 @@ +--- +author_profile: false +categories: +- Data Science +classes: wide +date: '2020-01-06' +excerpt: Explore the role of data science in predictive maintenance, from forecasting equipment failure to optimizing maintenance schedules using techniques like regression and anomaly detection. +header: + image: /assets/images/data_science_7.jpg + og_image: /assets/images/data_science_7.jpg + overlay_image: /assets/images/data_science_7.jpg + show_overlay_excerpt: false + teaser: /assets/images/data_science_7.jpg + twitter_image: /assets/images/data_science_7.jpg +keywords: +- Predictive Maintenance +- Data Science +- Industrial IoT +- Machine Learning +- Predictive Analytics +- Industrial Analytics +seo_description: Discover how data science techniques such as regression, clustering, and anomaly detection optimize predictive maintenance, helping organizations forecast failures and enhance operational efficiency. +seo_title: How Data Science Powers Predictive Maintenance +seo_type: article +summary: An in-depth look at how data science techniques such as regression, clustering, anomaly detection, and machine learning are transforming predictive maintenance across various industries. +tags: +- Predictive Maintenance +- Machine Learning +- Industrial IoT +- Industrial Analytics +- Predictive Analytics +title: Leveraging Data Science Techniques for Predictive Maintenance +--- + +## Table of Contents + +1. Introduction to Predictive Maintenance (PdM) +2. The Importance of Data Science in PdM +3. Key Data Science Techniques in Predictive Maintenance + 1. Regression Analysis + 2. Anomaly Detection + 3. Clustering Algorithms +4. Data Requirements and Challenges in PdM +5. Role of Machine Learning in Predictive Maintenance +6. Applications of PdM Across Industries +7. Future of Data Science in Predictive Maintenance +8. Conclusion + +--- + +## 1. Introduction to Predictive Maintenance (PdM) + +Predictive maintenance (PdM) refers to the practice of using data-driven techniques to predict when equipment will fail, allowing for timely and efficient maintenance. This proactive approach aims to reduce downtime, optimize equipment lifespan, and minimize maintenance costs. Unlike traditional maintenance strategies, such as reactive (fixing after failure) or preventive (servicing at regular intervals), PdM leverages real-time data, statistical analysis, and predictive models to forecast equipment degradation and identify the optimal time for intervention. + +As industries become increasingly reliant on machinery and automation, the ability to predict equipment failure becomes critical. PdM harnesses vast amounts of data collected from sensors, machines, and historical records to forecast failures before they happen. The result is significant cost savings, improved asset reliability, and better overall operational efficiency. + +## 2. The Importance of Data Science in PdM + +Data science is the backbone of predictive maintenance. By applying advanced techniques from statistics, machine learning, and artificial intelligence, data science enables the extraction of meaningful patterns from complex datasets, allowing organizations to detect early warning signs of failure. Data science techniques are used to analyze sensor data, operational logs, and environmental factors to create predictive models capable of anticipating equipment failures with high precision. + +In this context, data science serves several key roles: + +- **Failure Prediction**: Leveraging historical and real-time data to predict the likelihood of failure. +- **Condition Monitoring**: Tracking equipment health based on data collected from sensors and control systems. +- **Optimization**: Determining the most cost-effective time for maintenance, balancing reliability and performance. + +Modern PdM systems rely heavily on the ability to process vast amounts of data, using data science methods to transform raw data into actionable insights. This is where techniques like regression analysis, anomaly detection, and clustering come into play, helping organizations identify patterns that indicate impending failure. + +## 3. Key Data Science Techniques in Predictive Maintenance + +### 3.1 Regression Analysis + +Regression analysis is a fundamental data science technique used to predict future outcomes based on historical data. In predictive maintenance, regression models are used to estimate the remaining useful life (RUL) of equipment, based on various input variables such as temperature, pressure, vibration, and operational history. + +There are different types of regression models applied in PdM: + +- **Linear Regression**: This technique assumes a linear relationship between the dependent variable (e.g., time until failure) and one or more independent variables (e.g., temperature or vibration data). Linear regression is widely used for simple, time-based maintenance predictions. + +- **Polynomial Regression**: In cases where the relationship between variables is non-linear, polynomial regression can be used. It extends the linear model by considering higher-degree terms of the independent variables, offering greater flexibility in modeling complex equipment behaviors. + +- **Logistic Regression**: Logistic regression is applied when the goal is to predict a binary outcome, such as whether a failure will occur within a certain time frame. This method is useful in classifying equipment into “likely to fail” or “not likely to fail” categories. + +These regression techniques enable organizations to create predictive models that estimate how equipment health deteriorates over time, allowing for timely maintenance scheduling. + +### 3.2 Anomaly Detection + +Anomaly detection is another critical technique in predictive maintenance. It involves identifying patterns in the data that deviate from the norm, which could indicate early signs of equipment malfunction or failure. In PdM, anomalies may represent unusual behaviors such as spikes in temperature, pressure irregularities, or abnormal vibration levels. + +There are several approaches to anomaly detection: + +- **Statistical Methods**: Traditional statistical approaches such as z-scores, moving averages, and control charts can be used to flag data points that fall outside predefined thresholds. These methods are simple but effective for detecting gross anomalies in equipment performance. + +- **Machine Learning-based Anomaly Detection**: More sophisticated methods use machine learning algorithms such as clustering, isolation forests, and neural networks to detect subtle anomalies. These techniques learn from historical data to identify normal behavior patterns and can flag deviations that may be precursors to failure. + +- **Time-series Anomaly Detection**: Equipment data is often collected as time-series data (e.g., temperature over time). Time-series anomaly detection techniques, such as autoregressive models and Long Short-Term Memory (LSTM) networks, are specifically designed to capture temporal dependencies and detect outliers in the data stream. + +Anomaly detection provides an early warning system for maintenance teams, allowing them to investigate and address potential issues before they result in failure. This reduces downtime and prevents costly repairs. + +### 3.3 Clustering Algorithms + +Clustering is a data science technique used to group similar data points together, which can be extremely valuable in predictive maintenance. Clustering algorithms help in segmenting equipment based on operating conditions, failure patterns, or usage characteristics, allowing maintenance teams to target specific groups of assets for more focused maintenance strategies. + +Popular clustering methods used in PdM include: + +- **K-means Clustering**: K-means is a popular unsupervised learning algorithm that partitions the dataset into K clusters, where each point is assigned to the nearest cluster centroid. This method is useful for identifying distinct operating modes or failure patterns in equipment. + +- **Hierarchical Clustering**: Unlike K-means, hierarchical clustering builds a tree of clusters based on similarity measures. It is particularly effective in situations where there are multiple levels of similarity between equipment behaviors, allowing for more nuanced groupings. + +- **Density-Based Clustering (DBSCAN)**: DBSCAN is particularly useful for identifying clusters of varying shapes and densities. It is ideal for PdM applications where there are complex operational states and intermittent failures, as it can find patterns that other algorithms might miss. + +Clustering is widely used in PdM to group machines based on similar usage patterns or failure tendencies, enabling predictive models to be fine-tuned for specific asset groups. This can greatly improve prediction accuracy and help maintenance teams focus on the most critical assets. + +## 4. Data Requirements and Challenges in PdM + +For predictive maintenance to be effective, it requires vast amounts of high-quality data. This data typically comes from multiple sources, including: + +- **Sensor Data**: Information gathered from sensors monitoring temperature, vibration, pressure, and other operational parameters. +- **Operational Data**: Historical records of machine performance, maintenance logs, and failure incidents. +- **Environmental Data**: External factors like humidity, weather conditions, and operational environment, which can influence equipment degradation. + +While the availability of data has increased with the advent of IoT and industrial sensors, there are several challenges associated with data collection and usage in PdM: + +- **Data Integration**: Combining data from different sources (e.g., sensors, maintenance logs, and operational data) can be complex, especially when dealing with heterogeneous data formats and systems. + +- **Data Quality**: Noisy or incomplete data can lead to inaccurate predictions. Ensuring data quality through proper preprocessing and validation is essential for the reliability of PdM systems. + +- **Data Labeling**: For supervised machine learning techniques, labeled data (i.e., data tagged with failure outcomes) is critical. However, obtaining labeled data can be difficult, as failures are often rare events, and historical records may be incomplete or inaccurate. + +- **Real-time Processing**: PdM requires real-time data analysis to provide timely predictions. Processing and analyzing large volumes of data in real-time poses significant computational challenges. + +Despite these challenges, advancements in data storage, processing, and machine learning techniques are making it increasingly feasible to implement robust PdM systems. + +## 5. Role of Machine Learning in Predictive Maintenance + +Machine learning plays a pivotal role in modern predictive maintenance, enabling the automation of pattern recognition, failure prediction, and decision-making processes. By learning from historical data, machine learning models can identify complex relationships between variables that would be difficult to discern using traditional methods. + +Machine learning techniques commonly used in PdM include: + +- **Supervised Learning**: In supervised learning, models are trained on labeled datasets, where the outcome (e.g., failure or no failure) is known. Techniques such as decision trees, support vector machines (SVM), and neural networks can be used to predict future failures based on past patterns. + +- **Unsupervised Learning**: In situations where labeled data is unavailable, unsupervised learning techniques like clustering and anomaly detection are applied to uncover hidden patterns in the data. These methods are particularly useful for identifying unknown failure modes or grouping similar types of equipment for maintenance. + +- **Reinforcement Learning**: Reinforcement learning algorithms can optimize maintenance schedules by learning through trial and error. These models adjust their behavior based on feedback from the environment, allowing them to discover the most effective maintenance strategies over time. + +Machine learning models in PdM are continually refined as more data becomes available, allowing for increasingly accurate predictions and more efficient maintenance planning. + +## 6. Applications of PdM Across Industries + +Predictive maintenance has found applications across a wide range of industries where operational efficiency, equipment longevity, and reduced downtime are crucial. Let's examine some key industries and how PdM is transforming their maintenance strategies. + +### 6.1 Manufacturing + +In the manufacturing sector, downtime caused by machine failures can lead to significant production delays and financial losses. PdM helps manufacturers optimize maintenance schedules and improve machine availability. By monitoring critical machines and equipment, such as CNC machines, conveyors, and robotic arms, data-driven models can predict potential failures before they occur, reducing unplanned downtime. + +Some key benefits of PdM in manufacturing include: + +- **Increased Machine Uptime**: Real-time monitoring of equipment conditions, such as vibration, temperature, and lubrication, allows for timely maintenance. + +- **Cost Reduction**: Avoiding unnecessary preventive maintenance reduces operational costs, and minimizing the risk of machine failure ensures fewer disruptions in production lines. + +- **Optimized Supply Chain**: With fewer unplanned stoppages, the entire supply chain becomes more reliable, helping businesses meet production deadlines. + +One real-world example of PdM in manufacturing is Rolls-Royce’s use of data analytics and machine learning to monitor the health of their engines. Their PdM system predicts potential failures by analyzing sensor data from thousands of aircraft engines, enabling better maintenance planning and reducing costly engine failures during operation. + +### 6.2 Energy and Utilities + +The energy and utilities sector relies on critical infrastructure such as power plants, wind turbines, pipelines, and electrical grids. Any unplanned downtime or failure can lead to power outages, environmental risks, and financial losses. PdM plays a crucial role in ensuring the reliability and efficiency of these assets by predicting when maintenance is needed, avoiding catastrophic failures. + +Key applications of PdM in the energy and utilities sector include: + +- **Power Generation**: PdM is used to monitor turbines, generators, and transformers. Sensors collect data on temperature, pressure, and vibration, allowing operators to predict equipment failure and plan maintenance during off-peak hours, minimizing service interruptions. + +- **Wind Energy**: Wind turbines are often located in remote areas, making maintenance costly and challenging. PdM helps monitor turbine performance and predict mechanical issues, reducing the need for frequent inspections and minimizing downtimes due to failures. + +- **Oil and Gas Pipelines**: PdM techniques, including anomaly detection and machine learning, can identify pipeline leaks, corrosion, and pressure anomalies, allowing operators to take proactive measures to prevent accidents or environmental damage. + +A leading example of PdM in energy is General Electric’s (GE) use of machine learning algorithms to predict failures in wind turbines. By analyzing real-time data from turbine sensors, GE can optimize maintenance schedules, reducing downtime and improving the efficiency of wind farms. + +### 6.3 Transportation and Logistics + +The transportation and logistics industry relies heavily on fleet management and infrastructure maintenance to keep operations running smoothly. Vehicle breakdowns, unplanned maintenance, and infrastructure failures (e.g., railway tracks or bridges) can disrupt the flow of goods and services, causing significant financial losses. + +PdM is used in the following ways within transportation: + +- **Fleet Maintenance**: Fleet operators use PdM to monitor vehicle health, including engine performance, tire pressure, brake conditions, and fuel efficiency. Predictive models can forecast when a vehicle will need maintenance, reducing the risk of breakdowns and extending the lifespan of fleet assets. + +- **Railways**: In rail transportation, PdM helps monitor critical infrastructure such as tracks, switches, and rolling stock. Real-time data from sensors installed on trains and tracks can identify anomalies, allowing for preventive maintenance and minimizing service interruptions. + +- **Aviation**: Airlines use PdM to predict aircraft component failures, from engines to landing gear. PdM systems analyze sensor data from various aircraft parts, ensuring timely repairs, minimizing the risk of in-flight failures, and enhancing passenger safety. + +A well-known example is FedEx’s implementation of PdM to monitor its fleet of delivery vehicles. By collecting real-time data on vehicle performance, FedEx can predict when maintenance is needed, reduce vehicle breakdowns, and improve operational efficiency across its logistics network. + +### 6.4 Healthcare + +In the healthcare sector, medical equipment and devices are critical to patient care. Equipment failures can delay procedures, disrupt patient care, and in some cases, even pose life-threatening risks. Predictive maintenance can ensure that medical devices, such as MRI machines, ventilators, and patient monitors, remain operational and reliable. + +Some applications of PdM in healthcare include: + +- **Medical Imaging Equipment**: Hospitals use PdM to monitor the performance of MRI, CT, and X-ray machines. These machines generate a significant amount of data during operation, which can be analyzed to predict potential failures or performance degradation. + +- **Patient Monitoring Devices**: PdM helps in monitoring patient devices, ensuring that they function properly and providing early warning for potential issues. This reduces the risk of device failure during critical patient care moments. + +By ensuring the reliability of medical equipment, PdM not only helps reduce operational costs for healthcare facilities but also enhances patient outcomes by minimizing delays in treatment due to equipment malfunction. + +## 7. Future of Data Science in Predictive Maintenance + +As industries increasingly adopt digital transformation strategies, the future of predictive maintenance is set to evolve significantly. Emerging technologies, such as the Internet of Things (IoT), artificial intelligence (AI), and cloud computing, will further enhance the capabilities of PdM systems, providing greater accuracy, scalability, and predictive power. + +### 7.1 The Impact of IoT + +The integration of IoT devices and sensors into equipment provides unprecedented levels of data collection in real time. IoT sensors can continuously monitor machine performance and environmental conditions, enabling a constant stream of data for predictive models. The proliferation of IoT devices will make predictive maintenance more accessible, especially for smaller businesses and industries that have traditionally relied on reactive or preventive maintenance strategies. + +Additionally, IoT-enabled PdM systems can lead to the following advancements: + +- **Edge Computing**: With IoT, data can be processed closer to the source (at the edge), reducing the latency and bandwidth requirements for sending data to a centralized server. This allows for real-time PdM and faster decision-making. + +- **Cloud Integration**: Cloud computing allows companies to store and process large amounts of PdM data without needing significant on-premise infrastructure. Cloud-based PdM platforms enable more scalable solutions and allow companies to integrate data from multiple locations seamlessly. + +### 7.2 Artificial Intelligence and Advanced Machine Learning + +Artificial intelligence and advanced machine learning techniques will play an increasingly critical role in the future of predictive maintenance. AI-based models can handle complex datasets with many variables, offering more accurate predictions and deeper insights into equipment health. + +Some potential advancements include: + +- **Deep Learning**: While traditional machine learning models have shown great success in PdM, deep learning models, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), are being used to analyze complex data patterns. These models are particularly effective in handling high-dimensional data, such as sensor data or image data from industrial equipment. + +- **Reinforcement Learning**: Reinforcement learning could be used to optimize maintenance schedules dynamically, by learning the best actions to take based on feedback from the environment. This would enable more adaptive maintenance strategies that improve over time. + +- **Explainable AI (XAI)**: As AI models become more sophisticated, there is a growing need for explainable AI techniques that can provide insights into why a particular failure prediction was made. XAI will improve trust in PdM systems, especially in industries with stringent regulatory requirements, such as healthcare and aerospace. + +### 7.3 Predictive Maintenance as a Service (PdMaaS) + +As PdM technology advances, companies will increasingly adopt "Predictive Maintenance as a Service" (PdMaaS) models, where PdM solutions are offered as a subscription service by third-party vendors. These services will include cloud-based platforms that collect and analyze data from industrial assets, providing companies with predictive insights without the need to build and maintain their own infrastructure. + +PdMaaS will lower the barrier to entry for companies, especially small and medium-sized businesses, enabling them to leverage the power of predictive maintenance without the high costs associated with on-premise solutions. + +### 7.4 Integration with Digital Twins + +The concept of digital twins – virtual replicas of physical assets – is poised to revolutionize predictive maintenance. By creating a digital replica of a machine or system, companies can simulate the operation of their equipment in real-time and predict how different factors, such as wear and tear or changes in operating conditions, will affect performance. + +Digital twins, combined with data science techniques, can enable more precise maintenance predictions and optimization strategies. For example, simulations could show how changes in operational parameters might extend the lifespan of equipment, allowing maintenance schedules to be adjusted accordingly. + +## 8. Conclusion + +Data science has emerged as a critical enabler of predictive maintenance, transforming how industries maintain and optimize their assets. By leveraging techniques such as regression analysis, anomaly detection, clustering, and machine learning, organizations can predict equipment failures before they occur, reduce downtime, and extend the lifespan of their machinery. + +The applications of predictive maintenance span across various industries, including manufacturing, energy, healthcare, and transportation, each benefiting from improved operational efficiency and cost savings. As technology continues to evolve, advancements in IoT, AI, cloud computing, and digital twin technology will further enhance PdM capabilities, making it more accurate, accessible, and scalable. + +The future of PdM lies in the integration of these cutting-edge technologies, providing companies with predictive insights that will not only prevent equipment failures but also drive continuous improvements in operational performance. By embracing predictive maintenance, organizations can move towards a future where downtime is minimized, maintenance costs are reduced, and asset utilization is maximized. From dbb23cb5b0e894f87930c1e34ce5da35ecc45a22 Mon Sep 17 00:00:00 2001 From: Diogo Ribeiro Date: Sun, 13 Oct 2024 23:40:55 +0100 Subject: [PATCH 4/7] fix --- fix_frontmatter.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/fix_frontmatter.py b/fix_frontmatter.py index a32088c9..411a6aae 100644 --- a/fix_frontmatter.py +++ b/fix_frontmatter.py @@ -3,6 +3,8 @@ import frontmatter import random +TOTAL_FILES = 14 + def extract_date_from_filename(filename): # Assuming the filename format is 'YYYY-MM-DD-some-title.md' match = re.match(r'(\d{4}-\d{2}-\d{2})-', filename) @@ -48,7 +50,7 @@ def process_markdown_file(filepath): post['toc'] = False # Insert 'header' block if not already present - random_number = random.randint(1, 9) + random_number = random.randint(1, TOTAL_FILES) if 'header' not in post: print(f"Inserting 'header' block in {filename}") post['header'] = { From a2e31d5889de8e0779a226a676609e3b1e957270 Mon Sep 17 00:00:00 2001 From: Diogo Ribeiro Date: Sun, 13 Oct 2024 23:41:07 +0100 Subject: [PATCH 5/7] feat: new article --- ...ata_transforming_predictive_maintenance.md | 163 ++++++++++++++++++ 1 file changed, 163 insertions(+) create mode 100644 _posts/2020-01-07-how_big_data_transforming_predictive_maintenance.md diff --git a/_posts/2020-01-07-how_big_data_transforming_predictive_maintenance.md b/_posts/2020-01-07-how_big_data_transforming_predictive_maintenance.md new file mode 100644 index 00000000..2d2b3a6e --- /dev/null +++ b/_posts/2020-01-07-how_big_data_transforming_predictive_maintenance.md @@ -0,0 +1,163 @@ +--- +author_profile: false +categories: +- Big Data +classes: wide +date: '2020-01-07' +excerpt: Big Data is revolutionizing predictive maintenance by offering unprecedented insights into equipment health. Learn about the challenges and opportunities in managing and analyzing large-scale data for more accurate failure predictions. +header: + image: /assets/images/data_science_7.jpg + og_image: /assets/images/data_science_7.jpg + overlay_image: /assets/images/data_science_7.jpg + show_overlay_excerpt: false + teaser: /assets/images/data_science_7.jpg + twitter_image: /assets/images/data_science_7.jpg +keywords: +- Predictive Maintenance +- Big Data +- Industrial IoT +- Data Integration +- Machine Learning +seo_description: Explore how Big Data from IoT sensors, machinery, and operational systems enhances predictive maintenance accuracy and decision-making, while addressing challenges in data storage, cleaning, and integration. +seo_title: Big Data's Impact on Predictive Maintenance +seo_type: article +summary: Big Data is key to predictive maintenance, enabling more precise equipment failure predictions and optimization. This article discusses the role of data from IoT sensors and operational systems, as well as the challenges of data storage, cleaning, and integration. +tags: +- Predictive Maintenance +- Data Science +- Big Data +- Industrial IoT +- Predictive Analytics +title: How Big Data is Transforming Predictive Maintenance +--- + +## Table of Contents + +1. The Rise of Big Data in Predictive Maintenance +2. The Role of IoT in Generating Big Data +3. Opportunities Offered by Big Data in PdM + 1. Improved Failure Predictions + 2. Real-time Monitoring and Alerts + 3. Data-Driven Decision Making +4. Challenges in Managing and Analyzing Big Data + 1. Data Storage and Scalability + 2. Data Cleaning and Preprocessing + 3. Data Integration from Multiple Sources +5. The Future of Big Data in Predictive Maintenance +6. Conclusion + +--- + +## 1. The Rise of Big Data in Predictive Maintenance + +In recent years, predictive maintenance (PdM) has undergone a significant transformation, primarily driven by the explosion of big data. Traditionally, maintenance strategies relied on fixed schedules or reactive interventions. However, as more data becomes available from machines, sensors, and operational systems, organizations are leveraging this data to predict failures before they happen, allowing for timely and efficient maintenance. This data-centric approach, often referred to as predictive maintenance, is now evolving into a more advanced system powered by big data analytics. + +Big data in PdM refers to the vast amounts of structured and unstructured data generated from multiple sources, including Internet of Things (IoT) devices, machinery, control systems, and historical maintenance records. This data, when analyzed effectively, provides valuable insights into equipment health, operating conditions, and failure patterns, enabling more accurate failure predictions and better maintenance decision-making. + +The shift towards big data-driven PdM marks a new era where data, rather than guesswork or scheduled interventions, dictates maintenance activities. With the proliferation of IoT sensors and advanced analytics, organizations now have access to a wealth of information that can help them optimize maintenance processes, reduce downtime, and extend equipment life. + +## 2. The Role of IoT in Generating Big Data + +The rapid growth of the Internet of Things (IoT) is a key factor in the rise of big data in predictive maintenance. IoT devices, including sensors and connected machines, continuously generate massive volumes of data about equipment status, operational parameters, and environmental conditions. These sensors monitor variables such as temperature, pressure, vibration, and humidity, offering real-time insights into the health and performance of industrial assets. + +Key IoT contributions to big data in PdM include: + +- **Real-time Data Generation**: IoT sensors collect data in real time, providing a continuous stream of information that can be used to monitor equipment conditions and detect early warning signs of failure. This allows for more proactive interventions. + +- **Diverse Data Sources**: IoT-enabled devices generate data from various sources, including operational machinery, environmental sensors, and even human inputs (e.g., maintenance logs). The sheer variety of data collected helps create a comprehensive picture of equipment health. + +- **Historical Data**: IoT devices can store historical performance data, enabling comparisons over time. This helps identify trends and patterns that could indicate gradual equipment degradation or the likelihood of future failure. + +With IoT, the volume of data generated in industrial environments has skyrocketed. While this data provides valuable opportunities for PdM, managing and analyzing it effectively presents significant challenges, as explored in the following sections. + +## 3. Opportunities Offered by Big Data in PdM + +Big data presents enormous potential for improving predictive maintenance outcomes. As organizations collect and analyze more data, they gain the ability to make more accurate predictions, respond faster to emerging issues, and optimize maintenance schedules based on actual equipment conditions rather than arbitrary timelines. + +### 3.1 Improved Failure Predictions + +One of the most significant opportunities offered by big data is the ability to improve failure predictions. With access to vast amounts of data, predictive models can be trained to identify patterns and trends that signal an impending failure. The more data these models are exposed to, the more accurate their predictions become, as they can account for a wide range of variables, including operational conditions, wear-and-tear patterns, and environmental factors. + +By analyzing historical data alongside real-time sensor data, companies can develop sophisticated predictive algorithms that offer high accuracy in forecasting when a specific machine or component is likely to fail. This leads to more informed maintenance decisions and reduced instances of unexpected downtime. + +### 3.2 Real-time Monitoring and Alerts + +Big data, coupled with IoT, allows for real-time monitoring of equipment health. This real-time data stream enables immediate detection of anomalies or deviations from normal operating conditions. For example, if a machine’s vibration or temperature exceeds predefined thresholds, an alert can be triggered, allowing maintenance teams to investigate the issue before it leads to failure. + +Real-time alerts help reduce the time between the detection of an issue and corrective action, thereby minimizing equipment downtime and preventing larger, more costly failures. + +### 3.3 Data-Driven Decision Making + +With big data, organizations can move towards data-driven decision-making processes in their maintenance operations. Rather than relying on intuition or fixed maintenance schedules, maintenance teams can use data analytics to make decisions based on actual equipment performance. + +This shift allows organizations to: + +- **Optimize Maintenance Schedules**: By analyzing patterns in failure data and equipment usage, organizations can schedule maintenance activities more effectively, minimizing unnecessary interventions while avoiding breakdowns. + +- **Extend Equipment Lifespan**: Data-driven insights into equipment performance enable more precise interventions, which can help extend the lifespan of critical assets. + +- **Reduce Costs**: By performing maintenance only when needed, organizations can avoid the costs associated with over-maintenance or emergency repairs. + +## 4. Challenges in Managing and Analyzing Big Data + +While big data offers significant opportunities for improving predictive maintenance, it also presents several challenges. The sheer volume, velocity, and variety of data generated from IoT devices and industrial machinery can be difficult to manage, store, and analyze effectively. + +### 4.1 Data Storage and Scalability + +One of the primary challenges of big data in PdM is data storage. IoT sensors and machines generate large volumes of data continuously, and organizations must have the infrastructure in place to store this data. Traditional data storage systems may not be able to handle the scalability requirements of big data. + +Cloud-based storage solutions have become a popular option, offering scalability and flexibility to accommodate the growing amounts of data. However, these solutions also present challenges in terms of security, data access, and latency. Organizations must balance the need for scalable storage with the need for fast access to data for real-time monitoring and analysis. + +### 4.2 Data Cleaning and Preprocessing + +Another major challenge in working with big data is ensuring data quality. Raw data from sensors and machinery can be noisy, incomplete, or inconsistent, which can lead to inaccurate predictions if not properly cleaned and preprocessed. For example, sensors may malfunction, resulting in erroneous readings, or data may be missing due to connectivity issues. + +Before data can be used in predictive models, it must undergo several preprocessing steps: + +- **Data Cleaning**: This involves removing or correcting erroneous data points and filling in missing values. + +- **Normalization**: Data from different sources may have different formats or units, so it must be normalized to ensure consistency across the dataset. + +- **Outlier Detection**: Outliers, or data points that deviate significantly from the norm, must be identified and analyzed to determine whether they represent a true anomaly or a sensor error. + +Data cleaning and preprocessing are critical steps in ensuring that big data is usable for predictive maintenance, but these tasks can be time-consuming and resource-intensive. + +### 4.3 Data Integration from Multiple Sources + +Predictive maintenance requires data from multiple sources, including sensors, machinery, maintenance logs, and environmental factors. Integrating these disparate data sources into a unified system is a significant challenge, especially when dealing with heterogeneous data formats, protocols, and structures. + +For example, data from a temperature sensor may need to be integrated with maintenance logs stored in a different format or even on a different system. Achieving seamless integration between these diverse data sources requires robust data integration frameworks that can handle large volumes of data in real-time. + +### 4.4 Real-time Data Processing + +With the advent of IoT, organizations now have access to real-time data streams from their equipment. However, processing this data in real-time and deriving actionable insights from it can be a challenge, especially when dealing with high-frequency data from numerous sensors. + +Organizations must invest in real-time analytics platforms that can process large volumes of data with low latency. These platforms often rely on technologies like edge computing, which enables data to be processed closer to the source, reducing the time it takes to detect anomalies and trigger maintenance actions. + +## 5. The Future of Big Data in Predictive Maintenance + +The future of big data in predictive maintenance is set to evolve rapidly as technology advances. Emerging trends such as edge computing, artificial intelligence (AI), and machine learning will play an increasingly important role in managing and analyzing big data for PdM. + +### 5.1 Edge Computing for Faster Data Processing + +As mentioned earlier, edge computing allows data to be processed closer to the source, reducing the need to transmit large volumes of data to centralized servers. This results in faster data processing and quicker responses to equipment anomalies. Edge computing will become increasingly important in PdM, especially as more organizations adopt IoT devices that generate high-frequency data. + +### 5.2 AI and Machine Learning for Advanced Analytics + +AI and machine learning will continue to transform the field of predictive maintenance by enabling more advanced analytics. Machine learning algorithms can analyze complex datasets to detect subtle patterns that may indicate an impending failure. As these algorithms are exposed to more data, their predictive accuracy will improve, leading to even more precise maintenance schedules. + +Additionally, AI-powered systems can automate decision-making processes, allowing organizations to move from reactive or preventive maintenance strategies to fully autonomous maintenance systems. + +### 5.3 Predictive Maintenance in Smart Factories + +The rise of Industry 4.0 and the concept of smart factories will further integrate big data into predictive maintenance. In smart factories, all equipment is connected, and data is continuously collected and analyzed in real-time. Predictive maintenance will be an integral part of these operations, using big data to ensure that machines operate efficiently and with minimal downtime. + +## 6. Conclusion + +Big data is playing a transformative role in predictive maintenance by providing organizations with the insights they need to predict equipment failures and optimize maintenance activities. The vast amounts of data generated by IoT sensors, machinery, and operational systems offer unparalleled opportunities for more accurate failure predictions, real-time monitoring, and data-driven decision-making. + +However, managing and analyzing big data also comes with challenges, including data storage, cleaning, integration, and real-time processing. As technology continues to evolve, new solutions such as edge computing and AI-powered analytics will help overcome these challenges, making big data-driven predictive maintenance more accessible and effective across industries. + +By harnessing the power of big data, organizations can move towards a future where maintenance is proactive, costs are reduced, and equipment reliability is maximized. + +--- From 8348b0dfbfb7158a8a649b8ae716d80a82703fab Mon Sep 17 00:00:00 2001 From: Diogo Ribeiro Date: Sun, 13 Oct 2024 23:41:18 +0100 Subject: [PATCH 6/7] feat: new article --- _posts/-_ideas/2030-01-01-new_articles_topics.md | 3 --- 1 file changed, 3 deletions(-) diff --git a/_posts/-_ideas/2030-01-01-new_articles_topics.md b/_posts/-_ideas/2030-01-01-new_articles_topics.md index 84633cd5..958f6bea 100644 --- a/_posts/-_ideas/2030-01-01-new_articles_topics.md +++ b/_posts/-_ideas/2030-01-01-new_articles_topics.md @@ -20,9 +20,6 @@ There are several interesting article topics you can explore under the umbrella - **Focus**: Basic techniques and traditional approaches to predictive maintenance, including time-based and condition-based maintenance strategies. -### 3. How Big Data is Transforming Predictive Maintenance - - **Overview**: Discuss how the vast amounts of data generated by IoT sensors, machinery, and operational systems contribute to more accurate predictions and better decision-making. - - **Focus**: Challenges and opportunities in managing and analyzing big data in PdM, such as data storage, cleaning, and integration. ### 4. Machine Learning Models for Predictive Maintenance - **Overview**: An in-depth guide on how machine learning models are applied in PdM, covering supervised, unsupervised, and reinforcement learning techniques. From ffc2a80ff3fc66d84918c1d85165887d53cb7436 Mon Sep 17 00:00:00 2001 From: Diogo Ribeiro Date: Sun, 13 Oct 2024 23:41:33 +0100 Subject: [PATCH 7/7] chore: new images --- assets/images/data_science_10.jpg | Bin 0 -> 80285 bytes assets/images/data_science_11.jpg | Bin 0 -> 109043 bytes assets/images/data_science_12.jpg | Bin 0 -> 101270 bytes assets/images/data_science_13.jpg | Bin 0 -> 101040 bytes assets/images/data_science_14.jpg | Bin 0 -> 405672 bytes 5 files changed, 0 insertions(+), 0 deletions(-) create mode 100644 assets/images/data_science_10.jpg create mode 100644 assets/images/data_science_11.jpg create mode 100644 assets/images/data_science_12.jpg create mode 100644 assets/images/data_science_13.jpg create mode 100644 assets/images/data_science_14.jpg diff --git a/assets/images/data_science_10.jpg b/assets/images/data_science_10.jpg new file mode 100644 index 0000000000000000000000000000000000000000..924b24cfd93439288232cce986f61a3aec9af911 GIT binary patch literal 80285 zcmXuJV~j9N(=@uaZQHhO+qP}*UTfR7ZQHhOo9BMsldmV4s_OaClgy7!ss{i70K?41 z)4|Zq(hT6gu(dQ}w6!$+A78XJbus*3?!O?kFtKs^|0w`~gQc;{|KtA;;T$YoZ2wOH z{R5lS;eGf6|m14$S+Mg;fQ*uizIq_4vk zoX!e0Rh!Ny5CeL8C2&(n&f`~UyCEZ&%0h{s%6XIYWt`b_DR67uvEz1iI$#xa=P#*S zmQ^5gG=>dfaSlYarrsjG&^_!DD*a|7`W7<;=RSty+xT%k@wVXCnv4-oCb$r#-jYlvX&f`PgKE+b9 zfp|aWo~!GAC)txp?QrWl1_mlHC;RdU$ z-g@GK#8tDg#39tI;h&(yfztt~HG{a6m72OeVcDA`0eI$FryCPAPHdIgf_W_4Ax{2$ z`7|NKr1oHdmr0Ea>(@DUIatuSMhu33)|F`OL3uMkxss%GC_yqN6HJjjZ<(}6undkC zVXlw<0c>BmE=!OJu|)K}9WvA5_@Q9KL37`B1Onyh;fG7&S#>~>;w&93_$`h-3i@PP zZ4sINENT8W9|oeuV^zrxgsVE_i3vsO3R)_;@|bAl>9SZ`#NSx8(Dm+Q_8F`f^RpOU zwgV<|_wy{_Dc^y)Ch|LQ<2Jr=%7oQT*h1|GRKze*o` zg+)8O8?i8Qo_o3#my)%-)yJ`p2Z9?7MbfqSyRrtP;ZgvMJDTC#?7StWbtqd*2e6sO z=Ju>Fm#2GMyvrHOL|c*GRJrqW*$+DU*CYi{4?MXlRm*1)T0_~=V3rya2N-O>-c!WT z$Ikn9or%D8evF-vcwW}wE#hWi-&!Bml=O<%8dpO}0k^kCRXIHI3G;nkX7->6EZ+`| z@_ZjID>cc|+sG^c9(zTb@-X1j0kf72Px5s&N-6kG&)8H~NC0U%j+B{?OTU~;R+xW5 znG_l`GgydZLJg*w>cbiCx1ibt^s>}SD4$<_`XZUi2IjbgC<-Z{;P~e$%J*X0Y-Avd zI}-4bcHX?CaejJMEB5Yf4C5QU7t0gPK^f5JX8U_nZL)leJ|mOsa3z!aNcKm_9{v<~ znxKwImvkVT9r(z=!~xFTPK_aT)H435guJB~fMBOOFVLhtm1x|l!vi{C=d)JGD0TO| z!b3!3T9JJ_HHnVdSz6k%B!XA=%-@1=+Y!qHJJUW_@zFwf5KD{-T_Ruw5=osoevx26 z6+3iPV1BfAC$siWR*MdaUsnH7K^z<$tTGz)bMsB*slv4Q@1S0$d%hbe9RPU7eHj}_ z_d&6XQ$~31eSRsq%2#?npUHoS6LsReW`87t;n!~pwR)Q2??ER`Cn(|T(fu4S9CBjP zr1O>>1N&%UxIp*z#o(RpSf!1~lJbIm%~;HKrS&#ew2j9r$zCmbnr#=t7Uc0?O$rri zeqYh#s*YV{XaZO)1dh%}-F!qZN*y?|u~SViB$y^Pg)U+x$^bTeUSGF}57kUGQh&&A zj>=kLgeGi9j`nGea=N*5{(G15X5J|kjO(1==*oUJhK zfGi=eSGxG@?h0~uElcsWVpgqyABYwTq&I2u8J(Hb?HO;S7=hUmkuHO*qcckjK^JTh zJV@S4Ac1#r5VGkv93YMUApth9eVYJlBlV@}5WTTrrGvurTEb{ME>Uk%y z`^1<4D0eb9xl)`2%GpY>BP0M+A`_L6jH>rM`i-@yRkC%X>Zj%2Imd#TWjHT+c@!L; za%j^`*ja0ijv}#8OpLOk_P_HKX{l5XYYhY}uytE{t9e`)D^OW`JilO>=wG_2X2ORH z)~PrO%%n2^e0TB9=j*3Ok5!(T*Yg>Q&JAMK#q!e^OM{+H%i4d@}hF2OnaXv#yG{+y=^!Ca1w zqpP$5irr^s+-Y+{RVOy}F$a>E>Ea_lh8kPq-&XHcJtb?T`3^!^$R?zCOjAnPXib46 z+sL-M-7T!x3BcoV9j?j=4bvG6$$KU>gH@<2s;bFrTUdGzHvw!FRxcFi(QmI?1)M55 zis@7eSp|iTkavaIcAN`}4$@q}=N{>*1(a#Cps0-Pi+`mG*blglNz-T>Ewj_pW1T6g z-elUckV?b(5T9X5i^UCzJ4fefS^FfAbrEc7^&mA(wK&K@X!89IL=_8 zX8PPTf)dZaqHcC@Z$c3?d63boYT`-t$BjW zxOw~xLmohq+_I4W_9a(u!{E3!oJ91)_m(1%-e1^G@Ws`HBEkL{-}=%iHY$ht9TvJV z2pPy9UZ32o?Lorq(^_;A1%(7izeZG|Q~_?QEEWTnn3s4_t3Fj0uH zGC&)z_j(vs+18>V9a?}(_v)!In_07m?>GXO2#bKA!ob`}55`Gb3qZktay->@L*gvF zqw@u+!Jrz;uUgxd4K7AqI=LcH{E#3Isi<-1?mIPm4hT}YfAnm5O4!s%(-T(AVWY<~b~Yu0CXf~$)WRZePt z&wJTD;eFiBVcuF0Go?DN>26 zNsRs8Qod`KRBW#|I)vFaH?DyS2y#;$AJZ|c+)}eMr2zNAMv^i$T-lBBMoJ+Pe1k+U zi?M_QSt^x15W$gcnOTHsmp1J;{rf7@j$w*-_f4h!-I$aSe(Y(?6^`Z(o&mQ|K70&T z^!x)SUEWtGJnr!{;I!I@66!;*-&MN@QmB@Bei#~&-1M(eFG@97zFy_7ik5pjsAj`} z%mjS3E@P0m-@69^xim7+fxoj?SD~c4A3jYvmvhUP%IA>mHjA)^WPNUaB7xYbYL)`G zo2YuIQg%->F!qK0?yklX?@VF?Va>m!XlL}P5|e91}ZT-nZS3B zl+~I{6H+U)8y{1_X7^k*$eeI6LzuvXqs`pe8t6dmp*>^qSvJjj-j};ONKoDyBEK7D zJkXK)GNvD+hsLzfbeBOzz-r}|bU~aRtr%h&PnM;sp$2 zq3G(PRwLoZ_>;Q_j5*el#Ik=!Lwt|8#8}s0Q*J;AIKMAbToI{6wDS6C}XoVdUq*H;4ZHWmKtW2*g@ zRHMLQgfHHJ&){zn%)C!PshN9S(JbUP_w<}vFAp~4VZ_|dL=2g$Zwc$W>0;}WDwQZK zXLp!hGD~*4Id5NP1G7V5*N5j*#Q0W#s{LdBn)@9acwlx(|6*c>XeyLA{3IhxH0c|k z?H`u_00j2~Z+;*lap|t*e-IYWo>Yg}x$J?MYG~7)AN>nMMh0@F)rEo(ziC=ALfLL# zanTURz`|Jb#gP}%nm}sNy88158RIE1s!!s#(VTc`TKQb5T{)XsL&$P4`_~%u{*`N* zLFkm3wTG%fD0NcEPDw=0kHpPJwG=;gYi&O;C(y~I3@r=34*&pViL>bHhYJ9DwI#lo zmb3OUP12Q<4#eMJlh@|_FwUOCoYGpk&xIywQk$NjigSe^hkWuJl&sDkREwpR{-ILF zRjsehI!AA8wTgBPs=K_WI8aSlps&mOyZ=QAfapf7QQ)_9ViZ zrlbSGAhsnUCugv1KY1 zEs1Gvz&*tHA^zXW7ClE2AfDA_isNg#;)YGxE~!Tbs4!UQ zL~q4Irq~c3W%}Q}W!i6(A(l6PUjltOC3iMqWy9mZ15>v{iLQ&kCKdTHedXG)E^3ME z`~h`Gs5ur=k{K`&kGt-^%cgpkJe_xVpD2pkU$%1TFSTkYBLBjimH?f%)sHxYGp7Y{ zmZ&l4A^aikXQFCZ;aND%K2OVZ$ArS!A|5k1YPQs020Wo(;98$U!U4hAC=3O=M=bxj zjs-ltH(!E%F4h`@%j>du8KPCpk$ph1LU`<{qIXs>WC+2>$KT(dp2($2{oS3^aSA1? z#WD4G)5zp?$UUj`?@(r6HYfC8qBUXfRL_cF$sPbt-Lgl~(#;U< z^=`YM15w{cqMI_hL#Rocv4?I&>(?mH#xa?}+IVCWSMomv%*O)2-e5`w8a#mEO;Tl3 zOnas_%gk8;Zj0M$(qXuu+m&va9t9Ryp&2BtUb+7qsNId8)r|-! z2j%#VgIp5gl}Khqc6({9!YtsW(977akR~_*n^-u_IdxVTf!=lruU<30(2w4Ny2H2d z8sg+C#TO$ME@Fm8%CGkEtx~dZ2`*F0cy}?lQFEE?lpvr^0_oexVk$&L@!R)M1es`w zdBBm4Qio}~D1Ke3--hp;WQ=DNdK8?0%_$6gl~xt>XtfqVH|yl)1t_k1#DI38E~0Qz zZ^59*px5dps-zyZn2s%mzn;SnI~g9EiRpnl9_5`xPkoa_jgK!hUYd%1qM%5_2FENS zd)GaF&vhVRjuWryO&;PS=vr`TNGdG=K99N7@7FPHtQ`!{%n{?~Zf}*+MDl}D5%Lrm z$-sN)jAD~$qLeV+yf#9aA@@(L;o3QuDa6n}D)8H}a6g zQzG`Eh0CDuUY@we7UwU|AAx-(s+6p3iM?1DDi)m|Yn;4-HinbWPAyu)_%f(d2=oDs zD6$7v1Kqrl3QJx>L}Y_NgB_8%ryS7oSVUF!ra0dRjFkz|+nuoUzpPmB6*~P_x9up} zGH{`XT9T-%5agC$_uZBat^6Ct)&vU_ZC7^TyQYbaJ{4VnR9$KV?=>-MG1)%q&JKTB zNWF7g-l{_0OB!|0eWX?W3P82Ay@2mrxSoJ__^%&;0_03K`nApUs`m|QYpn3F_o6LJ zH+A=5wGO6INNsl{BVX!L0mo3qRCZLePhEq&*}A#{IPcI^=B)Bmguw9Hyri zj4^COO^;e_T&}@=)ED5Ati+!2IvVgv_AMQfTD$8Rues6{z?2-<2BUbxhj78NtdZfx z!GUT~q_H2Gh)rrGl>npaB#5H81`%OB4w;bkg`+&Z3RaQ1xuauXpF4CHTURSG zo8LUf(>8EW{1JZUuy-QlqxDx5wHcy-2SMZD*>vCp-p#D-IRsr0)Ms$Oyh+hNceE8V zn~(~p3YJ2v1cd1PLeeR#1Gc1FyJ#pqFnY6R zET$P4*Th zM&SXW>szz}a)itj-IMvJ5C|_p=w1%Y*L0 zKnsQsrPlxP;i?z7ccDPTD>7x_F9`PjCbZr?RM(R2F{h3 zAuMBXA?yhmmK=#$XnGvJ2I`p@Y_cwX;<9kQ%BqE~dLx4GMIVz~U&S+0K^##mu~>3#I)#HQH|8lg8r=r`|U|(3>whnba;H>%Z4L9HL({> zZdBN(SDxKfY9Ehu=aZ4ZT4gA`AA9?DoibK9>{-$1EBd6}k`QS`5fMWNJgk%AnpW)x z;dZHNEa%rMR`Mu~jakErEw$Z3cs{}euXFdGVc5N7eZT{P4jyLfC_(uW#NwJ$Ma!RSlz0qzb_z#f zn$7%6e>5X#g<$nm0JhGOcdE|Jv08hOo&0000N3e$O2W$PRG{(yB?~9AU);ig&pM05) zvv-@CSiwHBcUK0zyL?RHh)i%X*$)a~ARig5%D5#Z(7AtDR)V$_Q}yRrE7Y=|@+*6^ z+Gxio+LBn*4gE(BZ+WV+htW?LGlGM}3w3tMa@Ewcr*@+D*`;VG*g9n=9tV=$)aQ!4 z7G+E>2xRfn`H&Cp?G3ciGl-}m|8J$m8$TRWfb$(sbUriWcZRGcKqq~KE~~Ujc4Zh3 zF@ncMP(H`VzWQ+{8XMEDh>YBhcM|AG->1U6;VoKNqq&DS8R{eZqD{ktNsSyH>k6U> zLl4%D2`K|<5Afm#IjyLi0yEs&r#t9jWC!~D#+=$59{ia)B~sw45iX(THnGXTKEB|a za=0weQvq8qVv3_u?qBqs{={q)BO|(BKtTz`6jeKZi1(MS<*>X3C?o{>qDm5S5d`w; zRJB4VK<4Q;VNLGE!TN5EbI@VZSgYk=eM`NfSpoXeY08}4LRvvC0Y`t=rfu`NJd7k~* ziyzv*WE;JY%R1$8*Tfb6V{qCt-9qoK_)s8lezT6XL2mhGSznTITw>i%)k>z5q&ScG zl-LFbOezLsQ~upc-OxYd>X{cdqX5A%#w4XStdEI3Tqr{s-q}e-iA>!oY#{5*Jpq(d z?q<|_^|956i7x}S8q8SJNa9ErOCOi;FerUx8 zaoK#dlu@9|s5-pZ7icb%`aXfg#hPFLO)tZu`Ry?LFD$2Y4+i(8L;LC--==?1`0QSs zb3MQB;KUOTdTd(IsT_LVLA38lK^9h>ErZ!OCf(!tZ@l4> z+NqUOAR0Ou9oJ6$y$41PrdZ>9T~%*UTsnP!wYDdQ!} zHl#B|qBL72!EAjm6yR(bYhAhTTnb7rb-7%Z1Z3fmBSMkUV}hTb=dL7CLA(W4uge)L z4v+T`QlSQ-vutDoEHYMO6YIbpV8$UnUNRXEdzR|fVwJh9>EjO3imNSceT}eO2GO<- z7+JxmlLqF2dXvbGKb8OZ11Ov0EU(a&v&U=NIN$R{utC_nke>=E*S4AG*|k$u(oDLg z71L$Y55Gmwy&gM>lPi?i^E&j~5n}Py-O5&}7u+yK(evbne5&(2B+sRht+u<6Op5*< zLOJWw(;?V+ZN;bSo;TA}|}rdWm=p5m_)hD0Q%XodA2*-NytSHuq%y~B zTMQX{G-O?BMr%_??<|y$5L*_rlh5a3ddE6}!evd72uSOtldDhaJRm0mOCnh7$};{>>0yVRJP zS7h%DdN@05AU3fGE~c)j_h(-6`6?P6$!cosn+OshY@4GnaPTZBCIR3EwRT&=FlcO* zfE}{O5Klj$%2Eazy!Oz0K#S-#epLCUA{)udh)F z>C3lMPeZB2fl=}^P;Jt+S-9wxMkZ1SXlBg0lRM+u5Id;y{96|_Zi9z6OfZb;ZX^;H zI)7dWSz*z9iksLErw3yF~lQ~$;! z6ZStKfG9?Y7$E=}gZ9qBc1U9-{XSL+ z9zh#KDTp18HaLCC{*$fjfYx|Ly|9)nEA9cDWdu zl}Q#RfjMl)?Vw)D{SwRA5%9-0-REx$u{9cRd+Zsw)Q)y5!P6(4T^D36tsG#M(VwAK z=UMf)BC53#14EuO!2lUlx;Ct8}Dmu&Dtin zA93_@b|V7w=z`Y7&g5o~d1+Xf%UwWLN9e*Yokkfu6%BSC(`U=E3rCbf1~PKaq~&2F zeeMoPY|h*;yQVAt?lSNd>>f~4*UTl@5zS*x&|gnYW|^-mIjN58;e|SY*Hp0xY!pq# zyY&t7Pb|;$o&*TPhs9*P@ST<&B}vNV{^#m{Ufw%2=fxAPeJha9) zROUkgj3#^Kko+Pf6@zMrVJq)AQYC2TNrIdFFQ{k5gL=N^sQeJ)9b!(|wxsX$@!_}VM>E3ga2rAT%ed7 zq5o8hM$OqPk@|Fo1EE&5)fjyY!ti-#A&Tsn+VNlGmNDDSPw#rvCvJ|_aK?*YT!@Sw z&Oe76Cp3dfssyW_!LprBd+h1<|C(LZf=+@5t>~DJaQk4cYa)(&wwQW!7SL6(LpI)3 zAy+-|Z&_RF?p2|(^f4|l?g7Cvy>Kk3M(|dbpQi962Dv|smHR!v&DccF0OtAK&*7F+ zd6V+d#iJuKauU5Dq>ad@Z+7G|8HfBI+9%n=VdE%NaaCj9nAk-6ELOqp*~M~N`GB8@ z9mkuX{6j~X;7^3!cLIU*tgO^>NzsDYU233s)20=Ay2@FBgJz;8S_V^@HKV$ZQnJAk zfofgX3>t7s+~J~(%I}uYkz2N5YkdPiXMJdsiRtEC^+^CCsWRb48^R;vU3#phlm1t(=Om4DaeNCbik0U}DW%;w;s0iqB z;xxr(M?hHPC!5BSSz}}WIF7^3O8K8oSh~m-&(-q$scM7b&N#%_w^O+}Sg=fQn&3B? zNLv`3sWF6rC;X$22lf2oheK~hF^;UhhN7Zd48|4>NAp@0#jMi`-Q2tdmKMdWnPAyqoH3BZz5)K>g^YZhXDLpe#5B-wp2Y(EQ#42&xiv`w zwXICSiU5o#jTHHmtiqCFuII0pG@HJFd5UphBJ@&}I-Q@DYY|z4l{n~#HHbzXr)aPg z6EiF=_?33ur!9m&fnGXMZY2@fnfnFk@wu>5$*&u9jyM(1H)@ztk|&x+Y9Rko`qQgJ2_3?{UPfU_6p?O$5zQTK4hxaGgzsN`V3rupw|K> z)AlRXG2~a>4MYR8v-FyakRX{D-z1R{4~Vr-wb`6S-HF;+HEk^5lr0xnHLXJK+}FXc zJ7`ytZHX6eK+GdHO)?{Qs!+unD5897b&)M8F2yX9o3csezPhvzw;wD64;Wl{M!`!w zUf;s@`O&jejuvZ@hH)k=dN0Iub0eIw9widGX=jHouE-a|fq`Hxb;tlT0A9LRggGze zL$FZVQ4u(-F?nO2Mj!^YY+bNrogPbT;YgYeiI2NMj*-}GM@0?OM2h%c zH_Br#L-nR;vstyp4#{95J6@XZ(prR!h#pb|liSkLFG96qhmLS@kqR>E#-#>fcES@# z8swD|l(k8LE*|{#<=VXll0B%_-nrj$VC(tl01G9xll=0R;dGjKuDp6Q&^i~zSO%$No&2v9JH&L2*1qV0cZ=i-h^Jo4*;#=QDV znWCcL2QMnLHdAWhn3Pzf^_OJBYC|=s2L&H=(diGE-s5d_gVUN;pF=z)J(Z>U{18jI z`1C!NfrBnu>?Q##9RBJioR_h2(ZHj-t_ea36LooY24yV)QUZkp2~MqU){s^DWU;cs zaNT&##!7SV;_YhtSTt--rjqyHWiWXtO>^_j`vP!jpJuvX0tAC9&KM#S3dIZbZ zR9QV=c?#qb8y?9<6P7vyi5$~FFs5g;yxXxIb=%TBC`Ge(h6^HL9GBKw^k2`w=lNaf zL58TI>3RMEiHh}QQOEL=8qsdXgDVKkE_DTL&c5=l2UL=kpwYe3(I0cSI78yps@V*{YQbmCCgE*IdDxchH1g;J z*wPE~+fyb6;FU(wb?)3nTV^9nv--Mg#+)A;Te${X6=B^K?)rx3fTIf8I*HQ_kcHY= zUb&0E%r0^dnxao2VZ&>(tgAs{A$vSw{Z4{8cK`7vu4&f8*4N#eS5Syl>c>?xBBVohj< z*M<47dk6mYt-bP6lN&17MK3$joIIHwl>@esjUhBm{FW!9i9E>hbN2U^f;=`-JdKjv zV-3n#>nSVLA|ElIWAm>a4bXh8qn?CyApSA77#nIsW#`rGF$Gl58#46s>2iNX?=6BsK%AXh18Ln?R%KTp2@4N4xloYFXfJ&cTO$|`(|%AaeP`|jFFm& z_{i%jMXuh`#z6Wm7vxCt)Nu^>b~&{Uaw}i16f9Qv`*9{}^TOYEG|f^|F2;>ZB_!@I zDPb=j4E{gg$vJv)*AJ3)jz~yhUZ3KgZVqSd6R7FY2kRHgJ0p{Ub$@1OOGy2Apm!>s z_gJ&z#P0&#G1DM7nXrRRx(>_qwW*n-la1&)97Y~g-;y;r1^JB&@+6o&uzolU^1ud> zy*7e;d!^2a;Qp4}@F-Dqw_KCpNBkVn8lj#l7zx8htLwg96u)F^L4f(rQ**LAf_#3ty2MRJX@A*d>V^wFOq6{7;%dwNpFqetwftIG*!2olnCmnS&vYo zGg}oDIteDA37+3^koqC!xUW>5&u#-O3iY`-%UCEtIhGh_WSoIbC%C_wNAyphAf2H z$-MSYHiMrvXeAx9Tzm^@0DQ%a7=pv+?dI@|T@&eM_D$)efn2DEWtL^4XRTpbI*)7v zk7XearqnOWUdrfu3d{9f2ugRS$SW6Rys|jd6K#phL|M0RwX`K++-HbDpu_a0dX$ep`x2@ncO>9SR>=17M`vey1{=wV{ z5hm1K7St${eU+Juc5>N8Vs#R!|3c)i`7P<=ei{_}xco2RWKI+iGk$^~@)86SHarve zByS)1uOoYek1y5JB39|VU2_$ETX^Tg{y1lUUv~Ddq!&XR#KFbi+&}ZoqEbElUyO8t zg74z%dq~m|riUme-C8X0NnE>Ajoy!X4G0pZ-Xkdp^a5VT>eFI@(13CvAw4PzAjg{t zu>TX(@h=K`woA^9Q07#6^so8)bOcRkI)Hrzvr(}WAL=hlB>GfnU_ImlUv;ZhTvnG~ zynqGcNIk!jsCqJGXCtGP1pr6&R21SKKQ&5jGVY&Zu~Zvl!ECET_hp$_Ho(b##;CU=GVzsCQi-*$ZXO*P2!w{X zk_>!>x(ZH>yMlO92tJgbFbvZ8h|igDsHOuZ!UW~~<5dZy1Rui$z3Gur6bLaSU(5`% zAD7mf9@{`J8I-ry^~pJ$kQ$i)^!OG;9obp@I5_!`m!dun;d{yqxE2ozZAqmrgePYr z`ILs7)f7?nhL@lEU)9%9Xw~{dpvA9Y5l_|Ycst7j$C~8=$jllrmmqRG6RPt)rA?fL zmn}nI5uFp~bDSlOPD*Z|L`n<&_CmOUKZ5b_9okyq4nQL-c9wivpRCk0Xxdia4pYo* zhr@JuKPTJY|OAQ|x@F1e@ zRcW51`>+pGawBsMNuOQCdHh5}2F3yU{Y@D~8S%H!1q?gWlZ@%q!Wr zR#-EKCj^R#%?<|d{PLZbi;BFrSZ|@lXcURoprS^|Jejs)4VF2e(7yhif6oxWs|ghXUBb%ANKyrV_P+`6ag5t9AU386oJ>q&e5B(% z_8qY>1qnnj$1vu1#3>L>_~acoNAYm0+0*N^RS83m)T(Ecco zQa@Y_ceWC4_&qL0vH!}M@xp;mk;-R^+R)u}r0>~iz@9g*sXK-sKTwu%X#kY#o0j&Q z&!uC&zY=qpIaS`**&qe=n9b@i>*^Y2nxvmHG6-oU&PI6=;;nvc8FFEWRMxlFJz$9P z0p`f~2Py*J64f#Tg(VxY2MrENH&*{zkyvgPWx9XWSUY7Rg2F9=a%Xw-Al0pB5+%F}f*~sH|60#RL<%Jc3Q$QeS>>)~+w6Gki+$t^X zv-|u#2`f`yr`fy&kYxzbDGA`1@V;&v{u4IFv73hkVZU`mt?3uQ0ix+V2A3-KIq$TZ zAY4TFpETQR-Z%+${_AWrQ-^UuD6F40C`Rz*5gN34JL;H#GtYYsT3Xo69uj^+Md`uo zflqN6?Xde7zLTTKf%bvuc&V>P0gy$I7KMArDmNT)4X?9e6!q?Y57~mD53Nn3)$Aac zclpN8cDVYfp;KW6x%gz6>NRyBbbaZG1|LI5%*!VwC0S}*GVIS&?%Ct6znf7xz7oqM zj;E!RwJw_vLSB|hbb4Y343t0~bx*o%`<*@b4G%BZ_#kcyWjW!YJXabChi=NrH;+Bv zbbXzNYW?Ts6UOF1PcPdFg2sE5jw$H_)pTyJ3^*v0hil1wgH}?wSzkIzTDL_*VrU6@ zWa4JJs@&6&PMX(5#}8M}XmRXm{IhU|QZv&+fiRjX{Q)kUj!3+nhQ_RI$z=iw(%cN& z$FT9nhzCg)briC)c4ZW;p{nv2MsnQ-bNie!io|jh$sIE@wImZM63S+2bK9B8_3$el zeQc3Y&j-Eo=03A-;JKdTpAP~lPE6*;ZN2bX4#^n`D-g;n2T8P!GEyW!60eBj=&YFh zwykPX@~Tg}QH-d0D^}%gs^d_zWb+Z>1)TJN1_*~(88v-|YkSR#k2Dq#8r6KPF{F;X zsGr3e=pn&I+3Wk{;3fo1w9and(RCkI>|ZoSqxNeNfYw-AW#K@23_LC^y@p-d{Sk9X z{Ri_U4R1s#!L!{gw#M{Uf}X?D2Hj-Akavx~be{vP{T}rf9z?+|t0%F^(X>0Y`*Ymd z$z3pMPq=CxL_MbdC-1$$jJ_%J2*#(x{Q1S?(cXh0Had(>9P7(k?cQJFjmAyiB)pFQ0 zX=ES1a76~!!_THSpv-AZ8p&YPd%ts!j5s@|gyS@ibaA?a=&M3*ICfiOCqD{>xjNzvTLR>y69J ze@Fc&^FO8?)cM?9ZKMb|=<^$kEUo)x-NF2iASyl3Lgc<2j3!K6bT9%bR#Ro5L?o3IXheYD=(Xr?+W zLn7)%V-)eZ*!L#Cvc-}I=G~wPcZsQ)?&$wOIlFV)qC3MYTK>7{TGhAb*0iltL*UrV zoXi?XA^W0Ktk*!jPGcKivZb;2SU|KFwmu5TR63+G+~*HGsDqCPaSkG*=oIy%q4D75 zkPtC=P=7%>u;65&HXx3m+I-m6To>nGN69FjsXIJp&o_6Os%Hac=l%;JxFs*^ygq`J zPs`==uSv^Isrml^LqNR04ryA-^cM<>5sns$?|~=unSGFf!dU6afZ)}WL^}K7Ud|Y>|u;(Xr$}sCv0$qdgvE;-53v= zUrMsjMoalWX(d{Izh85)fD0aiye>hipso90K5}~Z+4;BjEdR)rLs_IuEt1)jlVs4I zQU=LX8{@ZBs;XkV?1awB_s`;8Bg>n_7v7PBF_n?-ZhZlWEz#_cy=6AD5refetn;@= z>zW@nudO@GIr5SR=U}DsEw2Gi6M!B{s{YbuJ= zvDiNa8Lh3q(<1>~O}mCVNJCBN_v|2~h=F9F8r`u?%}#S;{BlAvj=PE8b@H0N^rU;> zCcQ*BfRUL!i9Y!trSj1 zaAdPal2CLI9qbJK`^xx!zUNr~R@CPhCb;??fRlY~o`_Ij~0rjz5>cHj-Hk(~Ps3uALp9#&$kuo)(r(UZo z6yS%+5nA?_ce1z2F``9sk&gIiY+(uxGhYaYlB4!mPQW&J%uCqoP*s;$SH(Z)t} z5ASmlZ6$_eWC^nZ4P{`f+L*3YH#jnq##?%9*~s7UClITK4~?uV-{-TM=G)D^TbH%{ zylNYbeR8*dpApKtamXLk*c=7ti&0DzT#C z2tILw)VB_6^h;8xQnW8BU}Id6L#HTRpufH`lAp-E0SY%z>6LWy=uv)deEeq^a5qHD zD0cfOl`+~cL~h0bzOd3So0JBMnZZQ^i_9NM{%as21#YwvSrV77RD_xJFo-u2@}ABQ z;nx6bx)?Lg4(HvF9dl?;CCKsy>LxzNUS)(2V+XH0zb z2LF|jrm1UNjPS*0V6Uw=;C=OS?$*iZC}uAMGnXLr*cu>iT-o)`B_D11vAzxF z8D-StFuWy0j@H{)ohD6r4E&sE(q|$T%fkcW`>mg4Q;c12X;zIl4np z-J<68@(2o_Z0eJq5~9=#mV)H392q%kK)QMWVHw2wN|>20C0QPjxCanTiyg$O5p5_UQ$tk zFxE?;TJ0YQU+FaIT|ciLJKuB4Yyt1Gcn9e7;vDG{ylFU8ZN{gvhw{hD=kPWBUE$nU z<#=-}FcUh*dP^k5| zYVWv6eIM@#?ImroZ|C`auBz=)ioLa^^xTHr$Q5y7Ab=(VDW|Om-o#5?{3WKdQa29=0ek1x&=O{2&8?`V8@qFdLs&0(ZsuxCA=Nt9LZl5W>(oKBUT5{3s!EHFJ=M} zp!m5Q><*It>f5%ljh93ROwTf2_gV>Cn7Pecl-=b$D;+dRUjF~j%$yz<7sRiCApZ}+2laA`D7Vw4ykV^NSYD$c!|{#r+x5yKrylP zr10W-lt!xkx6%EB&+c8fkj3sG0`fLEcvR!x&)Eieq3LZoA(wi-0PINyq#E&$2nm?p zwwDqD>v<=TbHxM`KX1_2BQeklb)hAC^a16owGPD4^0nP$g03^x$VO!xLyfF3x|f9g z*B2e5&J7%Jv~$iZnBIEbprq>lUNoeP<`Q_bW-ELAcnmwfQ8!`h^hodynu6c!MBp;y z%VhhkS3A={3uKi4EIji{m($;07aTo~bS(FGXQzIe=wH^HzMekYxL?GOKU|E==ioGL z^q}|rvE)4`6gU`&X|dM@&JdDp!K%c7aDOSwSvt>)t-FYG>k^$n7Pt1!05x{~(-*g$ z(X`S@cPGi>{gP_Dz<-dGiBt|49Gu@_{zAjK|AK6k3#9NDOZgjoEV{5p=}cd@oL8mQ zg1jX$YbSe;zH4l;GiD*QtBoESbZaI8)3lS>i`H8a_d2K>Jz%I09qZ|Hbo(ID6*A{- z5ooWt?KG#RO7ji{l0K8*%w#sJKxmX5&Mj;#X%bb|S)Ne)9YbLI$oky$!cb z+ae}QRSgpa%{`YCb{LS&r4My3kte9Xdt#|<8dh4l00>e2o9e@r1o6w@d^=m5H~}oz zCs1CTzJnNnX%NBR`Dp9dS}#>O6HLQ-405TCl4@dA_7foVFUYR zLL*@_eLC!HS!_q=Y@yj1uYxXb^QSo~=%B;-{Jtw~aXUUuu4@}99?zT>Q>YEY04m(f z(4;bMz#b%J9_9*x4G3)O^1XefP`WBd^3A!}rc`#L<51z1Pf2&5_X}2vRw6&BCwgHZ zDK@3A%$UCjmHftV0X1;}#U7Jie6sW}Y%OO=OpEXo3h3;>Kqp66*tzs1bmwLQMqbWF zi*z$9EL#rB=gEoFE`c}07@$6Z)g*&AL1MePh+0`ZilM-gAxViqHipsN8935Ge?Mf; zn|%!=V8s6CJ{bVqFD%O!1GYTKO*4v@q@S=(a6iXK=JI}`j+3YxRRN9yje?cQ6&xE~ zJGMl87m-cw8rFJqE$x_+m$ zw~Hw^KvMp+9yBQXn9x?oFkE$Nej({Ji>2a`uy|#^Q6k>}yvSOfo@ZU_6q4KaPNuwb zK*0Pn!!=`xmX8E)(Vb$!`A=uFWoLS>_8=++m1K=o+QbGr%FKq$FdHu!fWYhCJji# z&aKxaaI5u%`kNL$LvkbEKGl5k=+Vh(OkcrqfVv#957cEjA1U0CpM91`Bwk4C`T%)h zg&=)?57D>J^lSqgf|Z!WFmK;lIEin!h+&fp3-;454Oo{UoqPUOd&!d7^la9`d3@L{ zN+MQ8r84yO^w)Euj!;3JR9dLmmU%&XvEre2L`LpaC>fTp5{*y$%cg+c#+_@ACkM;muvS#XI})U*3J?r_Ls zNA)O;2?>dtuDy14Ao0nHVwtWY@sIWpg|8Sb73V2~5hR^0%Yfs~9zADD%`_~y_I{Ce)NpayV@yIRV(`W$q zc{rj-&cs(=ZU?z-QB%uNsX^~;&xHCdcG|1^k4G=R^$wAWUn~g-#7&}a3K#00&`e$u zC|{Hmn-zD_)r3Zp?rFlFp4CA8w`PV~7x<7v;}C$THI`0scd<8@L!O_s?V1=Fz>y)z z3;2JO^_i{fdauD#gdT5S6>*js^fg$7rw@0Ug%<;5Z*pY~+5l)8CNEXeA<9X)u|R%+!G@6$ zgFBRp7~Yh!t4S*q4uG0KWGqkirM}$U2h6N3>>_;~LU!Qy%_`AJY!1BeRiLFO{YyBH zxo!TK*R%ryw;*JIw_*?+5FzIV1{fp{BeLR3GK(wsG&6K~#G$=jhy>CNT^+e~Jj*gq zjK=nI9zr#pkU2Anc|gk1Y;H(jxDC!82Bj_kJe9E`2{p(oa5Qna@f|FkfF_wrxzpPa z2s_ot7@s-Q1%ZFi8?Hs(EhPPE$+nOPe25%%nSJgPp zqb}6!DU~0WZCZE?*v5`X$1ouzD{nq3Jq4BJKP{MteW$2bFa+Q;16wp2qZhD!gwfj1 zEf~aXk6Sgtvz)5#MxOcC;A08OBD+r78W-CHU}gA81zNEmais8$IZ-u$1h|F&6T1}aDVkn5=QKzkX`TVha!6!^Q# zf|J#kqIz)QT}~F$2|aKbmvy78yNgA{^W3&EDU-Put2u-7&H>i22hl-KgP^7qfFvJV zCC>-L5C6hdId9#NrgND2T2HR3;*K;eVg>z)(-lQqqyg4Do-B+kzpaRyO#7SIm&I;$ z|0!;QI4bT*LSG7tl6Bp+8ts*8OEvK#G6*U39yL$;wE$9WGXJNG*Y2nWqD*C`M~&=T z=!T~@mE}zp&jWfWN4*!bb6`~vZGiZ|=~_l?a|Vpsd3U`X=|7&f@-ZEFJlz z4pPL=>r!|^KLU2RG7eW9VwAQDZonVQ7tY5?&P5s1LO0UmmV?c}EQ9&J(tMx+s+6Wd ziN?+WLokIPH1V*Yh~uNr`P5vTKWQ23vaQkU-*iv`4NZeJ<4wey-S=#&rvru8n0Kwv z$W>aa*6R7}7(aA6Ix;-hmUv1or`*c}k{tEcjE2j(P1!iPS)c>=HCeYbipfj@1Lncg zcFuu1Q^&y)1;l?-k+dI{F|dduF^EVZ{#{?t`~DsM+gh4LyghadxGRMfr$X64KY1(+ zuc^ImOP87N7nTbJzLGMwV8A|4FCo;ClIGixz2*x0-JX?-Vda{Q<22?0ia)GR3}->= zgyD+}uuJ^SOD+(}WoY%AwJ?w#W_N205&y7N;5?(vDknEt;8C8PUt-9`w6Fon5w(wS z(uoY1Q#2RGSP#0%na|a4&)Ud*bHdPEECr5EfrF^8QFo2U#tWU=azU?KZ$(Clk;M<> zxV&Y(_qjUtC}(6&(y^kfQ1N_Sg}E0CFo&^fviw}Ty=1l?ivY-^`=Hve90$~dzEh#d z0)G%gSF#O`^dGUD2ooKcP}6(bQt z#Lh88JOYUC=?2RbYm*-b$6e>byrx%eP|WESW{;o2^!Mbx2uWMko<39!a+s1<1y~7)epsQqd*2+ zoQFGL!^G-wF(r6rynOJ2z9uk^_sU_Jc3W4Z$>%x@>e72}spju7(BC%-cSOKr5H@e3 zFKqWJL`5P9vre+O#sgr1;EXQKxu2^WIW1>T>s^etjz>D}{$sK3N!??V_R*Z{Lj1;G zgFD|pi^DlSgBuF6LP$9xaHUEjy5rJhsFJm6-Bt-OyicI|OG@dg=TtOa<(zXK0vq;I z3;fY(9A$4oGHinK7qfmEOgpxGz99TBu^6^!MR`zaQRqL63cuA7trK8Rd_@GYK8 zL4V5#Tp0Dtl1tZKid%vg0k(CA)U!e!O;$Ys`c4n-BA)`6W1dCI2qWhp6QObs@7AI|CmqfBi0GfwzHG< zz@$RVBv!*y+vory3hAI{Ub*|I-AZm7eu=v8_xV4+f>ZrRiPwPxPzOn}=?!WY+6e0! zCElGq^FH+FpEXk#V{T;6fm_hu^&K{w|tgg!j7h!*D`$(T7+G1w< z0#aSc|KS|^@BJFeD@2NBpWN_l~QdFuEw>RDDp^8T$ zilkJ4Ld{eTnXLD1l?Dc0V$KHgucDv$-|F5bN)O}zZ{K)J`e2&q!k8GWu+gh4D-5nW zByTT|+cnAa_Tft;Cr*yRsyG#&FC93lzm)6u$NW#`3iO8B%+h$K0LSk_!+q}W7HwVE zH27q$m{kup3JA9NLB9}cOz3;-Rg+B@-Np!v6~N>dUmO9vjB+5gFCv{T^5IGsDt>WJ zk|b=yMKG1ouU_0gRwTunmV*uY_eK&zEZ7gd=v{%R^y8$qFn@ja(wS*hjzBj^9{KJ_k5)w|U(0B0UF`oLlDaW;CF`66$x?k?xvv z+kI0YD%$wkFo3}+!rY!YPR?Y;Ks z4V2DpC%aEv^23iI-Xv-dX5Y^%3Trc(BE`a&DzrKJ+)igZj65n=BxzOeGGSlvOo_fz zP;5iIl%9cL00FYV;6%44VGwEV0Yt8HE*;^2w<7BbFVRa?@V8D7X^4RrgD9-4gxj+T zAj?eC;&Xq}&Kxmg^GQguQk#t_dy$p>)fzfq0{mApF2cON?Bc!1qvD)Ee@6>gtby$m*cT${Z$Jmh!l?JAt@cx`2(dX)m9V=)Qc$V z^x$wX=r1kO{0>#hWLq=$L}J;}M1@F6ecO3vyIRsi0pI*v$7V7mx7B8ihZQ+8RChoC z8mnVU@@_QpX67pPn^Ss(nB75AtRtQy!$0P2=6i}Zy5C$AdVkQQq?l(}Jn(31df7VL z<+MxUAy#^;BGpobDyp^X9QBWO=HA7lF+>WV5yM<(>;EbUqHi{(GcFBvXLvlpOHEyg zt}hO-O1)G5{$QSiO=iU$=^9bGf(8ramvW40c308+JRa5u>C zZ}|V4xMG-Z(w?>GLP{^Q)*BRSj!1ljyhq3_fL1Sls;dG8Q&Bf6y?h2vh3zBEep$jo1#8Bx-waHVmA@Flml{_!s)4=8n; z^EWNU;G(yPn>(SIvqchoqL|cLY+vWJTr1u;#MQlyYFu zp9Sb0cu1F|u!-vm%sxn$9j%bIG+#_Xx3y{1e&-boNs$bB12xD<<9N8<)i)w)hZ(_Rebzs&osmG(!a!->VhCy~s z7KqQkD?YMxSSst}3u2ws74Z8g!X5}A9n~cJwigcT5}XoB^0Ci@Zyi3sz8K7rV+Yj_ zi1NXloBrrv$fZ(%7GJ-5q<_0A5dtK(sLGMf8>*dMrEsy2f8(6rZxOJHT8q=DZ?BsIbUMyE>60H8c@N0TzFdF_jBb{(dbIW zVANdc1gm1I5pZ(z|an4q|#-XKq?=VjXgb!c6Oy$HaueIq&JoY!C>7qn_vMr zy1fR0#6P=6--SehuQ5)jufHM+~Z$9XrgGcNRQmPH!@9kJ*na^q*|Q8bUcTt#G}z-@FOt z%sE*OLyDJhxpoBS^GUaR4tZJe2QvWm5Lj$%%)tX=&n*7c? z*@ylIw={te-wH)pywZQRZj~It+erT%67^TCvVGa(=<*R$mzi>jnh8pt7u8%Z4LV5y zESt$^UBB%zaJPs*2JK+vVlG=(TXkKZ+azK_pkEC5Mm6L2_l0!|@cGCf#I#QF+!4&6{Y^{PiQhgh>NfDLeTtY02b@>WGww4d&`>i85!J2gbWl)ox=D zyGS;CfPL(_S+AG?*t^U|M|}k=!?&!i{eC}MT&L!5;?8U{Q8cB38US;-ODGkceS)i` z0=mjb!OX5#rN(m81q%+I`Vf}bp1#lc8@5&^9rmlM;{bc47t4CAn~~_}#gsrA9f@|J z4a0mTf5^VRqBstp5om)wSlKJi$Tro6%J*00o_o6;-eq?i9fg>tIbX>|bY;6Em!ziR zPq(%8F6Kj`LVO3ogkKIKWjx!a&~@rPX((Mj-#=AtVHXrPF_gfx`}T{jhE_oUiPcm5 zA6-+73@>0*6v%kI7kWe<6V7{}2oaAXvtSM!-Y|d7fJHcS%yN1~lyd6+l37af?xe~D zuR&eqT`Z)WYp^2DK|{TtKkSGp80;N%7AA0v9%Who>Z-mUY0#QqnE|F+3CRVkm<75H zR7)Lte;U+>Q*($JufSA-Paws97+$&pPpkYyi{h76fHMFRp-n==7BKS>#<$oY;GG?% zR0pi<-PBPltVcSN_s#P&*d+dF^cjDd0hl=Z4!&hQZ^xL3fy}k{J+T@{5*Zh|)zxoP z&P?Y77aEd5_n(G${L}G-XrAQ-?{kUn6;Y>v)Fi%6-gCuh5T{+}?O6r1Ih?IGowgB} zx^!ZYZi7v#tzbZKn|%0LrcShQ5;RJ$3@eeO5RJ7cd%nz*V!jrmg)h{*P6yE4=7)r! zh{dV$5?k&6zKp_%cYIE72^0SG-#5yZp38sKsIFXi1WC=}QVlE*8x6f9Ev^ee$Xp7h zj%h4Ca!j+3##I;D%&&(D0Ni==rdnsjfA<97C)mLU`W4u0VQDRLRdJHSzo|kFK|*5* z#g3ac)+p7m)S$7xl3;d?q!fqsNp(qEC%&qR@9@u5+y`zFU->~N`KxqCv(H>N^#yb{ z|3S<7UjjW7km=GAhQy+~_2HeL>#u;pvL6zN`L8;dSKgmGW;*F+(2C!&W2UESg&c=P zziuvN8_x!x?$mXD4IS51oH$)T5smOg*Df|kTO$zHgGk0+zw}6%F)<1#Gy_&KJ zG316vyh^xG2{<`vy~0Q|wQS6A_EIhxPS;J=3t9Gsu zBF{^tTV;#v82Q9fj0*hzQFsrzy3bWcrsCu-R?iRmczO$foIB>-fBYj70OJi)YkgDY zO7(=7t0L1r`SLb6KFKWkpu&5u0K|(Pp$8)i_YG90fRpY?#^RR5492AWVR4TFpdp3% zo%=Y_e^>Kgd^DP3&j#Pnx6Hn!$Y;YKYoNCfAt|>S5p$D^%(d>-(RZ3p^de^air90b zK9f-)yt)_H0?}0~y&NuO!Z5@Ksv!~X+rw>@={Ka-d;VEFuLNoVe_L)ct`CBJl5O`G1{F~pa4d~AG`=aoURpZ(&j(7 zC8c_lO&fJer`&QaYpY-ZC|C>_B84iotE>PRYA6hGy@|l(abSq>xm0^#lK4Z>*~k+I zMbU`JBDyVO7##AXP1;RVY>UTN5 z^RPaD;zc$!RnpCBv1^7s+%?RIFb6s)8Cc%INS$)DB50FyLpzJ^HBNM>()> z3cSRl4ebxng%CfhaP_r|Ei)D{a7Fgm^6XvjloNn^JUtfAmdL)$#A^H=-Odpq~czW3#WhX zmT+C1Olg5AnK9}|%N*0`ZjxE>@m869qtynFCs_?2K1SD<8&e-myr%Loe{fPlDwfBr z7RL$an7;@k!qcT#FVO%6$O02?6nHR{3KXmeE;yqqmIxX-!>Vc98UM0)tW2h{8L_rW>eS^)~DnG(+jRj9;d8Gz%$DzWW||;7;?S;hLo(`&Tgn@TcaR9bQB)- zf-4;~ooZGf^>Y!VKr!USA(Ej@fKmC&g(Mff^Ihxb3LeBzH1u@sH*YVdH3*AQJD=0J zgLVfI1v8v-l|@IZITiSZf+0Ci#u&5G&-qMZi!Eg9?h%o$?v=BvjuSs~DOM9MsH+8yH?_3QDLxm=X!#0@is8!J)IAq< zX@i8Fv+_uD@TT zbo`p%JjPEECms@~ZyK^PgrDZt#clQj4k3p7v}U2oW~6ieIs$@WvUM_*3w`cXaOc6M zL=dFOS(bNqhej6;w;@CdTufeeo(C=cu|j#%gO+u~!C+SpEO}=IwO1Q zDii4p@a&g;n@j>mT}bp*MW!GKYOAz#qsHQC*x{vvo&q(J7Us$`!=}^h-reNkZS>BP zVM6x+<@9u|(YwaM*rZUQfognwy9Yl{`a^0TJ8e(5a_piCi39yRWm(M~Z7GoyUe`ZR ze$z+V5C2Qw&P8bz2DGfSka(aqo&IM6KHxXg8BQGHl^Iy>)s}s}kbrd?N}4f4Y%4>6 zm_4>8_a=bztQvp#lZ_c&H}77*4MCdyZs1J7q3XEdH3ID03r3=kd4I@X;4b;Gua4?RyD5z~C+mzS49G&7 zKVrk$Ji6<%R5Oasyzr~jB-Q-w-cfh5ue55(S>*S}jQ_!42vg}Hb8ta|^37ZsZTo_r z$tFy??s7qyE9T;&*Eo|pr-Ppo8?i6?JTzWhViJob=5X?rkmbSx^fw+m;>t8tWfd3d z1lJ0g_@mt$DPwwP@&tBWjb&RcdPcyu}uyYRRDD*GEe2md~rRwf9zRJz8oqYS~*HfeKx zW?SzK!dIS$p&&L0>!2q4`N?u+HuQE3XaB)Cjce*|ayj5<>@bFk)e*mvsP(3;hY8HJ zpMDFEi~npj&y!gEg$he&=BrVaB2AC6x8*mNSMo}4@fR zZju+MXM;T<(>#G>`)PsvUnN(FMZ9nAaYO*Daod{>&<%;;yI$$xG?r_dyVIynML(#Q z6}^9mD#xjNMhr>=)P-x}5F3*zN>;`mF^zh%kiB2v@d*z|&PZ)yueT*eRm*zO37tA9i!5tT zYN=9ob@eiFo`VP>Tw>!(l{)-ILz+3OC5>u0NjVdUh0~1D<|Anu0Pgw-1IXZKzLU(!%+!)#n)(q-JL&x%iP&nP6~#YUd;}4!RRLLoM$7}G)ri!6_JlL7&v(z0WA$Ig-X5!LsV!|9!b7+ zG_TQWTF|CmT&9=c`D=MjPZKa>RbMW2Dubb-SjfWDgh2Lm=BTp7u+Lj|Np`y;Wi@E< zFS2=@#t0JpCL%t@>P`qJeTeD+q&>9hRq&*bw>R(MHK?U8ss#Q3V9+TWO)p_YEQXkT*t`FG1kB$;>UJDFx)6}AA$-Y{A23F0x z-1261W1N@Ab2~_?{1G(!BE)DjfmxLl)M1L@;=A&{@Ru74rOHDwZ)cuf5@R4WD)UI8 zS=f10EbrFW<|ohC90m6y#S1DEYZL;1CYdzrs*JYk%Gi}g{>I5RW}_2n+j!A9tyvgi zPlLw2g~G;PwN)1+rwX?;x(^|=_fjU*Q6q+5uQEE4Yh6;mO(%}}Ck=06ihJNW*;*8q z*=tHyu8=&sZz+dGM->ViKT4I%lukJ(c=g7jAGVZQ8weQ}9$nk7O9ntA?WFL-hjVKXd8Eu>+u&S8jDSkRehujCKRk;waA z1Ir$kc?K8Ttu+NcrPh99Fpsrvzp<6!dmQ2g#zJL)w7GPo7}t1+Eyp_pYLv?^;FUIR z8%JI&Wa^V~pEYW_;%Re_RxB?aB<=2+?y@S=G$XK*sE~7jsr#jw{{cc3k;+L1ol9#e zw`@TzsSe^~<+<*iJZ!2#5atyN#bE#LTso4Rz=94{994n>$LXT;4xL&5^!$sK2dUwa zRT{m+(&`%iv`5!#p3m3_@pxLun#l>DS;5C`*qh#nQ=%P*k^a1lkvHtlJhd)eDx|TJ zmsj?C(cl2amm|$dWo*asl#5V35>PL7CT|IkD_oX`LacJ(^%sbILl7LDmt$Dq)!_No z13hFE)neCTq8;{Lz#l%;8QK^27G%e7*vx?1xkn0g!ho$) zgFi(i?v|>(19WmTDx~=sPJ;P2AHJLA=#K4jCplq8$e0^5vY2hdxZPAMN5b5JRpCg{ zeCtduGA~=}c#-t_G2enPb!Y9ys^WAeKBV$mQKamH(i-2Ho)oDbPm6jNiw=hd#w1G| z&7LU8LI^CE@SDthn>M1|gGgA^^8A&?6efzdm~?rF4#(^q`jEmF{381gd*=Kd(To8i zaEG95RyxBi+s}P;*6j-ae?HV8K5raR}Sl3q~=$gaCaQ?)8RKDv-^kfObz;r z)}O+{NXBk)+LLZGR-eu4&+QXMJM9~v`!T9(nb;~R#oktTm5N}xb5$ny^*2}+;{d<* z^u9W)kJlc3LF{fBhIA2Y_7ZuZP%h5$@u<|t`l_m#6;vK9)II`E{LP|`rJvY$OxbNp zWxY99`i)Dz%~wEw*1^UWu2ly7l(9=3r_|B?1fjq@dlBs9S-3&HwX=4=qcAVVx+(vC z_(*xk{`1)lS!zDlj2ST3p&hJ<`N)>I%Tg z1~)NtV(Wk+)Ra3CRb?TD1uWllioreOCWABc`pI5q4>Fi7(cw35*$%Zx@{0e;TiFl~ z2*%w|j#vVY=~vDmn`1|eMkLsX(9~sRs+`ToRD*azG|1lQC9%59j!9w`ASM4YO-H6; zmhiT_&2}#cwi|#uTyE1K|DuhQv)&k4J4fmbg%?u^*q=9Zqr1K4Xgzy9EN3PfQ2cmb z5MuMJ$E(e)m0|kQ^%NY#9GLSw@);CveA?fd9{Q>DRoA4Nu3=u7Ao2{6y4uuD%q5WdPPpF!e~L2KL1YUYG6%e-6Xd zv%kdsy~?owBm0|x>L~cSB|C%x7an*Ckp6LO6n?=?G)`!23|YPbX@Qb8`x?Ic$s{RdUn`?~1`r3(3NZ)0#8E7~;j(f)P4^Jebr?SDF@( zzA;d58P>%m;&*BfWd$U@D`E%KFJMN!WfW@`igj*?8m*LE8a6=hV{gZjm_c-tN(xS6ORD!y8F_(c3#f#48!)az(})t z3L6b@>EZV<<4vMj&e<5z8BB&MX3vSubSmtPcft~iW9P?v>KYhN9uvK{6k7H~(MFvI zd+UXz60=6qErdI_bP86ONs&Cg7ybrX8{(XyXJ@M$#weK1h za)LSNN@5cGs7QYiC3cX2!#fd(`|&W~%?xM3DX9B{L*~ocEl}6luZDTLx%4%$qB8uG z{pX!QT|B4{h*Z51`A#tUEy#>bg^xL(;r_WQ8eZuTy}hS9Z$2RQF-qMg{cj`L9CtLS zg(zfzm+WtyFY6JdIvK2Q<>ae9%X-kU@bg;M%-A)kW4ShMLtcIo3_w;9xCM9$x5!lO zDDhj(24vlX^Z>_KchD&;ft6HxI8W*s&*d29IU!gFYN*H8${0J7e<6+>EY{NT5~p35rf~X{dP~gLAo$e`h3S4m+`y94b{GpJeJG&BZ15 zgtV}v_oD#moWGBDeUX0cVlb8HswRY48qNsG-eLOS=LKE^%(H*1FFn%758J>Tu8iw~ zta9%tJ75XqQ0L52HH*HS9a$CAO&o)N9QQhD=LPiT0q`_vfdLl>8JVcP=@-i{ctF~3 zd%OL2Qt&I0`ob~*v(yq%D{c#zBqxoyOjiyCu#CxkI4^#T>m9392oh&U3jgzJ>7zKO zQFM%cqB03ay{*LKp?QF03jhGnG~!s4aT?UQpPN3|Y}D$$?h32uomFlc5Wvgv6UK^` z&#>GjCa!nf03U$h#EJ8hG22Kv@gIb{m_BZrb<{92kus~PsgThE_kMyaLyw|Pf%TQB z>*gX#_$*DXF0C2`de_8{SW8gVfxRVD1kL0|3c>_DVMwu|e07SIXEjYxq2uWp&e>5G z5~Qfc?&*kFv`CP&#BQJe|Ns6iTd(^yJ?HA7*f)2I@;*6sT?S9)(Y8}oFlaee)i0qC zeAEs-FB}cMv6{&MCtNYya4Z&GAc41Sv(b-pPU_brC7G$ycJK@SXLQpp7orkjW{5)n zhd|YX3;PY5SCuH8o4y9Mu5^guh!+Bkr9D%W6sdj#htOc#{@WhMR}H!-VId7-b|d^hn$nXr0P7 z(+rFNdjZvKP0!v5WV8(!h6zlwX5;;kp_XceuWjv4^@;9*IXg9CsNXIAE^C~vxY`7v zMaX1l+M8+8>acW5($#~+x4x7*V5^VF1f=SQ32)6kS#ewuc~8NNE5r}Rs!AO$8>Ac6 zGTZb1C^rrN3$P#=riAbjgBxU_yR3%8ToOOI1vo1ZJ>e6-j4wB)T^7)%xvEllh>`5> zWrD0#^zod*Q#Y3Uoq&O!EDng!?HZ6W&K*88+2_w}rt8#dUGwSVE$A6cqv?S>DLv}j z=k5?eCk5O`m;FvXlXa>}251KGA2yLkn!A2_3kX+P7E`wcqY+ev?wFF?G|?5dGdHX< z8-KTWG4#AqG82*({Mt^`)a(S9s^}Gmtf*s~xjwXZIU zNEHm>8~1L5!TwHaKY)ll6J$l`i7w2XYuO1DJ7x64#G4_3;)ENO1-p(W!2E~~b5M== zND}=dbM!_GSOL0vT=Q+hujERACx@TS8#YFG|p^`v!C%47LnV$(QwY?O$Ss5zE zT;C1vC#(djuxfrN!Aw!VxjY$TQNf{xu1`A82>{UTES@k3PGm*8A_*YU!xbgHy>LUE zI9Ijjnq=GX+$1mX>zS^r{b>X)`g`w$nj_|Q_?oIKKv%l>oEAL3;zwbg{#B7h&WgT( zp23V@32cw!XqGa`6@N}XP^}dMD{ASFYej+0MlbJR0{@LNGV(+T_1?rh7@?$)eEU$9 zkq5JY>jsw*Bmj>-4&wYB4yN{wI3#)gV4U^vDBq@T* zmDb6TAt;+Nq(kO|S=w3DD6H<~pV>O~#tq3%UzRwL&Gzh3=68nKa4-X-c-(9F z+rQ@6NG>ZujTh;Xr1U!HH2FapNnA|cC{hcPZKaKb_vHqV$Qy?yXZ#Yl852hD{K|EZ zG`J`~B;(`V00-B_rIL~EL1j9h{?34p`1jgT4k7~gaN_J_0tJtISh4Cs3w>^C`}^)t_T zo*VsB+DL?vbm8$6h;@+S0E#St!M0;=Qq0`Mdorb!b4wvj#)2VEbv|iJiS5p-j|fW5 z56v5xJ+4BjKSoU_Ua4U{92(IP*E%~UM&wEpA$FN<_n6$lp10;S3IwnjFUx}33DNvl zcafX-OZ02V*=&m!J6aCt@+(FR__^o*&Kwe`jTA`+-@&>7VnCh0&e{e5isG|j|U8& z1s${fqhCT_J9x0O&%GS2JTjE|4Y6S|bvNF1cFw6e8p<~45~M>_D#44NAyeMX(^0Jm{He$>_Jy zMr>%JtZ}GQ=ox+c?O|kuT#q_`ceimJ)1#&3j>tKsR#YX0?ZmN{>Y1~zxpRt1mT`HK zZ5D?#s0@DlE1Q#k(+|d>McUY;Vm;Q+q`wUOIzci`P|}_*QZ;MNrv@fbP%K}!s8_Zae2gZW+}J`d#&H!Y%(B;vIjnfISRT+dBsgC6hWeejx}Kj2=xsXv zW>#?96Ag3t3*2e1j;ag76E2&?+pfb#T;@m;hd0@H8cWG}X) z__qaGjM@cr5Fs=}!CoLq`OKVcypXeWuuWH)AC9lO<-QL2I)^tX6bQ+!JJ0`VxALZS zq6})rN;3Qon+PZ{{?%H-DsT4!*+{6~hYxb{xKoybYe(V*J+V+H1NV2uR6BJ%W2623 zm5vmXN)Z-A)hY1}Hho=DlIf$YIBtco*DxN7{idLk#c1xoM2oH|=P{*|t~Fo4m7vk2 zKji&(7g(aXJqr~G^7*4X!I=fi0fK20K~o2 zP7g(*60$Ho72*2>9pu*MKGOL-niZKd@ykfHyu#$XNmId{jUhK*R|VU(4f||$q+|H` zFCqD#F%l(&8qca{AKb9BRhPC4!*Nf@VKD?30Kig@RZztloRC0@U|+V%Nln^o^V6$A zi(9jL$Ec$%xn1deSAIK0C_>cG=YZ4aO$om}0j2awWnQ}QHJs}5@=_I%crW$SV+0k+ zEvt0MEy8aMiG{y?W(e*Dy!G7WZgndtE1y#SJV)sMn`1br{dFRbKS9G}H$jo4P&$xfjxN@QKCS`Hqu3B3&m> z2#$nW!NmZ>D|>QuL+AQUSG?yJP!gyMk8d@@!Zho+(4BxzFZH_28km@z*Q+3MGtcqn z+3+fecRp{ykov;|ro7!kQzBUvp0tKo@E#fcAFhH>1zP&XuVfSJdjY!U2qJ<}>W&Wi zN(a^Iz(U?J2!z@SI?shpXh?Qi$I;+FOJLI6F^gxaXVarG z_Spwtit%z}Ed1Ws@CD5QlaPnTZf4kdaDIv1Y%fc;kYSx2WKFjO!G0`Z$2X0X1AduY~jY3YNnxncIakk0oY}+$) z*IwxW#u6H&c;kubu?hX3z^?r5?Nt7?+g4h@A6^QhmzZ~J-hx4>6h`#wYK@W|wiogV zFm=fvQKM?njojT;Wb1-CgIaaq-@($7=b<_Y0w<$7e?x)eBo$dd zm>!a0lN%$Ov!oDD70$v$q2L%``v;G?i>tTTS*j&?Z@I&IMSic=4q$=G?jiH-5%GkF zF%Un>67;EQ?!}%uNmz&P&y)n|*GiMcL4v2huD@VdCaZ>Xuvwg*r(^@_q{q37~&RohDh5`cR zwA>OH_M*3ZD#(urkzyl|;VKJw=ZGSCRU0y2aBpik|0b0w;V~#=Zfb=(B{v2b=E&djHF&Y+LiV zX#7QaFS67;jI<DF^F!d?Di$P-7=Bq+Qyx&D|Q`|4ylJ&nHOVcf*rM-uKf+v($a4 zcC{O$Q)|xu%0n;qiR3ul9fD~FODu*tfDre&amXh&j63~PB;tJ>ELRYpr*jDK8(e|K zq+>taCAC)RrOB8BfOinIgpJdV&ZY3EEE~yo%%q}pHir$p8|Q1ARC>$f!$mv{SQl$x zw*R~Q3ncM~eU;D^U3E|{DPe=M6n@2P>c^Plg4X{i{co%VJu#YlwE$%a6xvxV8x^Ys zL?`q9%gRVu#p1}!Q6mX{-?T0({G&PC#)ZNjdp)a99n6|P-?05ehm|*wbfSWnTs&D= zRgA0$;CnW-DzEAjx#}Qkr(ARa{8h?^-8%teK5YJ2&0ce}JyT5(!_ca{eCUW3Y0mXF z-$|BRG}{NJg4xdBy8Y|90-%v>v|dEuC(l>EZ3g{)8Z1JDRA$89BO}y4H#Z6AEm&P* z=31=U8g&Fk2&mVvt_ip`{(VO54*I`Xab^dQ@s+NZ$hCjO2THvD;+bz8n2_bL#%-vZQ*>?MD7j^z zkU0hgVf|JZ`b{S-6rD)~+SRUY07^VBJVZVLT0I zGwZ6}Rd|kEnTbOWDHJPF&3=vZSIlW)CM$t8Kx7_IR4lkQIa$DmeL}QF;eu}!rHU4} zg$k!ijjn^Y#}_~V036`||NsC0|NsA&K8KOmz<7NL-XKx64EU`YK3C)_baQF@scIkD zUPq@VAz%=P_!N8{xyZ7z9JdC!?^_TcPVB%ut^fh-5C8xF|NsC0@&m~T;9`phZ~vF! z0(zqANeu7F#wdz$7~KKo%_$&t?6IyN05&^F>5Y$n%#W^6Uw5dq0R4$3KWE}Ip8Dx1 z{7*2#lI+|qlzZ;81zi8l^{@ZEAo`xQpQz$Uyg5h?Lz>vn3Ly`|j-N`Or^`0=oC_K* zq?dh0%>b9(xLQ{eIyfn} zRFE$U|F%Qn#Ai88q!2HEOs)T=*(SVDzcjg9}_ z_FC4o9-n;m3~G9}QvRDJF%exb+aS^7#C~vtPGJ=9djq?<_A1ff|8k~-GjAIjTPQ7} z)rqmeHAT=f20ArTh%<+cr8uY1=j9aK0cQG);a~s{jGzDi|NsBaJ^%m!096(fZNj<$ zUgzx>`hShiH|y~Rk0qo2Sdn+fS$Q~4k19X^)LN=bQy2gL!|cS+nzt2fOQs~bKVzfN ziyS=M}uee~4cPlRw*gx`_TK26xKzbaMBwzaF=c zfe?{M`U@a4ig4h~cXi0(hteX0JO0nmSA~|OdE(neCtCgBKLtvyTGc95ZvE;&;PU%- zE`lTO^pYXV1Z3z`I9C6t958kCwJvDljMQDi+xde^*mCzZE+{3tgnppIFI+1k&XQm; zrOc$4i>W_hdI7Q;3punBt-?V>nm_~Lrt(;-FD}#n3gNazdXbBEbQ%QI- zO{BgJY7@m1r4hB_V19FY_H1JwiG6s|lcW+bR&!3+c!q%maav}#|K8a9kKP+xGx#7c zFy8U{VAQ89dgN)m@)&JhRiLO|%RlK_yv?HVIj@FizmS&c5jHQn6bS?*|Cec@7g7NA>^l!z#@(BONB3BgEu`|!? ztmt8-3@fI1L4Ny34@6{KN_{`s*%y#tJ?drQp(t#+$tU~aP?1bZc~`{UEG|##qyO7+ z1nN5f-Jxoa)@748B*f|%{dBPcn$0y59`g!?VHU7uJNXOZ-w-)I5)-{aAxl5eY*Qsh zp_&Is+UzjeWnA`doZMi?7x0KWvd86|Bu%56bP)M$Tzxjxnf;W58C7GJ_(HT47_I8k zc9N;<>`&W;8$+cRw&neAUqJ7pC=OT_-|HXU7f~1PeYE|R=9U*|hwe{R7brQ18I41+ zo^PTxw5Dq!@x#z3nFCpF#meJNQfx`!zDL#+N*_{A$L0SEpHniw`|eg+y##E+&gm>Q zUvDaTIUR}rX?PQ)_t34r&>p7+-|94>x^@F$!SUVwPrs&mY!r%oVXvf`0n;!818NUL zC>MTiun&o?@7rTOCcRaJmoj8G`{^|V?J_Bq<_mg<4XCnjI7H$KL@c~&4@)Q zrQWdq1S6xW9YwwRZqP5GNJ6k$SFElKE12a1s6I64_`1vyOs`8=F*wWP@2DyQfm(Z4 zEi--YBoSVC%OBQ7zzg|7d+RZ z%K+`X9PW|MEg`a~?g$ee<9&-KPg z9kkD^lbp?Jc0yUXou$;%kk2ZPgj#@1CBoM$gJUl5e|ZBH7p`y%j2(Jz)2u&K<0}Q} ztJvsXENB}@KZR?0=qC$@<(zkfy*7#|rq7)s)pb>mz}&pAGvd&_&WdhdPswYfO(a09_zlJyNmJ{tM@WtwSXA&i5?u)(>sFDyt@Y-Ox4sjjbsCdbG&;E3vgetk!n|NMJPEzNTal z4;(ak&IMl?h&)@9q0t-^fay1`(uX#m0vyUmG#tGgz3{rydU-qDzvIPTCKR4sZT!O1 z`q&xPB}Aq=g4imk*B%=~23W4(_2ZBcEekNDqM9k)L_cwY?4@4?-%nUACX*7fHJa*q zM!BvQGqqIa-lqUGWB>}nx;tLO??73#mVA4kg8ga6gJ{d~=o@U*R+?Gb8aSqxwo6q< zn)xAF?q9o!6Gm0S985xC%Ptqs7p>)H#ua7bI--MT@PdoP9;7{pzjd5i)tVP;1W0Sg z-hCQGqH3)_qKxAf(wHm|5XJI#p~~L~S`$3gMAN-2wFt}n?Yj=$Q)!eHVVZww3I(z> zX;pn3CT{n-**nV{X<;H?_a&M8e;sDRqQz<BP z4!eptGW`UcPw$%K@`cNbYV5_O7C@PZ#W;HuAS-Q6!gK|={?w0*Ej|YV3#vQxCh1KU zjBNdoqNO!J*;A+h#_IQvB-vLwpXWWEg6IJ_i?M}pCQkt3u&iWWQZYc#X1yxJ$;bip z_>|v*+r)+(MN5#_$#wE??Z<|7i`;&?s>rRe%;}jNNHn#jqT@{~B;%9%3T7O`sj=$` zTTOFl-+unq9QldYfKIzmgGT7yytn^6l3R1|H@*w&1K_{^&;8DA9psqdguraQ?PHN4 zKQ@_Cvz9-Bm6e=CXtf`u7?W7E(ctr-;1J$t^L(<39lY;S_#Oz=x4)aJn%ZE!bp2?& z7L67OmvS#{IKWF= z`G0}nI`yr8?lN@3c+Hj*In-}^5F4(jYgjn%H0XI2s7%kei?fpryPC*hRD_?h91>m; z5XIMdk8MO7x`s(a2bu8QrN(a^yta<)8XF@%THc*_>IdBRCF!Ie&4^C&4<~0<%w5+f z0b&T6Q=@>YKwQzs#5;3KgFR2&i~;O-k12MKaFNyaVKR>g$5scbaTrlUI#la{k@p7iwK-9jnk;90`an-^uBz&5BNc^}5h+X@PHl1`>yl5e_Nx-zxj^LCRnTr~bE7v)Kv5me-j|L|_xssuIf z*p7HsGYAe0_tXQXk3mn)N*OteR2b&j;OFeFp;LeXk1|~|z9-e3O6TA+bRlJoDy^p^ z#B{2W8W{+z(z7ARU4}n}%^*7J%+_P$>fc;F0U!UVqdhk8o8-GY1;h$ze&sB-?OO<9~yK}H$oIq+R*;91K=Au`00W$?m^!^h$dIsuHv2U zI_b=qcK0w%+l5)K&}B)nl0nU>wl+zVImw3r{yN+N)b zdLC#YoGRi}7RJVCc_An{37NCP+b#X{i#C2aa0Axq;n1G>s=p%s8{DmbL+3ZYspQE+ z!&WR|pvxdmcraYKX;h%gY9j`9%POJlWYWpW+(hhMv|g5V$2DcEbi*ugo@Mlo4i@Hy zCt%02*_2GcYDD?LDB|v7)$7PTeZ~g+IE8OjRaI4OOnqvqhACckhgjR_Cns*)On} zK*pwWHtFSqxgB#fj{*<3;1O&*3~#mEYHaX$1KspW+?3kY;8pN5fa6)jV`J)vN~UW6 zn*9xdI&9pt@@ynlU|1BfYK%^t`1d?P-c_fkV$sPY^}qXRZ}!PV&k6ntI3FH`@02EL zMZiG^crz(+WFx?M`bRGI!o@nptlhk{mA1dW)oA;kuk6`qDSim>&k>#KIj0Bg}87_fib& zU2ZhRnB(M|UxG62rIyW)$ug8Jkq`j?65ASyuno%n8+-DE9)VI5`L!!H&=iHqFWpJD6VB4BMdGfBz+Gu{@Kj71CMFs-qYokD_{U>+7jNRX6* zIuOeUPC|4#pefElN%v;Y;m8Us8o&coP*#nh+cyDg;p#+ zS0Mfa^oRHkEak(2SQners~PAk2mZD6 z^+ToCmu$O^cdsi6Qx+0W!kYgZODkBzOxX_1?7Rq?};1`M*VM`TDR*Xp{}D z0xvrICY_0Udg9R(R2Y&~T4B{g#xUT${L|!{{xQn-Da#KS83h;}1jK*nq{i_khKq~8 zLt2YFZeoAc9ZMy9(K>GveZ>XPESD6aR^r%WN>x8Q_qUkUOeF277MF%B7Rm1V3@w4k zU$XMU8m%hACYb{^C8ktgy@sPdq_o-=rpVUGUfJ!6DD0~~C{~6YHr}6Vvw>r{B=2#8 z<*({$VTQDbidG{PVhB1NlvP^=1dDEahSO1OjvM_$EK405wEft)j zuxzJIMwD`ckAm#b-B(G=y8}W|xi)qHz}RMbC1a}p)}?Vn zvX=TlGtA`<)({_cIul)V)y{=fnh8!p9}FftO5^;`=6`I*_sx_3Z`bn-Ud%InEFM}7 zPcMNX9142wuE#;ZR6BFo+6&;VbH={ioqUnL>*mI>CyICR^;MW5jDQ7lG%7MpM_RKv zB}pNBlPFwto>U?!68-Q+^vzItq?gO%y<96gcod57DkZ0fFx+VT@gPo{6uwB)I8XHlO}tW?-I^j31-1{qK|9e z9ROFDM_2g%WvBL&M5aJ*=YfE}bg1jo3wajyIxaa^{^7{loPhKq!R?qss2mCKMV}`j z)t3UFG0T6r=aXav@p0So>whc}d1p-jH#Ndk^Qu~9rQKC4I6vcS1T)H041aHR;7(ho zQF!z)3o|B@`C#(rKvLA`t6x(#OU27=;kaf%C6aK^Ug^HbRAJllt#&4+wxp=Lr<0)X zkUfRFxn0j2uza!)2Mwb$co}k^j~!LOUJWC-SO8Mp!*>n zcYz%w+Qx7}_Lu+doah&QHWli*)2b20g`!k7B-L|V@Cw#~r-a{d``eo15o&Jlkp;NE zw*sW{U65NcK}Pk~#A$;_6yr0cs)uRiZ#TZ;)5_QpF)~P&s^;YC!PaM1pY0TI-jLaa z5+?g94s`_Dlncut3KpE+1bW{barfoxwg&_ald7%47M`|w8J_Pl$GwE`P~4-|<`6P? za!WjWQy-;hBRG=#?$0y<#dbAGmsC)IV%WkeVv({@Q84;VgSk~T%p5p$ish51_L?k; zv97nI63l~pZ!|7a_&yGG;nU2>Jx)1SXl@DB!E0>R0Pon#H3i<*fysDsE0YTwEAxV* zJnz;j#nbhq@0$E4$qED`N7K2NS)OK3>cCKc{U4npFdQ}lpm=NvuO8bMqh=sOosOoR z`-!X^0xd5rLs-#H1-6)g@EiFk|Fjq8_Q**&LM)s}@)FB_iII%DSPdEcpcPHMTkOj} zU`#=ZxD%r9{0m)zK_5na^K}WvC*t|X4FA%RjuHc3aFLGFWCB_9%>ZG8s2&6c zW8EM~Pi^ifo`X>vD^M-2uATf3X{+BqRGQh^tAH+Q!xY%4>|yt-F=yp(lDNMW$X@1D zHtcn8){_5#?yW|#!S@*WMRZ_5`1md_33k9c`O?O?;_=TJUjvmS@g!LT=~xmPg|B7C z2l`m?*Oa4)0^1xQmz(q*=RRin>LOA=Xu*I;T5LeWN|HwW6{>|sK z7T~{ucBj1H_#=Yw(Is5cW@=+=?-!rCPE}fNfJ!i_qu-jCrm|Y{>>o>Hw-Pb@C}ea~ zA8ocE-ZqcbDqro!8UD>66bc)}egl=MJ){ZyA)6<#>gDobG};X;orEs=EA_3w|1YEk z!8{KtqaYqKn(H__wE*t}oB~OgD7%C00Q;QEqP2>KA2q618Y5aikWxyRyQ+yCduT+g zc$>RoZa3>kzU_~oK9#Vli^r{e71RjDV4Ls=w~fk?13lHA_J$JEqJP0`WVu7D#9ZY* zhE?XOGS6W@^2OxX_DWZ|E)K^0NM&acH^SU`%>^}6F1B&40{j$_`MtMFluMe+190SC zAYUM1p^r4)#V-;pfemx-Kdxzhh9_o6l{rYjY|UeAnR3JNVhpVNAv*^(SedzQAIvv~ z=oysmCNJ09lV9mGlPy}%wboLIB#LQsU7pIR<>_rIem)5$~8rx1CTB6}) z#1(WTWT|BjS+M);s#(6PWvpzCZ00cJLvvk_q30c+ zXacY+vfR~zbgt*N_2)Bc6naujR|j|wmQfjbNk18qsW+#iWRc<%dAsc*%&5&>2syw3 z{-Bk6Fkbp)+tZ{{^_fLEX!^V|yj|XrN(D)IKy>Q#9i^6maiY2A zjZFi}*S5is=|KVPX00zpQR|wBcq7`EUGaM}<`b5~jVN~e;*-NcVB+8CzTKazE3OKhc6vw7kOJ>* z%WKG8=yDXuy$j(z%bPD;+;^EmdZwicpM)XA?kSmeRWi+zP4K2K83sov4rwo8)0mL- zxrfIDf$){{r~?i^>41w!QJ)o8>UA>h-DxjKh!Ev=bFHb&s3SN;7<`!MNfe%v3u_{^+*Efz{WVT(M?DvWTOtEl>$RI-TrvI zXnNlevp%7$6x->@0kN%%Y)<8~+HD&icU#whVKhA2k)$f&NPDWfUhKZNLlUkIcsO|> zl)yP@lcIU>v!D^s3InF{yE4fE3Sc#9(%+~pbGGiuP@Q&zExAh{orL-aL>YyVjet2= zd^kOG2&6XoD=Zosivcen7?uzsnA=EYA6@wn3jbnTdfUJw z260|Yk@?FG-drOp#koFaGB|G+xOo|LN19+M<)~gl<_?q92YnTBX0&>=hK~FnH$>9adpK zD2EuFqWz&w@1b5QM)c)QNKsZFqPt=8C+3Dfso3afZ9EHy1ZJXr(ygJ864QX@&i_m0 zYK#9WwkV@f=}w?j=(4e)o<9lz_h6~DcTm%IS$5R>Y3b?yGn^g63|6oKN8)otQbbO) za!DIy1Vk4fj_tzB{{BzW#NBQ7sR}~g4=;%xDSc=xv65G3X1Ee^tJQp7W(p=Mx}VIe zWkZ3t!ya^7jBkb4No=oHiaFk>)Uq9bR|rYt-c5N+gb|^NXlrO|HI@UpG7Ehpp^cuj z;A%wKJ+TC6wnJ&KNN7SQ*N>O*sb8Ge=i)J+A$*8no6qGy$%2qJN&HDUYu`U0NS<}j zpKEK5)_vk|%s%b|-7Q~_u2c{ERu*nd(zM%MJ5Gsmtro%IK%|T`#So2eXu#Cl+EhaF3Xs+#dfJu9o3n9 zs>_Vj3G#UZrgLqjMfKq{ztA6%&mjcNUekIi2QIWM#xCf$hY#~goScFBPGRm*A^U;y zMHf5byBz8O6+NbVI5p}C;V+b?^VUh1Rt4QwGB%)UattTY`o3^un29~WE4NZ-4pADR zUS|*g2R>sd2>42bL1MzBSxrPA&Y%_73v|}ilI4lG+aIBD<|HU&E=)_p7YYYFeVYDv zX4q+T!p`5$pD=mB&Q zqxpE|TKgCzFz#gTL&z==XaZ*yh|Oc`LPB~g;k>SL)1a*8hj*gFQY(dy!l#78ReS%PHh+HQjk*uv* zp|iXu)+$4p6!wvKdm>;QNC3?HR}vXgi@v;=7X3%!<6NcJw6y&iyQ#n6#&>tip+ccS zm|;iFDpoV5nUY{%bsu}|y1CNYALu?Tvf}}{5J|oSG2hKuM`3xiqWpfdrOkbMB+T+p z^9-6}YpUwlqeDIoQGl@-IrL;?>YV;b>r>0`XD ziT_DbMjRW&Plp04-8hCxU!B^CnB;<4Jq=k)_$wZ<`hJ_h*1#kscSXz%UcNBzYanK~ zH?6*@N>PA7-1V%nAVtNQssR~W<-o%hPXz{7PkPA23^Yk{&2T8|ePNcibv#-Uv|EYQ zS?a4d6cd2D*GYFEVtFC8Vp;ebI!h8QgiyqVIT-1R2qeDq;tC3R^t2&ItN7*JV1TgL z^Zi8@QMyM#rcE4>ph$cKiIcQaiZbbGpUNCX4*dlC->raWzc!#N87y3_jOO(J6F_?DR_7eoTWKg|5p-Jw$L{S<|>UtJ%*kp+HXaLgtv` zq%KubvJ?@w$5BmW=sWjCz=ki~2u1mBYep*#IlNp-eW(-&7nuU_;UA^!SsYld_)%d_ zQbp6q#E(Gh)9z=YlfGsR==V=}m!3%-&9~d$F`M5kx`I5!KP#?pf+f z&T-W2PyAk_xDTGThpXYoqEgcT3dqW6%Q>?->t(M+X*0tbklGhW;E4*kw4Ya^b7w=} zSpOOJ;l2%98vf>jLmIgmvzZ3jiLfaKO7D5+wo!Q)VAnc9m{kUEFfJ*fmz&K^sWu_H zDYwY-b&Z^6c5=><8B?HBZvG1Z)H2YlQ;KO$rwQ4w%)%!zXo`-AYbn2e!iSwu;mI&v z1BLMmshx?MvVM=TYfee+R$vbcM;dGPR8t8TDfGS)9F*=-O5qvn#^N%4u{4*~&WY=l z9S#?Fecy0&aFoDtP6IQw!?V-?IEr_h+pon?MfSF?lFm8;(f(q6a9bCqDhl2QR+ z!!LXlX~c?f6Mby8Z~gRKk{MCK%=>uC2G=Y@-;uQR&~(b4TEPSNfVXgrp@x2N%)huE zG*=(}_UGcd!|88Hp*d*}LZ&-b8-Oe62ts<}9=rOKzFr`EYN4*pgUxr*hD9M&+G=tI?Jzj^u$VD|;Ohe#MS=9&n zmq>EY6Qli-bn27srOvN!ZsF*BLpn*@7AN-IzT9Fs5l+duQEH_33teb7f@6X_PwYh6 zl`V4eG;*$bhqe|-0;Y_$%MIScu3P8=!377`bTi?^iGVvM6LE-l(V~D|?w>%&Rv)^6 zKlLb;-b`wEBxCCP?g5}C#bt65d4h z48{`h6_1C$5=j>~qUyy&B7Gk@GeMhi^%+TfCXcOa@j4e%@R8G}XNSPAW-m)oCwxAZ zhr}DIW%v3h8BIRny-uN^j{po#mjK78esLmh<_F$R;Sz#Fls-1Y%0ix}A5qRvuj>13 zV&x5;KRUfpEl(|-7i*ZlNZnUCT_F;tCjw_EC4(auGv>poVB|Bt3-@Gd z*{77%H~1EVcpxmV6D9C&!i3k7x{c;)QW!D`yjx?0tu6;kWF0l%%awsL2t=Qg& z``^WEa_A_;fUM2b>5EV&58{-(4BV({OlzD&rW`+IN{XM@M3;kpuFn6XM?-I05ERYJ zJM z*49^a+rqw*rshr5fU497?Qv**&u4J%H~XSjf8c0h?iGXrG_>@2Dms=9gZdZ{j-HAn z)gG3R0eVpwJ*oJQQo38-T`$bF`<}~BGz0;zdumR3s4y{Gdp;?a=Z?%;IB$eC^fyS~ z*tNRd*Lpsw^l7g4)^K1So`H6IS&EsB)w8vBkko;vl_beUTGd#<#xDK0kojK4b<=*G z2Um0?Oht9AAr`;`Z;&o;%^uaMSMXLg;*D|C`CvO)Z|!a zOWBzf1|#`yT_*|S_u}#i9#r}B7FZj%{jho%Oa2?xjq{J``ShvQ{FB+i* zv1co7s^07-C9nP-1!Q-q(Ub15*A3U*^Lt{oh!$WdZyDE|GbGAvKUX)IEq{av*Pluh z`j+(%3yM9UFd0VkKiXB2@uSl%C6237qLjE=kZgXxc}c?*xhwbikQ6WD*m`PjNXu7qQ0q)`@@s000004H_Fuvvu#fPg&g&Tg?n|;9c?f6%?0_u(?O@9{Q~z zLvB~A?SJ#EkTb{|@piM^S%A;0KPc5u!f*fs1(~Ap0~@H_9wuK~IztwiJ>i+N2KWLU!w1@3-ISWPH% zitU|-p7CVm;gD%QeEWy7e0QGrl*MJeu=8FziC?q{^1@pELxk`TKp)dr9F> z1MzghLo(72 zVkP8Rrts6rf64*6RZ4vj_FpO0n%5$thzA-dHwqZ>U#1Tt3;#)8noV;1lA+Tyc9@Y* zrLpFY$HYwC_#(KauiXT(pNRKkb0(EiNf5pGBxB%;O06j#;2iaF_@K`MEXs}lvg}=z z;XPu_-n?98PJpGVEjP!Iv zoMQJJmU-6H8SWVm;`#1VrDhnz9Zhfnh>d)0#uvI*GmkR}U9*j~l*>i3Hje+ zYzh3;*3W2tq^v*C^u(9jm$?cfoW@izK+C;l0%Ar%lrfc#tUh|Sh3txxA@rHi7E`O? zGoG@WAq1-ps#q^v%AG5I4O94EvJpPZalVFhXy0EX0;!7rL>->RD62KbMO0DU)3C}R z=!d(&F#_tQzN}k;P-!SL3r}ijz6k%a=h23}Bs!Imn>?gYu5Aso-_(WXy1gU33Z8H} z_NwY$hMEbII=>XOKPh?C(^z~W`AlL&c`FHuRn%HpEj_znnqY9QjN<{E%I4==fT%5u z7)8zhA^Gy+$S(t2&-ci_CKXJ?`8w5o#a3z#PW>~o`VXZWCUMW7W?-CA`59ZczGu~e zD1UB%YN$r2LLx->{3zYtp7OO7AoW-o0*@T(C*X9)ihDAjOZ>8}en~c^-OR10ISaXV z7xZro*l5xnI8tXrp&1B1-1`00rm9w-JRNRd1KDm30sbuNU1R3|Y)2okj}zdmgMmjE0JwEaJ(k5@|C_wOUqxNp4V2QP(hEx*lhq76~kIxJju7yM-HD{p-8vlI8IKr+W7|<@H^LD7y%`J+=I>~r znn7J~!$f%*{wn%CCJZs$ys4|RKk&;D;^`ixv(-VTR*9f)Cd!Kn!bh5wD`)+W+4O1L;)Dq2GkfT-`uuiwQqtNA=vTt`O@;b%{m>B^2Y6Y zMgHvd%%Ca;GvG6UBt?iVPvEKTX=kkr4KI~FdEXEQw+Zk+0^cw#qd;>rlpkg=?uR@D zXs=@1_wSSKt#b_ZW`BhFCQZsusI&F6-pI(|ODK`N>rWz2^C8kyh)xYdjKixeIN*hh8h1R{06@Uu4HuYZLq!Xs^ z9K%n)J+!}V(;^B`L%=629*SM;m(f`8$EFzuGsPJ&4Foc)sN)ziraPtC*xm_d-5T?y zM3nikzDTzG_Gg>7M8aha)GJ?GiWyUxs7^6IkbXcFfqoRZh!OIDv7d~=0G9?4x_g)vbp88HOs6I^|1V6v4;;am==BBz{aOWF79v z6oq-WU*HZ%KT*=RCG4r^3tZe7>!sum@BbOoz~7+LNZXKDa2aM{=vzP`PYuJb9NXR7 zW*L{hQ(G!fbixlWR`a1Symp^IOUZ;a(Zb{7EDOG7BE&(6Z6g!)B*-y3QA82^ z?l4F{n=T{6OhxD+O@CF}zmogm>kvgjPBq$Yw{B!#gXiXVV}5`$dCdp58>^%21&z%+ zb-}x=5QXMeDOJGEiptcuJLH!Z@y@3c^4)rqX_WA6!XIRsR#3G(5&&xuE*(Q6EAM$I zUq!_38470oMj!JQr|nGJWpx+j`kaU_0*vGs z41Cb%Jk0M-rf(_7?|`o2m(kP}CrncwN|EkBdMu_ zP>YNu{PN$aiEl2gZF$k>A_~@6eg(k}WY)Y?ou-8?W4~QMR|V;aQwj1ce|fJq$j*6C zp4!$ocx-oL%BmajhA>g^f*HyjET8LZF`eL%Ch~(x#7T+tQ-DM)!@ck`s#L_`yG$t# z;t-b%iUz$b9;x)B`NWSm;lcZHekYJ*#Z&MV4y6m1e(t_ISz4!|wk%^nd>-ivl)pd( zk834!TG}7fJH=t-W9AnhnDF~E+E+fgp`;(AMII`1eAUVI{P>8a z{Y8I(Yu#&Rsw#KPo`VW`GMQFx(R^ndkvKZelq!5FyfC{_t=saw!%e1HSd>0n*OrC) zwO--@Dh<7!3TT2oPw2(bNCzfH!Qhr9s_g+FfC?|BW-Sa+=}jJ!W?y}BY%jhexivR= ztpRU~JAQlBMxP{aWK&F=ed$tc!b6Xi!w9%sV$Y3Ge)OPgHLN#I`?shiFye#N9y|T_ zUy*0kL8#9bC#Ly*cSwc50N|>#AzTgX%`QF{Vzz}>ICnBmih7lIgDRj%NC*~Ou>){g(ym>VIA`%M|N9^sznv+#nB&ZG6HnqbRTEOafU~u}kz78b$jGUxn9H`QYJ{M` zS`RAmi9#XlR~@Yr>AgPi5NX+FNkk_LWBv8yTh|ekoVQ+~4n{y;RDG6*ulX+shjW%% zC&r;WW7_x%sZ%+QKLE}@c^!}EygGszRbosKbWiPjBnaSf0xsmPhWMRBe%lxjyQz*K z29}jlglpD7mp7Pm7!2raW?0_MP-Hhbq2U45s)84pKuww_o~i%Gzy z-6~PSJOh(86|$~3D`)5dx=%uR=;7WOYq(}RO|o3splbU8)Rvm1^%kjGmr0hD|E8BN z(RyB#nTG%pi$4nw6;TVGH=LJz5Ad>J?o^Yl>`s8K(roU!>GJsH&`v*oC1U$`r_Q(< zpxWZEN{6x$#>BAcsd=5!1s3XUD32*j3pdegl`(zhA8F$_nTJbmWJwet62GR<>t0~0 zp)E;Jtk~>?L{*Fnr>qX^4e)6z|AN;6kV8BI)IJ>qb{i0q$1{3)=7SS4(B;prQ^SV4 z_1rgH9P~El-5T(1H7`XnA}%q}Ko)++D^q?O#FS7yL6>jN-m*{0{+RF))$)EO_^58v zIWE-(k87y;hJGb=cCsZr19g2H5^C!Ow?aS`C>=s26>k@c*$a9lKK4~{_yRv zR{5QSFq#a5%0G-t5)e2oE-+IipikbDHq2c`4S>4;O_p#erI%g(=Md=91j}NTx;z9d zf{6~etEs}SH(EXszWrZmgV?`ZA9Ped8+TtUeb1S)x9ly@C6O~Xhl&L@O_ckA>BH!OSu>FX3+5TICwSx2PoH1#!EL&+EXAHkG55ek&n#cl%DtzZqpE+3HBP&CCjs8D7uXww z4{bjy(~V-E&hX5*Yzwq3-lg<>R3>eyy&HQEZKQNln7Z9%$`iTAN5n7~S=C*Z*A6(2 zGm57sq=R@N4W|}WT5ySZ{KfW=!)as&$V@=cL6Ci;Wyq4(=rw`9I1>XJwN#@nxO~e5 z;QLM87+G}&D>5!>0rAgpXC)y%9XRP4ETL@g#_5`VO9#Ek)u6HLnbrhwfGUb-BV!DJ zI3;K7D^YNMI^`x70w1OIe3Ke98D|ty!l!*2BS~K0s?T0=ojk2BKaB8HW+o6R5OnI* z>qr-LCE#;kpny2vyj%u`UnWChG_Sl)$#y`0F2%=etNZ~PTdI5}0SKk?2|c|>NOBr& zQC@=QtK2*$EqjI7L0Mll!r%8TDxUZqY!g^D1Y@r<`H_PZ$1bhA0ckb(Q2nijx7~}j z4>XU3pI0K9O00zh3T(O`IlunGc>;OMR+ZTaJ17)t`{9k2ZP59Pt_ma;M0%(c@yA?L zO%dce$j@bb<`~H%Ur&zCo_Y$gnH1lisALZCb4Kwj!*~P(2B%zfie)LI4+g7?9PLI`JC;R)XD#$C%)6S z3+fN-9HF>{!61|S)^>nsP$7vT_XXkElnrG!22hjm0ABFp>yDd#9>!=XHH+Z=PKy=F~Rc|qc5l+F1MyQcLOF`D7d;?`}s6v zm%Uoeq~lR}&6FKBY-eW(tTsN9{g3Zd$?%X}&C%%M_T2ex@yMGa_y7;!e;npjF)_Jw zwTB-V^obfzSL;x6?a|+{K8>$pye!uk2B{bd=v;AH z9%Q6g3)m<%Nv{=O8Ge>qI$-XyHHhrVYe{<=TrNPZ9-L`WOMoDF$yYdstV)K8`B}uH z%Ug4&EI3UmRAojeuR8)K;v9&3YsA0Xm>#ya!3)6Bwu9OS@%gq3w-;$z?x`-vaGqAZ zgUQnAs!nPD%eaO*kwz~}%3h1^pZlcghD-Qg2xPl=3++SbZZRki(NhqyrwOZshM;p; zl7YU8q)jCMy>!cLtCgkP^o{AS&vs0ro9?LR`G#4;F6WSs&q!Mg`wcKz`Z*#**HIG* z9ybw2`Vcjq0cfKpZ|cQ<3)W5PY^Bq9XOcJW5eSvJXK>6}kveH4u(qyg;v>-DF`0O3 z(irU96+kX&5fJlJv>~OWBMr5Fj^u_fVI(M(7y8!8XedoV(-*}40S4bX(chmR>IoJ% zT}U~B3b4+{c82jDegX0id2|=mgdQ8Tty>=g#{-1Oz8lDo8yj0oeAJfs`@eFPE*|5Z zeMeqPS}t?CZbk=}7iCJb%{*;BEuC&D@NEIU|6p{Q=U1@axJ8+`B zHN!@P_#d7}aIgzhM2#k8MB@p&oxOAe;)ZxBVt_zA z6U(2AO{ZuDu3_t&0^ImNn@WyWc{JXmK`PBH)qJ@MZd9S{-;bH{!mKLS7m1sw=naB% z@IU~ZdtH9HGb7-e{OsowS-gVKdAc%FT44HWhPktI=_0g}2zo)OsyNftl$q1VV31)Y@X<);um@;bJBI);Y85K^b zuYkN?R#)`qg}jzWS1H#7$+rHw9}98piO3%^6$&eHIXr=4qI#=T6Praw@|p|y__o|A zN@}Uh93hp9yFgME=c9DphUxNmad9mwK>=Hj!b5ar+kxP{HbKnLU$V0t?+w2f#NYPu zX0Z1I;hHv>V(YO@gw$k~yi5e1Az(AwwC9=Q9+bKIyS zWPR-!QYgBQoA1ccB+2P*r~fhcJFn}=tnrn=oD9%M{{rQp#Qw2xZ(wrWsg~=3z95Bk z6Nsv>isRDR9INud1nI@RPUIG=btTp=*HvO@^xYOHb+ZQIgO=B$=(sy(< zn=UBn*_@Gk>qD)3g-q|K(ri2_)6KosP8yK{Sr(jRHbv<8X(S?~ag0w*PN~MDN}ubZ zCW6^Qv{WzboGd*(Z$oiv%fZdEK>^@!p}K;()l7VJ3^m>e;q8OHSh#eY?tY44sTL^EAQYlKNeAnKwhbWyR&y zb^&3f4BO>wfrZ>|zH|aBUCYAhg+cqU_zw;+&AMz{a|N`e2*gFS9}iyRfpgM*`9(4$ z^GSk&)kzrA+&B3A*kBuglI6b%Uu*(L93+ke%!p6}SiOvK<~AeCv> zFk!&dEVkY2Z9~N#2dPh_o9BIAW}k4wc%-oqDXcwC`nl1-m?dHZEja)55j@^q598AO zUbfG>TAd)x+2=KMtgpEHn%~EnV~V?ucQLq1uyD2M(=GJ>*oirKlI>&e$wf7nmx2LCqW#t-(}o&rr?vBN-#|#m%;M^Q6nSI|(wJ zGxws~AL%nPdDnr4;&K(ESP^w&e%Jk@yEVNw$Rv`eL`698lqK`1(oxOlg4Mo~`v2y5 z%{s^|a0grlu?*Q{&}*S8lIjluj><4o>~*{E;U0cmk}96{`EX~8v9dI;iYT`ua2I3X z0tU7sK}H+jXX)76SwzBKv%7WC+s491X}X3MkNZEz5K!H4AIz?L%~&8lOMUH+vzP)fV4cl39Hrl#6d!oVVn%PLQ1EH@1N4$Rprwg>4HJtdzGnNpq-I*+y^c*# z)`sc)u-@Clf-}%+eC3OR|>Q#Ul9hb8*E8dI!^r(wc#YzwihHlPwPWggyhz(SA z;D`;+94s#v+Pend^EYb!&$QFCqUHLKMA&BMg*xbIBbK258n#db0y^Bh1MT5X0;Y5_ zwVww_91ciqR-6Hr?xB$voE6n8&$=uUlyxRX+<*Mv5q&;eV&QJL(%q{i#YYJ*@>ss? zMZV;aeSpT{abXq3VAK%z4FRo9vVt}W8Vp47+>}>o+`~dpaG}~+G`+ugRD&D?5zEub z46EV~cyB&DRE#SvD|k;;AmgmbOuXx;KmbJ$gt4y18M4A$J%HA&C|e;JdeVLnpZx0{ zXLFNpLcALhRV33)TFVLWwj7L^%9q4Qjo@{SZ`j0rY0OP)oeTdU>Z>s=b4fBX2AYu@ z|MCU>+$NH>k`+)USW!p3dG}2TvDoIL5r&jOLK(Ilx=e^SoVSa1q7SdG8+H1o6%3t4 zpuwy{E87_p@uH7Ec&&fWSE)G7l8^QiKf&RS8gBW8HaGoCqPl#xpde+NV(35&O2E2; zMnN`$w?^x4&C~oTvg=+iPFt1>6<(@AD6vZ7Dp1>WGa1p|8T&GvuIfXRPuTq1qi>7? zX(ptV+J|HX=u3jag?{i(`Sd9M#`6Hh!J&Y(q$sEV z^OFxAlQl;}_y?EEJGto+iN~>A%_{LGxH?)-oSGin<>5ETErq?j20reI0+GgojR%rg zUYDYcss1fKlE9VMv%O|SoTN+mWh^`ebV77{)GD9)oIz1W+&p;*SM$b6@QQ6U z&M~coFN*=UU@I&ITO%4rjDW;qE@?gz_2ZSpodgm6$g|df9V1XGAG&IsnV?sh5=Dy? zNs6h|gssK_5t9ryUpgEaO!TMhHXF*V<%MHa^}wteX*<@l^F@Uj-Aovu;`JPLtQmRy zmyRZae>sUKbY-0GFOy}COy}`XNDv-8DJ9?`F-(im%Q}poadcr20McGiHvld1kAegm zjGQ&<(>r8&6+nQCNVY9a04fN~NS!0`lN8Fy>EpYmaoO5Jc7R9A`BEr;sfevzlfNb2$1Lz`BV_3)HSt3p z8F?W43s?zXIFb#?zB%(wTnj`+|3=UoT>zFPaBuZ`#ElzU0H&{nzbbkE28y-bB{Hpw zHHu5bZf~LIlYDMXvh?hgsZp4EKc3pxD5EHBx}mF{)qu~dPdWUA{*C;LzFuT40#+SW zpBM)66`5n(B{_I+-}&kHXxunktbjFFAyxDeYD2u(bM4W-t^Xh4P#9^`87_Xx=qa6I zRp;~Wv*@|ZcMO(tNjjC^3ithL>iN&-rZU-Zk7w6WZsfG&z-)0J)6DF+RoRF`I(&t? zR5%X|1SUjoo<#|B)N~dO<6b{MOy7{BppBC&=M)+&ZA3U%NGq&!pv$VNjPpu68vQYdwK8d%Bj&ZpZS z)O}q#y4>C#3D_^USc#?gE2f_nCg}OF(;&Bw1TxP+&U5P0GJr_RxuV(Yj?mX3@N>#& zamZY?BSRvv$QKU_nKBnM#X%2PZEvbKB)KO7xxSXqa~ld-}Xs|p7gVFm#FR&1GOD7KP(-QS89_8_o= z0bKc23(H4A#&E$h&;mKeN+zSL@`{#o-^YsD;m#~c+?@fRW77rnCi4>SO5-S71q=R= zmBeQ@U-nYX^-rCt>d}-Ed}7ekrrw~w`VbIsr{FhgvvG$*=_fI5PK_l_FxuNhFuBRD2w;8_*sjfXiL_Dek-6txEj$vn ziqhnM*T@4${9EX8HRho_6_2(THYgTbcPYVz-EKwy{)m@v@zJZe$E&3y42pQBj)AQ+ zPX@LSUlmtU2j+lBMo<5n9U7Cb=71UyP9RK8Q)5h_z^7r|B-2$>Kf~O1y#O5!pwLiW zS7TY=UylS?W*yBkwNHkBQfCbu)iR;`%3D#KD(>h%E2Fd%aHdv{GrhQgn(%t^APhmG zKsd`={8|s8OT{OT&++(BFX=9DbUK!8yy1F8W2wz8uUr|mtPXCrFiv7Nn&>F$2rT;HfK!Z{bUMJH&f21afvv7lfk{8^BSr2IBzDNgt@f1r5!?2m2zz z!ncGUAL6IM2!E$eIxxqQINdsYmVEpUp0vPon)b6L zmkZhwE{tO{v9Dluv%-RyCg(3O^=sh6!{Q0@`-^8L>N*sRlN3~uckvZtje9Icd;a2o zTCyM*VFXKd|BT$(1GbT7oy9PC{Y(p4L5Fp@ctZ)-Wef`V*((Zz;z#P&-(`6oZBaIs z^q``7W#f4({qrWgs6hna<)lyxx7j*=_LNsh-9S@eF_p_xjHHI%BTf0^L~&l1%b=~+ z3uR!ptKk%sEwPcteQsae%;M! zyIA3l4wDgShVTZLlId z)M9EvAO2Wve)y0kjOzzXH(K*0OJa@ZuyT{jd5h!+R3T>Y(dEmS1h_W=p!fg)09588 z^iNp7z5ZN4Vyus!m0|iLg?y7dW7#`7O&|kx{*8!)!?_H(A<2*41nDB>Wd4&^Fhkjb z?CoGJLeu>e>H($PGAbwkS|I|!|UG<&@vktYWDS@&_YKct9; zmxawUE)G)3pqL&sW(2f_F*c!@s9p3wb;LU@f|NH2Hznv0An4A^Z}ohwj0TNQtD31$ z>_|@6M6esB1^lb$`FFK#6InSK=2=O^Z;Tv0e+=eaajq=0w_F9q5eng9;})6l`)eVn za7xfNL*gv3Kdn+R=Tw2~%B^*(0P%AX!P;!&ax(KxLo5xjC*?+smtm5L6^RU_X#Lw6 zRljxqV^FOkYHaZqb(G52d}e;D_z-*Dd}fo2NO2$26ucM*qq%p1Dzu>yhiw?a-&Fz5 z!QW{;7UJuaW28p3l@YQ~b5CDMq6fw#b3E@y=*q@|2XN2veM|s0&cW21f@O#&*oMSR z62-_LKvto^xAFFIbSZ2|>Y&cm++Nv8jToo|-&%&;j^k=Cuf2-v~ux)#US+@tpfqZubdx z7v&STcNO@?ga=!JK9Nj%-3U}Bl0At72IpQ2O$F!z9yDUz3Jaq#7K7^4NHwFCCnj)Z z5J5%}y8wF*TSM@Rrq(m=(qk%^`E-wJN{wwKe{qRIxOp2=)N(>SgCs?7kfND&C?wwM z9tZ4`FkdBRXz;2P`sf{ZflE#X=#YsqSXu)NCcyjQEQs&plv2W!-tx6QkB|>rWWD(}cx*S{L4XYDx@05`~mz)oK ze+cBL)&}PVx-6=)jq~IRWl8bR63>j;90EhLLc=e(2d2Aj87-xYI;(9io>}rmd9@Ry0lq|+-!^?%0Wc*vump_uSS7~+m_c< z)x|QM*iVtI4sa9fT&$6T?&eN-AgkN5{Xh1TOXG?CC-4n@Ii{m_Yvg*6Kl6T?apOsFA1d+%3U9%KFIiW~rA`b-#O{ z`d?mH+l%?b0Gs5}XMZZ!7|Yu%>KLvxT{<8tah;)#6#pf4iWN|0BiQhJx$kxcwWj%P zg>tmr&K3wlsH${O<*?f8l|`DY#8pWzrPcB;11U;*bR!(6LiU9(r)`fx1;n8E$;lmo zh9oUouNlF9M^1xc38L9=JD+1)Fj*mjjtisZCi4peyRQ2<@h_gZstlNB*1;{ZxrrI6 z95`0?d~@+>@rG1|R=~&<#8GFTN>J3PG#?J!7CLtYo~~>kHb@P}XKzWz12`8ib2Bwr zZrp@sjG>PYXTQggV=w4{h{d)SGz}l#a(Qkrhj`t;1A?d5WMYz}ZT^QW4nwpxT~9x< zVMS2150?sE=KiW1TM_j`90GCzqLX#36hD9TY6@aSPlGwE3KVq`vO)^k&8ic`6sMyG z_&Pl}b%MoL-Ls4J$BVvgOuO;Zc6l1_7{RT2%6Z4I(W{uUlZRp|Aa4jQr#FYe=3n=v zTceH~!S;q4WQ)yG%Xbmp00ga0tgL`kFqPQQf{K)wn?s}HtBtE4IKD$`kHaT-UW<*M zk*IsdIyqG_xEbS~S@8X%;}WYvk;Z;UQasH2q;Y{0RXyJ}C<@+0O8&3QikACV2lYWX zZ~y!`QGmR9a-NlV3KL)q_g!o?i>mA$FAyAf1Bn>BsBpGAquD^2t&QUe34r|0ury^D zjj!Q$y{VBK-PEP+KaQQCH1?yI=N)XJt-ZJ83a7Rz%BT5!(9&BX%rScQ9$#I_p`eQ3 z{wtScBo`_^P%b4=O0TEI#rupy#>3{G27D-9R3-Rl1*Wcsutj2fpD*&v5NDJ~m{?~^ z=Uv+Qo*QbwH80_sWG zVtH^=EaQo#bW&YO`-PSbU02BP86F8(j3Rwp$l(8?<(Ff(&;Lz2*BaTzV%mt*Y3M|e zDCPubE+H|h?9eC(9+L}ur|=lh&ngPUrHu?dwF`y|4`@T>&!pM1lOLh(P2<{98aJ#K3&YH}tz*(4jU&;40O8>2 zgGI#RI9NBH3^=!QbrgyO8+k8xl|L)}X6OxEU0hMBRAFnkpsfAzMI@ z5W0VU^E%naXkjc={W6h7ZPE%xJ4C!}fF1)jMb#}p?8>LK_O|2s%{dmC>J)Ha1`&O9 zGD@pHI;10@8kt%b6*vxRXbpfcxlW}fk0~37u+io?@ zisO6(I9$>xT5`D>-kO?%qp;gWGI9*rT0j@-o zQcXlQH{kd8Hs*4+U=SNZ+SMJU_K*%DV&y}8GYeiPuZq51lGqaq_Yt|g}E$1u|}Gk zft^p;)9e@dXTT!u~N7y7u)tsI7`)&=_5&{OZDgu1QC&w-ER#E;c zmeYjO7HHx_@HA~VU#b`oISI(YCUa9U6E!)1;w4SfkxMl|-fBND*&Z?%$ktR=4_Tt*(|6?fw9KVX2bh6=JXyhKPCFu2) z!O}wZipFgP>_n9=Si0wo4jhFl%(bSvm3A1p?&>XdDy>?8ThknXui=_DGWRB(awy@w z44$nSf50F*%sR*gy_pHh25C~ij!p6un7rWS(>aJ4XLD`0<+hj|km-<)pn<;lKqhfP zkSU~pbK+7-_z-_BMh}Y-=3!pWweOwIHmB?TA2cmU4&tuO*Slg%lv%ax$Sz$KimlXt z+GHeHSv^D9P0ifKQ3Qnmb>%3{EloQX2`Li~F%w(Fze^X8d$z+Y$gS|ybwJhjQ3dyd zlHc;#d8v`kwh`i69w&8+A}I0OpoX&J%e~Sm1<(=pS8;GDIlV(9?ey7^&t|=~49TS1 z@}*Tr#hw*0ui^U{%~_e4u#eOi%`MV;&H_de1in}Ujz6Hujb@~l3+xi)N#om1&=bt} z_k_NXUTH+TT0j5O&JyCCvNu!ee^iMbdVOmZO~-k5i1c)64p>S3x*}h{*$J`s)c9?h&2bftLAO>be{P&?jJ*RPH>VuZ5a0EwptN`|s> zyV1D6_p4|JtW1Tnb5tlvAPC0GMj(Ag^(+1grB7CGK+)^f(Gl7`wVw1G--rhHS$EBv zK;9 zT|)a=n%-R3-UN9XLmU6w2R%aR_D9Ra-7fLRfHo9op1CZ&@#KR5_@Nf8p#H8o0Os_ z&VYerKKiw;xj)vGWCtkVJpaxgDo|Kr${Y+N`sQOaqB={2VyCjddrr zx&$nO@JILG51KKW9-DIbm*6P4zVTW7*UEwMfP%$jc`=7$Rt~b|&G#c>eUl zEljm|{UHOcBx7>8e$0}iZo5}|SYWuQe@6smI;bK22W|ak01PYqd163bq(Ah38VbrQ z-pcJ_5`u;0fR?z z?n_NWA!G7JLFTrjsAc*}IfHqfAT5qYyUS`xwL;jXUjLFQJ&qg;Q~BGM;%zWSg|<4l zSqe3K9hxe_~wCYMldemH=uRywf z0CitFy9B}~iOD^OZl*Fic_h&0>K29v?%IxUPMf(>)7SJFTA(4(!~?iyZ9W26g7xra z#xi?g`E4A;i?`53@CUTUq(=++vsSp#N*+`4xcJ3SF@U3nZNB4n2dXYV3L}OQM8F6{ zN=X0?aP^b)IshHI_z?RrWU(U-xAHR_aDrbcOGF8*WGu1WkGzM&k;tf?rmd!1)H;hC z?yABDsJiE#iWbe##5~(eJl1<6>4AU4#bVP);x@6@mj}*vn|cc|3YnKFnug` zK*eDp;#pAoS)G3#T+&d8nA2a9-RNKKws+#sU2JOy!<3Nrg>^3p6G~t;j7XINHE)Aa zr`u}gcFyBrk<1t!#h~~LeLvR%C_%hK?@e~=w^Ih>#;x$K^WkQ>RBZMe)CbS6=U6VSF(Eb4~)_NUo zNkalnxsQu&F87hgNKd=S4=q{il2P*K{A>6R&*5A}^ZFLHS&8-3q&o29OxKmHoV~qH zL=x0KVa>x3X*evJFRY%^pfhlrU*MtOx*=C2;Y4=Mx5^W+AJT@6yK~{UH z<71}&SUIm@`c{{%eUjg6wd~OAC$WX5#CBGlzR;6cZ=-I9P^84QnIexS_t`3m4mOYg zFS##Qae?~EQO8Rb|2fe58I-Z>{w^$ zx)~)q2ms;jM-qZCUA_iqtGMT+3BmO*79Rlld?ZX!BhqcvqQPhV=w;1uC;SJx=6X%Y zyU!cXvh3M2;~kIkS+cKLp8a#LuNxw&NqD2*i6!W_lf~Q0PIGoJl*%0EBpkIsa8xQXswnLS4BqHN_H%YdKR&o-Xo7 z)m7s{4XQ_72HxzoNRg1*~VHfUM6Yfw| zA!No|hkCq4@#Nn?2kUUQv5Tu=+%y@cA44vK>1tKu)LrWmJM0~}-AY-;=ASHj(VYd0 z%VmnxvEEuOF2Xs07I@9E86N*IT|rjZx)teH-a9m9f+%n`vbsZDPMp{aRzMu(Al(C3 z4U;kbtY9fOK#}82hs=XnLM#|nFNER<`(_)mb@P#2cv|*}<%w9%bqaVVubfEgb{r!Z zqOJ~(lDNOj26ZbI$JGNUyw2B1Sggx9efn;UCkM?bpi2ZB{DO zc$Pl{aZ@Zz42cNO3SC6ej=Sv-7!~Hk$*jj8@rLEz0&tn*q71)$h#*OEs`eRU7Q0gnx@O*llnY1)Qq|S2ur?qj=?l7AV1LHVLz&8L`RWiumIP24kRn2s84Eb?K zFrbI5#|pjS&Nh8p1n8(OWv{asG)&9ego$%>%_e_q8$7bmO2-c0fl-j$jYo zh82RmGJ5#mE;~GWAEi-W??rUWV3s1dAyT&7wPA)?GywLXO-OY}D>gVHl4gZRQ$S2vCbSN%QGO3#4}l%E4VnaQ#AA{Z4e;BE^& z;p{q(e$<6@fmM%Qp(@HFxhT(y=vpp28O&%N)Z6DHm!it331RKKfoCiQkAnbY9M~rq z%bZs@4xy8N4012Gb~ zdh}Qw=LFB!72qj>Y5P9Kas3nIzH?e4Py6Xkc?e`!s=Zg$#G)?|pSX^gVNXVd(v14J z(B!4q!{%wrJ#J)4voA)EuEE(uc%X6kwF)29?3H14tOZWVE|G5-EAyyc%l-24g_8dt zGl1MKcK&SKSaR=%_zx;M7!cUrK!VQ$`gEwKTU8fnROa1zn!&dt@r;UzG(=eWg{XYKLrQlW9=E_ zLZ-NnF3{gR!^fA7x-V^Kj#|Gc|LUrFP{ocLpqS8YbS88Z`GQU3xylQn8ZOxdZc|qt zZoaV?jA#31~1lKvXJ*)Gx2?N>oT;UhQ{9_pfKmn|oCfktnsAhv0B*izr zr~Q(7Wnj4u3x_M|N(qvAnQ1xb%f;*0ZoBAv+s}EN~C) zWYnc`$kN0Ot|++KqtL=}bmiz{6J052($X;Pj8*nI=zI9bL-S@wH>fgPE(3f(V?Dn6 z?fSEdj!RKoTD3QFhE+oZXYH@!H#OWv=0e*oj!eC zIQq{vSF8h4ekyK>b{P%#=Qd4c`C8scX`HjGzHEHbWoZF&5SOI%pU4a(znPy;Y)iB^ z8`lQsU|D8+A{4Q+2RO1iG}!da8)Kg~KmTy5_Zp7VTTfFrA#NtrBtA`q8w{p+?o%#O zp%#NOS{cNtT^O3cpx{l`XkXtv&gv#hX?(z^!x2iH(;Ank8i{d_(p#-K6<`l><5fD- zqf3N&x3~tBb`0XAM~>pU8qAnvkpk*yU`8p#L^Ct?b+1VBDGw$`0awu+AKhy{7V}w` zL^duQo9t@I5hL#AFoDaKn9AqekhcEs{B_&3&MgO$G+=M*t)Tsk^b!~QbG*MA%nUN5 zr(d|B4Hv)lwrRCZ`B=WTr)c60!~EC!}Lh;TY<3v!F8$jZ~*VC z^W{k2iLa90JW(e?;x&fDWpIn@$mq3xEDj#m$Y)pFpuUOY5>q5cELGNZON@By-zVoq zh*;Ea{~#>N{RfW&c?3r%<~=F?!)=bzP}pvK8eO+Sr_%?J7A_+!flLy$fXa`(=#tux9S5U^w1`6C?Zwuxkb-T>nyV^8wTmh>0B*Sb*4R>AJh$({c$H) z@_X0Jy5XuAYEDZIC^b_WHUsrs_1Z_OibG!NQDQ?3cM9J;-CwS;$_dMD2Itrl;re;M z+hMK}f8hty4f$+5AYrglzI(TME-dM|59@x;qx$rsneOld*aYV4+0vq6H4(O+1&YTa z5Rdt2vi@%5@Yj!$?}m6q%XyWr_YO=p##U;Z>yd9L`+*=&`n>vT+zL|#{InsJuGX?3 z?OlY7NK23+h)hT(hV56D)s_Mhag?>u1gIcZRZW<=Ag+Z~aylM>063xdyl?OsD(-=) zkIT4E>S7!Is`02MuGffx?pyVBS~_3$9kWGHmjeB}Fu7Jm;`Pv%5r7R!+ zys1s8DwHNI-Z-S!b*)jFmq(79X69{9+sDkt!}BkZ~&k5=ft zji|e+dEW(<<8BbQc5KK=aub52sb`eiVkbaof=6y71$$u?QP(Q-=&)WAE zfxCcBh`&nN*4#|I*1GL6`RRqup-slAj+a)jzjp@TzJlTc`y^b1x9hdSpTXA6^@A`=gt~D!68t9{uB+! zq*wOVeQ<~w$@2yCDNPG=A4ZQ9pCt#uk$CTnQ*cmn<&Sw8CtN72(@l=DTRp_7bEV@1 zc_3_BQ=4CPGH(YiwW0eJv#4aUhm|SzoBut9kH3{$LiYYy4dYrF3GJMcm%~7#5&q_DLj6Nmha{I zxPmC72W~LL3kP9j0j_lAejcBfP0syhL_^q;Nq20G)H->S_iRWF-el77s1asa%ouz`_x2~XREAT%afV@J8?3HC8H~BhxYT;*ggYjWIu|uWB>Fj z57|Vxn5as+iPt);6^=RGqUGl3z9ttk@hDlf&c6Z_e)-ATWq7Hte5f|j%bIGaZmjKo z;NS--)vGmQ*!yr({0yJyL?{+WFzFyE4`rEAluO;TB5aq`_+OGbl+V0zIr*hb|N56` z9>PClg%YkIB+xN%NRl8W_2zky3F39U2fdV)+UUsm`jrA++i+NEKV6S>o}E(O4l&L_;PqD&%K=t~5=I@D8Mi8+>@Z8EWbRd)$-n7N#zBrgoJDv< z>XKc5OUH=^SkL@+@4}zjHXuZX$?R8IXi7g97oCBhuZh#gj~+aC?oQhngd zz-HWkoz6pOn4-_U7O~2UY3H_xS0?YHwDhel`P>v#4>D!Eph!UBm7$+$C;%QJ3-|pN zVPT?2CeeZ8XC_-6oy9w5EOs)0L%UfsOdNkS@gmMX`8BmCl}lat?6+Hxo3g2y$YSV= zPrit6WKOO>ZV~R57EgsQF$SYRT9#BPd#IdOVk*(2))IH&e_kTZ2$9fKL1Y+9@?j$? z?D7`QF&Jg-b>TCv(LH-f?gw5U6f-3L?LjTGp5|K7&Ny}h6pf;mI$X+?5y}S-d)j#&jF7@^g`~MHR;rf*bZ8^Fg?))rTwMS>9!Y+92<4vwgVU zb$-YG)A-jnIs$uUS=$YHl4zCE$pqg+eSxq|xDDmpge>+@1x*n}6>ht=^`n<iwlR6)6V0c?WzAEe=#=B5H(z{ zDe!&o%>K=L2J8nQwH?em>yocO&PkF_n1I%pRuL@2_4=+A?AKx<4_VhM?L?#}(h#*>0WSf#ptn4hMke&q`3vdp7~g&YQ8f`U}0r`cH&vjre;#3{hvu_IY~C zRUYiVY5!avd$?5TSTBrfIyiL+yR>sk?9QN^vA%*%lu(XASI~R21ri$TD5WAuKLa1( zCtzYkBNjyw$&1bFjzBqp2$Lofg}wYAtJ&=SoH`b-73E}pV~ELE6~CSBNvyAoB2a)N z`-r~zl_7&hxleETWMap%6x5;~IP|`60%SwopInoRTH0g^m&_r`e@#!KwHuaHbU1v~ zdPzRDQo~qC^vsw1myX4*^JE&9AO>mSEr@Phn8g0(exk8(eK=kHW@7IR z#;byG)&T%ev2v51EiSYJQW`>&MX{XEy44eUET(6`S-g1EKMNs4EL}pXgIDOfxy+qQ zicL3@2tJ{KNy(oBHMk^F;^k{Jk>ek4mmxBK!b=a?tkS2uPgmqqv~T2N6Pdg@PoWi_ z!dKza255eFr^rhdD`W{2obVSTbRND8WUPYB6Z}&2*9N{B*^xx;*5a=bq%|>)2KpiI z3jb|*5k?%?o!)g!s7aef7Yict$AKcfWU=a-4rohIdHbfm^HRO29%uw>G)lYI_%}A& zgp!?O2Qz$Ng_PcbT0b{mu%uSwS;o6nP3y*=ld5ap^Sv9X&LBFHxQue>3ItJXF}X8D z*kcPQH1stEvNGk*oBaf{f3lWq1LIoYrXf+aNSrBNmIq(L>qr(_q~b6DtFxsdb=BC+ zB9^ZS>}*!AuNN%JsNwfH0Vk*_P>M0AyUXfrn$S4v2%MZ!0uZseC393V)o;tW)1rkm zlE}z$_3|^m&eHbz*P@?xngh-K8-M=TPJ#(?og*K7uwIj^^hU381ir*4!RWrMTzyc_ z>zXJ5d?#05$i_+cJosuc^=kkkFinEHS5E=|$Vi*y#+X#aM|h5104Fu=C$`-NPX3Nau@ zqO-#PI8!FVI@bv{{42H#`l|dl=~)I}c$)b!I7L~Mez{ZdG6lpOsf7P&`77BY*T=bM6 zJ3bL&p`JFwBMiXh4A#6O(FY6@#o=g7!FW;fEGAy|O*tB#hwdwZY?e@B19MIdR~SD- z^zg#q`5bO0x+%8Oo-SF0j_1d@`PD94k32@$4U9@)4WQLD8YC@f1HywdJ}~GbSw4uu z@cWd`-P!Hs#-H<)^%$6>_>TSy6RE&KqMqzDvV@nydOxjmfpOwh*T{SYt0bH~^9+)T z`1-TwI8x>mglSPQBRnzv#jMM18SKgOLrRK5IJeOoQ+bEe@Ckt@s>_@*rmE7HkIc_a z3OEz5W0=d%t{V|_t;bn1J&UPabNRYAxQPbcc`)(~6EqyAlopXKCuMWNE{#vR;up@H zVKb%?Djze0w`E!#mV=DXH7Qx*-ZVP7$o}A=QCwIZvboqy)9( z>(#ZJJa7#o*;pCK*sGdgF%QMQC=V+b(5C2-491yWsIyK^79y$(eeVOF=l;lr5;a9- z-nV=J0T;T6UVLpdrI}~Ygkr+NO4F)-P#B|GLF*z{>w=5YdVB&fFM=;!vQ;+Qb#BJF)Nc0j4Q4KM@F z5@H?xEDE?n-*>WyC~aYsdUxkvJ10*3SDVUF@6-95jk_vY(Iy8)y*Ky7zU9!Az)g4- za9e=CKI)a16CPzDO>!gK6rJ6kiKa4EGCNWDp@Y*lL@gl=E#ZI%|Lgm1b z!Q+k-omZr{EkBw;C9$bMmuecIdb%+07;I=5722&F`*&#vEE_q#9?p&8r?Y2P0yJ$Hr8pCOy#iY+&A|}|KFK1U@i4b`8Q75RIB144vKx$dogOPR`5c zl1|k~LG2`Z)NH~C%aaJQ{mzE2eSY4qH(tOwvdsXw3uoit8(fEz{O1siov>)Uu^TQs z`q(GH(yC%@;m0CDM~f8GXC2DnJhmt2@qBv?IwPG*AgVVla2;a%1df79ilwO8abc}e z_b(L^xboZuElhGZTHuuxu@7{Zp$|Q-&5uqY?9Z{Q6SrJv7TrDP-^WYbGwe(wUTQal z-1i0_mHq0s2S#Q;m3EU8es`gTQ#+)PZDKTaAOZB*epz&ODcr67i*7)jfm=HnY^o{> zsAXV={>AqD>Gf5UpK9jAic%DEonQhr7!hg#vwmL=4B05Xek=NeP+{8E-N= z`IQjSleQh!LXP?HKN>J&ZMjcku8bpNo=ymN(^G)g((Y`9F^mm&fJc)L zg5KT0s&UwfqwF2e@wb9)0TQ+pD^lUv1IqN)1J);QOz%-Y>y&tBJd7R;_2b6w`p0Gs zT6NPs6>)@lmDEWCo%qnqCvLkUwf7>V0p)}an% zgQ2x)dGq5wTq~t7`3u0XISx!a_l>;y)L>jS0031yS%x$O2hb9cNs&xNF`bHT)Hp>_ zY}fmJHfnU_1P0nF!gL-(=AQr`cA?b00en`*Okai}GS-AWZ4e3j={1$le{{2EZB5-% z({ydy>MF%Od`D+7>-&YN{wZo&1nM$^X$OzYX5ed!Q;EDJI(y{f!f*7-GfHVhr|IAB za}IxNa32jPQHW933hz8Dw=CR_Ax@syp!BGW=MP9GqSx+Da1%mklw|17SO3hAK#oUC zUxw4$h)dlr-x)yEvu>TKTJFG*2$nkWzR%J$ef|6BSn7?>Cx&d2g=v4ih?;L)C#s&@ z#sN7tl~gQu%K=1uWZ+koY6tjDTkf9V2ZYg&3uhaagAc3C&dgyo@Q$tC*)Nw8r%hxn zVY`?%CsV-}E%gR8v&&gjmC<7klUbrwtpw9xVWN@Z279$xPS*Lvi%`)<2cyLP zGh?%MV1(IM>;B1OMdCJOdB2!fEf5tZbM_ zzJ?_DM6^pW-<_;^N{rlUe%kisFM1g?o3u>jCQgQVEEFGEm#AIrp!&3(FDaiGLIj@g zX)SZgtCOPK#Z<&;{svw_*rP-8oSW3oeg12;-aoInqFk>W%UL2Tg&;h_K721>33IK| zt&7`-%AklPB3}vZ@zVJIgebP#Xp*L_feG%U&TZIo{@$()mg$ew9}?l`+Od^R7rdh zN)0E%G0O;@F$C8E(xAhO<}UsPcv}me063j)b(6z9LHfFMgK47%qo7zjI+AB=k1oFV zK}O#HUDwgsv@?EzAe)AOmyEIaCIc(_%gUVjQ;ao)2ga3EQkR?$-hvRDpQ8vlWY{vz zLFfCt*mu5961-i~O3W+v`6&Ogh_Rz7A07?4x zl~yPOWcg9aS=_W92z@?m@vH$dVXU`;oPhFJCG$yUA<&>(1VBl1&nCd>(=)mEa%`h# z4g_@isEs&%+^iNH<}=8lLy(elt*!Kb#CJ9pN69pBeac7EWJ2rPHPcI<@nh7~q)5rM zH>-e_MG&2c)6Sn>^ts=D{})*K?!zdC#+|f934FzSl?<;{L%#mjXVBH#VW6mfWt~WR zs`>V$t<&)A#wIM%t|l*yNqQSaNDS4!5{t+x#RvW^!I+*Ceo-r1*7}&qzGA5;UqosXUU7-FIzr3)lmvhM(~=nqLAwzxlhAz6Uec+BG3l0(X|UW%3jvVG)B>|%>wQs@77(X|CX^e@A{>7^VooM* zMFw}8R}Qv^2G?q!&mA!{(~KmtltncBVh=dadK>GetahSN z=27sZ@oo{}EB#@?>C%u&GZ2BL`i|`3!z5?Dp@tF5A?Dd#9e`MBRJpf}x(@KfS(kcl znDW*$XA|OmNOf;P*V5R@dD;%uw#GLCgOH$z37eok=$r;QoCf;V+W^zKa{mdUPomI^DMh#-Z~-P6O6 zi8o)>LvKSFIo1=xc4Uz-kzmp2qD~Z%M)mkRgjfp)e`nQ%Rq@F9(!FeHJiV!g)2@1y z=J=cE{n<=DX=1Mf?ADb)%l0?pk_@&PUIMmg2+KkH0Y|E=%>mU@tmQKj!@sQaY;^EUNw>#Lot@N<};u_sfTOG(cEic${{cI9-e9yN`wE z6++Av6SkU&%36&3!_m7(KLboQD?YDfRRmJJ{@t5!Tx5jx`^Pc*;g;K8?g{{jEK|hB z!wWl`ZUGSqb5`{k?iT2U68ywzg3NcF0F^3qS~ z(uh^moQ`Pg_TkV&_4dG+Y1_{bB&vdv6BjvoTotKD4;2}+tpH2m!}}^hbT0Z|qS(I) zfSdeGdrIt;`rc_WzfCa*BAU?*s8DXaYxi#NWQQHRRcA|y#IN|Q1N;Gxu7&))nui0Q z-y%*;%(OKz%RX(@o2({1U+@8%WzV9tJWuJyFg-Azx)BAQw04e+sS?{eL8m;LG`u-r zC$Ni&1Q}r%fH+zFl!J|ZA<6QkYK>fepmc68Rvu}EvXrNbj*l~&x_KQqVx0yC<-xNu zG+i<-7wK;Y@=Lc}Kp=eV^AS}ifq9T@r!{zU9*>dLo2PxTh%K1ti1iSkV~?%_--ZHN z?n2;V0!6LAP~UA*9l#JW*SeOnvYE1}m*XOjaC71j9B=)NNtr`Ry>MX!qa18RjScqQ z1JY$+OTWtnTHY74)^5dvx9|AtzVz)D&J7Uvu2TJ_7<+`?3^Lr;-a#Cbp>5Tr=8p$t z4a?n1;5IZHBsiA|ZVON98^ztW!_Y}eL3VMOr(J|7PN#wWz|U@$8$V%@{K*?tjDFOj z;cd+v^St?Q-1dFc9%^)ZSk3Azpm z*0_hFS~-a!RPJ+wTkn*cs%t8g?vP`h0tfw0 znY8fNxgAcbczhRG%fHZ_0w3(BYM!t+@{nY)Cs?0ME>TV{e^3JML)Su{VL9Z_=$d}I zf#g$7h@0z~dx!TQ8=e}*XpOfhOr2cZTWFONNgApd&nqpJLTw+tpe&_Kg?i5-waq(B z-4ENBpsPjXOmDKQq>diOj3M=jTF3Hbr5sf-ylE3cFv!pvsT}6;)vmeDEe5a1obv{) z-w7ivUSrXw>BM#aAr$OemR=h&f({<5*$5> zCT~ZUb-cWStJ^k-%%t0qs-QQf^_bTLXjYH+eTq;)lib@$&Kb4K_99>a>P)5r%&eu9CV`3v+Z;T^u?XF*m>nFyN!QXynqOS$Yq z;Il6SYf!f`lcr^0Qhc5(SV5jf1=;M;kz~b2dKZ14Zh#*({jgj@OFmt9y^9z(!vc@Wgli}}9b;Yyih9h3JHzrIk7inglNJ>%-&rd+H zob=aU>f@Rgb*q&5xpMmizVa>|ne+|EA4tii_Ys-L=70$mfuMM~DBWHkc&&N!$I7Mn zWo>eAI@xpmx%wCJdVT}OCEN>#0JF1|(=a-IZ}cYM=gSmp5*FBul~dn%ci1w3XQQPPd#d`6gvwD%i4Eb}i`Q?cwd71l1f4tr85gU01fQTBgXsf@IfuMzO8pEsfbED^ z+IgXA7S}u7wQ!Xnug7Qr5hK9iZeR*H5f2pP~r;{`u!fU_deZ#Fi7Ax)1ltARtc z20;pSV>ZzCr5XcItKvp)*UNqMc_>rJ1~gXBp5ibq9=*G;7>>xzmbO6t9wg0|wQ$m~ z3C1sZ@i`!xzuWKDP4*gnDlw~8h9E6poRu^Kx8sz?S@H{B^64EL^^0@#l>nbGB94aR zn->6|mz1m}kA=mZAj{CwGH!Y%SYy6%d!Y3p!)Q9t&8000000001)cL|=T1%PEK6^+E7 z$o)zeRU;P@Hd(CY6(q`8D(HzE9*MiXtHa({tdA*;Q=}++3wjVT3cYPI4B_nF>B)nS zxy}kRvmii14Ky`YNAS{+VTyC26F5Z>;G*N(V_^rN=Z) ze*Pvw=QAi!AQ!bJDo6<82faL>wH0HB@a$t%5cx4=ly4<3)j|484lIVP;iDepRSqFa zhj!<+xLY7DD@qd90_j-{eTJ4U-I3CGPt#))tTdWO3JPZclJUu^vh)=t5uB{eBwf|3 zt%i@O)mF6#2|q(T!%Kzig(a`!{E?-b$(H_seDdXmDv720TO@uATDf{%bO9Nfe=9XQ zQ0?2UQ-h4-$8-m8!dxLP}TuFi%n7+UQXl5CIyc)Ydm(K-H2(8G21y5T1f$273U zwr87P;H%1I{JNfZyU-?fo$8%G*8daxd8}YC`%gmIf=T7Y9%-^r0cPyO#N>2jdF zB$E|QviM#3{2#4Zin83lzPnwXNU8DhXDORhr^bC-?o!d}K1Mel(Qk1lAqnciXJr_% zg$#Ds)H*>H8fzsYACbsUpkYfh*nfYDPrfzhXqKLd`#rghgxqnz!`6BE0}8}!T?U2OPVrR0I~02Eajzi`hRjjG`ThG;APUqyKmxK zRYD1b>BW}#x@!yB1gM^CuzK{gnj8ftEd6excyA#q{{yhZ0UN-CVcgyNF_XCu=$IfX zu#d(5ByUqRckz-P)E7Sr^c$gqwx?un!R&UQwb_i+d%Iz4qWS9}0ruvyW1qK}b&w3P zL~=ItJP-PlY?$rQt34+-7*gVlcfL#EboY>Ue*!NtegYtadji1844(G$(h3hyi*?_* zCRn4wN#K~|)I>j3jT8!3Hv%<1AXMPmlThB>a1ovO_N2juDa)(!LzvFQ3c(T8r}^(T zccgJkP#sN?A6038#&NyL%p+vks%WYvY+Gjwu=_6~G+67wHAHqyYHqs`hBw`EvGgIG zm=Sc&XezK#3y%)G%z?{_isXX~LSPavEzbpy1GnQ`P44m5*8$mE>*f&@X7;(b&N*8C zyjhD)M*D-iBF01sRbEy6{?i*XcfoH@>n9<&7{TEg z+&mc&y8q0_gE=F;_8pO5yT|ls+1IbcE7liAzqPud!^MCcZ@yk&{W`q3mIp>8Gl)2Q z|EygzbatB=lq|?BRBZsuQE}3rs}1ZlSSg^RaF4~(iP!B4=}!CzB@F3dsR#RQ2hg)_ z#^j|676p`hH$GVXqt% zPqEt=StTf3=-Z=TM(O($&-`rInw&{E==d{H@?aFm{B-g&q5D=a=V!`^y{XiW#)Yc@ z`E*xE)kDl5363A)d(h(>gR5Y{fnE-PU~}=;j7DE;GS*m0KlMmovEOBpIJ=|I1YvvA z_q|#%qG%)JYHn`5Llug*4k(FBJzeE+C8Z9luPkuGP~kq-@JNmVRUTX4roJ=F+bqnR z5a$D)76RYicz|6$Qi*}jY}UK`3hZ;b0S+ajPh3+X4Sv6=-sk?9_nxL*z9oeboZWr= zY@tQpyZ_`7-T=nf0kYHTUmpUX!AU*wQ+k(e6(PX`2^*p>i#qZ^ywdRunc5Ns5bi++ z^DfqA-Ax3zfjDgOS;-S5AItDyJqAgklrz|7!GHhg7R8DR*Hk5&YJ&8d4xqxC5}bkU zF4Wp9)k=66EV#C|g}xSYw}3-1${!F5HM?8g#V~9i6CKfChU=10b%FLVQ%$2wmE^n$ zvPjl~yGk{?K-Rr?i}VlA_2c_rihq$z3;~O~=y|dCNS0!34#~9i)PN!F+WK_eOM-kf z4;7=!?U_t>@tVBV2#1x)?}!i0?Gs+Q_<2H#Mzc)^UfrM7CJK(^f)D*KeTM$ha1x($0Gnu%BDzrSd30QWNm=4w&D(LrCAeNVt{kyc7cDmxl)?Vi zhe9rbv6Zn*?rCMI7bi+13Y8dfpF%iMO5~nisSkOmGgI8&<~TS+-E=nb)~l!VohQ#0 z{BL%X%K$;(4Q0#y_6XX+FNdlma0Ud(*%rsGqljhzaP)sunOobUw#RUabX8u-$$+(G zIH&W*1eH)+B(W)579Ca6y_9~mnU`J)tgL^X%TV1EP|I-esCQJAs^;s>f%CV|JE)rG zvOd?+1GG~T^JQgkYBCL#chAKHdIhXxe69UzPz3GuS$mtuw{+pgg7*G#HiJ+}fwC|i zkJ0f|*$S`(&`swn%$Iy9`;8S&#$XFN#@7eM-0+R>^Z)J7{$;_A^fogbLej&%6hPHD z{GKI?G`fJUMV_?{QIM9D9+&dXsQz=Ev3`umf6xew&M*@AjOT7ZU1M^vaDTo0)ay*< zBiQ^#LrV?(JJKXqyEmg4TcLr20FF`mTytvITZ=u4(u>sV*0us}mRBzNa(SYCirk6bNhzQ*fN4nr(TaX7Bd)Ajp_fHj6;$dKm6#vw9*Gd(~W@R`dTBx)!pSM#Wut5gIld1V=uZ^Px$jZ{QXQKpNZHpDfH3GWs1O=a3ZS5*Z=O_+jB?r= zC%E6Q|62H&n%9_GG-(%t+;}*gTU_ZEtf?U{Tx9=xNz0ay)A{ z@gE3NTCinq6)oYB@?`Hl6vdZ;_A0;K;k1f<;LDTn1ZPcYQLFw`t6TR3$YX-h*qQD` z0%IRdzt1#sQt>TPmNXmi{GGF*CR-O$%AkEZn|7h|zUBD!dRw55W7PSoMo8xiB?Yrn zFQaZ9In@@vMrv69_QbD%00000003%eDH_ffux@#J;FSPFOHAa%Q7u>*EVl8f`y`vw9lM6@Gz$q?!Z=7|Ag`T<~mfznDfmEIMjbe;Xi^jmiBu29TW zLi33t0ooG7u#&$cA=!^ygv9<=LbyIs?IA5}S)jpxaLR|}#+wW_&>NirduuL>U%V=- zL{QPZSl~Ay4~}o1{mz)X`u#c`sTE5r_9T&ge?kHx+rM=2z4O=bmBH;Mus6-yO{0KN z)t~#QB%*~CdEj|HKD~;yE`k#*3a58k706)UzM2 zp)7msdnp>JHvIh{3oGG`d~;d-xJNolq^DpNb2-whn-K*}ptatM{*NS9<4sEX-fvnb zDqr+n{|hhLIUmeK(~CtQEqpO9&WF%;p>whlgP$VoQ(xjNaFkk*_o@;vsq7v9htx~*X!u&Wch(3YMGL*mzNvt99Zav0u09z3 z;9*1T-y7qnG9olr{P9#_EFr_R+IPhpY{dJvgUh=~*@~AD1D?1500009TYvwwxZNw! zzOW6K0000Hj4CnX*WpkPJd-(O6W&l2BnjQXZBUpZ>iM#Vhv24vNip8J+T5PmDIuqc z-|<3yv)Qy}K58V#c1`iIrEfEFW>RU%dNxJVbCYPN5?`~&c*32A7B>OeFlS--xknZy zaXyxy{{$Jh+fncsl;e*M%amXPYefg4buW|K+!G|)ASJ#6(M5-@CsxeX9%I?UG40G! z#1Q;JZb!-?%3QNgwIqZCZ=?0L-#L;w37Nu*46xXG&0asY zS!3Y7^{w|noKi^hgO&@R?WE*IwoZN;kIN#4cbS$vW8(x<>@qS*{_PRGw2mLW&lpCRfOVuW#$QHUieRQroCT z=H!FfYJ}Dxa^e1g=8Vks$e%AI37Nf{#|zwMlb%?eZdu3zyBhxo2#_{WOr3}q!~h7T zeJlIqvu}nAu^o_AjKi%JJ0Dz37SqffJ$kGg$gf+E3R6O`j@b~Xp~OQ@A*t(L&EJER zP(dI%d>3H^&{`*$So-{8854i#)&cgoKiTgDMdwu`qBN?Ml~-r~=+~`%TMUVTGF-eY zl)I|D*h7|J06f9}|NsC0|L!hrleT@NV>L`l^Z2zP>k+ZJg}T9l3^OB;7_DvU(;S|? zk4y$^^wR@}@ko+d-=z9{xZs-c8W~m89?HuZkSBKm6x6P}-1G5+_)pngj`R=}z^z#a z{d>#xQN*Owyn~4>(&c*cXq+4SlWm3PA|>DsGaG zE9d2k@!9|Q3z+NM^%L1zRI8@4g%AOo`JIe&B5n4OwTy|zaqnbwAPC`Ju<(iv4X$|SUHU%gFISs2V0`G%W!uHCI%}Fc5IM!azH+utTBw#2{))k& z&-zLA&zskzn*&x-38JU5P%Sz*TcNsFFLnu=D+pO}k7q!;6*O>xESjm51t5qQFUUKx zS;~7*BAO(%N9AFm&t73VvPZ8PC9+ySTAeP0#2o1*{K<` zK3B#5L3i@(-m8oG31rN;fGInJ@lOtm+Y?^dM^>8S8}arhwypo)N^?X#7~b7qgOdaqXhuPBPgSu>>Tb;A|Ns8@d&6g<*g}B=ML>4P zx^*+%w#oWL)HZHO?R;H+$Ly|#6WeF6V}yGAK~VI~mQKlr&z3S4p`bn8xK`^04gSRuyPQ9(hxWCJ0N__G~0O)fxEl z56}-!%jX#vJJ%7Ta0yC!ihZp;`pAQxe+=^Ok$tM=Z4_TT{@w zE6>WJrz}B>k|-sYXX#(3;*sS)-jD(O;hflK7`rJUitL~8Y%k)Js=lc-Hn*lunA1Ss z&e%?6+jtarMdErcD2KA=76iV&aA{8DWQ|Gi{;E&#a{$Ny00MLe@W~nO+VmUHeeT^? zrRpX5Bhj%ZK~I?BP0^hCoikG)a}D$1$$IlC%wOPOXQ| zGY8I@evSXpD^?^@aMPCe?om(*tlXd`SY) zaQo6M#kYNiP7X_LebdI z>|dygM&4_0`}l$680Sy0UDmRf&|TCm?RlrTzY~<=d2GX?MR;1WmRUM`==XA;Lnb*} z`GEV?`5jl_#7Hf&U+;_xHPt1XPZP*CGneE3dZc>Vog%?`4?5va!J zJEo2bVW|~}YOB)&$f7}}zz0~s$QX!GKwM}3C*pByoR#)6=LXRSW>-$Rq{pt#};AH99x6vaPtG%a=J)YVFV&4ioFZFX+i4z6o{dq|@x=OxS} z$W{*>sPX5a`pEq7n#J6pdGioUqE;Ccz=U%sn;byje!<@7^5T=E%55dCPhVWaN%x!PbO!&adK#OI;tx zTZ|G1qx>6TnhS8tWxVM_uvVNe$Kx=T!AOqmi-0?aCHLf%?O2qTDC-<^pc!U9xc1W~ zzpD%|?K3}OBG=bX&Crh+G%J*utWN3~LPs>Q8&KX@CC}kFw&-POPt=h!c0*&`iN)3F z_#jm6MT~UxgNt&X=C$Q(TZEmNAd;T~-UK;|0rUDMDH;2th{)q~^~^9!b$>T)mR?$# zG^4#e$;~pe!|s})DmTxBY&2qG1)|+Xj~AmLCO(%^twQCse6`X}z!`B2tpu{D%@$@p z{Ky8ld08ck9~b}t0du_L^vwhH9_jX{`o4qDYW4DNm8XrG4iY74+=go+KIC(?HUoc% zfh?GAHOn z#l{Nv+y&sAN7LSEkDT|$7OfZg68MEJM=ECC%90rIml|78N3Yqb literal 0 HcmV?d00001 diff --git a/assets/images/data_science_11.jpg b/assets/images/data_science_11.jpg new file mode 100644 index 0000000000000000000000000000000000000000..3d6a2126871ad1cff0636657abbc7e33d9823825 GIT binary patch literal 109043 zcmXuJV{k4^(={5~wrwXnwr$(CZQHi9V<$VdamBW6p8I)EeLXe3*6Nz6>G?4~rn-TE zfbh&+y&R3)t;~V`v;U!;l{vGWmC^t7g`JtJ(f{KA8L_3Qt@Hm+fq)#XOkDpT|9^dSuXXMT-5(EeL9|c0Qa&a{K z9{~b_`ENoT{pXMztvs#%7XgKWg8C0ItsL$C-#EIEEZZH{NZ7C!;A~cTB~n(~u@zThzzNIpttMClU;WB)TQL4mt0R=a1-~ zP9%4X>!v*(1)VmpG!y~FC6v9))_U$r6RRlsa5>&1NugQmJ3+6EggorYNN3kbl?a!rtYRX*OA|^5#m&=S>=k`X=OcTwO8>$z{E@5 zpi0O>lvsBRc6Thm&`a%M)$mmH>DEk;c}$d#bwocjGIIf+`yfWcq8`D_Ou|^~WkrdJ zd`=}5{;!Hn9lN3jWKKM3T3iX{m?uaDF3yC_RV$Ddhz1||Vyp@GSj1?$2IsqYvd ziy>`(*Kz<@$6FbV@kTV;9@0^y&68)(CX zJ)PWs27KvVxQ{i2Dab0D$JKJxsnf!%WOYa=?!w!)5;{*+oIoAf5ffSi4>%YuT77mp z6Umcy9A{S4dNRkOF$Fco7>NtRChY@l{l%kJK<0&CJaiM~W;3sk8u#V={eqtn)A&on z`bP#}h-rO-H1+)2{$msglFC+do1h~maZk|}om8ZUW3Z)=S**?7ozwX$ zGuxH+R{;4RsT-IeBH#)u3@YXCbWMdSn0*^HIl?t^{250fqw zn~r}gmti0|>Qn05ww(qaO%}k%*J?lP&S^O^#Lv?J3jy%=D?j)Tw8$R3NRD#yh?e z-@puX30?@QFjE?m#Fv5e6$t_le$uzJ%^^L?3m5$G91_1^>KPSSPLvZ&TBJ|aN;1(S zw^%Uza+rM`u1_k!Ib3c_kvaFfz(9Cw=l3F>36gg$v9cC1eYLXTGe9=p6E?dHp(apf ziY}=khIya9STs^)%O*ci$R+v^0^ed1vaMH>4VE)jL5Tf^9tGRq&5}-}qY%2zldN*= z1=M-RuIL9zQTH(~n(qk_pXOXl0CwpeLdO}i(&2e{U0dA&q(RZD529=0SzYiO>P3`J z69uAIBQKKuL;-4&4t;&ty#GrkU$2zN@pPfDb*Co>5wJ|-DW_(8=1p}=TVmLgcDh(3skC|q_6p%7jk_Zt;nN&k!-N7&TPCEH zENssWVd7W*P?5UUvY?qK=YD_e0-L%@z9R3WA%#QcOv)4`-|kPpdo>_#mRhBfs$1Be`d7Bk_N7mv1i_A)X(v#9)Q)g!DD2^-1S4WK(QW*2 zRKFQzTY>e^s&$>{7+%v+&mlt~2=x4EEp19ME%R3(TH;>@omQcUKVyg>?1L*m0W2Pps0fXILe%V9|c$llQpLD~&yM zhVD4+(Vnc%?QjgCnyORhNw}A%Q|QBht`C+w4n8uWU-zCxE>M=Q%%iKGetv zY3LE{Lt==Uh1kieI|SiRBsno(F^o_0_)%9IjnorMfvdIYmVjT+w8}jzTaMYS&BHfO z;x=s&r~y82bU&9UQe(?n%UZ5P7n9EAmO=`6QYzsFH67JpfEkuV7y~-b&1zbJLe>#5 zI0z!YH+4&$EKH9&CBSxyBUSWXNihLwPy5bA3YLjG6K0kW28LmSMxm{twM0vkU8^$G z@tEMdSVJvKdVFLl{_j)mV>($jPD#LQBR!Mj>`9m&7ml&(v;7ve&+aA>3>brxSvn5H zC_^%t?(&B5R-j(Ul|54ZB#C|jHw~ea&N^FHt0?+o+4le@W8H{#S^bF zx4LOTK-Y&76P%g3=@HC(9rSCdRp%pet|xqsg&bPY&6}>zogoN<=wKmK%l2ka8%QnWAqMms!LCvj6kj?p^D(&cf~<=54~9Eu*69qdW=o?GU)ku%;3#oR5OVyM?e=K^AXmo_y)S4Hm3PNmwG+ZXOvQSeL zsk(vLGc`NJfmpT|YzXM-b?&cYsH(NMhgAY^K377)Q# zp%h4*>p8n@oad&KXUVxV-IgaV4Yc=Bv(6!ZJrG0Y#`hR4IGi^jMWzB0&iCGI1X$PH z0}C%Io+w!wpB-ME!v{+4MAomK!QM-9%92Si(vhbu8~pXZ)jrt6iBF)&o@ynO&U4u! z3HzTR^YN2k0lz<$;=TF_f1`deNam;VSb9@GeB#F!4q1-rPvBMgX>3>Qm*T=K7~|j} z0w;{9@Y*Jn&)e<{Iy_~6&D5w9tJnITi;ZZekn~0Lau0YWyoXt~7s*(ZUG|H6K#jRK z)*g=HYd{*47Gr{Jm+nExM+^Z~P!oWD|T6#n!%dDsqkH2J7&Vy^gv*X zkNy#pUAC|j|EiOmtV{PurcUo;1o{gt91A*gzoJKA0k%anG!-`~Lem^I(-S7TV7?Ty zE+D_$85YSMRL2j|T*mez@V?ZtCTz}05&N2~s0aeV$;Jei-a)MW56W-zvI>w6YH{LV zN?_o%XWYeYC9r|kNrW5bkZ;LSuLV1`f$cgOV@M+=zWTzR130EBgibpmFlMEX{iNS9wj*TWI0%+H1F@6wB6?{P49{w}_o9+!)j5o}=G&Za!1hzCd9v9Ms(59i-LtE+8E511C z2asn0A@T1W&!)uY^qNTneii6#qUlNTYWVn+<9}zyt7y@7nG1$scbL#^VDPkUIL7x`lc)(0nv9z?)5gf7abhOmzL5&i`70)!L9L zAiq(I1H0e4CmTTI0usT=bJDweD&rg3b)`>op(CyF-m(qD^^q|-el~n&)!<@!{pm-l zW}fbmz7CxAGeb~G9l)d(X>>PF(-`>PBW!ur^z+R(CtABS&KyZEt?S2$AI*NHsw*pF zu)KH!X=*f&PGz6F)3uj5(pJS>V$>-R5*8s4lp!KPrY}JRj|eG3wtlqX>Yv=i8ir2A zV#R5Cj)cKi9kK+kgH|6~`*a8dLNGru#dr?NzCn*ow%Oju*Il(`?(JqNe&EXmBp%`>+0O>m0JDoM%!cbS9qK%$s>1;N8JXsP{PL$;COX0|2FF%?0>IomTf=Rsq8Lq~2^GW_1CIDfR-+MO)127) zw*DCgJ85zVNGTNSjsPNJ%3r-{T}0ouHLFCRj*`SOX>yfW4j>JL9g73{D7CH^(oj$` z-Cz!BpECG03(!E+E|rB zmXd;F+46nt$L(I_o$0)#4(O~|CRU0gM(m>PDlvC~d!k_L)({)uDNA>Ka0;RV-k?Qt z8m61T0n)N4_~5gsV8G z_`!~seync>XNz}MilQ`m1(VDf63?LFs#L$(67Gbxmg91OYF;B%wLqAn4c0e9nSzK- zbVouI5mop@B{ta8oGNju+R^DQb>=lX@0t8KwQjfj1}!usuI@PA8M#F_$nUs>n!a^i z8RV8{HoQg&lOoB4hH;yvO{DkvZRBmm?)V>l0;TokjkfAM_B0QPaE=E~?vH>(F)DvQ z-PSILvfWp2Y>rzaJ=`2L)rQD|6-aoI78eY`dd+=Xv^KJh3bbyO)ODl-ngF0K7ex3L4N`*#KB_7JOW=Fr|tJw%>$-65~ z%N?#^9DShfFK=CvhQc@=d~U^1&d%Tu)*S`Ufe4751fE%lh}JRraY-4?ATK_Hx3D|d zG>`;Wg%i?@%gC;!6D2qEvq587L3>B%@#gJQLNUDZ0q34Jcz=BUYXJxD{a@@eXbV48>?7+{-X!1HIGd z;W($_)?QZn0u!07NGEf%T?Ct>Rok1}+Zq?_B_IeEHE8q)dbh3T8%1NGIHvMz923gDTq1dPr@?bnx*D4-Y&1Wr3!16m;202Mx>s# z{I1AyMU`mHA%4aKRw252Lf~siG#WROUJ>cyY!<`cz-Qo{Cf!2VwpSuzDd!YlD`vgV zq$;G;q2^Rf+2BYjs;Kfvl|>|?c|nIylqiN=nqZz!5Gnk8PHWIXl#~EjKMmB^gN{cW zNc}~*Y{-!AP{0QP*KSR8PD(w5K|`&-O*{k5PzXJ~Yet`5>L`ZmpBk|%gPU!1J49K= zwy_=uqaZG6)}R(b457_v6F1cZK$O{5G$XttMSSJFvUBIy{cuwGxQ?da3${uu(#+gm ze@|jHscBEIB1fY&O;nv@oZjF#JYDxon6qYMbP9#sva zqW{h@!7~-(S>LFa48O?tfdVLyb$HMEQn-Ik2lnG?C{vfSLEtaUood(e9m;$>oiLK& zxP8Xmv>Rx)a-0L|YpNQ^Ocu=Cauco64+ToZ7nq}du-mkB&ey~uOSo^|$`2VV&HfIM z`f#08=inQ;@{?GXP>OCcSkUg%gG>QCZ)sENCI}aKOuH=LyD)x;OLrB#{5qM7o->V1C{x)KA|bD&yh_|b zMm7^tr?I4d&BMgcw|wYp>NbEMtHD(YS?2O1slYHeMG$HE?1q&t;16hny0Uon1W<=+ zoTg+RzpbC>)R}nXDG_HRd+w&rJ)uHNb4_EhWU)6z*O;MekY`Kk!LL>fKdEs3+c0yD zjK^9%h5c1#H(jv-Cq~l&iz1IhHJxNKs4#{;F9|I_?-~d}0*qu5H)A{x>G&3s0IGXw zrYht(Q&70S8p^~AB$>;-x4bK>S5cbt8*ZCrw${YY6j_O#Zb~@T5Aq%%35%CckZ*w%(D8ab z;9G9`N_zM`brh8L=~41b`a{x{Xf$2K_C4Csr6Z3yiFb$Z!shHX-?3U|~$-btR9@eyi~lZ04~zJwp@ddH?r`DqWPEMxnfH1FU!1N?RT@ zdfB+p1}Vyb`fXfZJPj38BsE!U%m*wQG6vaH2vW3mdS(_>`KlUTs?FTfs_@sZj146F zBMj%4cl*|0P^Da6Z@p#Fwp>92CW%WKqxdG0sane#A9&^n;`@h}94@-(ry@+Q8(gzt zctb>FmXL7ZZ8d{h@FK8*4I21^X*F!7UhbF!(5{U{vX_cxFY>cb zhs_h#`KDynV@;!ZZcjd5qm7!w$-!+pDfHx&Mk4B7dzMBFmY8yIE&JEn$>*q}Z!si( zdAGor_+q@rG4C`=GcD!)Q09+nn$|~Tg)o0_A_cp8HeCJuTdsflZ~4v+<5{cjiZ7?- z5Xn;)PZr~;*qlu8rxR6t-u>xfyg7|z!I#{rCT>%0YI^&SMO-de%nN$;PFSEfvQVJ-kF6Vi@dW#Cgck?=-SxXM&uaX23Uh>1 za9sPA+4uY95;R?e73an5jLcOTldc{B5|k_JJT`eGuG)D&0Cfat;AA~`{g(}aov*Q> z6}(qeKE@e0vgMjFjeF3nC3O0eoL|mBjS=6|Eq^5Z=?@$oY$L3vh}S)aV_f=Jo&V>y zdFlBdPN&fjVzWn_Rju10Jvv+8WuAh6vXcQiZ0%ItO%uaNL;1yoQD$u#N*RtesGz#Y zb238~^t(7t?u-Wdht;TlRER40H*KuDEJ;7iM3gDWE(eyr@E<`3n@5tY-lwC4HRyGB zrb4+P{H{(S_vukqk5;FqJs?qGEZk5xck?8s>JBk(kM}e8%8LGjqRvKrKqP@6MOyF| zrXd*V0V18>$94COO_S9TFm_hWP^!`KIYsv5Y0xVHTToXi@jzZhWrL965h(pAl&`mM z&bVX4N}~7LZMjt)9mNp}Jq)Q$JIf84f)PFn$NnV5cv&Hq$+s=&RHStqvm{GmVEpnj z38@WfUVk3p3*o8Ti`qDg6mGodjJ6s2R_Ox0orp6fh_R%a_|Xn^3I0O=(-ZO%9QOMj z%cNYR$a($4bpq0Il`jy-SCSswW8gEfEN62LLv~G(rle6df9BfJGmEG=h{CD;la~l5 z!h<0wVoLcNk2uPikYFckl*j6#qY5XRxBOKotN&GI6qU-kXlGiGV!DUIt>E9m)#@Xg zZjk@$VH5hPq^zjY){|0r)H@?>u*cXdG?Q|B?dPIT5(zO3mM*=7*W=z&@1#xu^_2qR zyP`zUzteT3wN!YHBY3a<6R?E%w=9|LuRwI3g8@yvIHJ9#1OX`u^>n)97`@O!y)TXN zghc8WKQ&47hnHNLTTNlH0=0IJ0L|BZ%JaJ#Zpu%zkqR9IAq%>&|0l{!O?#RL$|Hm& zTc(WBMzUzgDs}oaPT)dvE|cE{WVwGWr7awg>CyJOp?ZovN{b=D+{Q5esOktpcS1Vm z298W5dO=d>_AO0np)}#V84=dyvEga8-uLQTv{f2134yDIOM=+RXm3eLY|B~a!NXs) zn`_X7W)wP$uBHYfw2pSFG#me9=6QX-2#cDLA%|~y(cWl_HHs6o#Xcm(yFM*EQN|2U z+C!Lgc7e8c&&O`ko$H(+IJ^F?!ib5Tf`#XWpvfW`!#~YXSTW@BCs2R1pH-(mljv{G z#z~z}RB`S%WSa2@tZ2w>zvh+(V5k|~$0ql;%Xsw>%wK9Z3khA^uz9S(qoY;p@NYR{ zt_Q=;o4F}&EL1CSy`N52?8HDgSO!rntDpCymXkaC`GiHrIeG!|!-?AJzwoO_2@RZt zQ3b3@92#&_G8<@$@s{0lOv{#sT1SMuEmaRoZvLd4{)Inkm|23C4)OTP@d2}I$vGbr zob^x+rRKOTwgZ%v^~&7r%PGnPXtHU@jkBAP>Y+tQeEd7YP=0gK?C-p4z;Wq4t`kJ# zwSY=lv4;CxZOSrBG(7v1Qh17@DYlsr0}av%ljqKo(dPw)&!{~A5Teota8NpQ{E}i2 zGuIKg&!uyx0b?yht*jfvDvFXB&N7n&;2rInF<%x8#$>?Tr;?-FI+OJKi-TzRU|k?t zo~}&bwAG!I%0LY$jau&u^0V)BoIeAZ;oI^1wNY{K%+sDkEcZ=X6_VU5`Z(A*-VL;A zw!(ZAG`#|^zB1=5dEp)gH0zld9YzWjGTJ-KR_?!tu|W!fCl7SXfab)Lm1kE{!8ooa z@231kPEOapEv-9*Sg$|pl3lVNb`Z+g4GDK^*|&`es$yirW{zlKJu!kughNw@(Jrt? z1jvCoQw@`q%U=#afz3_yJLS&5$gE@44I|nUIh>6%Sli}pZ{$X%h6>;=qxzd0?xiT+ zXagjO)ixmd!w;-=Z1y)XcofM5x)Gy3yYSWaf=sUL;n$BMC|>i0+E961=zUxs3Fl}y z=dXQ*6Jh@(^QmDOpbBzkewW8>jX{IQNUo>wzWU+MjMiBlzq}>B=VxSclJiU~VjwX* zz#AaXL}DRxQP(Eg&$k=KUqxjsNd6d(d5(6bk{u`ft|8gEgoLEEgg1sl5EU;5SEqkS zNTTPR1Z+{67l2GyyOB}A58w)3j9#Q2j0PUhVx(BM;r@azZHE`}AQ!meKiXEEm!#SN z!I0R0{8{k-H{@!!rk?m>*YRW0p2mMvwJ_>2!w-zmDu_m3Rh18u}B^Nj(2Zb>YcUen3R;?X)J)fw&oq#sfx1?$1qiFV!AoALD79{^pBF`0-^9c zKRNYz6!|Q$2CiSfr@2peKbW?Cp(u+#JYafW$lyn4tMLm;;ULAVzd!hK#PK)z+8GtJ zW#%%GSH?Q2{(L#izUyGmS>$T;uQz)C0fsdf;6XH@dHFQDv^?E1murh~~)8PiPIp%6-c!CPJZWoI8xhj2Q>nUphlYg$+w0xpvN54i)2E)sbLSA6(boR0x}dI)a0sFa!8xX zpJ}k?8XA=v?9O#>7ZGkcQlM6T^5;YFXMV-l*aptIA7Xq*6JJ&YexOO=JF8g-UB&1g z-@ym)YVx?N*vTGvA@VvrW@@0lZrm&<_?7UmS-0jJn~mN^TUR}g4$8PKLkULHeLz0N zsho;-5qN8DF-DvqNS)3aWz{xe8#Fgtw(Dd-Nj4YFtjJt3P7gZwkEcm zKm4u!)_F*JhymqRm%&brn3J@%xAy5)<8)f_1PgYbryEG<7qvu&MX3LV8WmQE4_q5Q z-!fo4ZtNRp!h-)~!~wUn3Elaq(1He5_N~hBqU;@IALBc)K(r zYET0pYuLhmQ z5px>=PC_2OM>WBsm6>UXGR=reDS>r=(;ve0pyWbCJ0gMUV2!TvC-W*5lhdL5d@z)4 z%bJf`C?}%8fU{LDm76W6#b<`Gl*QFG?#Yn$r>x}bqE~s&XbqS)QBxsk>Ypb?$`RC_ z)-b;kmuo3$G}gUrHw%@pjF$heZk>SDPSDZKr>vb+!%J_I$63lhCv&stnR^0EOcvco`1~CL&R|xrD;?2kjJ0yUq#Uk=>P3<$yfS zDXx63Zr?9&1~y=k_hvt-Enf|NyG64N7WabeMjf{|HQoarHjl6Mxj=O>AOAND4Y_x2p=I!G$ZoD&{TEk$Jh^ zP1!t57QOB5NbX4hV9bdke;CC4mPHuuevW%paXajk>h@6zf`wGysI;k6bazIOslz4* z->QXwwtU@Ny_Kh{?~nSNuz9(BljZz<^;L=sL9zTA*w!&1ZD0ii{q3t>&ukB&y#VKL zVT?{c5>&lTNn5o&MUbQ3#Gg_kR|k<--bKm-X>Ek#yH20?vf))-ATXG`_PnO~72%;$ z1{@T^RZ60N4!JWN$_=x9$=SFiC3Dmy7ryfSL@@=%SjSG1W4W!NcfA6QzRIy#dEomw z#^YZ`>NPrD(_Im>l>*bi$ZD<6#W=*#Wm12#P9`k96c zfXrUvhRoBS>`T*7rbo<+H#tdXw)C&^MSpvx zH+>QglV%#|NXy~ATN)=BZS{8CsYmHrXEQp;6*nNi4qtQL@5!UJh@lgrh zj-lR?i>6O@NUP5GHyZhHpt%;Z%J6sOj2Y~kPtgpu;mpz-YRy@7;z_f!0fJ{nmxJa% zTF0wHS_egag#CI~Xjx%VmYkjy1SdZkN;B4yc)QzC5L!}*w}wkX`zm5T-x%oO!sx6v zF*nRg_yfKKcaZxlvk-j7kvaT~-)lxK9zE7}`f(QftZq|y9PaY{t2R$!(@iDKlEcZ5 zp}akNNg7?U4*J45W&FqW0elqcpAb#0(;q&lS*hrmFo8j7<5s@$Dm8o?FN3ZS67d`6 z^N9eMiRi&*V>Z*_AD@h&cbM$O?Z^d$I-M<$zask6qW#eznfHQn-0Rw>e=dq0&@VoG z9*Y0`Ln51P!)kh4TXy5dm*MD7D{|+{6^m&C3rdBahz02dmz!_!378uI*WciJjrb}S zHKPO@=)z2qQ&gISOpN6aAt0~3)6$6YM^bR?gDPcU zgXyJWklXI9*<`A!{v!^_K_@YECGrRSmMb+NT zn!GJ9q{;!yZ8G>gL+_XKXTgyCSD>Fl8@6HId_srBxO&wXJFz0fl7zi6*E$ZBh=}M( zdkKo4P7q(&1M)7wut0kEi4e-1(@K}%3ajY9>~2(P=~j^ zh9iKQxyeyOInR#?2SAO!`ze#tge~ zL;&Hg>7E&}7J6Xo@Xn(u907O`;*!Vp&s%9BI;sVIOvfUpY!25d*Umh@=Xl*3og|iE zfLV)ZI!D+AtI1ki8D&puhm4=lXKB9=S-OH-o2Hw5@qdr|PQ4`SQr;U$Zp471Z$c=r z3wJ~Y7v6i58SBybD{xqLI*H9?D`V(l>Woj*`~^n3;UFxNe&gi)LSu0-@{wWpW{)^h znJ+0&8#?6naY0lc``V>r$O^{CH?S@l0i`apvCPaE>OD;Wt)F}$Bvg)Vfk70=VF7vVnA5To(?011pl z8RlEhBZBzVWLV{GZUafwngv1wBi&Yi>(X+I`?I1a+!xPgUx;f!=R_Xgts>-Qo1GPf8AK!VIsiMr8*h` zGv(B6Aj54Ku5&PM+5FC}fs@#QXxQOw4e~yCLZ;5gcB8b*Qp)yA3cm0Yn)XodI;J0Q*;k@rQDhW;oLmOrKIlONNscvr z3w}Utb?xneo}11euU@KiscY#b=X^ALjlIvRul6@66xx!4=N}~=Qg>VGmtD`QGpk^~ z+#oWCt%B+)2v^TQ{pcSxbzD|KV;<}%&;zmrMZVxQIV+xGJ^akkw}Al?I9l(UZDhed zfMjhf5sL2ne;pR2R4c$t>0t{dTZg}&j%9w)fJF=M)cj`~-gpp6huC|@va!5O3=Wi2 zMZYCG2e8&xp1Bbf7ec(b)WA{Z=(yOHiiUF2Q7uo_24ZwJhhU$G7%eQ{xGYVg@K!;z z)9XB%5^g)o_N^Z|r0~ickJTmQiVtq}Qsw3uvV}vJ^MN>NV(-SxO!j7D(R}CSPzWI1 z0IVVP3G@~m9P%n`)4mTMuI?G> zeK?9E`-wYA!}NIt2dr@wm&%Bx5B~IjRe=BU8iVR;UR_ER=TMrDs==NfFk|3j`|w=& zRu3Fv74d$58-SO-iEm}cVZ?^Zr@r9kb?mRs{g_5!r<}}Gs`@6W?uBH!Tgop1CRS)M zWoqHIc;g%BnzI-OCV+mov%OtmUjtk3h#2r`N`LO>prPk;qqlkoe;;;NZ*Z04?A$rY zP4#9J?Bbumzjbw_64C(`+JtgN{sc+Zr?5ObRi>HMO^(&`q3}}hy|HPmi+!K2epR?X zBrMhe;~-=$2C~xVFEGx$eJ+-0OVW8~k0NH#*InD_vK8O=F%VX|P64F|vCRR^G{hj- z^w5P|Lv+mz05$MFUo-n%`I$UnB73IcKaUz7i(1n}?k}2gwCXQg^J*nTMwAGqlrp9P z1UJ@ZkXrG=bP@r4Jj~6V4f+0L2!Hs5+QHQmb#AmE+14J*?gRhC9v=z1P=Pd|Si)U7 zP%1S?ejAfvdDd(P+@Tj;o_(2(WuC7W-xxrP1w@1k7*2l|Uux=~xJAfS5a}2=0ys+L z8H5;K_X4O2(rY84sBzlSE+`|9!Lh%>5gPEhZgPFAT#4p@lNG>GUsGzS;t zAMlBgwOg!97dLXL2BW|DN1a9JrCih4%$9r{yz6wM<1ALf>B3Sq@3LJK6ht&cP9gpO z+SyaGpt$}cz_Y>KMLed1%je!*W|w-R8@5%OvDJu4MSdAq=KU({y>gS( zUAbE5MfhFg+5f;AXvj)njETjA1Q-qZ~RrT#H5GDX81KWfBe zJ}<7LcUi~QHtIqpB?5K*ph-HK?KO)3Xon=$li9;K7 zF?`L^;Q?Tp+x625(IF~3Mz{txm+4$G(Q;$lt_H@PcEpP-vOWP1`<8=h^C|e$=|ksn zr;nz9o%KHF6TA6*zeTl8zgXC2cItnCIh%J6Hw&iHSJ{jUX&o>P>00 zHyrrZZODw5+9N%dhaM7}UX%jVlXxz-mvP4ma4hpPm@YA)_wM0kl*r0tp&2Uk5pJLV?~WnEJi5-8P#8l3#>JXOR43(ty>; zQ>{E4ra2c@$Q;eqlYGH_d<;pw7O8dkXA;$9N)ueR2gkoa*N=XR@$ZdmG%}>W50{0z zuI`>|H3<`rE*i)u(#NlApb5n;+!8E+;QRg&;r-EYWquF|#{mF0-*0S@ZwA0pg#@$Z zV&JNx47JN0aof^m(!gCEbKJxd661rhAnkh86<(f36mOooBjCU+|5ilf*Z(^u{b?8 z21$Atmj(99V52#yQ^0~xu{MS7e$|hMahh-iTdHeRa*(P0l_BZAIt(uc!Km$Gb~P?> zcnpFfdSb8VK{$mPYNe<>`4|D(OVu<6{gUM9pp4^cyp;I@(_M{-N#QwJqs#h{%n!k( z(hGTdrhf>iCRrUi75+{tLrH>~Y^dyd$)8a(z7F9jcE7?cTw=;J_WQI6m%&(Zo;Kqd zP^)71HjYup_w33g8S|~0_hi_uSZojR=^?Hwad&57xkCVKQV;I>e&B!>lg?oHw<5}s zOzkpR+Sb2Dg*SKlGMafMc)dHnuPHD*B|D|wfEMMAVBm7`bhqy{=f@Fo@c7)3kq7rFq zFTq|;u}J%~O}M|v zz_At+TvzZNsQnrV?>Jy1@S1W;ZnT{jZo*_en@>T9`f?pk;TI?f4<+^q{4GVOEZ>Pv ze5RbVmC4&C86ci!gAyv}?g>arJ8;NbPOme0XW{ZOP~w@uOnj+=I_@i)w$>T^U}CIN z2=gG!-%*n^|2ZfG(z8+$U9|<`a)!RBrnh5{G4&;Bhjw~fz`jzj&m$@fWJulTGeZ)Z z8}+;P^uQ zoLbkt9GBKB)O+BSK|<`Ql_Ti~M|8;{(jBmVG@!s~J|Hl{mE8B6lf3BA6;)m?Hb@@p z0KYro4rvgkaWDWRR1YNp+1r$caZyLmle*+GhgB8Ki_0)M@#zd8yJZC@`5qusnj_~P zg{zd7UUYyZ6!qNZp0eI$i?S%AP}gfvdXSGo*`18 ze45;_@}g^V?-{^L#4kJE5c(Dw3wJgqPKn-(rP(uP5!RNN^}y;Q9v;R){m5LJ#C(cQ zu!!_QnJl}PY)sfCceA4*hRF8N-KcV@>JE%Ia4LfJbr`wwdBg)yIydq?5**ZukL-w- zQ9>L0Dn|Yzi?paQ&317;$~v7Etzp*tS>>tCER5EdHkjLR(8;nllwfHAa43RX^|a}J ziUwpdVBoorU<%LMN^bZ^b9mtmgaVI?j5l zqw-hDgU?>#^te7fNH^*pHFkc$Y zy%Tx4nU0+wx3pd993&Nlz4_kudlkwer@&HhcSC#q1K+~leG9Jsz+x<7(-+UT%%!4> zw82{A(3x2HvL=IIqq@(*Lf3wx1r`5sqv*T4l`2-G(Q$i1pe4@j10&!hfyRqG=xwmO zEX1T`q1Q`=JY7hIR~JEh(sj!V_`%I8vsl}Dc?WYG2D=}`O@5Tc0pUDI8tE`ouW6VU zb8M;PBAIS#aN-nz*A2DbEZjAuTCM*|)+gedK77bh*zzn#e|(37aHFD4qs&>K?L?yw z%W{kr?M#FGqlJi2QGnxVXl|_avmg4->U!A=O0v zc`qa0;VGal`&fu(FLbUUbwxscR`T)g=gfZ&S-tf4wA%)mXN^W?)zKZju70hgV=#69 zdX`VsLVCErSUa8V^5-c?{DWmWbKl#k<9My5C}U(WDUT$B^i}Wv?src!&#WuKIh5;` z#z=a>b3njX%QjgJlI54-m-!olo=_cl7?br{8@IIWb6Nyr!W-+;35Tg!+j}BY zi`f@e{mq`kukk{Mk_#8LY!t;m`blAkbh%w3DRH#HQZ%$yX~VPgVC1Wr)tOWaS-`x+ zc-R5gAPz)UG(+ih0Lt+BOZdZ|)(+OsG0Ak$B5U`|q;w@&T70 ziy%dX!%Z{~?x9_{XO`ly%$!&UXU!H`fIgvDIloxEg68P~)298??Kf*sd1@M$MSD;v z6^7#@*XKCcZmzrn(P^);Q7fv0Sgt5L=Ee57F4i~Fr3F%`j_qf998#PuAi$XKV0S=O zuj=y2eu~oya0Cdi#0Nw{lp;)@D$|tj7mo%ss*Ceh+OTERd@y->Sf1anN9l4 z@48&QbnuEG&JW=i|M={+k{J92n|z4M`HOM!@G6F}QR-qgdnqv2UPoV-IUFVi5$vN_Y@u zE#v*Q?vk`xt33^bez`Ac9?M2RQsuP`+v&{3hma1O6kGC-)>qdGL^U%y`t6|Knb;6j zy|?ZL>FhvhcU@!}?%%03NoHzV3@{bx$N2?Fv|v%^g14@J?b+$tqA1U7 zR{w(X(uz-`1a572T-aL1CqD^MWpDVhb_rbhT(gooQ=ZXR)GdBUK%7=k(0BJaRC;|?7jppAedXx*PmQ6$L(MRO zL2wCH?&~pEug4wmY97>#v)89{T3$A{-S^2TN@|NJdnf70K$woSs$w!Vt$9)REW3&) z|Icbj(=`54=>%vW4XY2zG+73yphiiOCk>48fY9dKB{(}{XR9&;CrsbK)6SD=mNHw>`f5%OhO~MeQCZnItl&uMCP|P^46%cS0QwISPs~o zyo*Mxk|1fBKZ*Db&9H060XoY{gk;=!P#Q$$UoSDG-}G>-K59o3xbr?W)I3`FdK1Emk37%j8|- z#3D#y&UT0~108~!xoVT>>E*9(={Ks=g_{EMW#J0PTR_0rN>sZmRY073_U=B|Fm&&) ze9kEs(BnADhJ-wwpHP|-FQ@k=+3;5iro2jZ^yLm(atF@yEJZRlh9ldnQwlLq)G1<# zyQ!R)+OvPKmxUG?^}6j*ckPBx_~`(L6yK3=3p5$^6I8&!Gd*gVZ{dm6B_nIvqV!M01;x!ck3`T;Gn%pmv(w_6T^|)pH?iP2IqCWj0E+1aNqj12A|vo+}<9`5HEQ z;WDOp&z)`%ka4ZM7Y98e^JkFHP0zlb+08iD<>_}~kC|&rC~PYYreigA zu``Q%t%r`V1B|H?npB}_*c&4GiHCV!iN2DQu#B~Uvw)3dn`%nf+A=6reJ|WWPx@!L zEf-`^K+v(}JedBUr>zB04x9{7a`QD4&msGbjk_gnoEV6e$zII^prqP4 z_7gKU=xqsBb}*)vNlCibv@-*(h-uqAc+MSGtM9EC0e%>^{A4cC`W!-DK6$h;MW-Iw&7=Sf(ax3iLTUs7GKw!||U zuD>6O4D~pF$8ug*tC%tP@RbQOYYCGYzClilffbI-j zrD>rUmMjkP55pUwA|{Iq@(ZcYUOqPy=?#iT!Fhot$lEuK$vQH>O$_VYgo}e`O~`O5 z4m`m9PtOuM)PYB9IwcnA09^$8Axa5jGYnSh=mwq(q#U%%Mz;BZu<1_I;fInA98;0O zj+{d3Ge`lq4^dz@4_+`vdEiv-p-VoLXF3YToKSDWg0fg62tg)vtPwrm!+0Ul#lE@Z z?J78FJTF^z@h-tnD$Haz+}jM7_tOd;0U36J$J>PZMBL+WuI=*fXv|CEA9mo_ z8P#j73bGmO7?MsFsL=MoN0>OtNtz>JTbjQk+ifYO#C46s(LS=AX+#(kh|hA$7RX?PgFaL@w9|c`SD%q1wcf5&V?!v}ZK#PKCCex`drd@WC!l>DISiEuOCch-(r0-xi6JO`LOZoCx5-+DpCr>gX`g?7()5iE zp_tQ7ZUrtP2eSm6!JE+>Bj4&TsdUb238 z+{+np-T)!*>~L@W181mpXsc&mo@poW`bJlH6pwD3g#H@2TQp%UEu_qXX^nntcNDgJ zvQZg^s&QXGT5y_BK`%8p`IfzEKtsx>|G`Fjq6~NbQUem5kp;-~ z6ips?sEyG3!EJmul^FzS^GO?Y9wG;kO@@uM#Tl3?c1Lp_Q^}VOhSL*}XReJeJJqUh z<6fwQ?W5G8c@ytKz*~SA7ej??j=80X@linb;v#6GD1L4ew1~!DXI;(E2x=7Iod6D4 z-%_P}#id-xYvSch2lYn&Lh^gXSTi(_BAphFe5iy8sZ)_Xs);R__#OB)Eo{iT9;$=}wl zv;L z1x$=JS`?f$SyCf3)wH!UfUkfovvZ{G{3EnO7qX`dUJ;ldPW{fhU+3Mg&tqt3s& zD~LW9)EaY1{u77-2MwoS2HsmTPuK5^HcQjq>)_GI@Sj{H?<$yaAsD{@TOTBjdOE9# zzSL6LoqjFjqY9I+henokJD%Wu0tKtEViii+#D5+?%#NNd$#;Dq>0IDfNcaDOz^9g$DELE*O|8n-%} zqf}KO@zYSaqPE9UwM>_A+I9vK3tONCRY|!_RI;Fu4%0eNN(6A70EkC%!Q+$a8(b6n7AvP=osz*N7P7FK{uO^MwrZ|{ zX;W38sU1H>jV$gkQdQ{yh2uF^jk7WC+$?Fg_jT8Mp4a84234DG4gP<1t|jyL2OrGY zUv0mfqgpxye=T#uxYu?i^~RVcYD4SnhY{cul0+{0<=7dz|Hh$t47OqxhBPC!G%#<@ z_M?}6(X+(Rn@)>+_$1=%<}TweqQKD<9Hv@~UVJJe)w>dk~f( z0v5Tyk3#6e!Xmqn~ggw?z~N)9XqaJrP3v z36%=GM24X~{!xW^en9^N%<@-b(BWQ-3XsT0=zzX*eWz49g8(;7!lt+|B8%3mr4$jk zgIDq}^OvIO5{_C(l)w5%9DRdH8pYoi0W%OJ37psFHPjNO1pm$oJWcrjISg+?3HD%q zPDuFT$;tZ^oKz`s4g6kfPD-H^qJ59ejWC3Jw^e1{mfakOWfBc|)i5Drd+`8;Bv>Z9 zb4~*~rNi-^E&aGVXkRTa94-nf);0bM^8M%{z3`64^|{N~^2joQqFdn1q0f8(*BaSn zAlZ1#12hd`Iklc^k5O8pMfJtAuJ*^om@Y5(GU8TAdtJaHVL$1Fj*V8A^QcMOsAb;a zFI+E^c1r$QI|r5~<;;b8w46!A_}GN5uj`93ib`9s34QE?&=<}-u9NP=h@>41J#0~c zA-%g-WIMQAttvjR^^p3YcW;0cA8W;=3_X58)38kQZ zM!AY4i`Nw2v25(>0KIIAh3?7p|48;aCn7{9r#|V0kaqC-8~@jkO$2WffI5scvSrHM zh~)PA)mla&;E2cbJNQyp56^xf;wyyL1LDsV#)mmz>`Oqr+?uLEZ(<1_QmW97i?$)< z*r@(&_AD;zeOm)~?E}AL-R57|7&(P02u*8DjntJnZ14hcvlheB095vambrN&H%~5HUN~7?1q)CjmwWa7Fk2P|T#_y4u-g zdG!u417P068+}Sm6*+#Q>VqPcbKzOah1Z^78KyDII*1KQLLHhbR(o**CE2uNpv%<4 zVUR|&Tf>XmCH#=)UJqo1K>e?q&2JG0- zJ*4*a0vthrCs4xGm3SdM8`Qt1jG_F*pg*1m0eeZ<3(n*6aL;_rffE8UPrtQB8OvLp zj!JoF81F>NNky%8pLmCQp#ZfSpib}oY4l=J-zs6rXxj^vc$%+gT6`A-UobU@K{`|2 z4+fLns%YUSaG z&K2I0B}tKeJGxx;XXm5)=MXD`5id6cLIZvN;OG`>y3yLS&>zTCC-E=jsUSY$EYUoW z`fNKI+z00OJx0Y612R!)e6@J;wzS$23d5+)Ez2}IriWbU)m&9!n(`B3ycyUQ2(x1= zh5IK;yfma3WJeURr%*IJJ>R*&=!dH?q(qHxx^ibe(&?~Qz%i2;=(sONJSDwBgZ_+^ zuvaquJgeHwl$d*?)m6d_RaOB#MMo1n0t82CS9v8Q4+j;TbphI?k^~hYz5RrV>60iX z`4gCBpZLduJt`x;QBCASkPBEN`=gUVFTV_8QDBjct_kiaBNz#JGhuL#-#sX5?0nSy znugNrem-4TGKt0)#`$y&h4_Fd_OM;Z-rOSP@9^EPbbJ=LCxGRIcX<_nAFWva@q+L8ZM;0WuQ=(xUBAoRYb#y zcMqynbXGH>(JS#={E!iysS8r^1O1S@yP9-NNSC1R)SO~*-(d2h_RA-6_psT9-}&WX z0071q>ve7>9vN}n7bhtKbG_a2Bup=eC=QL!+zO9P4Fnuh9V#x#m8nSnAKPX}*&+{~ z(ghbNxWZvTxY^5lMvzsD8?^8+1!e|?9eP??mEfNw<%B$sEOiyi$5jup`2a%%SkT@vYj35 z5p3Fr#9bfvH|A#G)+<~^6@@k-A}#ifi(tqJ6`9&O2pZk((j>M>=(sHYj2l}1=hn~x zejq*n5jGX7z0+n|rz%&)4YHELj1r1E=Ake%W3OOcO_R)$;okzcYXul2oUxmgN-9`X zY3Ef*#Td*ygU$`s2|+Qy$G7w6Wv!D4n}hN9;MLmiTddG^#Hsnrk3(!-yw|k~&@PdV zD)v#0!E(UyV|-}_F9T@EW{>xx-!CD?_c^g=KKf;bk{Xtedx5WECcy4sntI;kWr;iY z{R1UU?JauDi=co^qK#lL|_qnNZ#!nsq$g%HR_y&_&Duv4?%Q? zx|z1dSh*R|2Ey5Xua-IB`3KyL2yRqjqJ{I+687hCif^`dnd8z{RrD470Q_P;pZ%?M zQ?x6H^^Tc77G&cPX>JOguw)D|=&cjCUt843$7l=piqSMB>{dg1%r8s2;AvU4m@+K1RY%|9DRwLlG?SGrf5B@i72f;@VBD8No<| zAu-wbQQ;sVnGLGG(p89f@T|^GHDwKW)C-LHU%$;|C2{Ol3awMKp(q#kd;ww4H+@=^ zZJVC2-%=47LN!+9Nvj75{OWBz?nu0IqSD(sec&x4IlqrqqGxpDTc1Z@aM3MJ2f6<` zY|?tcmn7TIX0g@$6)+K(apLT$i672hk3&Ag;>ryg+dHO% zwSx&KiUnaYq>d)c<*%5UJaS{qvY!i>l~%&DX11KNJH41W!P)X{+y z_c4=80;o}7>B8Fcc8cZ!>MJjbu|_YcwTNJA@+9(^SIy5z?i^xjTD}7sG3yG+k!tg; zD&;=ms-pMjSrQ^-pPx3=d|Tygs&Dg8SY+lp6*O6#lVl`VjbesHyN+Hk;APP3MzGy- zeK2VY<|-*c>m@MBlcJhM#;Mgm;WtwWwb&T@!4mJ1-81-^kOqskvvY7BbF!AzjzRc| z98DtA&`%l9$W>2!=sAZdkbT{JUdZf-M{YE}1|I}{6Dl1sXqy@z2fL2W}Q zU~o3Zj!y-2Sd?f9j<3`6T3kgOFmQ!A%Hif2Qm(&8dyty8aoUNQ{~-VDy`-*aily>s zB*c7Hk&vAZz?QY=V#LnLgp|`!fys-xBa&L?YaXmP!jr_1sEw0th@$sECb6W4K0^Ze zMIh`s)3W#t(jT>5Qd!SSNOfbE#h>1gT|;G;+X;v&Es*fxRJLD(V`Fp{m|j@EG=U7F z1&@}144HtIQz>WX`x}n8!pqU(Mlav&5k}VGza%6K&n?b%t3?C3V>@l*gDE}X3rOPD z1p7F@kARbWXwozl+M&k-RIh{Z=;i`Z(ikfLaeet&}=q!BD&x0?>0{323kDOp` zVqnzKXAT?>LN??h_-;Z|pQW7RorjiT1bXV)r>TphPq*B#(xPF~M4-3q?@IX|LKgo} zH_*qI2sF}?5p?&!rDGl z#K1T>4OQ-H;oZ?Q8W+6*_xWj@I@HTVk0f4LavW-9WbTio&ZpkATTLS&g4DiwIMloHKO0d$6 z$Z_U2Ry_7}Zjm)d9}uD}Xv}%#M2z%42)Q?F6X<7;TO~$l!~R+iHDt(9YGLj<|L}J^ zIh_UKZH(1y+~8rfIVT;xU;hFGrVRfX6@W8ZSPXO}QCWaAaCmkd6cwrn>#PkZSpNw1 zIK*eTd-7*Qe+sQZKJCvuQGUa=oLRg;#-2EvSUV&NTfIxrYaiPGagZFgd!jq6bGz+gDOhj2%9OCcZlnO7Dp9W5d$Dr-p@nCKuH1d8F~p-#4cWV(0}dKh zrP6y&A?9{5hRGMM$5t3Su@_Mid#Z@7UmaN)#A$JVOaxxm;RvkD9cnxntsn*}8&C0& zEwrC8xDmG2O6;{`0ZM+_;=3M0-Rz^Swh9A6AtYVYJ4pL8a0x&uf?w*gwHqbAPlw67 z8R>gCPoRAZSm90TkvFJy1a!|p=oAzkP+wu#tEzb zXq2tyiwH?rq8j@;NSP+-)O9^Vf*12$4D#*^CnQ)ur0Is*N3>;O8}wGlpBs&g?%Nus zmbgx#rn!W-0uU1@!e*zKs82%bcgVIkhzKlzEOcntDV;{lOP+0lprslK?n$fBOYZ^j zV`v=XRQYm)a;=5ypPlzguy&3GlRB&XRB)VEubJouKNTDWZSU%(?kRXeUb?&)mcYE&+X3v#S>9`SMg>$E7(s~E4v?#6$hRhje4 z3UO!^zH*2UP>U^J$UF=O2qxn~*@R_l*x3ok;P8(J{pk5EZ!}%DYDy{LhFt_KzRbFg zxkvqz{xq50v*k~CX_>3Eb?`kX1E>XM5L^(_jD3CkvJ)yp4#?L~dj>X8qZN1(er)@H zM&1q_pRdKS2!_WSob|Ni7yMR?0EmIC5RWsP6QRhLq$++cDp>v(fSsE%DtM(}WrTQi zmi?Ih-e_q0i#F*zV^md06=j`Ty5&GRIx^)1Ovt=MLuE`P zP1WMR*xX8|< zV6R-Bp?HZtysDR$!WSuW^+{c07J5G8*U&0S@ao2rw3K3t*?Jon)qk)92=tjwPn7)* zQ3NZQwv=x~0Nv|N)B_C#tx z*h`}2-#B>D9=Vu~>48ZAP>kJ8S%qle$lMh0SNS1$9hRiZcQI?E@)DIw`5g_RMpDyr zH5~o^auD(1-{q$^ApbZyHMFdFS_`Qg`C*5&?g1 z&3Vs#k*-IuV(_SYr-oJW>Yl#7s=PYD^^t-;mUf)%Q4+`# zYVzJ+V;u@O_ao%(n=*a_&s#+E6fZXZ1k>sagS?qER9=>cfDZGzLq{e7TfRCdh z3ZV+-ews>hHE&nctE@darcNt}{sZ%)%ieKNo%r%}YBI1T=-r@u@~?ayED{-NOvOXG zdWcgxZo3hqb|PH}H=^!+3zW-fDygY5UnMGg_Ri!Y-VOc3q}XY9s3oK|u&%8hMWJQD z7uCCN+~5#lb|-`m^|v3`#ykjTHuZJU%4%vGuJpN4#vVR*&*WvR%4(uwf~ngO+e)xo zb4NzST_GiU@p%5GWZ#iaetOMBY)_(9h&^_J20L*nL{#|0Ha8h%8N%m)Y--zGs9T+P z)&fD(1FX8x*;;d_Lyy+Y~ zdyB7Emj+#O`?i>Ld!N@*!vj;8pda|oo$A^yN%FIUudZn+v>=?E1sPPaBR-D6fxGZv zc`6U=`KPiFMfWiNCkIwXu_A2clZdl`7Qe)FJ-Q^N$%t)ZPcTJGwXq$d7PN%*>-nFR z?q}S5Els;T7;NEsH=!2g@&|gV1>n|_VqeGAM7*HUh2IW>_)visdF0c-CI|=l0M(Xx ztlcf5{f;Ics3y=ySL=D;IwOKA{e7^w-lp*7n8i1VGhW^8PrtIA*EHL6is;S@y~36> zD9S5Q{qIY`vb&z>``D7E8cun61zi^J!Tkmji;kicwSarty7?|^y!AjQ=ct>91UcmE zlKA5z-{Vp4Wj?Re9sQ!Y$R4DYmLe{0RyR;!#;)5jtfN?Uk!Q>_`X`=HA>&{=n|$?r zR)X4F)h3TGT!OxSw2aF{iOdEgKn(9T#uoyG(^#1P`jgID-JLsc5|AGxd2MOM^0;Ne zw7|{iM?U%r?_a&|!`CXb2D*{2zyPc&4fS(;aFRH`kk>=#uqjtFF>85d*_JszQ)nCw z^Nf8MYm>MKo<{8-@su{Dti-lJ#CiCOJr@nr^v-f4aeSQb$Z*Oh&SGY z`fBN(G6|D5eK1C}Hd1&s@R-f(ekkOpqFp?S{Kkj@3;w9%{&3>5z{LngE#)&`kZQ?l z^D^wp8~3%Uap!OC#jWS|K_r(Y8PzN`XoqCS+z#C1hhu)#y3JRXtr|d{fU0b_+Arr+t@FU)eIYpfZys0 z?@H>)%#b8lEj=L59O%!H+S*+Y?hC;JU@Z0$3`1Xxb|`Y-K!eN#F>V+~s_9J39SBv8 z*3$Vz{MzI032q(X3JtU5DCbj`UtWFcTaeG%Jj#)pfA0ovZqd|K{eR$n5y)=UbE{z< zC@H_OQ^ab92^eo6P#*-moNQJ$mq&obgSCaswp@2n)qI6T#cs2od@RO^>?rDGS{+ou zb)yF`6qW))80Q}>^z&KvLU^ew+Q2Cy;O-a=9(n6j0EBi^RfC~1@gn9&9G5IR(lkX$ z$hkyMo#jjfUDyMCO!}9sm%*kp0{kaLa4&SuN+||rSs>d(CY2q?0}$qj(T{H>Pk@-VIMNNvLwyVttP z{f)tqI-EmAh~n}5Q%0M0ZpHINJPYl|KoSopMWQ}41^r0MMC+rviz6ZbHwrmez9{z} z92lJTBQ&QBl(bN}eHHAYZt8keih7$(dj%j6)muBU<~LCcmn=k7)Vm*Jlx{lC#Qfdm z!Wu37J3Rv3dC`C^qUBp&iK{F02 zVLX%%D#0Qi5Gf$QgkZ0q$2R9;Hy%NB9dJ7T8febJraJ(Hr}xfd&t~!;ym(6IYqB|B zp2T1%B%W~)Gl(3n3vjL?0Ao&Ydv_&V*64s|6NI>>x{Ck8+;L~=0&V_ofyKS4QLX4C zT*ya&GS0p=4xu3AVH-4*r3JTb=ZT{iiE)xvM~@KPL|-1uQW->VG&+qr)`lt9C?q9o zLy6J{;j&x~kPc|R8}@v~9GEsXU>X;Fc zeFU~0?Ag{fSTQ+(q?t&muK?}XRgoXS@rZgw;U*`v2h6fW6OAiJ9%kbxYSykrbC9%} zmv$Q9679KFj(=|0<*8^Onk67F$faKhnkieCm@!O0-8D?$O-fq)457yRfUJYC) zaYE!gEw#Z?2YwI#W-%v-woCUSjDr#H)4~(n>8hGxm<+&ZBwng`9#%0|mm&jpZu+|3 zJj%A^r93Uph@^1X8;3HbVrZ@JAdGSj<z+6EHHT{#o2cO4=r(Kn!fQ8P% zd+d5{6s!pJ1w;=!!4XIWjG31|FDXpH0wu_Emv?zPYl3H)?l7Bit+@Dp5I)7W`O!PJ zY|FD64d}qQJhLW3d^0jQS{QjuFRW zSWJL+zVj`#ACehQitE#$Ta4h;X7~ehUsh{y61WS<@ye4P=<>m#sgduPM5LXLBB7fnnmm|G!~YH((Z_Sc(9qOUir4+rXtGFFCc>6RKsay-z5;;`$B zFrAdKRHo=z#B@65BH1G1!Z4%NRq1{9Kuo|p02*FylV$45wHiM@z-gy3{BGUGM}4LE z+gF1oAG)Xt{|o|IG)EVqx;xrUP>#&Fk{rNd(+Is&M=Euq%pE7Obn02?X1`mkb5eQ_ z{mH7*rG);4k>}^6?ogk#JXRdZgUi)CAv!oQ$)a6~nt>5gvEN6@#Owfe*18GQytXu!aH!|H6Y@e!=fvT6~=vmaD<-$WJsLKY5uy(CX9pILE5Zl!N_yqr>ar1fw+7 zLTv`7S}w;Lp#0_;wPM9yAI&Y;d(^dp{*DCtENG!1F1G(Fyv80|JZxPKYooU$#xM*; zj$4rV(@-yQaQlvmmnLsrZC)u@P1ol3&4R8xL_1B#)$x2YpV%pP^%Xme3Ro1|Ybofs zJofsUYQVJ~zNjxkOoUlDj1kwqIkB_afP=8UKMc;EVR_9hqY-}x4=h+%%eYM`bCBvW z>Ka-4p=y6vo%@LaIE2A?Igx&Hxcy|@NL(`28PZ)g9-%N(w4-WF1y-afeWX>+@k$X^ zu$W|F;sJj3QI7AMeR((7QHS)?B4?kOpT}JZ}hdWkF;5#PBJ2UwN>`Qq@M!zlA&N@uM}w9)XnRd@t{gj z{rMTUC|MP8Tve&q9g18b_2bMp$ zvC*&1#rZEbSp2ng#7Co~OHqU2ua>0N9@wxFYTQL}E-t~n$?9+J;|s#6owdf{VZDEX zaeBV|mdElEhc;f60o57v&kEvp!oFcb%>ml$qD3&`E%XACXo4fPQ$cL>W=_}00e0XmIW1Q)?5 zj{a#_28vREm#9cgZdY+M_A#BLzyt_rXEjVCD^Epluy&wR2nH$FQS4CoS%yD%41DjJ zmr)ZoH&Xl~SpBhtH`)?o=i!0TI#Bm7d;L$S-4(#8Sc6C7XMQM&d4bn^&m&%Q%5UoA zX9y;V+pebQQLQciw3C{-igWlGEa>dw$lta4+C&wGmI=$Qt!NMwYS&!m=DnSd?b#y# zLRwSWNoq$Gte%ge20crP@KjvcilFMn1dM~5o&da0>)5aqB(urmsbQf^w8M6l-K=>< zCKP7R2w9l`ijWS`ef|CGCO7aVKgv-luwix-#7YmKy0KU}FWgXpMN_3_BZams|4BHZhPhWo6QkCEhFiiT*87O{heHc&V7D?Lu=7<=01hTF z>a3lXB)}jqr$zlwL9};@?xT8f03c>&c|)%0F8rR?p8+k{t2!cXhu_l_6{}SGoJ*}KF0~gR1r{|}gw*af z|JTAlDG=aign&dr_?eIebLGcB;H7zk(e6{nMveED!RQ+=o5C^7!VFLjK~a{K2x(x> zC_UYwq|z0FRZypHHRxAcjU09mN@J68wEmeZKOjgspGJ3L+-9h@R9l?07tA6ln4bz1 zJ0@>3^S0VCTO}>iYLJIdhinB3|94|G=>la{;Pt1V3nAT~#E#Irg8RSZ)RI$)60{Ib zmXD!Qyiv*F6-DS>ZnINSF?msEFQ*_6!iqGKI|yse{>99}OGW#PpHe0n!fnv!>V@?M zn>t&YdOWi)Nh`;)!i5qS*K z@?PYFDMk^go86jf!*Il3^_V3%`JZxSB+Fd+`f)bAEBcbMR>R!1qd3C)ma#hxVeWpfl< zF3ZIr?jN(a{f1&M^`*aVpnsMiFV*r@?Rfz~XP>qpo+n+XLV?QQTZ|!mG7CYU!b3|> zS{8wAq={{)e?y;U-oye2Pj^Xyg5Rnt=|VAe)BLx%BNXNvUsbAmW#@Y*(h7}+6@@Ikua*YA2u zG~W?N*GqYHL#jkKzq$>pG=i2=>pZeVY<8l%j%nnM{_76begui(d;UZV#bBy@f-Mk}8MM5e*FZHHKS)3)s16Cj5TRRYmLd=+yK{|*HaS-k1dY~8A-G<9{^WZ zwr-|YLK#!6g;=VHM==5&rtfS7OF>y*reK*!>=GgjAj=$4P%M35NsA75A{6VS^5LPb zJhq0fea_gL7rs3v10$WhmusV6kHN1@a5qh-^mq^4S!f-{TB7y(yq=Vl_4ppwL0(8g z3QspMJfLb(sHO7bAN<*GUz9Z5JuKsju_qFThWhNe9?G*P+vX>dw`k+{m%cO=fr-aZ z58wzK48%PN7uEGX5O~C2=1+N(FPz4RlW(TX$J$9S$#fg){JvdcGi4HeuqqNtO>4Ef zXGa>Cn}=+TIvtq7Hd!6Yb?)bay8EXCQFaD&A#}ubejSzFflU$Z>|Cb9b&6KdqN<4* z*lEWrW%#DM+}+&_UxLFz!F*Bim%u|Km%3!UcqHQZ!7ZN*^~G#j8lXkn4PYZ>T_Nd% zM3h?gln8yP0z@GKdE`R4CW6R^kby3816oTh=_aQit19O{WfY0cp+Fd;~$VC3+ zfz|q&vqtaN%4eHFCoJ~zz&;|C=kvRXHczHFYOS1zZD*l7oMGY|l)5d)@qGZ>8=IAv z6aU_sJ|#(_Ff|!dY?Mmp$5I%bh<}TaV#vA+$Q45A%o@Max>Bofjh~%nS^1Kia%H>1 z)u6Cb2!?;nZsb!${% z8@@MGOn|CwT$x%3bZIl&h(2KFbJ4ku0OYWvrYzps{sG|4cz8~Y`!`Lqg7P?y(Urf; zEYs5}O$s8^ltj^P6d{}revmAaRG@1ma*n?v2e1=8BiMOsmx^7!aAtPXLAF45iU^O7 z{Z}pjs33SZZBxLERqK`NR!tVXr*8+mthO*dTDDJ{IJ&a%@6GSp?N)|!g*P5=)}Z8_ zC+%$R4{{Y_Q+n`aqBX1cE4a^F&)nU02PKeCfD}XLG~J}OxiQ>fvK3YsQriU&}wVJ%O{dtdxH;A!uV}soW#M#92 zf=i@g(9=1Xmgo;lb8PV8ah3dwc(qVu8G3;JX=+>>xV+XPQ9bFmNLiIJZ%Q$Hfyyk4 z&_J^qZvR9h(*&>UL&_;MxR4#hq6^Wd9bJcm4!W&Y4BRqredl8r%xvJ65Z(wz$9%rC z7w%0TbzaF9mMnJqZ-!m&16Csud1{ys*D2F-;-+K-%@UngZp!q);7^2mjPv7YAIDAD z4x?A8m&D=0xU0XZ0MysK(l#=uFSqhsc2$SRaAB0KfUk}NPMjj+-f<2_fPfv}Kc!&m zGB1{aRWPfLQ)OgjKgl`J_%BOj6bC2u9fD7SSbZ+?ry)}*(GcI+KKc?%?oh;YsO=kt zKmEn}*3QA)->_eW;^nG|S)P3!#$2}av2U2a4qeaVmqbtG!F&Z8YSi@Gi0vxRbCeQ- z#3Y4eQ+X4R9IU7S03dIT5ShQO@F(W{L3Zac>v~x)c460ea~Ps(wZG_>S7mgSD?+B* zZs&wF_a!5lbHs;fi0@52L#Rzqd{;~&0kqhXv)7HoUofaL;a+W^cY(USU@G)B0gYZ< zrzDAf#|ofo(RZH*%W~RMFLce>5rK8bLfhfo7>;GCx7wbEzi509w?)w@W>phN@V zyKCb?*we86-UIwK)fRRUDgzUd7+aN}z3c0W6d&Ji}({wqcorH42}Ic>UN6 zHI_C<+PE7}Xmwlu+y{hQp4N25Mo{9b+Q4l+EE4wV3qeK_M`p8)*Jw913^46gH=nSy z2~?pcnn0@=bdQAWDz)INA_)tU6#5ys|^-hSzDK;iHAPI5HLqbQ|Bf3pBjY)zG$ zumLc2_GHNkycaNA6^~g>K~agU-)5kymoxg;@XXoZ-%RJlVW!WGt_H9~k7;|9-#YSg z53WCw2J7WdjbH$?>xCBbn`@wLF+9b&2kS0=* z^i)|mgK-Y4eT}bFeXfideQZbrAx(mSkG*Z@ysK|j93Df18SKH*m$FfYhlsYE)N+(V z3#rAUS-q&*kcHXC=6Q%pu>0E`IXZ46s47;x{2wr8+$KVt2O3=Ahn`R{(yW9*D{#3&O|`8DWhv#qEZHB0bGP(E)>S}H>@kkNSg-twY+`V zTh$JGaLBQ#(E@2~j#Gl0?jSW3fH$7Nxz~=&_##s%$z8&9K-kc}3W2_o^lfn*XqOEz z9thuKO6TRPuy~qTfcJLk?p3hvhVwZoS-h^XAs)_8!d+XNyTQce!Jo6sYp(y5Ek>Xu z8t#X{<0Gj$SPqlHeo0P1>gyfn+v25Vd27XZftJfk z>EdgykX!<3a#+A-fMW49^?w9d_Yi2K_>w^^6$AV@tX$c0n9`lvfKPtw3O_t=G;Ip;&#m) zY{5aLH@DGLsB&?rVteMH?5w#DWprIGGn-}(J#Sgir4&#uknC{9%m-izcPkI~nY}L=`WscK*L<{r^}6#;`TNcGLMn( z26>&2Oph}ZaJ4L-f-mg$$%DCTmo@$4t?MGVj&!{qK#y6H{?)?Q7VyMMe!b!J)7EI0 z3U<(wi1)+~^U%Aa&g{})6_wPN7cG2}y^O-Z+%E2KJV5()@a>x+V>fH}c=YHSIsgkaN#>!c0*3$VQVexpAHXw zAPPCUP_g9r!x4{m$;lxA_0^r|zn4bu9)y5(`R`vu8!EVFmn(`LO0Ncq9@_3ru<4@B zAwG2^rPxph7m=3I4h}T|>Iq;A7JKt!Irn6f>2wRntiZ+){iNussR_VZ0hp*w01Tcd zXx`4cx*(#jzqlzPT-T8}A`Dh=C=pLtI$9>AxUbukL_yNSR=9ikMdW3n@}6Km;uxFuKkN%3YY!Z7qf|u>BJM1`nyGaF{8rh?X-0e4XI^dnA3Omib zj0Ub71P&jLQ*!mK^1+G$j9u$wCMuj7v-~#yw&}sOx`sO&CYco5z~qNZxjI&6{tC67 zZq*thO5tYS&GVy%^q~21{28=SqV148YkTY5nwrI2E>{kgTg;Tdq|qg*A`9*kkVgn( zxLSv~=hF}DgqOY62t(I;vxJ*aZtD+s(i9jBIk}UAFkt5vVx-JlEXet0a;VyoI5jY$ z<>iRU*;yZLzvmTEbz1`ihkF6LdcuEI9kz&?Oj5ZM&87N`{d=St6DKi&`VpR%^#4a& zy^{TKl8P5G2|jG}#IP-Y{(%;Gl=lKY{R(64Tzx=Wam)t2|;n|ncw_vT0#tu)Oq3#$sAHrlEYzYF4)pt^v5G@f8 z^W+rHo{$g^^o*5V7s%jRLR3x>YZ=RoQslZdXhn!KktnfwTR~Mo`l*C3&suP3n-w9p zvnESb5cbs+L5Jt4{UPIaC%Ce3CmVj(H|yFU%H;23S+H<~4FfVp{B)#FFtL=2cw!FU zG*C|Zf#e}nfcBRvrE-W9(PPQCfL@sqV_DrTRv2=7f=*A8FJ{;NK+~lurCDYR} z!j4(Bzrm@p@I23LVu8Pnl7RFt3o0 zctr<`f?{CUkms#61q}s2<->}BhWm%RAVA8+3y0Esp1)CgSWtyBiDvx>=(p78_>U4D zbSyZ`n)5<>?Z?i<$Vx*#6qbLh)brW-V0$Y3}Y5$ELOt1!s@MqpAr}yW@ih zyfq~rOR-lm1v_k`A`lnIG(r&z#x(BT#f{XjMI0ukiTI*H0)uuFBKdKQmroXr&V2XE z8YzB_2&yh+lIU9^W-k(0A#3&)EdT=YG=z265MgdZToJ?YOg3Sek-|BA;)Qbm1RlYS zX161GTaJn@+K+&@Ll<7VC(|FK6)UDZ=CbbW@-tt7^;PDZg11tk-MYVwEZou^aQvZ2 zMxBVe4?l+4&`6kzM}`{asl%eu*egnvk9EmAxHfqmojrN(wjtyn;vEKsXMj-uG<=SE z5zAmVehAF&RLt8S4|Z_m3_EDi4F(x1+Oi~C1Brw%B)|%(fxtSigSMA@Ca^b%mG=X- z1R^vt!r37wjiiD zkRoSg%5*R)QkAw9N3fHm4L@}4m3PAK-Uy|O_D^Yb;#-3vLRg01UEyLmyk%d}Pba3dBq8gGIm zebB$ue@tLH^)$sKj_&fE$C8rwnT+-&1Y-$6Wu~LQ#f>lga^3}ye1deja2S{M*>r-N z_*GP?t4{baP7GX%&rLd9{XENAs+X-OmfC6TzlsZB8|J3L7X7HT7(B;PZtJ?9auJ?& zQj)834*W5u{ddd3u|wOlP2J|Ow4L2NxBgdafYF8#r`-!i9<-#`s?m?99R8L9%BuZt ztF&k4{@1k_Eik82_g@|O){zj-7Rl7`rIa@X*4VV`sc5weVJc%uL4695HOn4Bn7S%C z6M8dVxH2ZW*S2dz_aWHclgT}#tUcdq-R(xx3@3WXw=mWq`U-11Jw`>{9KI;vPT~1X zszP-zf#(M*u?o@p{BkMts@Y=INS890#lg0}J`$t^VNMbbyn4jE3V&)_Ex7VmW!B|@ zdJ(a~0o9g63l<6R&-s3B{0~A&Z(eT+wF!mL!7@n4$l(p5rJ9cql=pn!KKr&%!39`R ziRW!dB3bWp1aCP_@=Q%=IQ`5x*ykW+9xq^wXU~fotFLdWDJSYAv(%!TM;k-Z-{^#^scVFv7z`#aS|kv-eJ(N?pg$$)5@@c=N0j zfH_Dig!trLb0MtE=wg}~v2c+|+duk)K|(QD2%JP&g6tdsW35m~3UM7vc{uYW%0^(*(&!Jo9yk>ex>=n?RwNw~i)`$?K&J0Wa>!69c}tsBb~t_GIw zw4IafjBn%)_U&`d`+35}^ex`FSt?U0@bCGu+k_FdvkJiz)d#`U4dwwQ_pJ8CCq_(< z7}JsC?($%i6*z#-x2!UIlJdxA;<4!!=5;xmOIs{^rbkmK(KWFm#WRqs`sAlfZ$U-M z%}q-uT+;6#p?v;9${MZc=>1>!D3Pscu!Vz$AEUlEtwHG5J}KhFl=-k11i><+ZM!(k zgs}HkO42p^)=r&6JOD;VPVV*P&l3Q%w@rGB#_$+I4cst1&*z)@s1%Jaij1%e2) z?5#8c82+BHt5I0`s|!<;wd%z^V@VUXO#g%vy+Ivco2UTE`zMZ**as|RMi9K1k^Jn# z5WqU=J7;Q?2{#3YDltteA{C8UUg1Uc*9+iG%GeZ;wjh2xb|FIfVsh_$tkyXj+%6ap zdNX_>PHIo=Y9dc6D0YMp%MZ=Y@*q4b;B{<;k_CTg|Xp?8|$y!eZwGD zIwrw2652NDshSbP0IJg|8Lt;^_vC5$K6Szhs*8~MEs=n`m{BX;4n}4!vMkx!FWL@?)0$MzNdT6pRAJyQCuQy_ zUoH)XjaR0@98th=mT^=Cnj>>W!Mr6>TB;IMDnoY+saFpzNq#NSS0cLjCNEzZVEB}J zHjQ0JbPIv;8L+sjUYb+1070Y0IRgRz-dtl!3EvdwoRbztpV>^f90VmXRZPuwb7~ov zP)MEMRw+p4_@tOX-!}R!2QtZKmjMl+1(fgC)1*v_Nt&edB&r&yNi0+jZ_0dUFK?2y zv0Sak3%QLHPK3s&>xY1-f1Tbvh})Y7|No?G5Dd#!hA?ycn0zKUED#@4kV+O$7VfuN z%G4=Hu{6!Y8W67gyr{|om6EdE_1xUFsQ4wy9<_ z1YI)#ZVF5#4qlozPXFVjQN&*zHeo}}^TBJwzR%(NipGCBKeA+jnuL&2c4~cG;n=?r zZ*&1vjyf_e`yy-adh6s#cz&`ZyIF{+_#agzSEX!w4- zc+I&OaZRq0ui}8_M}7`kY364MLC;EaI3N{#6?f6VFN|C<-3^6cV~Gy^Ktb(J<)Cf`1w}uWo456_hk^r1TEjELp06qHa~!wGCT;X1Q6gp}m09 zyEb&7L{Mxe=BFh8c?tmIa#TCnmIb{f@R^1>yTkyxIH80Z08^Lz1!N8!`fDV-RHv5% z4>PtfSAeG1iExLN#aqgI#d@o!ds_T$GigPzDZ zk3Hx$cuU1eN>0%zQhr52SzIyQdHO_UH&9Y_kbuK;Na6^I6K^Q~TM5^3G>WJr=LdLw z8N)}^N5u62t~HnkbB{SBp#@+hQ+#$9JJk%?*<4!aGQVBgm~Nd%E0M?Ax9HWQmauho@qoIMn|2b>}*Z&)7PEFr~O z@clu@_r$eoWi+M1H8RL9kGx;w3b|cr&sIT*@x(#kI=oV6hEROl2OG+Ju71#GxR@1$ zKIQ;bW@#VS$jAQWQW+#Lmg}TBDM#(l_4sUFh9@n;J`na#MQwT;3!Co?+jUAw~Nfd7B;o69F!;-%PdZYGT`ai7T6q8(rrF9tJX(@s^SohL8q)_-zQjUia ze?t0GtJ6huF3n@KTo54KS)k)OI;|h<$CU3FpfeI(t-(aVrImsu*s3SFSRjDik=xF0 zJ(|~0MgjBF)F(JlFX%v@t^=wy@+)J($a_2<6n6>)qng|Q??OzcVOz>dKiZw%fGcxd z?NM{R@k_4ho}CnM?Oc%((UJ?KYk>IFwN@evyH56`SRM?lbDxBCDRQMV1S4_Km(Dty z{^?NyRY4pKEw?{6Fi1n zmW}nej>YTD9Y;&aDRfmoT~p1T0XfhVf<$7I!)h`D5^7apJKN!5BN<38={qcLa_J#w z-B^=|nR5x4O3ID^M6Kz|A-b=&;++h24o+Q{&PtPoiloZ>zp@Cv>m6 z3Xf1A(LSgLuZtUk8o^&IM)CxdS0QT1inko7TI6m>QuGnzV+odTb9RlOaI$Bz`5#Jkka+NV%MN$UL;Ivd&)VIJ7 zGvcobm&8qyU{n*_Ge19ToEdZ)0UI9^QTNib1Y9$_h5W^fzH;Y3*m<1S{F1yrr7ETp zDw=dDy)^$UC(;$)nvYbUGl=lKDU(X-^Ruw?!_Pl?n4t`7!DFsp6)svXXydW3i9enx zP7Ot+>Yi(&!MI(Q$iNg9rnW4sH%hAMDuJsjwVlUc92OQ7P-K6E|Fs@BHlwj|>m;Tf zPZO>-J0b&YMWk8C{Y?_6s4lSvh*m8YMi(ZMYdJ=sLmr(TRf1KUHd;dk*k2Odq-Y`H zHN_Hl`Z$oLCiy~sb0U4q=XB9*uyVive8i;i4hPlfVqeqRZp5mc2Ev5F{Td1>lNU}> z#zy7?k(Hn9rmd3fjYAKSl?s*d3&%;q4{A+g)0ya-agW)HsadG;(l0OXA$^}?j`|51 zmB~%OAMMfaFv@WZ{}2wAz3{hDeiL*@399rhq5uHfJ4O=}{3P7&xeem>E~|6o+oH6J z%yjlot+Bq~NSdwnpU70?l88eOLs_yD$_lH(tXCzV!B=AzvS{oU+IiP)e|-;a_jUG6 zbI1)r1NO{awBzERlN>8-X^jg{zRRSNn?D2P@5DxCtnV*MNgn!76b7#|qJ+&-%usy9 z&c=nMDqZsf59g91=Q)_PUAxlc(q>FnHIe{0t`x=tv>00^*U{kr-KJNPvX@j>wLvXP z8Hd(S?jd9Gxv0n!C>E9zV943O2hpjb#q;4uDcb6q zCub2e;O8^4{u+S#nig-}^LFdNJEmmcjTlso8(jj=bd-tSF-*}sx2~YE90ZrXlj`PJ zy{@&V5-6yFS6X?PV)W9}YtuTa=F*@Ys@m)y-AXvr5HQaF-yuPYZY*!4sH*uwJ}Nka zUOo!qaXq_`;X2)b%~EL*Q?{a1H_=LAPu#xk4g#Oz!_a4FO|8}YgNf1d3+8*hI1N3L zN5AT{*RCYJ(6EubYoklb%JgclZQI9qt6rn1xw}aZ=y?_G!guEK#W?&*Mwo|-XOD|+ zDrwCcQji}s*U~@{0r2ynE&8W-AO(J3l4_e@^_T;bnsG~`Y>46`=);S8UXVHh(Cad5 zLT}s5C=7fF>CwBW#nDG8_K={76(!A^a@n#`sDaiJ&Sm;xYs#Q`!i%I~7uV)Ba2A(mk<| z7u#OS2{GTvwqMuc^uyOBwCe8jz+7phBR@h0VT34yD!MJRTY1H~-HWRBA%n)nO7y@w zTc=b15Q3eLqM4*gZglKO?MY&rC6FJ zsRbEznQ(;3B0fP2(Q9M`JV72T6jCsPZEI^XUKNlbcd<)bnyGptiM7ov1xaL~uE~C^ zI7~=Y0$RP{=J4O}x!3!WD8UpXZD5sDQ`}osm;HAU-W00XJq)ZmPg;~B*&YM}6bWv| zVWrv3|Eri!y*qb@uGgYIy#!}Xd8P0mY{BVov3}lK=$2rP=WAqNlYjq^t7ju!IXPyF zKux-#-h33eRzYI@Ng5S~M6nuFAyX3aoi2ow8gGImN*=PSj)R8rCF$Uz2!*3Zjrg-f z=Ms$JQ3#vNJK07;+9;qm^O44n`6(EuTjw;Q06eoH@gdCI|JB(3k15v*P&{NuU`gWE z#S>iBUFIb|&FO9up@lcAWYg*3{bp(Bz5n2ts;K*}Gl$B_~N!P2r4f=)@1(TIu%siNT8KI@2;5Cg-*r~&0LFsSazOe@r8(e z=kz1?cmL~M{_p6)WXp=klxK)`YSTR3q#|{h)V9Bqu;00jpY-6bzXQnM&+0{;sE;7P zqL4B z3WhaZFP5b?SwY~3AGW2~Y}Vb6{H2~t#|AR!AfJRIKOtE~7Vb(7PQtx|u)o_E5EA0A zOM7^CLKH}Hb(MCl+0Tk%#UhLV@97zG@+_k7d0!ngNcyO77NHP10#HocKM6_qRG|5-2`GN}Hnc^QqqS+5|Q z0teGizQRtCK8sqSKcsbh)h^5zP9DJGDa76P<2_6Vo>Svx)vgi0kZaBMgymLz<&g{^ zRN4!uA&{OW#^F{br^^9fR=@pUkrF9oC37$^wNx{*5#nQMA+QzL3xvcyXyeL=pNJot z#Z%*oB|dBl`#mP51MkYZg=Lu6?W?`-wzO)zJ7l-oZPa5siEeoTem@iuoz)Fpwn*F7 zb^c(tWi?sxQ-Ck1$9!a78MP>%-i}he(F=~VGwdS_tpwd(cu_c;GsL^wJ zR=Qn6e%yw0LA}=)hM=H1j+Xc5QmZ_f=Sh(~d^H9FKwmf)xV810aq&S}fQSc+pS%oW zY2W#oj`m*hJ8-vglo)0n&X_(H(I~G9&iF>;&g2q)8Lw8`nhAb0+DJh{h3lC8>87!` z!Iu0T+e9?`UdZ&)I+%>y7_cTBYV%~ALzN|`7}X0@+kQz`f%U=n2s4p%#sV<>w0t(I zxz`R1>k=FCNX;mhpSbdcc~ix@nFNo z!I+~oX;Q4gH>B4Dyf~=knz|7T<7ku2IlPL8Ae$vnL*_zN} z@R_v*ro$vG%GrO}sF{?j8ccZj^znuU*l+IQ-ywYfLMwYfn;-CjYHLzXonC%vTz(}D z2A?P`P`QX)9bzDxch4I-Ab}laNP?{loTkp5ex3TmkGnr8Jo= zA`A9en6;U9(&&_c-z4%rOR5vjJp^jAm-e;DhcTjqNXb|7v5c__g|T4dSTZl>1ii!1 zEpd=wW3Nd>EyK17otmjA4oCX>1csN}=yJJO>&RCLHU^L`(n7HGo+fg?f;={09MEcu zCzHb5sr`546~ZI&4;I_x%wX7e5Q<&J^@3$*M`Fe8Y4N8>aY_t0x}4YYlgnUN4VMnr6^=qAJ1(zoX^46Q0B2M}jS}9vq6@{lw%~orp+|QS)r1?dmYFhsvHAHQF%a z$?}BDoCcTd+}W&F5OR2ijhv#0;DNH44IH?fi~t&w&~#Gw4~Ny?&0NB9rBuK=z2M`^wi2EQT4}Ia$iv`#h3MmV2O+PL2-_W2r1~E)7Q5R?Jf|cGvbSzwr z$~JAHr4HIehntdopoQ(x`7Z`W1FwA_eG3_SasM=%UMKCD;5@Rpaw*ejKdgXtqX}Z+ z)u-@XsgBRN;NFKhFQ8%{4n~&0E`vT@nUks|N;N^|9>k>ar%+LgKI_ZyO`Nwgbtwm` z(Mr)FtKU%pve%X^JodV; zRNzk_CC6{0?3)L~_kA0#t|6D~*`N5_!(BvyZV2Z?%TKBQN2&>s(qx{)M}ZJy^;V4(;~R_kln zSBBN|xEmD;-qIA?Kv6KIDggSpGn{lbY_o_GT%5K^`}>$GI-L_UZ;)tXMP22=<{1gSM} zB0Z`tg+L?3+q2m_ns%Gqybt8($-}P|lu;WIz~-pm#p(+Ro_n`CxCkFl5}|ODo!*sM zUqTz2f)G!v2c0c`8~f{r%%Iqb2=ICkt?(eI2_6!;PXOZ6Q#Wj&`T&id!of6>Yde1) z8)6CEy|Ql7jBr_}syW7{k&q*OE~plrF=`5#Ne>> zl6z0wDlTjrs0DdUK`6bLCf-tw54YGY{=l*HGk9-m|$ z#MdUu{ib0HCEu^|!83EP9XADyb-dn=xQ%pH+blSAGnx1~`TSK#2+R+0VKNW!V4#1(+zMt?Mm*oi*_ zsqD@^#kV6#OLKv4n9Q4l3O13~nJWADa^P^*|k05%9R z>9OgDvLgC+D$IF>Tm4~Ak@x;J_=N82q85V>#jUm} zn#46J4|D=93##q&;{WYJ+xW&cY|jW2byQ@i8;-OQ<+rkk5AD>O{oil%DUVs`)3Vg) zxXUlUhLkzIa3W=2uyu5Z4B=hwNaAiU5*O~N)=Wbj(w4G(sRRlv4c0joZtKD-13qOT z=^AbP7trCAd)`cv<+P{$s6l;~IfDVx+&49~0N4+uHTO5E2WG|mgM9mZ&D%FajrMLr zYj8gcvjSyck(Rk&qe0OuWcUcfNUD`w`X|4vk@N9})?|l1hq+Dyo&LUc;-?I(sDAAL zN{EtYWt49*3{*(0N|c%7D60-4(H;3tCJK)GDAujLS|u5PTLGMZ1#L+*P4585oX`H( zbLmx-tJ!BY7MelpcX7)wAU_SOlorO*Wa6lf%3OpMg#{5q2E^4WYxps51(PzL`Bz*a zS@-IO+1ocjeK}NJoN$R6lkyAXig+~pC6}d==Noy-TNFDmSle%0Nx@n0(2<`@#VuNq znK@Qc{-$dM9V%~DF1KdDcDaAZTqZE0B%2G?$L%BRh3UCXkl|hG2?c_-Rz?|2e*24K zVSqrltk#p+jYdiJ=ZZ%b=9bUp3>(|WcMDj%us23%%vUmaMFHvl{ez@NP#|RtByl9B z8(Y4b=6~oWFBW54T(;yc}y&oTc{YkNpSRXtmMi!E>;JB^e^?-_Y|UZzY&T< z3@l@6Y>RQU+Bf$zK-YLYC~vip{?#WbBeNnEoY|qu0SbE{FV*>vv#&O$haTkS4UPu? z3lIi}ghi*m9#&Vpjd4*FE~Cp#SrY)^Jf2DYfJ#C&%ynJd!5<6}ZG(pa-HnXM`bqUa zLk;Z2XBOFTpe*oced9}2bwU4y8G`o3kzrWm!aj5BZ%R>~boBy9tlS489v$MZ4X#qO zPmRy^RYY5L>dcl3ABgSI*@3_8z|UE5#pYtF?*!wA92-75H??&goOk9S@dM3MRu=@W z15yu|Tn_F6r?xmWSDv~ycUhb5Slfu3Mr{7YR6@7Dc%?X%Ay$m9g4!nC&bh%88rI`J zi&2YVM<58b#Pzzp4Ei{>{hl#wrcy|QIG8a~KT6YAxu=>!P`^m;Oh(i!Cdbhgo?xQT zE4Xa^w=s6VIdY2$spm#Xqy%U1gFD8`Vq%X`i|*NrZ=QuJ8o`{E^xu$mu}ppVImOg+ z@0h|wi6axXGS`kM`ui5*-59X;PYx%a*)KcB{{e#5Arabm$u41c6fo(wJHie=K-Y@r zy@HmourvESF_`+?irz+6(HU%R`4;DjKkt5tvoq+~%HcJxV7UMV&ib_-4k&8VSTLdlK&A8q9@SRTzJjVE~EYHM?6(ibia>@788OeM- zW+(Gx-}0k0XS1tUh6TVK9AHHZp~8hT&zrQIxEmM0`d%4;taW}_i{pjYAEL(S5hRLFP!qKN%+YyQmI^C6pR@KBE3e|uz&=8XFKuW{M8-{_4rY%%8 z6XLxGTunhAB*j=(mZtiN#&FRYClgVX7(cHRhLp)L7u|DH2bC47%8FO=(y!tuMbQ$X zq6x^_H8#Wsz@h}9E@1$$%F{lM@Yi9!+C2y!q)Tig$_JlbZvWWqrnDP*H>NRL2ZEx? zxqwu{t1@#&n*Yk#&%u(Jg8Tei4jC|*#J?))TTRhSZPGTB9TFwl5USgiC)^hI9)!vy zQO-Y6Usi&U#`E+u)QOx2Q1Ta(BDUG_1~Qy*`(GKBV?!XBXyss~PBwRQH{ZoTBmX~M zfA&iolavFu$_(a1_b*2->(i&3jW86E1vTl+iG_^(QN5!x%j^%h@%`Yd((uGrxm{Uo zIll^-i@3oo)=G)ilZJ&<>wF}&ZD@nZ=xV;pXRo;BRon*Uw7i`Fd~G7T zaAXlr$4LI_S7-4-77)UFp&cxP+@iIK_b8#1$pS{WhA-^wR>6rgUGlGI1}iqV%kDD_c0X5R}Rn^udo%UK>z_wg~2lF}Vs(8s1495CHzSinIKle$`j7;1j8Yb4KZcPP zauXL%u?o1MZcE-cFSt*`l||h$pH<%l*?gyuJnnig0sYWqfNE`RTQ$xJe`5DO99#5B zWTV4+=6%x=mH~fB`?VLJWgjiI=E;{07p58)VJGMAnPJXoM`kMv z0~cE27)im7(Gvwft-fTHx{}=C+pFv>Y)Q{(ORU9VDe|M01I)SL$2CP@?!|tfCt&|@ zdPQiGrkYSq`qK)%jTMlvg4U+1421e5N2+oi`PEQ)UIsL#@Rq@Q)jLg21s7G5#ZxKi z$eFp1GmYMCNMtNXOJ^CBGISMwKwn^hVQd3G1g=|Wi7Gj0bBIJKAIZc`oP}|%)SkyQ zgup)XPTp%f5_Sd@NVCe}-U|eD7c-Owv`I7?B1a!~yi0ZFi5>WJMGrsLl~CNYyz+%r zs*klMV;OViU3?jBCT{`2g8$$m7|Gyi#^X_Gz|a4K+B9(mC|wFig553M-fN_sv5j`5 zPtTg8I&i~zxi-PuG=20@_@0+I$?hxJpA3?Hl%&(c91j2%Ei67;K1Y9bez|Q*pY{FM zZWyn+Bpoe*cMaW)ItOZ0tToLnK-FEq{U$bpL&>)?@yFvAhG!0Jmk+1wl`zoT1=^-5 z$vEUg75bMWv~^_d`N2YV8;lZ(O^|NC^b8&}^zGb25cT;jyr&1iQ4-e?st!*f~*6N|84$ z_(z%O$2obOj}2w^!F%UY#RFQMD)BV1sz|NB8`QKm%1j&H&0?>(`ESOo65Fukpl)Lk zKltemX5R%1q?|QSe6M4tQtFYDSC8o`sgJ>6kNFosycw|}tZlMBnRJ|D}8Vk42Ol&B+h ze+?jU$Cb6i{8UO#PangwLs*=24xAYK3e(iFoUZ}3z#pDyLy2C#*c@*ep^>t#MGa&f zYW+K<8s95TrTP@BhH6Zy!p^k`W5{}bw!x$AuRb1ynJ{5K8Frg=k9%H$*uq>SJ5#D=4A}Q7Gdtt~RUvmP))ekMU;Ypfu+S)2USh z_im85I#@QtEYe9+2cj14FFC2Sj&;jAN+xZ|H}lpW0vZyHx(zb-)94t^91^)%A5mQ2>(EQz?M~ z1?w${YY_HlwiIjYnw1%Kf;mi)bgagLAmu=5AW>Y-BH; zr03!^3ZH*TM2MkWV^FTai22eYLwh2{kP&9Aa}2guY=N=fyLKu0k>dbYvCHeL-4vDz zF7W>ao5FBA=p#tq>H5?k)x?injM|A-+S-%7@+kx`Q9#fR1{bwIcb9bMCNz)wxTM-% zqofCBA9XeTjN!KAwG-PAcy{mdcm_}#lBs?QmOmB&LP?4As`sAdyFtm?56_A@5X0#& zKj4i+*~O(lP?pC91&~fn*GI21`PnLWqb52(mJ!*@?Eaw4#kq!)%RDm?lQ-z`|5=}Cu<1SP#vM1T@<+b zobj0BWq#SJhrFiAcR*Pu3*Y6LfL)d7ZZCbFH1;ulaRQbNyy!}(uCg-$;8BSlu3`QO z@}6iTTBer%y(4s5iLS2dwAfjD2b&E zvrxW24Fq?A#=42*G&|!a7w1-4hc-xTAjppZYI`86WF{@{NE%Fv-QSTB`QA-*nM#rd zKVUJPJX`GeD+aWVZhwgifR3~q5hRK`^#Mpw@2+G;@1voVo8#hu18BxvZe`I=ls#|5csa;_9x%yuVhU(8$h_@Ayc{i<0V|Kxba^g4q!d&gg55-D zzE!LFBNsPzBY?uiPg|ra7SV!_BaaV~bt_s&+|8yAo=nO|suRag{WKmL(43-U>0TTn zqXVuAA}^{T#>-l}pT0A(mvqd}w7lAF(p+xiC2lh?acSWH(Hb3L`Zrp^J{yG243pAZ zrdJV{e8pNP3Add;MeR4YlN_&ys~2j!GUtMeJKXFjJ4U(1c`tkA;Yq5? zycpX3w@h1?^m(h4#VJGafbPvt>PZF>S&pw0j^m|3>Z}lrLJaL;q3DQddy$xdZEY0pHu<2){j^m5V($L*^+!dY zA4bIUzT-QzEOO6{?i`3m|B|Q7oUjMxsXOH4Ig#>MM3w_8cmTNGr$C1Cxw=M^T|4c1 z2s7ay_1d-ROaGemxa%)ih~&WZuO3!co`oMN!Ht$V2DxOpfm^R$y}wR5ZPql4z>V)U zk@Uc>St|zGU9p^nX@k;E$QH50Ii%>&X2S+c}UKv0m+ zLjriQtF3X5Hpt}0I6sif)wXi|FcZ zMgcHGKnJb{H-?KL)^LH^UE*@#EQ!xU)op^Uu0;?lMNu?aF6N02_gHlVD`jby5ueII zKXUov$syCn5gr)^h+=c5Q~zMH(Y|b(hnn7{BBkgqiN>vvcXaL(S_NlGwKQ3?OcOJh zO_oCa2Q(ga`*5D)&E`7V$~xcQOi}b@8)X^fvW@&gJ_XTdZ0ab>v|w!L!VJ$7Af6tp z@kDOi$%iVOvj=nVRW)&vR|6|;Q^-@62=rARsi<*V(YT+if{+iyc}>70DARdi%Hcgv zv)B|&_r|=O=6#sH$2ggb9a4iL`hGdg#>gFKihr0W{vD{3gEL_ip znusn|!6*_&sh{U30YqN=WYl_clZVxz5ektRr97zanrsM`>5;9HajNJy(12OO{cudv zVf#(l!f#-m=89QhROshza3JPylSYfgvj*a7E9@N)c8pq4{X|!)F^0%~H%HD=cWsa^ ztstrDE|M1o}X56`^s;SGSDua?V8m&IL88KM42Y(ciSTo;O5?liBhxV+BHjp1hQ zNn!ci^0;h4Z?q?Rvpb8RyddEW#aM}%rU9(wHzZzx)Vw?+mRcqYxeI7zDlk4|p;rU< z5cquzyg7epn$r{uPF@A48ZWTalr8u?1tJeRct2ivTX5jQyV^tUoP*T!2_v}wNSIL#+t?OjIri!}VE9Tuc15N)5g?bTZN6G;BA2~dU$ z7M^^(Dd5D^zexrzflweT)5sx9_Q^$%^O>~hk!#$U0)YMUmlWlmL#?OfIa`ILHZ~6N zP0(nSep+=vcqR@FSGm~T0EtIBpQu22N&dd%Sa=|sE$n03oghdZm6wnqtyr3 zJmuO>D{OE2n563{)*E$D&zv&jzw84(%i!I+r(|yX7MpkG^=zTsl@S;{vy^0*1*Zj$ z>!JUhe?&P@e}Hq#QL8c}z3^~>7Sw?Bes^707{lmUyi_vcoHv@bGEo@!1NhguLq{#+ zi%Qzp%Pm^986+QT+6nO0WLu*_i>g9WkX`=r@?>p5IT(ap8uI}kOw@?m1O#}D8B|tWQ~k(O>`a^9pf%TQlH8m^{6;^fHy6h6f?n!oC;GH|DI>Lb`a}|HT}2zYb@7qL1%I z>;3IMm1Mp5&KfC%uASyV9j22lA#tLG+$%`tdSzC&beg|JhY`Y;52$v+iEZIgkiU)% zu}MFQLr6ZR-DBsH?r^N35z;~9=Y6YU7}kI0sRA@o`Awe7PYbCLFzBo_k|Z<@+Wa?_ z1pY(*Jo6l``~emf9PB9ZyOK5zktRmX6ZDrvd$M+L^{!7W>T?fO$rdcnV5xcNKJ<;x zx9@j}|8$S{xgM_Hsia3>x{YOqy}SQHVl}uU(T<=rN`BGjah!k0Dq~}_~sqE zmJgTFhw&yJ0B}pF0Xo1+NrGCLC^&8eAeotTK=VVMw;baa3(`mOG3L~|P$hTDP!F>+&HjOIB<8_M@c-Hl;bGh}V?^P`4rc z11!{{BqrWlg>r5p+%6m58CHd?L8z~Oh3_Q_Z@ZHU8gpz7K~P^3=@@`3TLS2jJF*dNj-Fb z+?r8wH!_RW*Uu?bCCsd#78OE+EU6j5rrH}{NR!{%fsK{N%d0S0woY;Q`>JyS%+XeT~ir8qt% z*}esSv#5l8HncUlHKcg!L;^O_xj4;_>hGbr>eZy1+-pnkg14yVDj8|@iEQNtvQxKA zN+8mZaH`(OOO{f+qJyq7*jT=zj8sQwf@Lrq*) zP*uGIOXf6@XVPiMVV;zXYHoqrd6vv+;vpy^sT9jY}$3l+azf68ZRKYA^So?Tt_g zS%ezd?8I9;Im2hG?eLNko7hQNT24ZVmpCXuJuKxb%RPEahUV9^VoSbr$oRQ)sCx4pk_Lh zl5=cxWQZ;1#;m1d_$z_N9!x3t&{bE6p0r9kXdb29vY(pqMCodyk=N=|-dW)=hSH-s z4bZH0qL~2bPvi((Pjb7t5V?v-1x-mXZ&Vt(N^IQ&+TcrntfUH_FyK>XgL7F{pay#q zQ}M`N5?;Y@X8L8Ew6Vs!uKzLV`!~| zI^ZT>0AmPdZl>Hs>~C(Vjrs%iIY0ok%fKcD3Uao0T3U4be?Y+979kU4EHBxuA(n@m z>1>j0UbXcr zF7x9jH?xPWf0uLYLGo>@(^cq+b^~nKFl`(6JD{A+Fd`z19+@*s0wFcjsN4Fs7b2Ja zbVe{wXwdMXD>lms55jN`L7=H5!F|&;Ty%zc;RPdZFPodRXBpy*7bfum$x3#q=dsI?%EgTaA?vty4%4`LX48YAhG841#mA zaI2Yh2~fn&Ei*qSA)qy?pHKQ>g$>_ouyJjxtSItr-D(?WA^7sMgxlJnxTH~U6e3h6 z(90SNjUXn$Vpyte7!|9asIa>r8~F&M;mVkxsNM{A9nb+Es6L>QhHsNC z%Zm*!>oF?~(H}7~-Go-B0HOwUPw9|rhVS0{g$Td0yDg9cnMaBsLE&}{Gm><=>Os@w zf5dJB0AIc*XB@Ru9`qviLcZt{UN`-GdG+NZ8|AlH?XvyX`u?PSDQm? zPxbs{4jwTp%)+->B}d@i_ym4;Db8G!jr6|>un7icArRQn`O zzCI~PV7iE`904W-2uTqJjU=an9{NJEDqT6 zamuW%bW5IIFz1~&GxPcD&b?5>0ep+&|6SI%E_%GnsKK9^yyvg`*h6sW&NpVvb0AOy z{DML)fBsH_3z6$dsbnieNz_-?iIO|IEEV^ucoI9PuR2y7axrYz7rUcoQs4KCh3x@i zak*9l>z(OvfclcWA|i}n%qJX6%~G{kNKW6bZ-0MZ^=3Ii#a(gUCx!+cPSk@NB)zoq zl02f;I_;@MJPx))M9N=TB?WslzSu3{FVig$Th~BC6-WkXUVq0s{^66XiIt`H+81_7 zT75s#x|)M)FaqHwDXMlg3{17Rw7@z+r{Xi-;^;M%z=F}6lPp<7HNVr92n)N`-$$E? z#VtEzjeB7!>Vg^%pWvrsq27rB zj4AAZKzx3kl16Km#c8uW=a}TQENC1AMHY17`?~*1t&o^FbRymw$5I9QtxmQDUN^KC zz+vRiKGrUMYE19?be0;98)J6~{G#Fc=3m_=GjETBolSJU^w|(Ikjthn>rc%PZSj_5>~8bZs>`u^i>JXy1t3Ye8rL{Hy>|QU_u2vUHiS$q#ETX)Jj=qtH1h44`N|Rj}28; zR7%(wwR)6o;|PZ~HEjBBqeGAT!C`3D9z3MO@-h01;) zTs}S0kL$o$Tx@9tK522XC(I@&GlgzFK*VU^zfszcw4%otDriB${rm?LsOjL&CHe=d z!f1b+W$s68q{RFE%(vV|aR(?A+5$x+=3WOe@MS2xypEwgr`=v(Q1x8i{-Tgp_#0+Vy>ly=2i(O){08}SpdHBEE-~)6`34+ARphyMR zbq-y#RD_hyV+(l(++HSTlap+gU%hp{^_+a12jF*pL?D^GK+l{n*>KQBikon5Pd_iT z79=n?84zyof(1?b`Ln`2)i``DrYNXRhm`sQc0GUP@ExS`+v;6)SJt6cl{E`qsy!2u z2iwc>uhZH&J;Xr+1K=)oJxtl{3wmL=?Cphf0FZ|QOd156E{ekQmi=pi5*C{VBSAGv z;{7QbW>+o4WB+vi%(X)WEKR!2bi|FR0h{Bw>c;0f44WFa6vL4%7bPUy` zo^r--riCs%JK0UXK*ub&wkkxioT9c=FyqT~v5AGb9xhRB%(V}L88v=+9NV900Cg~a z4YpsjHE+-#b7#%SKTQFH1AW9Ubt9jlH7Hg9NgrpUPg+Kv5=#97rHo@cxcg+nRig?H zM`U5PAY912{m#ur3K04f#0hPEM?s}(zXgCE7b;f$tPfN8X9LwKE>0A*&h7??nXHt` zahaW|HDI+Ryo_!wK$z!rPLlHljaC?5GcaEm70eg&RiK59r8f$Ve`^a1b0C!8q>rno z<;oHCyFd6aU~3&tnJS;9HGq;%OkO|=n~Pi0=-b7)v=wWyy41rl)pd1+C=teYu6I9xm`b~w;U423}A-Wm)+a8hilQj#e*JT?^UiY)ybkVxe+uBMEc zSM=i8Su>MfG)UE{I3r$;Iyd+&BIS|>q2p)AdhG0=Kt^je3&+$Z-^PzdPNt1}DDEBx zbWZi!q>JMOH{_u~J8z(UPFJlbh%V33k|e_WGs&ddtk)kWB~8(Q8rDg4_rjw5F~Ui3 zbw0l>Xft7{nLF5oW=V)ZxF36K<^BsPV&)W|LPRb+=y-jZ7*Krk}T zpZc!h^5xzL_Zrj4G>*5KLFf0+b2A#Xb{Cd z2BtTl$eqBc{qZ5`jD@e(EOk*z`ovhp3W(T6##C2OQo<7kYhAt8(xd?h(Bw$%ACDU2 zjD?PeQJl1XC=F>?CZZiHv&`sOWC2{cRke2G3jdHL^N03>#V0MLdgS0dm*Z75y2Q_BD`z3xwq-7t>Omkbwr9@&ZYy(np=xg-mgwulqeBqT&eftxane ziHd6~Hs;NQ@#gsBwrpBHWdRZb!h=rB+&Tjpw#>ww7<_#Z$z3m57C-9F3tA0~R>zT% zvEk!}KdBlCik^_>lS){*<%S~g>?VBqA^wE}9`g;4FEA>MEkq_c_+;l+|5;!kDk1QV zEpbx+=1eSxW#G&`U(zl2YN{{iF2RG{}Fae2j^!mX*pDu4$V(93((|3a7+3C-TfWM z?Xf^1hxv9SiCU!10#e=*J3ssTI*GI@{_Ys`b3~zsPF)Ee9`k0jM#jv@Sge7zw3cxX zKT!$zQ_wW;RN2dEe%I!AOb&_zjTaTgMOZs8{Gs9t%Xgbc*x?MM_U6ipuhQ5lYTAF` zJ0D$F7F$U!ts@yAjT?Dh45?fQuVRspS~7{yp(24(T2YU&p?P}x)w{D~fzZuD{y3~= zJyi(nA%BzTY89T=fPrQbXp0Jg1(r!_2YV{H;SL$#TTYDOF|5W1Dt3{Th?tlVDe?9-?;t3nGDDlC8-) z=`v;Y8mugqvzfk}_McKZ z>fg0jeTr3_Hc+!eziCf4Zng&bw@p;yGC@uF&@>UFbrtr50p+er1(n0()g?x_VIxN{dS~sy* zEuTA3gW5!T5^SFJ7IJ7kMEbk==(SlZOE3}S@>jwJGenLuY!2z5(MSIm_Ph56$tN z;F86P`mH(#1UmAWh9ocqk#4Cp2kc6PMK`T@%Y3rp(E%=@BpI@@X&|CNKg^vKx z5LxIf(xb;P3%yMBhrd4}o$Q22zik2Kyksumgx+9W@$)%lWwMI+?3!YpuwI{)i>8V^ zzjgwpQIIOBPhjoZ3+7PqAtC`!!WW$5X^2+oH{foT-9_M!G|Sa9Gn-oqoK&e!h*wlk zXnYnrbmO9z0OQg@`l8+cu%>bNvPrfL?6|jkm1%{j?b5C4)*UxsZ!Tit*B(aFo^b$Ix0-n)7i_RNnLeVrGLYf-;ArUj*vWMkK>@Kt@37M-=K1EX)o zVB^uIbAcPU2w=)>?AO!#@lngKWzK? z^Fc;tiwEsYb9d&;0twRewkahB6YX$_7C1jic;hPQ`UjjeDlVP7g?;Boos-o}KRt~} z$IMctRl{QJFJQI(+^ObqCdIcV$+ZHyg?tN!(>h`H zvmu?b(u}3MigCaDZBzww;Q#44UNZE12f&GBoyD{3=zop?xY6@srGj~rDYmHztyR(f z^(czsjJA4s?V(Ks4UI%yNXzFtu2b($3jl+q);pRAzwkWA(+h4GPEg5GoLL|3Uc2Mv zKluBG3-PWvA2$QqU?HT|uCvNn!Z z_<^<gp(iAL*6xT=O>yY=(#Fo`3S3J_LWdUYth*t4)0=IiBXR zBU|OfFj1E zM1|EGCyhE3V#S`($33&|>f$LZwxKXJP_?T5LKirXerU0~UxfZF(-?q}(KIxVe#jW| zOMqHX1^RojU>HeOj?P<|?A?^8?yUs&@HJ&OWLyai%MqYbROL;{Jh+OGE!YPrr&cyK zJ}&RUClgu~w`JI>ux1k(X(}+RgP-q- zY-D$Gt@se(uDH-$GdX9?&0NnMk4WQU5rAdrpD9!v)SMz%PUCD`H(Y5h1aC4=qbOUmuP1>$<2v2j+>kD`sR4iSQ+EXsd!o<0nnwWlXpqfMjz+0e) zvMy#wY<|c#Q z08UV|%R4dIAQ7$~rX$ARl>FlhR)Wj5O7e#(Ad4!zu$dDOBV{=fGpm zxl7yX1!woMWyaUNR2KDQP40Zq=8C2BLs={H#R!M}&pzGS2)QT*;n45toRLDoEca zlDSQLU~_fXiDo&S7yg;l=Sq0!a?7{$35fhX$LWoyQ*#m)AC$wF43{`7x4mSEovloD zj)~5wirnKh;C7+J_S3^crAnUIKuyBL$}_W&5D@R%O+4)aEAr39p%!4FdBmA$jH+Ua z_>bak<4^Bicottbd}sq+*sWSP$;cI^7qm4ZDf`?3I2Q6%q-LDIaOA-@*FCRxLEV$p zGD90%|6(Eqqt*3t9q*3kHAbm=JVTF%?g&iUI8Qphefy*mZ^MIj;K6v4b`%RHb54|O zf>9L^_yj&%VzU6j!D5U<{Ew#Mdw4Ba7>)|n=6j{4+ptD81Z2f~SM{)`v3wQ-3 zv#1A79vAA_E+(7!oIMPo@e`CI%Yfk*8ynoN%P6W-w)qU$rZPettdBJyW3mRwu;c$- z@!wbjChDY*Nq_2Gfs`(#n_I4&PS7~SApQKQDOLvbS~#?#RBt6DhYx84nu-09qD^ag z$L0``Ff92PdNi#YLRH%^YdS#hse=+_9^DS8I-*Zo0v>v36y48#EKAtl)wFN%{_|Qf zdzko4tLXUNeId^Pd9dwn69+rSyOvn+5`r$Td85dhx&E5LlV8HA~p?T3Irx`15 zba+&=$S4Seq->5nZ4@1!xfIcp14Jw*H~AX|O;S=}(Wish*Rn~@H1&(vCEYg~&1dV! zE&cX!cfs;@)e)mWoI&^2wm1pq;heRrFK{&U64J8$oS|X=zsAvU=?c_>CRs6``a}wv zgJL93M)o}Cy~7cn{R}Gv*;MT_pbku9?_;QU&)Z~?e4|(IE9Etsf&<+vPoN{Vt9DMZ z+0SDbihN6e(dfR41wE=M6q9>45+79B|2HQo0G;H!sH4c&S;MkDk|Ey+ReI8@K9q2iEF6mF=kA=r@tlFALg zg!Dj}F1A8!)IqN>@UCmBs{y!^>^BaaM%daXOqi3)pb%zfWTo5wh%=97R#gWrMKA43 zMRH4|Ao@o5N!%WqV&w+wjv8T6lxtdQi(AVX28h4YdsP_sIl&6%Q15eHD_#gA{MI;O z%KfsQKxzRkqg0U~x(CdQhIc{LWoX>ZFY%#2lI7?HZS zyRifbN->;6mu>#M{TB>=s>JDC)BM$=n-toa1YIK8VtR48p{wZ$;JnDQ;Bp*Wu7>nS ziFjgxQy!1MYpBDm9~C`t=3SDWL9{L zz+xm)bu|K(oqSaS$P)HLJqfUs!h1@S_5m9;?9yX#iW;DF4`u;Z;;_v3mJ<}6c7wVnlEIAR>l51%|`K+Ayu*~FlR_%@DC>|Su7E5 z=!8abXtM3003N0wt3IN?v2Hw1fiGDr9_2%KX&?+4OIDWBtOXM zI7pt8pe|c9+sC9w=cQJdyeVwwO7<)z}+ z_(0)^h~t|V7-&y8l<5XBHJNXqSiw?R9!!?*(MZ5JTWOl+YYo1r?Vlm{PU&Fjy)lWV z>AOUU{%pyg79iOrXj~Xk!ch#6<*p+Ep%HYYisKjf|H}5(wSBmIN2a+Y?O6%#7!~3< z6@S`G1u|tILn3=418+q-#E;l>0h6Gzc>D3512Hl;XWme_VxVMI=ZJo{c}+88T%NVp zB1N{ODZ|8Uu62*PJ)!W9TcfE5qZ9dzOWo{?kcmwF%L2Ha0=mr@29v&8_X4Q*Nx2Ei zBm1R|2x>ERDfA>Ky~8r_&IT zG=_G;yAW+c9{TY<6~cg8V=nu~-vCMp$d->Zi`H60f=HLp;G>%p4i^rWW;P@q$YMrUx-#t1aU}NgcQJxQV&pyYys!tu?yDbyF6H*ALjb zD^{HoaP7>&ql@>C?5XIbAixg{Po-vM3@Mf8gioMRztul2IdMAz|FqK@l$AR#lTM1* z9xm#haR*Yphw`&>hfQmCNZx7wz|kW8@j}J?oWLXYyGy@x7x@`9Zr;0|v1`v%Ttv46 zkHZQU&)O40)E@ZEljGyV=}V=Xaejq?1P3860tX>8oyp}GG0H}tI1n!;q%O}6RhUL7 zrK9&RPp#ltb`J=xXSG~kQ+El`AN#x5z%~yQzx8Qn5y+3pc1jLu*+XZ|Ai5ad7M?SWDl@j1wa z9m#3YBs2Av>EWnqX2P+DTa3>JB4<#3ry>pQT*aT1`^`sS3s~}*Abb>)mnoF)>lQxR zbh0_)^UAYcP^I*81;A(;AUjOvkfPpJ8AO@cBI$t8wfz>$7@faAgD#oThwuHzdO31B zu8I|w7X(EaB%H=s=&>MldNH>WHzz327kMm4i-xCcxC}kdG7qExnJk6XG@cV6v;}eZ z&H*k<43rS?4mrL%0D%b$>(Xz^`00Qz+^T$l;x2LDa7QMr2f4JUF&!=<(X*Ts8U>GN zbbZBJQxyjce?@p`lx@R^dS0EDQlkv{V(qR?k}^&nUc?!Fafw2&5Es-9u(4k$8O^0J z_MB1x5R?OL>5lwH7j2#lVsJm9Sw{_0qUNiVc}HEa32ilzgMt%=%=HOALO3Uv&^OhL zSVt~*G8A2L4c>4Ic3|B5?sE;opzY)ZI#t83DD24_ldxrliGI-p;l7liO8zfgu+=3UjOcGs_xl?;ECNG*Z5xujBInqEprnkVXR#+` zDgVIBP8E3eezUe6D`sY=O%mppV*tVQQG>zhKF=)kkJ1p*AB^&;r<^o|rC_S7>al%P zxb@c^^UVRUT9ToG->A6@o`nE&xOMUi|O zN3i-(qxxQMc~6=^qw8U;-zT>T*vh|ah@*HY-;OHELf z$c2QL^OhP#`!ErGk@dQ^gm(Q<0Hj%VabNq=E^GohbbNJ_-(02$!>BLtRCYxgmIKym z+q}G!J@#6K=YF$U99$`H2dx|U0`^^RP-WCM^4%r<;)W61W3qTuDn_a)x&qu z$rJ6HnaH;NumFGiLw%cROwR)i;e;ppsY8P)abgvwTx;+|q*KpTHBx-K`s6-VY_2PT zb^-ELHHL0?3mXNIH_Na2o@}8NvJYOb0yqqFGA3$I)Yg&z{MB>E{8Zzuka2aV9Mo~K zz68#vOtLL!%AYAs9xB?}1i5P-+)oZEgR&Gq{1+r^ek7|pvIN9fM^a0fXhpd@CnQ05 zK)F&u@tPf(B@C!W{)!xLrGGN{XTvV)`P_iGo9?gC2kB~T9%Z=FBAms%>DNKr$B~@? zp667xIRfqyrN*Ub&G{$kNQgE|gJupsr)bA+=)|-)k>JC~Otj!&Kg8g4ri@iHEcjOByQ6k2#~#RK=^R|FL(02%*>z(QdBH=1(wOjIxI8M)OKV?kaM8XBTnqHsxtW zEMmMu0DPW0F$srU0tmItl{@DO22_yv!%sebY5jX{ zcxcwF`VjG(GZ4-hD!r1n7La!yomVH%+i1s=&oLGgZj4=@&bFvN$b)yuvi+FNlc9N5 zX$vB6K}4|yZQODw(;}_(n<_ zf-pO+QBcfRrbH*|apMI(6}sxG5Kx!EO`pgAQFby;l?{e;5Lgal0IefUUH(r>hWh}$ z2c*zVFD@)|cV2I*G=Z=PjWlk^Yw3;<^c9hMXwJ!TbJN)T1RwA4T7#Cw%5q!Yz)caX z*>83JW5KbWhQaERoFV_il?lHc5jzzJ)DpqCVK^9;7|nruZ4wXEiRmZ&+f*3<^$T_X z#jWAYl5_!#RR}}iOX(EUEr%xoCDk_ipK!W>5`Y$Kw(bC`BYOi(!qH)4+bRu1!DX-_ zu(P$dFa=k1{JK$p@Ntygm4sP*6Z(VOc6@SvE}p+tR{&1%w;!_-!9HYQ+bW?A$iw$t z)U*Aa$Zj`FsJD!_xGTmwOWgo^UcBEgArl@hkY4%l?H=iqz7lke?(&%|CQg;ia)fwf z7SKna?G4*AH&hJLy!-P?-d`yNwq{t|QZnnlkQ}Wen#Jp~(Fh6HKU-BmDpbQX1wghxYS-kq!r?haE2^S^(Gveq z|35P3yOOwZbBZ=%^}j788h3OB&8SkAW4aK#j8suO{(qZr**D^ING74ldQYzG(g#Bh z5qjWf9PfUMp|to$HZ>5FCAw+uPk|IS<#;xZIKPr_?jqML{b7wd1E0~-Lz4QfMx}Ye zZQDSY=xQPQA0JpLdlJ$IZ->pIqA?~FH@KM&aK;@kP4;1FW_6}d60{@#-)XAKvv^jq zm*zGkg%K~IV7YK)vi8X3jW72KyXz{q*r`f%+cjd)KG9uz=1%4gb+s3d z9&#F2X|63}>h8g<0Zg`;QI~DxqeYU#GFSumhD+;VsNgHN!RFP0kjk;y9IqeGguvzk zb=Wm2+C96im(Z?c<&;=>TZ)F=aFvwSiIcOCsg9RGH1Xb7N3fumlmAS4*?IE}QEvJ# zbgI|6C$UqNRbh(_he$a^QN-51iK*k`=PV|NB+$pgGBF^{2pdY2&_A})$oCzfjFr7h zAd8Vd;g%BQ;LCG-YfD#osaPgCa3E}M-<29`&>+E*0e}d?V5_spqke$NuwZ8TIot8S ziw0x_>L!c-UBZNRJTj7ue%a(Ig~_!0(!FUSX|^uTHmhIzO)%ttv4Scu4#QIkZzQle zYz+ogwS6EWv zWQtL@RAp}~!H$c(dR@VJPWli_hTXNeEg#`WB$5KZlMcM}4BNE%2(A2n6}eB@8i4~q zdgFB{TkD`^hH@-vp}vumUhs8SIyLY+~oLFaesG%_v?nT&CNf)<2xMZpHwe zxX__e_$-N3U+(3}4dfaffCNM zG0M3}hEGj(<-8J9Tglh3Xzd!0$;{;@V|el^Ax@L%Sl+3xID|v=7(p8Po1+n*_7S3X zt`lkg+JH%_aq$y6Nzw5py05VddgV|nbVqQOV6PA+U0isddRNN^^Y#$sr6WQ9srPS< zqhv52D;jOE2$=_r7EWlsH#|=Kj^;}s<8^0qVo)43v-UFs^AP zpTgy3PaZ#^E&V6eBSB%W^#S(fzE#Tis!oUE0xzCVlz~jn>WH!qtHVsFc>q@sZ?9I& z2@HA+#HPIJT%)T1VLFd#0szpdv-v6J(F*dqv%*6Km}O4;TS%~&|4qNSc z>SeWmX{zQ6k7$2DmF%!WC#$4xDAG}1#BNK%C&nyk%o;_+;GyI{=;9Vkys$!%oTnt7 z1Xv`R$uFBW1fi_&chbsI%-RP;ZCWi^i1Uw!>NT|BQ^%M>_FH6&p#mZS7p*rv80;s} zQko~n{#CY2GpL(I?>K0!iV$!?U(ZnaO{AH;z*u1et62Twcu06Mm4aV=34KPj;H;{X z7syW!v}CQ<`&2Y@aF2PESy@W2Bb$PJ#_T$3!m0=+pE3CKL7bsbnbLBi8=T`Sz~-PP zl)6(}A#6S=R|&f7xd+%*z!b}#F-8pIu+Qq*ZN@Z%ZzNmS@fD>c$bRYMUQhx$05uW# z{TKE7mozcG0?3zG1DxQRR`6zo0X;1rC|B6}2WRqdxVd>buI*K4krQO)*c~)3Gb&{# z)qf%0Do%uCE}*-tQSki-Q=4U(l}Pnh%~Z*|H?33R!g-Z9=JE91so23Tza#%Q z>b|gJBY3ta{C&_yKq>g@CHt$5pym{+7kyr64HU4H6DNKzKGWI`u4Xo~xK|nZP96y> zKpdh}f9j>`q07NyE07nOmn8|)dezLUVF$i9-(a^zxne0l{Dr%c|I8$iZ9Skwo)dyt zR@`qc`|b++V{NHA9*M%#6`3Xj_wSPzx)f+|oJdH4NBRV($Z7GZc&3Jh@Dm6)V;tW? z&}bqAOl}vcu^?G#68u;&(5D7xbF>EBY=n&8HC(&iDOWEg)izt&9^7K45$7Xy;0fdi zO7pLc5KdJyrtTqfsr|S6=QX$`MYsXWF3Xh>{K2}YO}IhOX*twa`&!JYI75D6?LGf4 z2i+h3A`MLi>q4&C@-HqIfvXh52(Ox(6zUO`7dIesbr!`tUua3w`u>&gF zm~j|D!Hqc(u6~f-fK0KdNCWVz35gEWOwVt!zUC`Fxem9d6oM+58^a< z9+&K-F8(rcy)qJ}*7(~3gXygXM&=9Kic*s8VSR{_UWI^1H*aWLO$g*@^`kaS&PG>3a|8W#Xzd_~Du&bVB1}wz(Lat6A#CSN zhl)Ftua%$bt(dP1dXe_fJ!20yb~&HF}irJZnc*qBZN z$CNm_r00pnfbCx6_vo~=QhE1e%2bfF0!SZHaDRQB`Up^61Jc0eSdj9Z%o3vrMdz?w*m=1Q*$%imR8&HYghJd9o08!<770QJ8sxPX}(*%W) z)uj)W!;+{mUi<%+%ujz9bXO>nJS$!pVZAZ@=xH&rBn3wDf>_=!V}_I=olNY7W*HA36;8R zZ~7J0Jl^2{LX|fYNV$Obyg*Kp8A`uW8&AMB_PhpM9naIWRxq014iJ{7!_jTLIJHe= z;AdNWmq$Bv_8N$XkP#OUt7e}7T{$l(Z`udf9I;$V0q66qNhj@oZ!7%&PAoYI7Sn(`Oh_}1c8x|*64I25MZc9P{T_hf z)jL432at}?9UJ{zT5T!B+OegqP|U$|Q%g}cd%I0ko;n%#RY}T9`Sf-`m|NC4(aaTZ zmo?t?CLw8FAVo?msFL>9E%gE9{L4LX_^W2vhP%@KC4+vB3I7S{N_L-;q9$xr3i_G` z1LQ&fUO+O%-6I-M)oRJB(``H5=_L!dOjO^SB&Ksh5f3#ho)P{Md-ov#BG_YLnH>(R zqbsHnL}Ve`FvP^PQ6yTNkv79ZrkfeJ#f~TKEAB&&dTjNN8{MPqJ#>)^NaKzke%|}8LrpD20DNL#0*<@*JWfD^z~OCHkm7kw z-A0F$y@8F(A{eY(rf>n06CAsIhyAJ$WQPsL7c41bR%Bw-&#AD@jrs#opm5BVMGpC) zAVu81vwCqquJ`EUlttb~7iCGCQp~B+yt>ag*KPQ(+aKjWu6h;%u5+>uw6LE?iN-5W z0)~`imeBG`H65_<@6#uXD_dI*E3UHvO2%Ea3HWp0P@q!1jy57E)QnGUQP&Wmb@2_G`flAT+q@fZJE&pv`j;gB`tD9j>h*V=6Ti z3~|7w$c-eLP!Gl03+fV=e3KVDz?7MNw6Ff7U8+PQgd`2f;W*Z;vbGjIB#2QkBn3n7l4$I?R~7BqV~ zYSXvx4l5<5!QN>O{$+RlUm{hZw(N-|_szp~^_acCDyr>fj$jpt@^rV2QK%r6xh>iQ zBE+~F5VPY-@G@{zXoXkjQ&DT%4nB({mwUi=BstD;#M0*FOZG5Vdf=I%#sbEvUNKFx zL@K~7^|Y8NlRnkf)Q_O-qENdXSwQE{LDCm`2ofF*Pw7^ZUi0N&%AoYlV-gGbM8@;g z36ysKpU;j;S(5NBL82x`*DJPB7(G~*Hs(DVGhUL9H^dPr3f%Illqd}bx?u@>i=$BA z#th9!8>!L+)9vwmMdZv~B{+oUJIXg?;TwhN*y#yy_>e?GmfuDw+ zFxTM%nw8isf3Xt81a(YwZtfP14s=7#`s z7ieo3wF|a(U}s60*;x~S8k|>fhaPBt9xJKamDD6dFt_*mp#oRyuaJ0d(2q-8(yM;^ zkcgmB$GI<^!MOrdpv>JoO6s->Yrm9;&79BlGUOU8ddoM);k073wmGur*|kR7AF_8r z2T|WuYshWFEl#OXJ>wG$3HQE}S7NrSX4hoyORKr8lp;j0ESQ>>Xc57CKbyIgtsKej z4)OWwG9zJV5;NX!+v^`6Ur>2@I4^k6I0t#re;j0VhhG;$Hg(W53v=K#rr-RM@nYJ! z&fF}#is2x>+^f8neXYj@g$(+qr?NOUw7|j!ob^X7$v@n%P%DhRM_=pjH)#@LAm)NK zYvE)7c00?;=A_zl0__x;c!3;0Z>UE$cSQ6CIx0B=%G7EJX6n7iPADLO@v3;3qrHPtZJiJq?? z0cd4(!nw}J6sufZ0A}N7k_nOf`B=~{Tdvt)ozBgdvw26l?B!>1?AmySGw35rkm}`ZrgpsY3o?OnWhvGT3T-T98xo{^oF=}4Zi*% zIopX6NI#=_8yRh2CCJBX^>pj|*m?-X^&*)1bjb)wRb$WbQiOz{;pjJQ<_T>kv(t|S zcnP4Q**^7Oo2l~Ub8!&%s$a`wTb9G-B>|M^Ef*58F9`z@;Uqh#lrR(;0g@g^*keRG z*d*(E8F1Yp6J7go{cte!{hlE0(tExj+ky1M?=GC=8v_X>)GL)Cr|voIS? z!Bhos;HcWLF<81-@+i5m%)cJYV5_VTlX~~UiLT)*`+~EcRI(Gk0wcj^7gSV97D(;{ zq>od8P=QV0LbNTOx#T4uEg4$&{>cFT5?KWuy+*s9et!>uuL~_ifbxLt%fW#l;<>yE z&mVNxlhplnOSu@JECuSh-bS902#pJTS@iqfxM&4)Jfo76IrELk+6xQ$tN73?{gam^ zH7|^xxm9-9*W^BH^|7p+C$KF0+}S}|6E@A{mW7^WG924ojHz@YQV z6rm7mkxWPaAqT3nCCh8DTjiyRUQDwvPg+G;j5X-Ptxnuwc^R}kr=v=I$0wA>$HTMV zw3)ixoF>F0nL|#-vKK^@3u2|)P+uNiko0&Cr6I^Ty3Gly;U(;d&W5=oe~qkc&Hx8| zUVe>u-fX`<2GtWje+*4znKi&tczX>-%<k;=noR;{mYw6(GJDnMw{$6n71N#Ewbon$ zGNA?>ay#st4L!QtDrD#;`5vo>1ZHaqk8^Hct%tPWM%v|as9Jn%Z=$j9@p`|Dj5FSD z3OCEU0E0lln@W5V+2M@|vE|?0I%J>h+j=MNMo)MqVy(yAF-+jYzR}8nges5w1FL-S z*m^~d`PM;#6~&2Oq8SPzk_}R-n(Lq-9>YuDJ`;bw(DPM+iC4haNozuN0}PGExTHzc zWi>*(<*OmTiS zDd-zzoFOIj(a}J^QW(@V?~Yv<5Qq@Ii{f95KE3T&R+VnfU4F4!3viydGo>tr2(ERA zJ|X_P;pM*!ZCb#E&2BI%)pASrJfd98f2VD#fr`&6n6-R5j7xb>W-)BvOx4{@_JMfN z57b6T;Ckw4y%FC;pm;cBEa(KFG_QeTnL9gT-*;@qVbNY+8uhKoF=0vNFYIppnngf5I3qYH53NZ#Ri1hg z>L`>ju{(MlW_9_moTw#}L$lG`WjCd~=@b9tX;%DaN!N1mB8GaqDUBk1Kr!8tBZUEa zHAJT3<(2)S%eF^71~0H_2^SOQCTE!~(dFwkz#a{ROVBs9h5DL(753gOS1{fIK}`@R z;)1cz(0}Q706Rd$zdIA!Y&0+1SM#ehEuMfnj{Ep^uS(fD=THk6xY%ZL<*W$qz*3#3 zSV48;vei=N7o~OsekG6Ya5Qlq)%wX>kH$+k@|3|+S1ZB-KWWPs+;0qdlt(rK0wY|Z zZUSD##jlxu@zlMPKEF8Qb!%i`in>S`gBtb^2+CDy_ZERN8CU0DMRB6*;n@TGRNuZf z$UYs~faP{@piXYM@)Yc5faMkc#8WYFsXGRp?^0*;$)u#{1$nbreQA(rY6B_BBncv_ zj>i+LdLC`;vz@r1EU$LhEMrCAq1Q*eh{3YA6#F*s`z3pr-!1BqI?;fBN>eM?Br-|Z zWF3tr!Tn*rLjVFO8=r3JGK16@kgooT14#IW(wn{(=*@JI0v2|MLpu2nR z@u6$e>OIL*m#QXX6W)%}B9H(iXk%qZ2_2wr7HWBGtPmc0-J&xM%igSGr_}kh>usPB z8(WPQA#{8j1lCfw4#^#13cOW8krv>w&IyJEA;FV_G9p6~upY7S?O{F)UCh8HiZ*3j z)d*lE<|?s_vB+GtA4&}55*5?mnqoXCn;XQh{;Ci&{HZz(WY2M#G7)WV6E$}!hJE$DS}^)MaR3?#Dy zb(iO>72j55ye_Gen+u!itafw?>Pm}i{)C`?94`kDwGm9D;JfJ8F&V|-Ub_>M@|?T~ z%(jB#n%XG`51K3X%ojBY*KJpaSYQX-p=-n={D~Q#9f!o5tn6(c3i+V$%U(qkJJD)Z z9@tx0SPey|PLk`D+UB&UD#_tujd0tfBsAo{XK!(faSo1^Bn)NVOsNd)b&9tCIW0Yv;%AtM_iA_7K>1ws6=P$E+v|Nm! zT5Mkq&A5O|DG<;jy%|pR6iW*y1_xFnG1xSpMV`FlYp?X*gm-gojGWB-2s6X6s27+$ zv@bAJ7b!6|B9hsvLw%SbLZAxdYQO_m;KXA9{3HoU>;P}x<6agMaf7*1p3gp_{c1Jf zBxd5`4?qb3n)HERzR93f9)}h-SY{BoKdguK$0}|i3wYNp|KpeB*))oEK;Vx>Bm!OE zRPqj3cv*mWZ_^m3JK#61K)J^XVLOavXPoSFkq^2I3Jzl*c zc7%XzD88D;+wyhBf=fxBiZZU@rwCoKWP@c%pauoK_sggpmZY-h8WM)s#14gsE1I7c z^`Twekr!IArELO)<9m8k9vRKU-xLc>7i8cfWV zsl1o8dnjqSa)l3e#5itpvZI2Pj(jJ47 zMp%BnqCAaUG??^AAao!n=>=qccphd&R}o2HU2{N}yAf!z_t!eFMkUuu(p-~2U==1D zw8J5QnrYX^3Vi!|xO~rR3~`Eemf!KB1GMS4Cr6c%)oPjfrw+vp=Ba<7iv)G(b!X%s zX4)P@K4B1UbLrKRPhZyO!f&r&3ASl!8W?HR#w%SxK=<3UyyfH(f`#)O>*y~7tBUhX zKIdmQYH9c%zYLyD`o6pI=kK>xjEd!W{zu8+N0<&ic$| zM{&>}V>+#VDlKyfYpWilH*WB61HHxkeRBrii#1mQ?;=@_q3nV_!mwREeOha@72d|o z3$%WB@Gl$L5&pUg!$0uZ-Zij^;U-JnOCJT5tacmHHn1-FPNu}@c)+L96=_4Q*mgX% zVXn&3>>vEY&j=J9qWY@dy>~;)7r45V!tq;jHrD+i$*R{0Nh1EU=XSj14*K;9L20+M z?`X_FypA)?3OYvNVt2&Mb8l@ zEUib{=T^JYpi<18yC_|l;@>OL8n-jzZOHo0P^TBy#=g%Ie7*(7yK0NfO458$1~n9P z-zAt8i3>mG=c)i^MvtpY67&>f4a&kjN#0L6-o>#}3O>?iwQ8bh1zFu~5h? zx^tb-D!x_c6Fl_x`y|#7c09Lmw4%~YGu@NLtqvE$-P5~c6ijL9xe2W zIm%N2uTE@JhIb$*m-W}bNSCTH{k9MtA0ctkWtI(|JD?%$bXN1@c6redSF5sM4MnQ8 z(LHW%*52DO9-hOaH&!`L%%2S@iV^hU$XGdB4Zccl2EOo*sijR`M;tfqD%DBeal4>< zf@S*ZdOXA3So+geHaz@r>Fl@u7q4B13+&y6qj>BxFw?H-LQbVn}2V8wLBNAR1Dhz1vf_{bewKYju^L=QF@>gYS`Aa#FrIuZW^eqbQnX-PDYs3=e8FQ6m7!Zbo{n zCv;BA%NED-Q^82pGL|15^g2^nJ?yKTf<73OKP?NlNe1FFF*t|=i*3t0hpDKi^W{T5 z!_D-TiExQT(}P@q&&uVMVb+jO%2{kVK2bUHD&AQEemkg>6NhF?p(fyutD$P+B(<^C zFo=n~a8>NPO%G2%v&2}<7=kTO^>x^nvm<$C?l2NZq0R_c-_sgJ1~EID%{>L2o?I$a za7;n2ak;yqi#NfoJ2%*Tes$p1&kVYWt^f}!B( z!glceeBX@jFXYNe$CI@q4}%4vSV@qDHZJBxC7a=wqR-TD@&?v zbD-(CZPp!?>ti@?&F^0e*a$EJp&|%F3@G0)iTl{F0kPb$GKFU`FyV9j*%*>`sNl7 z7g<9612mAEW|R~pY75iEC{k$6o?d-FYeASwziwCFD&ZZ;PdhpFJxKy?PEmrF|IKsj z?>V|0xC<5tEjM;))SM+OUXZuj&@)R!VKX>3s6L3Z8w$91;1&NbPC^qAtQG*T zfLW$*v?@#pA{haiB=X*6K|;-}Q*PZ)OFn#6bNsC*g>SEBxOgFDkQ}AS=NmVwZy?WW zRy;JiTDzN5;!kx{ET?nCf=8$*hZJd$f~;vo{H*Yc&PTKIVHS&B(!5=h*tixfGrqxZ zjE|o__N&bA9M5z!_uMnpo6C zN&WAEN@#o){12%o=A?y+ho_O)0Y4@22zb690ouU7F!@d$xJS+V>V{a!)hB~ak8Ljc z5@*KT`23jBv~(zmCmm4;4zc)lJ4|;eZpFsJZ>1*xh8`&@lUXK{*;Y)jCo+HocRd=(xZ;JMgPv zN43o9j~46|pBgO&!3^YyFgKO1aXwp#3(%RaG$^)~`u=pV9`X9YTnT$l)(QTDA_m1} zO=w?DD_>t+()_O^(F}-de7}CX;c3s*^87d~7OII)t1~5k?e#=>V%G;-v%|&CCH~)* zI!Eabse&Dk2IbZ_j)quZ4}#JyT}@$VY-nAhegg~njiji^p!n>vtL7gH)ZKFL9^WS? z0$H4*P)H)KrJUpXOXkhV9 zU0spG^zx$@@HU;pc2Vd!XRvwAfP+@&gOGHUjCEzrSPDxNj$`GwWr!0mJ-e~Voi_}W zOQ2irv%KEk;5B>t}GT4mWN_B@*FXAQ{HxZ^hWp*zDYWnBh)Im~&&9i9M| z`qvjvUbr3ND^Yy}zkT-x!e>z{vX z=^^;&xB*_544cQ-Tj-URmmx$_&$;S}VJ2$-Ln7-t9VGr%`yQZOW1nMwhOM=`M%&c= z?y?URSm9W;2P2i$mX(xCpFBC-FFI&}4=A|96pJbCoU~gKu^z512r#cGVrY(m1 zWqZv^}|i=DAfC$So*de7Lmd)y+W=wAL+&3MwV+jfpLdHe7g=F$O7$OpU z9ZXPd7~sZsX}R~PE_oU7F|?i`=O0$eNfk5@z?zIDq-|G0v!`FUAG2CoC`;7@r$W=Q z<5GKJo6J3>_ovjfuC@3Ep5Z-NvF&G!3GX@(BJLhA_dYKV6fAPHt!Um6$WXCqt=d*% zn2F2nCg#t#9B&TcF6)mM1+Nd^6$b0#-=32Eg&FsnN*EDPr~8s?@nImxf;i5DbsOVX z>1>x=nIc3@MhG<_&>ebF4DvaX?;=<*ku^dwiDOM2^9K_^Ghpkxy>#fUl_A8lswD*~ zaAcZJ%Pzwe3-^mTIhI$IXTek z-NFtbGQf9G;#`Ndqq?!AkYQ7DGw_D(I0857{m>)Pk*5CEfZT8uG~><4_b898OABsH zw?y3D-2b5YKogM?LHa(*h*Se(m21K#(_*vdlNS~vG~-xh9#_9)nZYkR0dk?1U1Ipm z`3~11tHyO5*>G}OSAL56oo>ZUT-B<00Rr*Z<$f*I@_x`D@)Pgz~&NcJQ#EqM*$vg-!|kgVK{$Hw1q13#8%Ywo9fQR%M_NUP&{+j5Dbl z-8o3^$c!Np)ZNic!V6|zc#MdIyYGFe2NI8Dnl0!h;!oyjC2RM*tpvs zjEh4nB9P%0iOvj4D%)ytZHp|X0}Bix)=K{sRM)sE8|}Z9i(o@-hYH3JA#|>De%=b_TclapT>L?N$z0PG4)*eMkM?S)b~{D|ktzIKcKH@tKwb~m zhU0v5sWak@<@lQ`0k@9-UK9K7bcnWJ9Ti3sb62XDTKa`Li#&g(w-PAF!gueTMD`Jy z*7D(%eAzwE__(EshU*HnYI)(8K-fT;3`}|F@Wv>-$-B6>(^u+h`*aABEv2qCAvIUn z(QKO!`+_ujqhqPKo%F@pG6*onYC?hwEMt`&*4)+_nRjL93J}`tSXC#~eNRLOQ;f;kWj(G-kvFi%Wd+IGb#-nK>7{ z1VLZp93^!_NH_Bg2E%aq6hFz^ArIFlGR-o+qr|YHCS!VPzcGs^eD2S54nYOY}c2|OP#+mnyny;IMK4ITbo*T5%ORGKJ8t_P6K&` z3lRef;1qKTn1>6V*iY4#M2Cou{IT#33vp3uc%l_#%*;!cvlO)N?g5_>})^c4q| z^UKL**e(*8yo5tei^yN~uC2(l8 zXyTy|jL=NEBN!6bxc~XXV(%*=;)Oa87-akq9Xu1e;^W)HgmQUcG2-`tM-^`_nuduO zZRp7S1qUMVqg;_2`%{~=NOJh505Qes@Rz=qkFoN)K*e%2LZOSvb%9x=mX0|a;qT&G z86sD_gPZQ8{xIy*9=KUxafUcWP?FRSUC?J?B)vA?rd>OSo`7(=(2EzI3w`MCh|s%| z7FjONx9-Y{#s?3CdzP3f)f1JrEh8*DK)1ZmRpAH<8%B{{^&F{5X1+BSseywO>X{hP zf4q?T73n~UH(_^T$NTqCdihRgpBaXSK@4_Bd#pOL4dHMkfS@g(nmSs)O6B1Vv0AWg zT5G}^5_2jAzvP)jT?83~hSFD{6nmc?$hzxbA);)@s)X(r;id#(+!_-_*nY^&w2TpD zczziYRe*c9%om|F>O0u(qmE)aMeEJC0X}?GHYuQ|fR#VCF!@GvhcFzWjD%^}yn3m8?GLBYIfu? z@GQ|ibwrElR+c58BmPV89AVVt?~TX%N+ z#)``vx8-*Ii#NGaSIH=kI#h!yQp{6kPLJ3(*L1Wu-dD^eMH+^f`>=hcxk*)TR_iRt z_<$A6f#7S>!2T%-M{!Ndh99ohWwitjX#vOGUXdj5!-Qbt9ybWggM_>55XFdUfLmt=l7hJtHTpQ2)I&&I-C7JSBEE4;)3 zNt*gF&X^BF9?#Yq+E#d5(}9WSFihT6mz{meIJ!7TPgrBq*_kL&M|KLx3J{{CxA?2q zjX@xRi)Ue&Ieos1Z}V01S&!hU+@Y#FAfV+$pt_o`{i5O${>048JCf;<7cANEtx3y$ zZS@@kj5UMQRaIVSPg;$;>P32tPYd?(-1@tuOIOJZ2z(aI0(lUCSOI!%+YjM`x!ko^ z>;g)K2}{(&-V3S35L({APX*?E>z1r@Q{50af6pRV-CVA>!LVMr1`U8pk`cP4#Y}G6 zyEFX`*ApZs%cIKG97LH775}Z;hJn^JM*Tm7IYK?ok(MKMC}Y)QLLq)!o)y)RTBC1vwZf z;%}jmXOJ(#Q_8vfPHA!A81rRZ-mwC-)YVYb)n2fX+7R>G_BpR3owbdvwzy!p#OUNPV~=2@?z{wW91dj-jtb%kN+N>(c+%nzkmpRtk>BB)-=Z1 zSMLS}lg~pfw%lzt37H@A(lf?n9xyuCM5$3^txa$K`=nohP;@y6sq=*dZ(6%1<;0}G zd<#-||X?xuOCF>&Nm+p~0sh+P8ug_I+dfmTtvV8E~i2I>ZC zvNY|!i2di9@b}IWVw!lgyz&g}DVUk;Wi4c?ye%pa?hto>!N97V8-hqh;zVD3_NZFp zysBhH2@x~CahauFT=ZSW0ZC8($!YUNwa^)ULzXLOxqka);`)I!ZjkmHoHX~W|1!Yy zQB#3|&JWK43m&IrWB4~k2ZNZ%jbP01rWh0R_tfT2V$zK6exM^ zhnV{cd)OzOiT%;Wegmowa~m^s)1hc#2$&@7z=H+w9kDaht-c;((H+5+rChcDA5 z{pkNCJ{f(Tj=TQiyp4B6@MXr;u=rut!dE7|y08&rtk6(YewGkQ@F%01i?;Irv<2@C zhBW`?nWbboGsFd%gJMo=C6vy^L9WUN3lz&Z~NebOF3(4csFUjA(W zs6O%If)9$5yOL47akc9{1TsmmS#UPvvsjktU0^clj;5w7nb&vP7QfNuoPDrD3DNGr zQwlC0VNR4{HP&JjZU+;z;5Q^kCPWc(CSa;2=GjQxCBpI`a7s0O&bqo7@^`m7#cPsN zgcS0(w)Si?`>0azMAcdG>h=xHGNu7cy@56Q!={l<%?ke#`N}UCpaMjx58DM3d$!3V8gfz5v=NVVge z*%s){S}Y7jUv&j$fGT)}(g)*_g`JA?Bu+gt_VJdng4xP;7(U=Gcl(JgV|!Ht;R{3( ziL)_tWUY;?NbbapWtmp?2r?>2-PjAv1PHye5Pu(<9S}Ef9h~Xvg9EJaIfxi(>oL;( z2DvUVu)fhv6ZD#gx&2OYgdQU>a}$@FlBn~qUr^C<0t zBI%2WNU768Z28bOP!_`k1gQ?!<91$+f5QCmh@9URbAMyVbrL5N;FrZdc5}0RHs}<0 zceP43vF6*A0KMD~@9f;EJ-8VCwyRFYf1AaNCRlNC_x&`O`MVyy`rB6Vy~o6nexX{2 zW&k9zmA&mC|A;(xmpA`u$YWO^EU?o{@w}$1?YJ}=N-(KGx@Y7(L-vfx>La9c4N|1> z{J>Xts}<`hE^+;OH0=PI4&%R^g6mO`ZC?a;n-IK8E1gxkW6MSOfdx=YY!Dli;+tUd{wtanICm zNKgBo`wq*#MS+ZQm3|g{+UgC?79<7aR>Cs2mfV7c87Q zBT{Yxw@m*4pJIoxhMDMd(f&A0K%PP@Lli%rHNXK?31#5B$}$;(WU$`B2I?aG4z=l1 zm7?>V^DnaFT=W&N*!epr1mD5=c@dYL_T0LTG3hS3`5lKQW%Gh_-b%jVDGv)iUhJjJ zpeEe|4*Ko;W`7g7ou38`_ZiUQ_VM_*cl=V!S&~`Gk>8Q*EFGov%~I`lfyk_&oMrFX zp+l3Qx)!DLum`|)X6!yr?F}^1xqJ-dkzUR9X`M;sSw^fb8?ZCrn&i^g6>>?qOu+%y-rYjFao7Ya zQ(fcl1I=BuF_0f_!NMe=KI)G*dqOp0H=6TL86HZ*$&k`gZL5j{AElgGEVIz~aPj3K zviz7DN>+;i;8~?LLG>7aNtI8s{DycUX3xh=ZOU9~aDO5Qt);RD#qx8FfeIRKivW?tC z1Ut-+IbhL7xku>_=bun>fb=7ZLIz9~5E=E&gV+5ms@(ys116&wz+jvXKy557a@HIU zRQV^Z=O}|tGTN3N(|px|T_#~{fk)(DV`Zdf0cGfN@2m`?!7`Mkolv=#o2J*FhJg0L zdJCR|r{NsENXMceMc??->s!Qg-9Mb!?-GIzYO!2YdV0vC zJP$CPwHuJ_bjph@SEU@p{tt%IE4Q%31B&C zP5zv$F%Wy~`Z=WO#)l}WHR$-!);yHsqIPY3;%xBhz{PJ<9L}lFo3n4xb20K`w~ca2 z(a3szeh`H+Y5i8QeV6O=xOx}x-%(=)Q#xQ8*KmKgSwLx7$q@-0q0YkDp78LaE2MT^ z6iAw#jv31|){C(QVHi%G+}xE?>54jQ$R>?gxS{v15b*tSx3+BwT4Zv~HEPsUhJ$8e zpL**_r%2_0`2DDiCxu=Que-A}741NO&qMOn%^M{7B93A3P5f=*mP|6lp*(06KX-Y$%R%5~=Eu7qZtF$LwYhL7^brj56J3ax^qupO)* zA#Z=Yud5y^=L|xpeU2pW`#;9fL8n=zQWOkSh~g_#yq_#lf@&AlT44}LnAz_0R;9&J z+^e07I%7Eq0MOcrS8~;n#F5kt1?&Gg(jG>Z_IC$_@s}+1@|`) za!sPpLHZ+bkyct8gRMervhvQ4&GMk+w_Vq}DERrQJ!|ivJsT@$1FQHcmrkUb1tHq^RM`vp9cdDUVoeev`yU=IV#gzUkYlr9=-ztj{nQg;7hq^Zq!hmmrc z!0r77Q?pF38ZC6~m8R&CB==;6<-2Dbdg*`f4l;yW{==3hT?v0A1}Ew-Jp8Jb=ti0F z1Zr_?+gR%lozHvuYEUF*16`cf5KqBBshUF{YypXJKu$Lv%>xCcz1-k{94@Z7M41Yv z#(#Y}+!&QsHq7CL=dr_|(2iJb#~jh9gCm4w8uS38wmEL^Uj!cqnAtA|4;HKI1u^kA zKXV|ek9^uitigLT!=+0Gw+9ME6Pqfb0sm1Pr}c$R2#*wCIiH%br%IC?j6FKj1p0&! zK{*!4N@3>IjX<6)5NMP$X0%|{tmn9n8JSh@61ZY$i>o_7E^-v!8{7(8mBIDjq$f4z zKU2O$QTPYO#xs8`+RFO6!zm1-H#)kbjQWqf z4#>34-U?^Vf4_&eRcbyA*n&EDh@G>+hnbYT;NZEjKa4Q*pz(v#uP2~YGA_0Rl9@*D z4(m);4Buk~KHPq%80}ku9g^njdeC)IobopQNzyK|SEZwCFzKr988$DWCx13kST0eY znWD#%05ERb`rVcR3NM(lg*0DYYcbe;lh6kUQZ*h^;9k<4QPs(R?{1UObZn4vX?wH? zpD}GjTL)do01cZ3@}9Cp+>_JPc(Yj4nMJ&yUnj4!)oGk6G?3of@0% zWa^C_ZdAyP)Ec{zMTFy9jPMUah?9=b@|c>PEaDCHfE@jQ-;oLcBqnuAgvOQ)MY+RH z!@qZ%-sBmHvr1Fv`t{5xht^&J%Oj&1CRd`;rs@nxZ%P(N-HNM0d=gOx2Lu=(Q}Lf7 zbSDA~0SLBY?KOE{a?B#%ED(O(_P!$$`~u`5>fl6x?z zeW1(l+mGv_8$glL4SeP4??f!o#O1cIVgy!|6LAYnhOdRjZ2H<@4c>+tdk>yabub{dOeUDWcnuL)4fs^lZ&~icr?s&`(>7y6=5kwGeb z@*x^rMj{>eN9NFvEp{U9G4`~j?SNZ>6t-paKWI%B^h?Lozyo0Gc+KcE6-8AL0~3~x zhB>pN=nnWt%23Krphmp0kq)}`URMv}Qh(jR6#`X+V*3ar+=k$rEm?vUbGf!{NlB44 zX4tS?@y)pXMG#8ltQf-{@DMw}<27=61kw?eCTycrUXIXn_J5H+4Q{!5rckk#v`I=9i(2Z|qkV2_J1Q(m~V4XW!XWmw*5c z@(+tlu%>u8K%p>lXbA;b0vTX2smQ%cCE6hUksvfbU9O~!So&(Z68M=NzSYC{A>jZ? zGbS;D5YmJ&P2UeaY1$J~vqEt+DMGqLz{KAnQ!v0h2@MPq$slj9XJ60t;4+_luqi+?Wy@VOh4*bN3pHYf4f-p|~cK5^-f zU!XRWEKoaSQ^Bxu))3?I{JWsJcetEfBaMiC-H@J%#2Fn8&ZCNCnQ~m{#XJ@vUD&Jo)l8O zV~!oY@<$t1FblWZjiab@xein*g8(4~vitzd~(y)e!rk3BSSXP2`SCU#8ac?NNlNqe@u2l=P0Lv5NE zn!v_(2hoH02{^`ew2VjD;NW*Rs+V15ai86&wa7txE?P6j@PjKh8xaSbT0luJbXL=8 zau2r54>k-THrk5qQXM6Vn`;D`j27sm$D>O4JamA2jYi1jDR2Cj1lz1`EAKcXDlS3E z(F&c0!0a!q;4>UfncfvF>v4)IWO|JkQtTo`rXe-+t!HsD&SF!*!%@Pzi@89^q*dln^Vrv29J=z z(rnwaK$+Qgi%Ul@ZoKZ@9A;zJ$7r$s54)pzp;7@U4lZ5!TaHT}Z7r;U@j2+>nJ%e7 z)tCg8Gj$4>;AsW_VPpcnkD_5;nDjy6e;3f%zQ`%6Z+&&*pjZ_kayU+i`ji&WEWuic zO>2nB2tSNjh*|^PXTWS0W?lgU)j2XJQ;Lql^SVQMH$GFz@ujjp6NzXR|IxkPsxg&l z)6$_LQaQf!9O>5NZ=^I-1r46G(c&)II5WT&5h;8aG{J2ZB)w~$xeX1|I#*c(wq7Jj zfs_R$pArpQWfarLHl~MqIny;yunZ#AkrozFC=VPT=ySLue#M!u)sLb3pJMP(PV2n3 zUI)YEdjtqmsp#~idCMgOnwe)*xRZZ42K{oX{bqr6J1z$S71 zsqPDIMmb&_n66wUr9B!U3_nK%jt+-Qv<1LQ^X&Q!!8vDl$Jo3{>w!E{9d}o7v4qT; zsYFZ87Q@*0#@}Ql7(~O|=1n|1WKdNSF1LmNKl9|_t)bEhpLFQYx*G8f-~KcB%O?m= zq36!mBD;b+%F~Y&o_}d@Cs4ty9(oa2TVlS6EmQP-!+RWXAk)IWsqQQUeyS z)1mJ~SC{7RdI*A~_+U$Uy`~dl?fEoj^SWf=)`_*f@JNs*kceA+Vu}j_xzyWr$=D@S zgFg1bY=G{P905c0q>F~)4y&BbOJlORhc0EiC^tFRINj#&+n8NgT&0E3I!3+y6aw#cMZmrVo?$3u#(Qv1}Ghxjh+MbU}ppt-w( z{5(|I^=cKiLnGZgH1t28@I~WVZ_*b9{b@GF@wcQnWl~ zFgdRyJUqs3Hp5bBrvuZ(t5iDQ6+<9 zpiAVB{qE*6lDQbMeb&RwT+UvGE#*l>RB;J~aHx1BIA|L{*;9bKu=F_M(u#y7;L~A7 zj6Y;xu?o;$(}kSMl9K12aTT8xK)b;7pNo8Xfc{|58*fZWj`Y7%{0wEu#~8efvad?< zT5vzyE^;rGf+D%S4gfzU_o_t2YYuPy*qtJ9OufXM6xvZPtMMI{1}9Isz|G)!Iw%Ct zQbwXx>zze_n))ja{!U4rcE&@2SN%|-Wzm4CP43Wy;6{Ki_`jrkUH>1#@kUrCsHea; zEC`r=bs}s6Rusujt^g|QWoCS&>1Dh`QwJc#TcE!<25Inp0i)S+)_Y7a?pp}euIuA( zm6pAW1;Fh?CO3yX%PP*|PKkFoH-jC*j%0S`fEmrkwjg{5-~PA2%zoz!sJw#lzVdh2 zv+Zbt6MKdoSa87wY>z(L`@bYeC=%y%0riT4IH>Ozlv7I}WwaYWsNx#W>62q6U21Jm z4F4q;Uqx}84(3g^^i|*W|=^9Pp z!iSIWCqIdiM5Y9(hZk#bx4*T=Vf65YG>YM<^``J#p$S42&5WaT`^Wlq8;kP+NGcVT zJ;udTfCm_Olc*zFSF*0pAv0X@?0giW`)7rppCrl@T$Q64?x6unoUN-nQY4z=v7Hot z;Vlq^+t%Y(uIWIBX{e`?15QTBQ{7W>Y%>-S8Q-4RwZP>?`}!J#24L*bSNVk!xlZSR zGhOiXwxe11<%D@Gh;p<k$1|0HoU^DW z*(OM%D?31HdX`1BKqeCypiUz3JbU7A#i^0V=(847sNE2L_)EyR88<_YrHK@4wdcS1 z(oksZK6`MPflM)`kDiLVo%a5{Y$3TE@}LXWx`Ex_mB=@Uv>^S19HhS>1EIp`Lt|Ma zsU-{;d8#rO(gdyO`ve7S7qMWi8*~q_5{Gm4x6w*HBCR?^_fDC8iCvz3;9y4)PPrl2 z&sG`RpUlRT6LlV(!_3w$L_Ah&?5TzZoiF%-ipoeL*a0Ga(h;B1w;7s7*pf#UsY-RA zb%;SxtzX!qj)*1n$?eV26Wtyq@IErFf0@Qhn|IiXR~abR>GTe4diX~IPS zN@yT3m&bT@I0lp`>b!jS+Q;p^ySwW?^4kX?FF{{el=e;;Z;o~y26t(u!WgjoGk9$0 zrGRC8eETt7j8UrJG9#krUPBP9{|J>?{585#3-9<-MSFEoA1%eeFe*i-jv+X{e~L(P zYsBgCLoD^kHx1Uw$mqY~dp|kg7Mxs2g$T*}bC`FXz3<~%IN)P;dT18c!H{uJE(55>KWsqGAhXy_4cNm1~@GM z*8PlF^u6cdO4C_#lC4RW*cwl}pJ?Rl zk=`zk52?_7e4a;GP0>Q^!l7)n1hFm-6SM z4{3jP`Al9Brh*=5|+! zK4(6up3YId>5Z>*5?l3F2G*B!S&k_$dMm+xe2y&Lg+I+lXkSr;db0s`*s!=rLz@X4 zP`EXuEIJBiqK&%xyoo$c9Z7o!WHaC$4Z#?1^KYAam^}VufDHEQ%6!iG$Qs9IlmUvL4E@#pPX(u$;N$t78OE=Mmdr^k!b=?6TS1dIRp+ zIEvBK0N<8CFsD2m)oco=orir!`^Mq@7LYa^nUsABd-2Z`lCAGeWTdCFIz)wpj&EzY zo;BfiS4F<{pzVYBHEwGp%67x5H8~;fKdUGA>(PT)sC-cL<%5mC@*gBPGx1vN&zA~g zpWaHk6!(!&s^V=33r;bKV!*k^ZK2<(gORLSw>xEgz--ST93}tGg)0Jz^gd>y$s875 z4pG-s6832^SeD7{&qZI3>S>|h1&5vy|FM=(4D9}Jxi@7kbF4x}+ZM#{=e=rQ z7t0NyZ3oO9@f6&Y)rbc_-0?g4*AhxX8Z5VCj2~-LD<8M+*USSi7tZ9oznp8iK!2yW zOsHdE@^j4%R9UQGH61h07g|e`+!$H{3}uBFUqK84R#U@1JEdmO8pr)vDzh?a)aD4h z$baY*C+r8*S|B9MM@$7dD#U#kyoe#Lpoi;?2b~;68@h^sY1*fazQ^crA_>T7!A%Os zsZ=Y_UeByb1P|?yjVx1255q>JefH`cEyAuoxt}xB(A+*q!{Cxpg*WtJp(}pXd#~K2 z?$%;nyXO$@0#-B(%G4sQfLKz|s9RRQb6GWsAnWbyx zLguw4G;@IDkPGpiC7%!cGdQu69vS&NB-^a{c_1Sn>!DcUY0bfGj#HXzA@AST;`MS) z=Ym@UWW_wQ2T#?L_n(a!9{;4Fh?;xkAy_TAGnNvYHhaM*RejPAW$c)Mg*FUlVTJ>V zfD57Pzx{Paturu&X3#MVYA)$pZi6n208>D$zgYAjm~fB7wB;R-Q}I(vaxuJY_6HqE zg%5=o)8g8uV=!dA2_Bt6TwtOWzRPprT@6rB@_Io!zZLP&<*4k~*ZW>dBn|Ibs}Vl1 z))rfvqk2&)ZF6$$e}oEF)BYd?{!`Ve4u|<5=>e#`1&{OV_^RstYhRksI>f9yGisBv z4Dc?7l6a6Vo64Rr&-_{FiV=B=JwBw}&Pf{a|C_ekK+Y0{PYIQEB`dA@ko+R_!Q&aU zYUPa=aD+1wVtOQ(O|_ozJy60bxKU<6iboQO#~x&cUB)lc>#FWY7VXeB-gj`5?4Z;VGb zWE2h=k`V7ij4A@6SDwA2nH~q!XGCH{&l7I)YhDY31h%(;g^A*DNJ-6iE(0N1gkHN~ zuX7bDblP&_IfCj_-A}(qK#5f(^HeC8bzf@6ME)6VP$&Jk1^u>gUKs&}uv5yXGX1$0 zMdRy_#eQF3E)o+r(w!+UA}@co)uH4*zWiAYS7{9?4vGm2@J4XQkmqJ<4+ms!8)OvD z&1{to6V3-}UPgGpE6VmPrc;u9n~Nk9Oc0U(%Hg#+Y~g1mBTz`aTQ}%K;C(p^v6IRt zW;&#`a&L&DK8=Z0@uT(BFq^^g3$r6q{yOSNBs<}B#CV=}wpI#Bk%E>B#*%z}J!t*4 z^NG&V6!z^_PAO3}_LwtO_-Q4UKzUhdR|EkSjG@a)C5F3Uo?X|78nO0ED%r**-!R@- zA)K`|86^?I?HMt;H*X}a&E%5l8Sy>AxP7&w0_@#zQXi$zqE&GfNErNbBIL{q+41o& zX5rR~oP*eTK)NI6niH`Z!9IAh|K>I*jYRCkRMIR#&4;AHB?0m^8gO>*T7b9(sZ#c8qGEOQ0j7b)6 zVA5F~`5s_9xZ9pR0J?W9w@ml%v-M~dw2Eju7ZJSEtv0}>kHJXax3@)|cGeACafzh$ zYdKB+ITM?rzM*Q_A8Sm6m2~{3($uhzy}A*b9k03%O80-7JXhpY33RaORuNbW5i@F( z4E`~#0d6WaXkoE^(n|R>PmR~(818#I2;|q}?aj%^@Q2t%xL6(9^2u;`-CFzcoDDRp$TMZ=dQh0AQ zlC3dI{sg1V*AR_*lNC*0%s|g^qn4Os!?iyc9q3mMyLcJcsA4zxdi{aVE&`@bzwzO z^o`?9zU+(Tm7U9%tsb=p15dj)yd);q|v6?9%ZRjDQU8gm0LPr(o(OsEZSYZC0@_UuH>S6b| zy5TZl%>~-hJj5+vK>>;)tf%Mb4DxXy&?QIg0XRg z1g>8#Kava2<81X718FVTr}-IB`eDjAT7P|t+E<|g)p4pBFxb>iOU9?DZ%olr(lv$# zUlR}$IYj|Yo9ZJz1c90eVV(YMw#t8$4*O+tcZv+KrJC)w!>g`ugUF3H^K}5>=kao= z*Jy>7rO<=yI$qO=RIgs>Xji#E(Ge8{RLRbNf=QV3PzuY&01$v-+vUQABk}pmn!y&5 zz`=w%8a8qq$+phl;){g+rX39?5eaQB<@03Fi&r45U(2w!E%-0ows&jK(*;n0m7;h- z*SsMSpWB|yS}~XZvfS9q$K}M95^VaN!GUqE$vT1&KOT8$2Jwh-zPR~u*`F{(g;x9j z;SK>`3t5;Q;sX72n>ovkTES09@0<7L5GOx%c6>D6k=PO&Rz^Insl4-{OzcRGnJV?0z6{(%3PK1sGOt0bcR<94%^?~9K`*Mw| zUBe4|PTz&4)hfRB$Vq=o*Ugp#=|+!2^ZVRAU?eNBz9iz?m1M>FTX~FJ*A)->m45Ob zy9|D5>Kt7qYl4-qG5+M~X7q|%nP+WW#c^BU1GhItOz{DwZ&p=u{TWX-=UV{~g~7?b zl0KC7m$fHi@gw+_CZmK!&dEF?nU-!{R+%*Y7BI=DSVJh7v`iPLxw*jN#vLakXN+cN zqEVSboq%N4DrTs?~mf`#!Jq^!`xx>_ET0V?0-p|V5hnP z;@|9>fG%z(w7&L@0~bp3(TuWC_;uFi~UZeMd;Q`DuWzq5LLm~WzjfZJ*p zjB$}8y(Bd+j}*J@vZQ1C=!wQTKX#lPBZr|iO_aJ4924WN^ksq4ScyyEl=8y3x^7&*KmNnj(!b!B+X;8sZW5oP=u9jdZ!j%c@XeA)+x2Fw;fMy|a zCV!ygxqf~xT>4agqM@O|r;w>WeN$5f%-Dh)AQS3OELw3wH3A{mQZwSs*;mk@zJbAt z^J5^~tB6Hc1fGa=n_sv&ybwKQZ3(_8Mug=I&h&jeVi}B|X_{({gb2yJk=c?`f_2k$ zNBLFs!pbh_r^O@6p2wxb zQTld+wBAl8I7lUu*&gWj(JmtkdMpzV<2wGTgkTR_x98HJSnOiwA}l?@@FY_AOa3uVK*^w8;?@#8RP0~=auyQ$hp8XzTaN?1r_7F(~=UdI)y8Od-m z>o69S7Be)3bn{siqL43FkLmd^n1Lp0Z5s2L-sGcXau~XCmZ~^~#g*>>9SJZ`9od^h zj-EP4SQwDbxcBcC+os6AaLQAg)ARcFcW9dD2bD4~Dvd`o&gFko8$o7hPA zHtas^IS@T^))d-(r1)l5c5!-?2wv^N84`VBZwtNRpEoSh@rs7p`kaJD`q9e9qq*U^ zq;`2Uy>R~eckPJ-Pka_*mh>r$dvPXv03s{i*9O4Qz8{FlFmAeV$;y8(6mjT90<`!m zX(+S=yB5Cs5Jcb?;kmhF-*&7Uv-%pH6?*jPT|~~@g5#p4szXX)!^zECzPfM^h`Gtg z|02<}zh4fRbFUqrCXRg_HCfP}y1a+^pVJ3J3tn=RtpiXUy6%lx@oH41l?Sy6>Evq$ zfI8i|UiaWVIs52!(%RA(+}1rTBBa(BeTg!h}&PP#;p(|G_n&%Z*-GP!^I=sCt)5eJy*u!n(1Zyu6*1 zI9p5V-+!r{d7>#)`sV*FO>q3pTRr`WEuhjLP)$DVj%J6VkA*Q+8@zV22XrkSMyClJ zN&pD`ZhO89zwLodozZWVOnuAk_9{do8e1|A6I>25?{nO5co5wk_J3wQXr1_TrIaHN z{5qi}6HUydnqu1%SXL8~-(BvAU`VyZoMPXN!rhADv8s^I1u?NfKk2C8^ApNtEiH;; zlWgJkeuU z+q$DPa!x4zf{e6-I=XV}o9RDt%){hSZQbKUSBK-!Tx)sJOn zjMhR2N(N`k&cA>VMGJJ!$<*CSb{p8J7no=~9@aPHA2u;FHAN}Yb^ZS)h>BqgHyBigvo>B4=AFcHFg7Z`f_bE%lp zJRoNoW3Ej@1o#yxgR%6^(S2*n%KRBx=OPx0ZawhZGmmYi!j{T^r=8I040?X8#Y=us z_X>+0Wan%RM}(V}ZRZV?k<4UVaate0rk?n~VgJ|%`$klWhrXD)s)-6M)R4)oPV$(; zwPA<*@YL+da+t?%z5A-da04wR)r~a(@U0Ah^!!J<6aBb?o`z*kE9?&ddCNTJ4<}Fn z{DICfwd=5u0HL0@co6EIIKm48W=AV+-cH*8z#17_K}e#{8rkF;+O6?Ka`P;n*Vjz08?X;8I@9CL^>|`LAsDW_e9=rU@q(f%3QBGEbvJkGHPJ+pAutwj6pZHE*Ct|VMp1utJ#}+z^#@X~ zfF6&dsmc*?by#gU5a71I*220R+b_hblhwv@IE6BpFm)o$H9qz-+{aP^6+~cmQZq_a zanYP(xmr(_Y*|C^M+V>GOZ6~a0IHe$QKE1-VK}Jnnd$5h;`{~bUBs^(7q#Z=+mQ&~ zV!!P!ZMJZ9O43Sk46>&3@Ojmtt`Ps4xQzi>RE=mxyR%(ajOw)S@P-^t zTl4n2@rN%%^MxTjpt`FD7V(0ft%kN`K%|Y0WLO_-1p~%f>Zt0aeu18et3foWdLYOI zI&+E8=OejCG+(`)<)7qb+W`W&P$R}!UUTpJeSyk$Ov@O~6QEeZwaKmC_fT9<(^Zm` z9Iw^$rw$chtHvyLlbez63AHQnq>)~tS&S?f{488L3X6jq{!n#by-lqPs$sGud2bujef%*i?^OWV%*qrkM6 z^{=JJt&j$}jbcwVRe5wOYuk(`S9SsyE;m@v>2+ znu8!C$7?+)N{FZ-h1q4aOP+8L6`1lzcHZQ2cIxY86b;hsH{6Kz5Cs+i&3JN=$ z!3xPZZ*Pfa6yC?fIG)d6e$;i*L@Jy=P2n6Pe0S?UzGhS>m;wl+Nwq z)o^rk=R^j-5sfWkmnbP9777z;L!(t%o;>&lk{OP^ffjX4yYuJ@3(1$Fl6=qm5{+I6 z*OaZ4CD(iH?E;uTAjYK}5yV?ecgeFR-NCQX;#5V!Rrk+Pg?3SG-O$@LTf1 z(~(*{0x8zt&dks6P?Q4)FMF|#**k!*$zpUf&N@1afBxkAx>7>`rqbA1p9Mvu?5HYb zH!sln3xUlaWS}jG;ws~FU%r zZoT(B*>kEr&5hG!@0O>hsd?m9FGiwS+!Hhi<^{QYgnpb9V6x zF4?~T?b*C{ywc8%W4U6p|7Vu3?5v!a2k}dj%p}w60~WH$>qk=?Q|o@zt^+wM6eHv` zCPwiuK#@Z3!%0uQIC&vOx1tRm^F_It+h3@p74u{Ln3tb{d<7I73va_+2J$|@{po=E5CNs! zhb)dqT06T@hG5iIud?4|_Zn0 zrSu}BW-h|^sS!F%0A@Mf7mybVBZgzEO}B{u`j96-7bHv5q6y=Kk0g?o0?i6r8B^YB zsKa@&51s={OF4|g^<{qRlzG(=to>|fdGz(~YFCPG$qi;br~%BX9LoUq|E|m^F#=8l z1pyjbqa*A6N2g!v_T@WjyIu%Lu~^|%WW%`A(+9S2p>vzZEC&#_Yo>|=i`WWka(Wwt zmDCcgnWO$G@KTS3P!0q-3BmPMb@w65GLBDd$~LK?74zH>G=Y(rBqtBvjY;wkk47E} zIaKi@t);NER*}u@06Mfl9$UzN0pg2LH=x$PGdFlLOhF!1<#igRFhIqK+%GGkk91Kl zAJ{&|A@w#tBWom(kc!}4!)-E3Q)o`l_MtHef~{OhiU?dF(OY+6w2Mtr=z05x1@4gq zIqMMVOMBdz-8!&nw(h(YcF`uM+0Vmw;FSE|6)cygx2sOdsOzs*BJ6JsTqT-m!W0?6@P(_W4@qX=!o% zQM-Pwjo>%*GE>Wp>j>fKEUWx!QueK^d;ZeF^=yokAcKz5#hy2c$Mw zPQHa|zk{l*e5T#+ZRrzHK8xm>x0$?+l7lBYCi2#6O|l<#8~zh7Z|)?$rfERh?NA0a z7!vOWh{O5Th9C-wM#aXg6y9~&x?IY_Fn`7vyjjMzV2$ltZ<$oQ?tCdfy|aV_XGb|A zGDXep!|E9#@;3AKWMe2B7oJ(rTYCJbphd*;z%9?BQy;n?B(P>bp-}NR%z&fPrCS@8mcrTqc>m!vR+BMsW6J#iO6B#vyYZZuqwRTLk18v zb+LEL@KMUok`26NTCr?3hrPe%0zwbwK$;Y~k$9=n+sgbizN?{-xOP&5~$_S3pkM8Lw;DvzqJ17p|CQV!S!hi~cb zTyKw3@7GRHY$cTFPs>ieGi1Jwps)}ewu7I$i~8#J~- zaWIgCI~smntn$pkdTbDmG-98(A79pb4&t-L{T3OdCBwhlu4onh{Yxu_(e7G9Qk^Yn z0KR0@PWai!?YPKD&UtK>P}{GK(q3qDm0u&P2-0gtmKh{}%CDYkf!o{cHXMn&hG}xZI+8iw|XLVpt(= z)2*`;hu8`8xN2r?ufNJ(kdQ|4 z_ykRvE?nClqg`#J7=1tWJ`@Y6eN-glUwz|O3be$A>!l4fJj~I(R5fwOqERd z%{xF{-3%)N-Sqw9cnEpp6R7DB^Txe2s&#%S)GqaEK7_o}WWqS?!}{E%;y%o!-cGmQ zU(y(7SDQc9uxccaaoSyNIAzF7nsn6P+Q{2H#|oqNs0@n5ZQYguJpjVMG5q6exx=Ow z6uLWqqaW`UWEIg)h*V$g`?SJ*yooEnn=4cB(5Tk@Us7151GZpU_KthudN6e5ZQmX1 zA<6`1*V7WkS(-x~U?8KBGDn*U_F6Y5YUqwmPP(wWt~J1{rOyB!j+~4e<+Z0HO*l;) z#tzujySMEuSO1ti`6#zeCeuKmh>}7AUONEdu5sZ{x$qSXy5kQ%9ut_I*6UQf*B9`c2I)d=fWa~Se^h=<+-)yjT=>@oFY`x+hp8`#S6I&3nq z@=h9mjz{~~mAvVmM`J;>$m)31T`-Zb<3>-xW(nDX6{WgV+DET~p*1#d+{Jg0@(bX0 zNVs`82zNanX70)) z)EJ8_4;F;FF$adHKimH!0}qp5VO^V;8f@ZUZWLJry{@~pHt?{;3G1g5ndEQjPd1ye zfNM5Bi#YHSFY_kL?yPUPWYBW4u2W^envKTz+^tgls^2Oe4lnRPU_TKS33EX%$YY7; z2|;W8Hp?0ILVhD>1}pzMdDwPpKPp<4MOoRo3*uG&1i6G{m+6 zPGRBD=ZeR_9FEZ&zK_o(6y^MHs|P&ACW=ohQx}4UmDP`j*TsAP*--5Lrr!dVg)dt= z)f@j~;##5(XuprpJ?Y;{3O!e)nenX?*Jl9@h*wH!nxxt?Az8+7q;SgcRUY|%GH$@{#C};lgifE)@h97{w>~Q2*d?F z+&d{mZ2Xe_zD|f?OCl%2T9XSNE?sC2U4$3A>x$CwQpJdB7mWMM&dUpNgvTI+PhEP5 ziutOyFJlT1r`D9$@Sg}?##_mdDAX&mCX%B+-OXS# zR!`j{T|(Ugy9zj&%pf^gFENfab?NC+-B%LBJrlv>;+aB}^ka*I>5d{K9((E_CAXu?`ny(l;H*4~l6&OE@=xh%7u zWi=r>z$-X`tfw4u@qRvsNISC3;ek4pLVYv?^B(Z+2%NBM=(D}lu#?&XD8c~xXaAdvzYgh z!VmupDw+!N&3jMK%n53eovf#oycV(?Ww~GGq8FZ;SvzQH#3jt)?KLFavyj<&DYDK%(VP zzG?;B%ht#05EDb6tBt{s{Ybq_S~JR^y~bs55ru8FZ%bQ!Wsj+on=-n|DhUHF6-CF= z1BK5EW-MW26dsUM3d4nbd>FXIhJ#`v)=CfvfX+D7!|Cl*0C4H0M_u|FPgIWF#3pF3 zM39?K2$||dU@(qgrx1UE=Y(5keiKl>p5`SfH5w7^@=~}yzyM9PN{R1K(`vXbE$cj- z5}hVE_b#`kkdge?f~n;ikGlBZ#^b84B!5gt`8SrY(Kp@rPA@$t4qlKeWHq`SYsZG} z*axGhWA*DKh)PZ1<;`kIz;>oj0V`7?>3AkwQ??b0fsxF2Z<{gbbE&vVKj)&`q`D=8 z2hrV)O)>7BVzW8jC_wFGQha+-A^0dbxIggsXihal-_DIFk0o`GzKuI!tENgJMg z5=eJ)?2%$nXk2bg7=3W-vA46$waC+PI$>~o9J-Oe5pdCGtu+}eVLNS&!KtNJ9uqx{ zaR4tIZ<&LzNmgYId@6K)FPWRulhNBpA{@6Z<40FIzl)NBPPB0R@Bx5jB(hH+eXtSF z4M-+ytWUa*ooQMjzK^r$c6b#0QLBx;L_Ai4an@hbb&YldWW)dOG>)WcqSfvT3yU3D zYg*aU-N5{JhcGg0q=|GyWdfT3v1CFR(>k1 z`T>N{Y79>~(cNa>8E?26(A1!vNP$#!Tw`mf)qzzJY_;kZKjSnE11vhwiODflp<>S9 z!jhp=!o&JCCwKu~b?jK46}~pkd|gg5_@++GFsd&YmGS5p9~`R^E)3)t)-hust~+%= z!O!hW4ke;^l^LW1dGU;(>TZxc=5dkFZEpdW^QG751i$R+edrB*_C%uA_&QIlQ66H; zL^l&HpT13=qQT7a1Qj=$mkzkzjZC%!mNCU>qx#^o=g(BjrffIH1Wsb2jcq_7@G59rF^^Qi9(raeycGGT&l?pHF zrqBtQ$cE(cW0-(be9(Th`Ecjirx=h-rXP#B-z>m$aEZAvTI0Ew7T&pL*ztiBR*+<#_K;wB`6Dux|40 z7lc}|z{>s;qb2G4JcGRiN8w@>_Oi&)w4ZDU6fnS`!LKDJv{}=-HYWi>qA)(|P{Ji1 z7!X5%=q8n8E}50`>x6LQk`+0SlPLDda)qn(l^cw$Tlt{Ht=xAf$pFdD);06>ei<+}$E zQ|AH=P0{Ng$vFA=Tw8;vcQKA%p3r9Qe)bO9Rx8?eR<$m6yCMk637cQ* z$r?eIQu{m{;<*zM!zbWxSjn8ObbM9__D>~hUjq4hEFRetM@j2q8>dKNu z0`p52|5KPD5NL{>TU%Aa%JZX`J=cg3g$m?NG3#C%2+4O04{qYYKJ>IL%$>|0>49b1 zk!y9bG^SN_atGGkIE}Ny=gD*kNiP$i7VMk{XeLwDu=ayHUwn3dm^p1H?_+N_+#F^J z(xDCZll@|gVPk3&??v-C`4&xkFcfT)wcRVQ=bIrJjda7{U=5#z#B{yBmC(g04|^fu zPc2p)Qfi&1%e(lBnQk=U^wCF_3#4d|Am?5{CiFv*4`x%GQ1ZgG9>L{+(R%}Ny?>M**fl1(US$&#JPjqcDMEdc9nT2<0`MvjVr0R zqFxvUNF*#?ayXyVLXEAd)+Pt`&$dvXO~=wdt|8fkOxbCUn_&IYC-hOysW%*p9v5g3 zv{0x(Y^e>Wk2Q_JC7dlR*q4nm&I!*x<})Uzv)KV3Ts8d8Y&N+K7kTa}9b?w=J2sFI zHLcy|iz0UQT{N>Z#-Jf7yUR)!yue@Zq589mMyL)i2ybZiO53cD+b6&K3i^c=0c%YK z-)Z-h^^u)W$?DrC}4bP8cT*)Pes?h4y|JWZ1m!a}0t6tZOHk zxEUdxWeDXfI9-C1=}t__o*!)+8PeJGwQREcuiC9G0x~T76~Ze!l!RBM6u+aN+wMFUh=^M&5V=aMQ6IrR~)zVc;tOUp334v6wy|KJxIRUm`K%?OJ3=wp|7wgpCAB zPVNc>UBS$z&AJr&6y;i}9j{+mrzhLQY6IP$<#x`$tyr6-}pwlt51))!8%IIXw)uxVtr&`>rYt3hbE8Z4R z1=F8T8*XF60?qk;Uot?mv-+PCjaY4=GI@E8T!`PdN6Q0!8MHGq{xsimPH{ii1GISz ztwEp&@6~&SMUpn=VQrsy93Hr$;mW6pEUZnI6DeMtiO+BSbfh?2Z-?X=kTMB^aFd1s zc{B{e8(K(iP-YtIFv;mbYJ2Q~=#%2&^~NkHe;F=sv?gtbqG{K!25PB>09dvG`U;>j zPp7jVPpp8MzGqV!K}X~)K{7#cFBld}d}u(yRLl?)&D}RxR+8tkxxzIzS&ftYh=39N zc7z4F&qqxzUHo3w@B{IC^3(F3EI;VGZ~K_}KG>DnurlK3bHRmm>rBiP!XND1x~HEL zUEAbVo?#T*YoI*d1k{KVyU@~}+p5{9#@xi1IpsN>e3$*#lb1hTC{>@HvR`kK1;Zp% zBU#jTXhgf$ves5XvI+v{^KKB*sM~_7h((w5a7VQpov}_@a{Iev(WIvxW%nC3#0TOp zFUuXeGMxB)Uq{Z9=S5l#b1r(}Py!`Mo<{r~fnMIxLJTfASG^Uq9UV}>Qfb1$`AZn< zhLwg;!>d{2?kSbk%W9k4FhG^PsMm>Q7ecf`D~RLnL>0LP8t4^k{5U;W`B=wV>)sET zPv$wHCnqY@Ue(8w$4=<-7v!EQx>ZA#%ly$w2f{?U$J+Jg03Qm$4!-(Xfsi<}E zGgMhrQ1;dA(YOJFwp=zrJDk@`0l9>DOS~K;3e0{=2sWNY90&EPJDh9~W%B^p@teQ} zJuGO&$7$j<1M9o9qEZ>UR2qEMUo7P8R&H|PV6S$hNdKi5ZSM+Q1XPR%~)E&nfhIg4)cin18OIMXLS3=Jg_ad82X2Z*a$QQmI1Cd zP-pO+_q0sv4^!o(`Xt%=(t+&4?i%o#q*E|&XIJpOm7HMWG)&Mkm*1}RW5BX83$_=jQ_N5o`6$xu?YB}&49Ric1lwtkc{ivH-A#>XwMM_K|LH>53uA2vD! z4N@g3(f5Q!mhEh3!wgCqFrUE!V`-SZT58IN4xRGZo>br}|B*xQ=fJLlbWh+G_jn0| zg$Mg~^aQ9KaniNv)VY`Nx(j+ExBm6DzwYLdS5sJ+zDv?ld0YjVmo~OelQ6zXQ7je# zg}xX=E5hgOrV)|E1aie&U2@in_wg(?D29Uh>Z?_8kF}%^Hf#>U|~;Fa2`>M==Nzn~L@L_yUa2Q5lzz*vfNg9Cd^>`tv+==&`glY!O~ z%tN_m=oLb1RcfnMUyPX-tC>(c>n<>V+=NJj0prx&S+Hs5zUYoLZu3N~Nz1_szSkwW zf!#?RVY*b-N2YtPoP{ zHm5XNGTOk@?A2$%b>V6y7XMJuT~8@l6CoJdqQ=++GV6;aVbkP>sf%3gYt?%<76 z=2In1X1rtc&wG7WqxD5NE0H7*h*u@O!mtW88f=AZ5Wl`bZ&%Z>%LLqAGR|w*&^%+{ zps>1g{LT-8?H%*FqOvXvF1ECm)q0p+TMiGEODsOoRn%2@P6yogltJW zbCu?)^;faGu}n97?xW;__Zp&trj7)4?3X~jar4rkPl~%*Z#cl^{K)04@Ihd$=_Q(F zl-r*(#5^1?wHnRE18VR4y)ta1Mbezbe=#de9)NeABs1E5tU_&0vZ}rrVXG`3Yk5{B z(Csbg|K(!esq5gr_VqVa_`x~liB;)Hs0YA1Bcbr=#>{#2cVF`toc%ztw-}kMeZ3XN zZnQobBj`nl^u%|!B*2bh;iZ=2QXI$)4UIfZBiYB*wzsrm=4{*{1bliS@^cKPuuvbW zob+gBP&QA(4a-@4!Rzn}_sK{?>KkbnKV6ZQa2upFAZ@{3w}GbBAkJxHPe=746GF0lGpawxG898^^i=myNzp`q&%e9iW&GQFYx2H;>b4zJe_R)3I z<-UWuKj~!T#s8gSiXHIHze$Qn`kG$>yLv(vxJ}*5r}Mk%z`cRIQJK{LLZHYzu&02` z(`YY`y7sa~@7F_XIjk`8cXguDJ6$t#$@H8Cu;&9zAZgNso`Wj{RS#Vv}fK2SN8PYbA zy`Jg*^xzbO!u?Z0vCb=Bf4Q?MJgsR?1rShv+@AOUxtAl;w_X5yYWNxFaw;H|H~)XN zkSnuOC`Onc)$HjrEdPDCM0qkR*N9Y-5uJ!p0EAh^lw%IRjYT=yaQW`1U~vkz+YA%3 z>;wk;Vl4Ud%TJmVlZ9JCx4U5=Bv?^iR4_d5uvEF}?1UjE^wNLuY>8!juV6!`I&_F! zIJ9NS9hkanpBjL6DE7Qu<#`C2q@ewH>TfReq6zp8jlHr31n_6jMwk z-5~7GqvPo@KSpQ$UY>B}8Ywb|T5>!j3OpG$bp+=&8V+p5eoIiJm}ILRCs@3hSzvf$ zf*v-(LxoUciC9IHJTYz>{wI0N_>pA>K{Wp5j`bc=ooeu|mlp|7fvp6j(OueXY{#GL4Qg5da^lE6J zOFmP(L8f*^8)6ddyreU<%=ISlVe-EH{v zR_~wo78OndCtr_mwIZ%J>_@=^0=72W6!?ZAmX%NSk1eR;!6HN#;|+2;MKe}s!!`Wu z14b5QZ}8C|$4E?!J?x)fBYYO9g41UXJuXiYvq8o8wOt+vMyk3+sJob;zL{(tT#D=g z0pTQ~i%|eZltT3ex;q6AUZM6IoL03XBw^RCt$vIETo&$eQCrNt2dp=ck()a2cIDc%_^^#0-Mx(swVCNyKana$svmW2|pC-{Ab z|Ni2F7P6FX(fq3Bm9X6Cf{h`$BmGip@bi83L`vtKngDy^`Y~oMhqQ!j^9^oSiAh|Q zjo_{D58HRXWRQ+P^-Br?y0fL>!i#gziT4gx1kp%R&Cm`tMMKMoFKft7b72P9yzNZ%S$(r5MH?$}gL|8jTwr>-H zWSb)tdcY~+2muQ80NPUL?PNPiA!@xNI+Emn6{*vMy%Gf~>{lu!MlA5r_?2`AbyYDUV!&?dNgj3P>_eP@w5E2f+_Q~<$rsnl9Xk7IRPINpc zm)C6^RFrKf_V02OIru79IK3?w#2mFt-OR?>!p>P_Z2tKEhg^<2ZpFNfoepmA*IU{44Ufq68q0jq@idDa5_U4)!ki=E=%Z^u`*=5^T|7D^w8bl$Ah8{8oh7DRe|x+F8b(ct4c7$ zcl=#9{VTP8YmMtNhTt(R8zWvY1QSPpJ~nEuByk1`POo0_4l=rc?zVBk2|~C0#nj)= z?vRd|Ed>nP?$t`v64m7QWzo|LipyyW5r7BZ$r;AietQZBDL3@z^Ca76W6eBhZx2ve z(t}l%_HtG>sWTd{S^dSY@;N_~ioW?p#wwgiLaan~I!#R?stC3Q%P{(Z?7y!pF=b+_ zFavFhGF&}}uVH^jGvym=pAzv`{1`;d{M=t62G|7RwyK8S(y_Kj zGLgm&XhLEiD1<#T6i)sos8n$KHKiz%RvS7TwWP=FmA`cFhv;GI>sVrGr>}+ZXOrXZ zTxGaSUs9Y`+&9|~uE*nC&3eGIUF%)Pu+>%!If=eSdg|v^t!ixjo`awy3Bj$67Y6KR zup?Bn0w%zim*GmRVQNmxS_th)V!VWhX>sZiXmjoGyM<_X09)E|TtkMpaR8g1iS#5} zVY2SzaH*#6uUOaK%Jk_Nw*EDVJMkdEt|96XDaaki=KurBxag7ZcC>mEeUknQfg3#SJ zKiNpy>JT;ZTgCGlH zdjB1SuP8CC7K4Y@=gn6U@__*$kFi(k%*G<>N0&?*_1Cx;w;K~CjT-*h{+!kd1oDNqyT3GeIP-{K-Mc0h9Hy&K zdMg37IRXckED;s)243u?2bRtc`q=>+hYeosboB)hip(mPO&3^QhC)x%3JTtzYI2~p zAnKzU&29#8Eh3s4Inu1KR4>KS`fdd^_#~s4WMRmXV7nk84 zeAN|6X~9B3%b~+-RE4+UpWgCROOeV};DOPRHmO)s zeoyselCjgtkLgL-^zBydLkP1!8cJ@L+4=jJYf})%j=>k`*8=MoB-XGK-hglqe0wkZQLToBO%Guij1EUK-38BTU7VvYj3@cfXf+_FI~> z!D9|qD(n=6hALSWLtYd%^{Hx;g%<+0*Vx|Wr0d+9g|}UqbGyd&Lcl$IpHM69nT220 z)S74&cimHhNB~x*e|lZ>?=2Q5qP{#bdAFYlNfeB4qeY*SLb5OhAG%*jQv*!t<$HC$ z(jJnsXR3FyS~d2|%G{SEQ*^ygFuI=zFHDxy;gwN%)my#R*+aZ`J1y-dkk%DwQv$*1 ziHAvoFqm|Um0STT{yE_{c2zSP+I@xKQT=;_bjz^Vqpg7c5=f0_ptx}%5Iwd7t zRZ9qTu4F58W>rdAHuD^yK6^n%Co;?G=6}Fkdk=jo|Bs!eBO%z(PY-5|Pvn*~orT7} zx!D46V;VW9C=GUP-=$~vs|up0w1Ty=cZ>O*O*W41eaG+A0#rH&@6AjKT z%p;Iqi#(CF0E^k+3q%TmaK)tHe4xUg_3|)`8REaQU9Nj_n|b&q;(rZIwi9_(Pecn8 z`*23i*qVt>TR=ML(`U;`H7s7m;zE@030Tia3}=JzhLCsl1@K!Gi?K5#$>Zj2 ze*d4N_%h7Wb-u*-xl{=$;o_*okxd5NTqT3=vq?(zc*=>Qaev9VN$)F0wT!OiqK8699%KWsqQUugY#MB z;8%iT<&^5$gDtlXgzTCE6`3UJN<>m^=M^A$SD)|E-clNEi1p4O5{io_p_Z`4Q=0qW z12`O}P1SS+f3!1LV}x+={dF)yFj^{q=ni%F1EygTJkE+e*VgD64v=<6D)Rc}x;I*T zEF@)1b?Lz>%mC8Ns#Q01D~ELm(62Gj91pujBW4mggi=&;K`5hzt!J>p4-F$qCVRKi+WzoE;L` zq&+Y9H3Psmh{M+_=!NV3pNskoArv94UnB+)!7})hz-(Rnq+U6b&V90JXRkPQE5WuI zB4$Jbn~LhKCF@ktf_=1w*%eBg{8%)$G** z;kGV0rnZyxG=25RsaT+>0HAo1#%zs9uft$1^ zt0mWtbB%I%gy)u0X^|uH{kQ_n9Dd>kNQhW1lgGeQ*jZgsXB|Ma;FY?1U0gR5A1Fx=tG5qMqUOUQ)MJtN$2ToH%#ww9wb8mLi7!Me7%lDC zcoA+PYX%P@8PS{b*y_@?4`1#vLp1!wS8aGq>!{^9g~xZ&SA%KRork`Ui3StC$epBdO7v6!*lSX9X&8(#`3;bg$)!KD6vX9bODb{L6*UB7u$W9Hen)a z@r6}qTG;6T_)-7sPPbw9C;&^@Ic0%I+_BQLm*u3$-VT&bdokVNg|b?f9FxF7%j$u; z+7w;nigKvFH;gq;Eg!Pul6P!V2yv6Wnwc4`pQGOQVtWX(r97!rLt`Gu!Zn&V)Ts zqb-j{&BtbcDWk0kA2UJ&r*z9e8W4TtQ3;3yh!(87Tkb-Hc z%a0JzYS2dsHFQ;tu;nv5X%IQATKO-^g z-%aZiTqCzbA11Il<+tgR)jF9u_a{TD>Ge%XqucJ}z7rMnQz`w6a5YP1cN+Xj((#F1 zJq=u^KYMP+4wd(^sFFI&wZ?$w)6Q@3&(3QWkWt%0G?qC-Eoj?WU3H-6eI}Ub1MUWhLI#nn=Yp?W!HhMHz|` z)MmKcT*0HKAdrc@VT<@Rjum#y|N8bk06rathGV^p^1Z*@f$`4tH>4Wm;9MANSU1f( zBbZjJC5GcY^`K!8h^2qx74!L7Nuc!Hfo`Is3XQL_j)O0AEtl3Gq~Flh*}PyvN_;RJ z9|bt7I`Pg@`4%74{j&&~WM@(ej?7bZ5QY9)3JHfg8AWdTElr@I#;_L@t>m&7Z0_%& zV;vvyKb)eJcXgjhlBMWY$<)RM^%lF4TgJ>-b54%@aw53$n z#D!?V6v)4M&N}~4Uffj}@2nKr4CDApwM8r7tWNw^k~g;D#qH~JFIT-4-iua2Y2w8e<0(@oakH>PMp+68o5 zBIQ^7o|9SgJry%yx%%AEx&imip~|y2wCu1P8DPdOUOx%|Pv?EQN^GYh8fYiCwWBB- z>v3FoOc}i%`-~Y3q$A`W8-EKN9OV_F+D{-;=kzYX)vU1N zxSFztd8Fq`^AehFl_g3UEvLetGIHe*dI)dW5@{QR(A8#VHw&FNTfyf$;GhPxhQ;Uk9f&H_}1zQaUga_)h)IQG~a zp1Jdfv#~#>uKf@G!S?Y(HgQ~A$9XyK)84KPOkTEB7D1}X!P=gxpLXfqQl)x!c46iP zyLaNP^uwV)GwuHSX%goyg`M7JpkiV~>LkNG5MVutzC6jrIEIyMMfKPS4HeTf{o_1o5?&P`D|WP%xs znYut0$}{5N^h1Rp_%ps0h40q`FlHUDFGu*P7(3jsNXqmbtQjPg-mF=yd%J z&-J~LsOG?Pdo$ zF#%@iHE#+Ia|n=|$lQ$1^QvN({CE0AHr0(Fxf*S7dS8Ru6ErCT3g+7gta{ zPfn^p{-h7DgvVS>4#hVThH#3u*Qsc4c#|o00P(I_pWMl=r;A7g;JD4^Cf>MPXg-RKX;TtnV?*+@O}W&{v6IUP!Hh378l{a16X7L$Si zk7?`75;_SY!NjWDo_t?TouMJlsDP3oS1^#QI~4MZ(QBE1<>x}5?58Qucvt1I$4aYT zNpuSg^JXQu9zyftIzT>xXEa`tRI@27$I$HYFI0q-cTj@DE=3hF7Cr6whil&01UDja zgQXtZH>x)d-^rR9W?qH#5PqKt0;_#PhytStE8>9&|`yLdSz=K zWqqj^WMDEtC88Wxp7OhU(lGd9xkftkz+v$_*@LXB?@2qvev>U32mT9euocbA!`Ap^ z3xa$%4F#{STIt7>%%FeR_D_C_&fAl6P}C_6@WfY?`iv3f71P%K`>nIO#lVx7j)`Nj zGnqs#h(gCE0oteea)yWcmtaP-t^!y7?A|+gIdh(>bNNv;LAWRl4VqnRA6$TpqbJtd zD7TF(YpuV|E27|k)Q1OPWtMsORhbHYs!X=uT+B7#B0e)p^7_ra1GdR`&;;&^XW$N28&+J^EzgY8`rI7oSuUeO-Wqej*m~n35(G%8TCOb3`!wR#3k6Ok zPyG?UGW=4mAu5!_v-Hg>9nl3Xgu8z_hmW)E_MDPO8Rb@N83!V9$!`d2LOWKMK<6gM z_o$gH!o#WFJlZ_F91sP49}T{@O8bizm6a+6Ro#INK~p{aqxLlT6nl@HFolp|7HrTe zRU!IROUz`7{PU^V&lwY^vQUB9A!n3Wd|t6$sC-d%Zr(mHJ~i|5EG$|n@b`G_?ho~s zdV}|S4YJ9@f|IoSnsuJf&?X0S@g^xgj#~Z^d4Ujc77WhkP%mVO9s>##mZtO$itFg= z>3zt#APR|+n%aT2VM^8(4aaRh`PVWvROL1iy&~2i<{&t5zcV)Tg=jiN4WGIzm-Omf zsDnR9;z80>*5DBi$`V5=!}R783LIc|DcnZQW(hYn4~7V}z~wul&)bEy5vP74{~vxd zdFK%y;b31UjzD!ccgC=uWiT~f_*M2tXDcsa0O;D!`X1$CYWiR4+;3#^(2qIpZ-dDh z1O;Sn4H0~$I3S>vmXBRlf2ydBxTkS6=L zs1knX4$%V*_=cQW^A{CKJ&I2sh076GSK0!6=OvhN7B_CnrIu_i!~`!OqSILzv38S( zyyyEocS@QVg~hHEi@4d^ir9$1#jVmt9zt3*L|8YpQ6XT-#s|Rs`7{;O9j(r*yT5kg zv1uB*?Y>e819*4u*}M86EKF)(jjO&QtY+LesUC`Xp^*HKhRV=tQIJxJ7naSpRZ|NE zB$7Qq1F?tydZKo*Bs3P|szItME3lH-&I}cC7i*<_UNr-Ze8}jtqGP_|9Rcn)QNB1x7q`K|YvbI;r92P2R{Sfi zZ0bJ56de)0ggeYwiEEOd}#IMf!A){Vr4t*CtD~k(wu6+s;Fwu?b#AK9BRcEDbSqv5gY<2hY7A*jU4HX#q+j`ku6 z4+K_Oe9&n~5v&#1626>8Ty#t@Z}}O)pNx?m6Wq<#FD7;bLvT}o z>ND)SFLC!bK4hj)DcPsCO{yptp;($yXL{q+Gn9zjmq649e&mv^miRO&59>#~wsY?R zRpaP0{br(QC+4rctcR%mgguut$8-edcSeTWiTAb9lE5)?UHytf9pS@~pQ}czO>RL+ z%taje6RTeyRNHU+>P&0t5QxLgMzoM;vgy*g(jYl_OthID`7 zcMv{!Pf5g70+)W67mKeB9$8%U+U@Tn?nk(g#eS{|nvD+y{sqEY1f}j3>1wl>@fcDZ1kmn zi5y~PI8nzQ_ETV}3!OT=HV7pPzM@+5e6UO+9tRA?Za_A$HD?aZm;~kTzuQ{!)dm}( zy*h0TCnJ;1a^gBo|1t96Gy3_=~D)Bu=I^k#>$@t z=T@#tE=cA#bB)8hJji~g#gRBqunch6x{9tR=yXEAC)`4L&n!xyWh?PJt;>Z*Xr>v_ zV6GwNER^a961f`Ra~-~mSip+b#v41{uy>3YyG zpA14vN}EOK?*=b0UYES8Gh+ljzjS(lH4A}b{~Q)?58uo~dJEeOK$cpX&SQ0~JzpTB z8Y3KfQ8J${6)iBgdc;rMFWkPd|HPy;7K&fGK)N-%4oC6h;HA0wwNyN@SpILd-NJ_n z`(U2Fo~M0Dk66pA+yZkIF$Gy!uv^=HrNCIbq-edF-q9yY6UT$b%-B0;X#jVycKPQLCb(43JQnXy0KdhZ=qHi)@M7 zEj`gleFBPV@DQ;{(V6H~c*__zR$=#BZTjX zf9Do7#*T5O)aZOoZ=yPF{nie@UFXe$LrHT?-E}JgDJFUW3&!+&oltkep=q6GW1$KI6`jK<>1x<3OYgFnUe zH_IM)IqYoz?eMZ8;%NKr-ROOw_bX=<_wD5=|b09b`6XGr>kQuIP!fYb8&Cn{U5$8%-Og=DtWtJF!B6BHbZWP;O_1FRiJ?EwR3Y1Od#{T34hpF z(4&Bx(6CMdDXWdq6cB)`Z`cT<>tIiea~phUilqOG+BIQh9K1;w9O1(ZxbU`TBWj2_ z#sC$E7UKzfrej=kM^~N0Q=oVx$liK2x z=5alIAtNNj20%wd)zzsfT|NT^dGts3GfExdc+mNzV8pGTo7%0oHn2+4T%s}g*++9F z1`u0U$lC;{75^mN`6Keq?t&a~4r(YA5o(a5PoZ_KXf;qNv{33(v?#z_<@{TZER3Vx z>y+6QMQn2p@EOW(!cqPz%zy}P)i`Ppm0?KzD~X)K!`<=Dc}$1fE8vveN{yYPqeXk{ zXc%S96q_y_1%J@86l0?5p@(9%g|{;DvoNF+Fg|rBhB7G-F;uRie&3g}>~WoyW@T=2 zKAg~MrNBY#jxGEYxykQDwd%+OIaRk`E zuVl)hBF&n}fyQ&82jmqY>_&XRQ)J#xJ1l~puS1A~o&@_px=OaaprE>IHjObtR-_ht zhJB^3dOZ~w5P5)4ZnVCyvY4zP6c^zY)y*f9H(9!D^Bi*k_bEMhCAsaBB4x13?K+O%l*HEIj2@}`*b3E(b&rcNiB-$|4QtB(@sz{3n?)_uF z3kdIa$0;z|B^5lRvLIU(11Xi(k$B4nAOM%r3jgr}%8{Fq3fLUiln&48#6~8d>@{CQ z$375ZJe~%I4y9)HJfrF3?wC99q=(^Hh%2tx$unVwqs5JP`|hZLhMUEiH&=WBTZs`* zKZyr2et{Eit9$%VUnzR>5pIoJX)`{?!|UVEQaf90dk15XUy$gfdY&G*Tio| zP^Vl&>$$U`?GrmUJ(OGp$f?V7Kvio^j?PY{*TzIfkp;`pHU;I9C+8Y&ay3XYZeb! zo~@gu*r%xr$_{NpFHbWWb$ShN_tTn5_2!JWFNrT2&0&N>X^;R~o_uXvYh7E|>hI4W zSDQd*>MLegMI#g0vtq!=(iOg&*wA*f6I}!I+b|2tWy|t!aem$nt94-jd12DWQ4Gly z1j`^MmHtn@Y->WGyv0X5cAtzqZL?9R4Wz4YQ!RTM#)|U5ne%H@?^q%(yUakE5-e=bjh}%ScuM6B5}}!TA_WHe zI-qm*E&^L`Rl=pZmJlJExee9GU3T#Q;YxM!h$emSf=*|J4G*9-S4pTfu{)~(u%-o1 zUrT_E@93BcX~4*dRZXV7w3SWaJK1O$FGrOXE>qxi55Ahi zC}o>RAWT)7_tmh{7sX4eWRLSaK`j%gDx7Q7X9A`2&MZqm;x5D)PZyx~CdLPZmo-dh zOFOAZ$@|4b_Rs_ZYPI*v>fS1WvSS~opo4Q_@S;v7q-v}q-%h^xPtcW}>uAvdIIj`= zlKv`8tz)i=*D++;;onX)o4Jt6q>Aeu&1u)e#o_&e#HgAY5z=jlpNKDahFOsP_f6+l z&YhOfm7TkIULp>y<-PPg*O%^ZPHcNc)a8(iP^I)-tqRZR{l-=>2-AT2qF#b}K2ut5 z9!71;$m-ES`M{7$vV?olimq0cNZ)}Bwpi`o3UVg2!i4zVc2p?s3O_FvdlL-aBw-`~czQX|~#BOyg`$?w$VJKzQygc^-@$$0yzrYPF&Kt39o0(evy_ z_R$)}WA}@HX|4h(brck5raw$p1ZU?gb6QcC%yLOZAy#`y36(g1kG0nonjW( zeSDOf@B&Rx&QcYsTy5L(7O)5+-f^(ke0^)Y%aC8JpI<}Q=6CzKYuwIC)f&-&V&b6_ za7j_`Z8v_hAKKa2Y5BUv@ zsbZB7FL)!p_-RlM1HWOh^)GSBD0{*HSL-^u2`~oGq=x(8&)!-ho;Ov{?Yb{LJH^DQ z;AY8~J5j$<<@?Vp5{FALohs+U@_u}qnLqM?g4`g~8Q<_IP&{8tky)WyW7DMs4rb9q zU~&kbwh<=G`hb#ZkJwf8u#9CF+<(^a$!6Lu(Vn0=spvtD^G~~bTD-Gx0s$Y|jRDYC zF=V%*Ob>U(-bDo{5}2odMd^&e);eorI3!53_~UGHk&W?6@s#6ooy*qV(SFEEyYJEj z6ztRBwE=@zlr*C-ry=&={tD%)5bH^+VuutAM%6jJXdoTki+2c8`iHd2s7EBGzzg^# z22d+}-53>lUVH5oFQH_t{-nlOzhYkDfv$^!z6sW`k3$PDxka$hy-{ugrKH{p z-n%rX0{A>|DIx2l;!zow5Q8)`zzyndr%(?($+WQp8ONw0&8)* zkoNk?X|`72Ug)yGkTyrAhG8|&*Zw>(CWg8a95!)|4vczxL zkTLRm+@LA0OWm(^J~rswdMv5>IQCEbWYRO4#)Ttxu&_MyVGaR*$Q~@Ht6WZBYbmqM XBwyHAMh(>&7Adbfg2IWscVI^lxk$Jk literal 0 HcmV?d00001 diff --git a/assets/images/data_science_12.jpg b/assets/images/data_science_12.jpg new file mode 100644 index 0000000000000000000000000000000000000000..02e8107554426c16c1102c9be2a4399d6614f4ae GIT binary patch literal 101270 zcmV(vKCFC~ ze?QOv760n+3*tZX`S1IWeE&}$*8N6*ssEJym-lJ^yV+NoFWp~QUrYbsy+*&kbwm0y zdCGdk|A+tI>AlzO(Lc;TO}`}mfBj$ikKDiTzMK9-^#9|3z<-nfZ~g!G8{~)mzwy3< zy3_t&`M>r)!~ZP%_xbU-_5boe(R&N~k@jEsf8jsUf9(G&@VW9o^}ok@;ri$J zzxf~H|IL2YeSv>C{=5F~{1@dH(ckZX^7$5iFa7`ikM`f1-(Y{wKfM3Z|0Vvn_PgY( z{=fA9?Z0vT#(tmvcmH4dZ}$JVpJAWQzqfzd{~7+D_Ur%u*7yIP`acH$b3gw7()eQi zoB#3tyZ(W1)l9C3Fss4;V7ab4@_-Zvm~nJOQLZ#sId7;ogleyp|6U{-gcdJgF96d{;GCohWVtl)u_o$D- zN}T_%>*kDML!ag1K2D`cI?$WsBptrpa)A=Mma|%jae3dnGkYrb-(nY+OzBmKST`Vm zPCY<*pH|syC&A-KiI0M(+Gl)#krpeL^kom~XA#yj(;;eXj`_Mvb z)Cb#g9x$Sqbw;by8!`*p9+uDhZlTsN)8%<1~{s(^_L7Xg!T-6s=IeXpySVQJ3rmg@rt}PNSNOYk`It^sx;kh*4pj24oKnFY;O!=_KSc+ zUVn%<70DqSF1ZjbcWcsvoxCGvK_p7+N8>|k#aBv=+n2zp>Po>og zxDAYN3r1HnuWPtRD3(m^&Zaf5px&yFVN5lc4xyS+qTk)Y!p7b=@#39Qjth9-`3}OI z#c7Z>HuV?eQgO%Zi^T-mjH}k0CcPD{xCso&ZlKax;7f(tx@Pxr{VV4>ie!2Es8^1w zkK=EL%})W;n44>Jy+)Vp67dCHK3AumM}@}O_IId@S;puhT@|%9yQ?f5 z8{G{4wm={izCrNTcdm6;ffT5bO-qDT40ZeKx&-Fmm`EsYiUuI&Me+fYfS|M#JknfW zfB708VRz;Q-;1$vX^&=EO5)RMZ>5)8lQ{bLbF#U(Ks8cpyF*Q*Coiv>?60@)8! zfgEtAOzOt^fh-&39 zmO-T;P1e4Gm(=PG6`q`VFEDm;flfO3p(l&Jx+pk)~*hU^_MPe?P5Mz-c1Yx;L*=!94)%7vB_T(^^O(G?V57cE4KiM=ykI;&EazX9H z(rt@#{t+||%5C1=^XHe10oE8H(=GyW$DX-oD{W!jKdH--p3bK${h27cagY@Oh{5IW zQ}?H4je>x%)c@$jJ1*ON!{kL6*_`PtXeJj$g`d7O`aDwZ0J%I({p50o?%7zHkf{^j z=Cwyq!~?kL#gqYdl7_NW*%u^YpqqK^xODtkn0ed11({hH??#|w0fYtT5a?%e2ifMb-vVCCl8gb!xJ9e;&V~YzMDq!d(f7%oF(P=l zMk<2(k@7~E;v9(Glm9L)SV>z?yGjmg!dM&+fPR0CXHMY6yo`klzaX{@H8G&Ke5Gg` z2JavQdd+F8h-70lkltYdGr%Z@n$>Mwck-{jb8n!aE zgK=G$`lJNvcBy0gHDoG*>?829W>-G_Ki>p(Nfk|v@KjF%w(1c?XU_m4tGOqw+nwk3 zL&6-(Fx(mzjz*$ZA<00v23_H?+jfr-;M0O35FS~(qD=#8epf*EC*(6f>sX{m*D01u z2ri9Y8!4I)`i?u|wFYLCK0=rAL{X?*z2ecu%ARa?fswDi({U>s4g!gR;ML2iBj-dW z0aLcY=qQkhp+yqzLmdsRO1{*%=MB{F2UAu*P3fwKPOUfi`;F)dovd@@Q25Xd&2EJ+(TdsB`r65YijU<%I`z%vJC#kt6>$OfG4NqVcetE!)NF|-qGZtio5 z)u&9Z6$;mu#yR6P;@8Z+6-+u$xC1pXkp$rZKO9jQo$#@nW#Ze@aIKDBI#~>7sa&2NA;ywFBMcjwF zK^I5}`q#kiMr{OinwO|`NeOX@-7_pDGkE8~3(*?Mdq{l{*Vj9^CsK<2X;c3_Vycl? zJ_|UO*ZFZA3Q_;KV4dsaveO2uOD!*OHykbRi5y}?a&yPOMC0LwHM8Irn-fR{*7?x8 zwlAFZg;CC%rrtT|vWP=`Z+99N7-5oFojLvnB zn)a*4@nr~i3V0TXp4Mx=bDv^v9FKUv^eJGv;+$1xZ+h?`+0&;n~;8Pg8a7pIJ08!|re>*;@Oq;$_7yA1^+&yp}=+3S1S5SY}KnvRFF0i_G z)fllF>X8-xSA`E??`8G=)|KogdRB&Qhe+gscHfQbjbIc3z{rX%CCV|FGPU=jUMwZs za5{{9A_aEIjsay}|2BH*k;pq80wgH3*ji_^SksQ3+1>d4{NrWnd61 zQfi{+4e}GqxtjCHd=gAJ;QU15`3#r@@^6uMxG99G!mb6r!}*>&$8oC!SAxai+qJ`C* zN-&BAZ!OkMoH@v59Dh?oa8kx6;h{71uhBUPHPi@+Ht5AqRYxJLjMpK3LZb9G@t}t^? z%gqA~z}OUU59t=A3RPTwi-kE8o?~gOtpBpi%wwlDvTf7LR$IWvM{JJ4bN}mqRqqhq zq+a}|nh&;`D)W!6y67Zm-hsIV*0fRCJRZ7O{$X9*0GG<-qtJ@poqxOr_Xl^+xtVmp9ixWZ1E!ea(bl0t4IDtvCKG z7c+AIh}S-vugt_~r!R5FHc%p-dK%k04r6r7f+p=hZdFN4EXgn(suog4Nvl`gy;a$# z^f4nKstaB5W;cjXH!AH{LsL$X6%(bmeInRxmRWsCv&5CCrN#o8n<<*kKDDm#+5hU#TauHH%UeAIuAf-j>qhTF{Kraz*zs%KK& zx=lCO9TfUJzrcBF+Fbs{!Qv#H`WP9)QdyQWqjZ6%tw&@K{k&zwcth7y=hP8#*G8}W znpPk0;tF}YmdvwkJmsc;?beHlmm)>(=)#Nzo2K{a@j>Z9XiSzIC-0P{a!)b=bJvw) z%Sm3uwbAR6mWsv=?48xSmUdS129d_$8K%(#u*6K)89S}&I>>QQc6sJ1tgLFOxJeJB zJ$#=~ts`EYlDFRQ(ci|**9tgA%N|(z;Qc6*dl3z-2y5V- zr+%$nMqH5?eq-kIpe#A@+du#Z`AaR51$;90KJu?s1r}dabAVE}v*OqQWQLcql#Kkw zaL87#XbUkM-Mg7Fg@90i|Ahm#VnzXk`1s4fWR=R1dRm(6-10az;2T(a&CmIIFhDgX zQD@Ees9Pm7gpLs^u_*+SSMf;AB_`aP-$@FlK}cBLmL(B(?`b=MB_|=9qVo&(8Y*B7 zP3*Wxb`4cWoTN;k)8W&=qsfxoi?#1>P1nT~0x~JqQx>+ch5fR5@kyfp&HhgJMOoQf zZ3n6ZWDDsS!7A;+MobX;pX(1|(d=^{shb^Z1hZKgRQpG4hRGG7vdK|7BV@@3X+5;t*EjEbQM*(mn2?P9)1%A3hC8ztzQI#I1=yk=taLq2a+A; zg!FGTq_M+;8vNJGhScZ|tgwUwl6R1VVwYg}Wyu-*o}qp@{8C7&ulRhr8^o*_&8oC= z6*xnjn>1GA73)T;9&-KO*`(Qne$p@12G~#2Kw9Ta(yd;ZAYDSh!R`kEm5$&cw0YN@ z7gNq92^GxvsR``Q#;Xc*%Vb-@)is^t_OZ}*^-g@m*>yNOgb@mMkmyUZ(5?-uWzqv4 zp|>eB_)1Z;g$pie$3wz1pU;Yj)!mD$?zQtKph%Tk_#Y$NljFl>^iHd+@)epJ3B)|a zGi3C;r7-p!V{cyI~WUL z($s2NGeD!}YgEHA08qgG*P$qtV)2dyPeJAgnVFPKjGUd9+ySJkGurF(=CGapRp^v>AktS?JKbndTveo4CC z%o_D>qcN0Qvf1(_8@`Im?h0etp9%8vrUbf+AM3^_cPv;Y5Sw+$;&ggQ;4aJ823 zm_SbRWy0B^BtjD*|DJ?Mzj_R{V%@v#B$G=x%A*5+fG|RI(yO^g@5ttwHD2*@`#O)z zYmQEB3>OBC_-JN)IsVzDok09M(4kzbWc64%1`7<+8~NKco0&(xl#@FOC&cQn4^UZQ zjD>cG$z+!Yh*guSR}dHH%lA`vzj%uVX6^xk)72c+^8jU^Z;#r&55hbr&y@yC|Dk5l#7QK?57= zOGnEbs_?#Xaq0OmP>O&>YJ z`_!<`Qai#+d#9~gtMZMxaD!n1N9f!DEY8Yi5lW-CzoCWD?=tx9fetuIF!+H9WAJ@$ zYu-|6q7+W76*`{(mg$(per2oq5fnPO<>WJ@N^#M}cj3&j)`GLJFD=vaWwzjv$G7V$ za^yJy#@MU#1MnzX$t~B%dcWB<0D<75{CO!fEeb=%x|s~bfcj!cpTHa+U;$o8^CK)c z0Oh>4)DPV*NebnD6pn5_==q5)NM@?c@GxXs8cRVnI^*7UmJNe$?*KOe;dw*>AdCoL z)Na{)K)I)9ThhjfPqAv1JX+?hsy|ha#4o)I@BG9UO|lrb8Q6&#?G`v}JbY8Riy>5> zQ0P;my~h4$S4ou6`pd0(dvhcl00C&p+hy5J#X-;u;bomy! z&2JF4z3tg`qR0sqt@aA)KfE7eIt>fA`)VW9J=cO+9{YUB;PW4f!Ao!c67N>rYPZsd zpr811`e)!ay4xh7k~FX-uqWB=85w-uXo}e1B8>iC0aA}5SVotb3~N%|P6Nc?nd}YQ zX(3QC=#(A`JpDkls(bkV-Kz_O9gJU>pJ^*VNJzF zO>{WLy&#m-7!&^>8_y9xFAVyMj?*Ox7*A@_I;_`UzLYHxL&4j{u~qJS!1P&NY|X;R zwZakk2?Y?8Ad2NI!}>erepB6yjyfD6Cj&G*X9(P3ytDyu4)%yBlad&0Y2WyEw?8{$ zHi(wNm|Mw)AF8MNV?aUqaBrG#wal8;?u2=b+kb zhtFB%34hnBHYN2i+z;c1A=fkH(|yMCi=IeaIM(hMgQzXUB~|-QFta@VNy1O$8-I4@ z#RcCxQ`Bv-`Ka(RwBqF8T*lTh_M4Y}UICw=#lY29_Uk+%RY`a-a5?rJsz5A088lbWBNv99te|2s%I1$wU1qW_NOWnyL>K%Y-2b+QbHVW)-vDVnl{G8YypkW+p)Vil$RBHjyUlTq9WFz z$r;vIYbJ?DIc414!N^H*H^c?OOjjqY3WVJ!=NyW5+?F)$wX^IMahm%}FNEm7+9kKutyB@fDeRw)@;I@uTY!k` zF^+EFSe#IC#VpJn4~3TFy8pae)Ap5Ij6F!_)&9Ti_|AsegBk6*g98z>>qGv$o|rG$ z&PJ?6kQDp%57yPTB=~cH0wBlO&4+Hf_KCpnVgt{ zvruxl0R%SshCata!DAF|YV~?!1)0q?CSsfGV+s>_&R>sq>X>X-EP(33*%P zLi9!aiW4~6=>@(R)A~u?SAZ~=z~LfYKXT#;{yG{)2RT#gLM{xJ`Se|w=CS!3Mb&_1 z3!qLTf4`l2k#im(kssdx#*WtdZFk1h_fE>@}so^E=1T1d|4H3@W zO|Sum8q|CC(nBdk)!3hNw;{67lDzrd9kZ)sVeXs30gj8o#5MzF(q3w6SH3}185iyqLRhzs`h&?^66{zFdm`IG-zFm!a>+Jc7vBrqg^+i=4-G@i?~wOaPQ7I2Bri zTH}JTQCc-H5{9?ABlX{@D@fiULwp(8=lJ+}CY76tc|7{d{x1OjNEGT?KN2C(l4J?g z& z8fCk}>Qj(eNV9ElA8LG@2jbh+`5T%^-oK0?gy)uKrbEszB7UyP=NZ*nGdP9PSLS#i zzca#53f1Tt<|kD>pT~ThkPUUsUX)NM#>q(UK{S(8gI_Pq`!y&C_wdY@{J4c=j2j2^d1>F7?a$9;7AI@=`J4Z-2C{3N)Z}PVZKlCrH^Hkw z>cvlhn|m5%{&f7&D&<*UL2kK{Fv2Rba4BdT#D%2(3~odw8)?;L+@c0W)|*dZ(7(}? zJOxjth*u_Mo7L6Tc64bc*}ur_7w`>0S03?}Jy}=XC*(?7WY8G53V*Z14*6cqV1GTs z_J@iOaDk$Ui_xyBP7AN#5?-2qz?iYJ7PZKW##&}3W#H#YYsJolgU8!0)9LHLRc(!m zb`T&Aj!_~qE0JLc>d#Y<`zv8@vRV&E8uIk9sTdYpSdctV!`ch&gR@oXOS7c7ng`ds z?%sRBUoZqjc~HK280Y6%YT;=9TX#w&pRloKgb5n57$P-BBViHM0`!M5t%#L zF22<5^N$%p9)F;b3df#OK@(0m86GZ!#aD+b5YlEUPecL2k%G;xevxoi6Gt``*`64w zSoi84rBhVo&`RD0%f(>K@-?J6w1Rts*Nas+2O2UiDrJ>8SXF^6{;B{i&3DC%Vkfct zoN|R7lwzy93E2IQJ;5uXe+$mIVb+gC9J^>g>pKfF6Vz7>)c?kAqx(x`bT5z1h5q|e zCut-efDAnVHSR-u)wpY~0Wg=Fv!m=to|6CTH1y7IFp8)LiM_29eUMRwhdCflI=I{_ zfW|v0GO>k`QMtx8db&jQtKw(O0$g$~Er^yHk$B$MR|C4czb*hzY z&=fU960T%g!EW(6%>B{&P?y8|5zpIeyZ22}nX2WR>`9g(Xl$b#K_3ute(Lnj@|QjRTuSVmBifqP`@I7350SWZhs+vK&&k~?nz zJ3i80$MP$xtK6jL_$rR2tcdTfvFGah2ZbAPYlQ+?u73caUlT#g+Tt%B6r*a?Es~NU zDK0!ckakEmM;Bf7ncYCuwhxV64Wjt^@da&}*6~CmShfRwZ6O%JYiO`>A7N)Ua#Hp+ zliMd6M>PkYX_NcuC^&2ZEAAh~wr%-`OS-C&v&e_$l{r+JYvgDnXh^UtqEr7wRpkiI zNO}^)Ktgo-EH@&!QP^cNwghEx0cozEXBR;=3r@Wh69sqM?rc}^%0<}A z#_z}fh90fwp?5~}WU?GlLQ;dX&=+Ai^}G?ZRTB9lfN#lo6L)?9DU2p))-KbJ0=xxS z^)VOpW=xh!k^4-a0vW}1LkAU}5MI}RRLUo>lzvWj=?mqG(;khf<}6z@_^m|ON8lkd zny||jpFzZuN^F%cbD*LagWEu;%GyE1M9r6C2U~t3swFoRZDk| znqJ|kQ7ly^cI7Ac{ampBmftsivQR@Y@YZBie0w({iWT%J+EcxrX#z1mJ#Iq0f*~0; zEc>5P@nY8M;7(fg9b02IN!R%nF|Ec?aAn-Z?H$Kg2KJe`kBv+d{nKObLy1GD=Q= zxYUzb8QD@`=9;5-)YU-xNV~x_s?YTB0D=F#E6m6k*JSij1ECdGA5mxvHn#k6RojCG zs$cb<@|tV~>MLR(CL_9z(UPK~V!J5a(ujU4Z4F=bo%uLemW$hJSTjR5s!^&xYs|su zF7~RaD+x%GKgdq?BdKN`Zp-}_vb;O?Q)K0$ z1b`!xg4TIa-nFz$=wS;?Yk;x(tFXdNdhnz2^rNg4`|HJn6v1d-7k!)A(de}6@g7PV z%Ty2nJAxOZ=`o8Y^ z-Bg8eJVQb?D9OlRy9yo#B`P>{C*Z3Fqb}q6b5PMWy7?PWl84gAsm)JvEe^M<7R9;ONz;r?|vcF7T#9KXBa=citK_mZkn5X-RJj0 z@r-+aLJw*ydY0Ni$o3NGV&*+6!pj)4p$#s#04D1IY*w)&Ynpuj3&msEw{a9XBQ{sb z`)jh7mHa&95mIEyo2GlIFP3gGt6WyePukDhWfBrVNcoYe7%tXi+WO*o>1;D@}2CGcs=;;>UNJ`CmHd5ZB zOOK;(z{Q~a-f#=NT2fV?d$lRWC71H{?1N<(kFcfvLJ!(|9zxp?tz_K{UaVhII*}UE zSSO>ef4koDl!rDAy9uIR1U>e2T$ZV zG5qbfS-2I}v0{>u;i^vx7Fy62ByPj_)*8Xc65zCytNK;J4qxf@<-T<1uiMfke{C50 zB3uyWxu6$5MocJp`q#M78?7^<5bbIvk!S1_LqoPZa8g_`OdLLGlO0Q36m8QV{i93v zInkYuO_-(yPWQ|Om{e|d`s!tq-z8okn9Vsa(FAFkB$1kCZ;g+;!@j*%4f_4+Q*`fP zUs`_~TKIKm0^ARUkMgh~r5#}*1F(cBp%ohPfbd@(j+Ry0g~x#JFEQi*{~cLfG-rZa zRH`%$Oi|eumyy+}tOf{e(&bIveXeWFr;V=t1MF9+d#fT&I$)FMzOH@9h_X!*havC% z`;3(|V%VM(v%qcbryA~zP5ztpZ30x>^@nj}Od$?mZP=b8Drh9gR(za+2Z_7G$5vU9 z{gK$naFutFcZ2`BBI}+*?IkTO19<7pIlGX9qIxfKmV3dm2E$+0ifGN0K_j7c#+bM4 ze({q!qcR*WsULE~^j?#TZs1gZFEjY1xqPhdQ}aw{TaJvhHe~6-CU&YXhw^=})7!Qq zi1t3AQ}!!%Ca)C@+dl45&PK53F40fZ+3F-k1l2GEOVjPASDLtm_bX0d&=6|dBwXd! zn1Dg%-Q2d^8}}h5M(Q=lEZ~|s6>l>8+@IZvt|(tR8aC^Dil^M8KEBmxsU$vhvV*X0 zj0dagyWjm0P5(OGp~WaDd|}woPdSbr5Mf%J4uIaJxw%qAjGp^o6eT7i1D41ZK-0ec ztK$$z0q|8~^OH0@wx4&jusZ%RjJ5$dX*57a9DIJ5ud!>PWT~2U^dj&)8%GA1cf(7# z_Z0}^oVr;!a57!oQVPGXXI|Gpj}!GdRk(Kwa_1i{qQIkGfskp15|NaOF@0?V4UodU ziZu6GI;mF|Y+s*9|AvJrK7_qK=;WUByvBcRuR0gWBl zDn>3tM1{VO5*$`3csj98l@A^aeBLG|N@Dea+%@ab`dT`o7tIZvd>r7412A!JZ|>u8 z?Wx>mnnh>beaL=dQc`!kfWZCX9(1OqwWc>&x=3Qsmnt*A8w&Y>v!1Rd-|FZYF?|lr zCe1-?v=My?t-t*uRs>lwM7NT_abJ>W_vy}?i4M(V{-HD=R+a(+NOJXe^d9y#(>Ju9 zx?tmN3D`O$clQuP*)@AK(jfh)5ce$mK$5s%-;6z|WrhhL>JQMXL7un#f|4dr%o8v2 zZVPwf5==S=VkggWbs-*A#!C{>XT4zy`80h}^R8BbE(k;`p z2L4k!Et?#w$v8HJPnq9i*_Hne@a^kc@8mcZ$@gW;SO| z%cSgd0Q&vvX^1|RAQ2+cmJ2+yHruh1KGc8!0RI0XLelLDHRX9tg;^Q^h8RcGnD7)} z=u8bEC?UX@^ny)b+X)6*0s(jgJ~>v0FD+&!XZtjCLsKQ2MI{d4`)?jMem4#-;lMJ@ zorM|5N3Zb!?fa$>!@Gd9N;wHrw}gM*LjA)w8S~v_q7Luw$iCnuliirq%D-bq)DV1a`ihfZvpMRj{Ew}o5j`XaX!BLBYhT$dg5>Gd(##>|y z+j2{SLj;Fkc|`I?;GxT;19H?69aMEcQ!9%-fa%OaBVNf@~MDWK(XoeuwRt z$T67ve_hCAMeW;~LwgF4^*@{gpNl$oDbFKRD4ebf=tgGtDP{%$+>!g_Yo8q;OWAG9 z5g6%~#R5z$5>cA%0sBp!->2%kafev467-?t=j{ILY{7(Q`cQ75Zd(K-jltg!Fle;% zJ36Pp$CZf3T?cqrd{gueOI-7JukJn~UK$FAF5Pkd@G9-Bv1}dtoAP3`>&}ROrRfT+ zSdY}4ECqTER2`)=8yf!brry{)y#<1`@QYZ-J?veHn+OJUmKO2iATOfCNTN^tQN28f z9~J|5_G*SnSgi^8^ioH8pG7b+^(AKcjrj3G;IxsfGaHN5Q3R;T=oR(|RZw|2jGcVE zhXMDIc5o9`0Rw<}nAqn3zA#obBRF6-Vde|{bzSIrESj9Y`Pa?(@M^@V512sDCO2Z( z=`dLb=0BDH*IQl?{yfL{(XtN|f!Q%kv@H=F7XmuxcRF{4?yMxmqTnni4AMOhz7KLhzTb~+JZmCJWK9fWsCmc^Bi@#Amq9R1JH}?4 zYXnt&a``o0SMe@WC-1(zP`zlHaq8_j#HsQDX01q6Z2K5z9Wo7M;RHH!+odFL`icMT z@x=u6r_HJOSB<3IxNuxv4$jx(NCvV1}wyCOvjvd5yi>hH)23q4Rzpk%;NHhm-ys}_N_zv|a^3h1a; zsD`xKEs$g-T~%WNM)aCtaXPiiBj5V_$~f=Vg$+$Ml0S4gH1`|-R;pWFvLsSJZ5>N> zOwP-AZeieE#kan1okOZz=`@Cz`=X%#{rhn7e(u?+BGl~(ea4IJJ6)!Ho3U5ga-hWB zD3&1SkuJurg7rkgB>V>kAUe&9@SK2)a}7fw=l~}hBpw^~R^1C$T08%eS_#Z>_FPvDFqwNRFT9DjO4NLf-_ zFi1~##~)=_W>-?JeU>EMk+Y0n1|sOcWc}0XJQ|uvxHD?`9;4p`SqLd=8~sT&8f|qu z?PSQ&MK=^>N;&eyI9 zZ`<~Izznf}xYL3``Hr9R5VvrZy%=qz!fW;o!Ea5hxJOv*a@-K|;Wi`S7Wc&WaSTob z__Ozf6`F0K8>|p1%+%2o2>VDYR!FhB$IRrwVxMNV$%o%OQAuhCatMrwHh1oW5&OV# z*4vGG=v7`#axABM00Sj$0xBz=z?>ELC0_gD+VUy(l=l~f8Nq1C1}Z$}(5^BkfM%yQ zw4OjqSLi`bt2^`V3!)Esz;v3nH48!DZV9vwVKzYOWx-X`C{h6GclaH})(DPvDeBYi zS~TNDz@xu+s@LLxCzeG1@F}Ly$iD5}IQEgbmahUI_uMe0A~MFENOy?sKo3|BCUGik z>+VX(M?iE$W(GwdUqMcNX2tPi_He@{rFUwyA5a&$z&@J`(DDJNM`0zgaBS#W1Qd3- zAQ1XRjb#|yujGVEh;TKDraC`FBo9pap|n(T^Ni9@#$=>(cGw7ce_lwI6A%lIoKkyf zo-Hh}18$s*xq2n(dvi-M4Z*utTfVg* z3x$N9DbC&@_NCkM)Uds*AIg?=V~R`Z5w`KYSHY~qDt47T2*o-vph-4_ILUsWJpUh; zi~}vmS0>xc6k!*oD+7DYW*s)#`wq^2)XNz*X+4qM5b@Fx?@?6&d3VUC zrq;{P+{k*$ei+RHi1In1ad8+5!tR467R9Z7s%IvLuT8YKLXb#4Pb{4?|ZRb48WM? zj-%!%(04xk>+YdqzlYnvZ1GZuxX@n+%gjVB$P)Yo)oR(PUT>; zm6LKCG(gu44<{Hwi!@A@q2Em;(0*IoMbqk8_0_vjo)5iCjrDE(rd}8B>2qjka%s0^ z3>Grhn@w@>ECB6m^QX)P(ifqzqe-Rza`#x!B(n8_gw zzyVa z+^t5bH5740uM2J^xWPe&Y*C3r`18fI0|uo!S4?mdL{kzI{4KbPRXTYp=>b2zO`z=C zz05@8?dqGb8XI=5H^8vvfF7;G(J)BmZR*94kpx*?A((8nnwRx|bGLxQOeoS6wiY;V z9HkP7h}_zLl+<)INt{5nOVzk$m6O1%c!XhrvxdG(icy+$Y{5@D{7}m;6V4! zbtG;C1!#;Ty65c)N*CD^>RK?dzXcua9a;ZP#8?I#c*2LIthw!%nf!2@*d*o>kIyGB zf_bG0oJ3v*D@E(qB<^|HwP`?J_CrO`p*KN0qE*2rvupL-9PW>_I;5a_*92%2N9FIQ(71sjCFn??w1;1VG*};lGC*T4Zk0 z9bRwVB38>?ZwUKcwAQoc_9u^aM5k@Vg~F-G8?&*(9qwotLfPFyhM8|}jQRilRe|0? zvc7%sDsPZjf8xG`8gNth9MfE@0=Ex{d&rOVjH^qfD)_CVho zz-oaE5R_T9up49h`-i~*3t`hz=7~-`*ExQFI*S+6@c@}inQEj%cGi;2pY7(LMsqFQ z1pF%qlQUT{I-L;AWbS2iYq|B(&RCIL-q7V&nJXytC9eL83bgv?=_<7;#k8v2^_;tw zqJUKBW}uLVB0aCi?2P~oHF-9ZK45#$Rytw@s&m{9)|wF0&C2~gu8sOY;_z70LP^H7 zWNpM-!%~`fmvFJn7sBXvbjqoUGB*Hv{{G0}DS4 zP+uhRkZjBp@&MdJ=_!Z)c^my7i&&}-uylN!p7MjOV5>Vdq~%ah0DkBQ{3+kag7sfs zt1ufUWB1*H%_4TmWkgwd8-Z&bxhB;M4Lcnrhh z=aV1UJy4Arx9r|V*X7^zr)9(s97`5ggI9?8syCs%!As2~2Xf2ahn?|QziMO`Yw}2n z7e>r~NSz)}TPV^Sg(MQ?ov*vvfxSeL)rmGSHa0V}9_tJVU6|}#a_$OSH7E-_x!VHG z_HnJg#x+}zdHOMB7kH-8xMMOfn-PI3b`E$2oob6O zxED_2)6@{&GQP+QRvE{2SPu-2O!;xf6wN?o*-UA6!A^m{#`aE2td(!F)RW0Mvbz{? z$So@fGnn(;Iprxg(tM$X-#4($MkGhghD15-MlN(gr=HkFPhj;XW4B}JFFgiv2gG9; z0k2}{B5Jre6L6l~_`0^Nlfkurd&XDa!4|!&G#|V|s(+kIw%{w=*%&?4x1v`0Dy6KF zxW>GDy4}8{3e>xQeipPrp>G(H5d>1V$=DIDpCAlu__HEP)Ecm$`l%J*eKb1M(HS6; z;;u^chtcJlz3Q7)(v08*07KNjpyZo3m*JuQ@h)I$FrE~aZIw1Xx!=i1o5HY`yF|tx zv}-lu)I9b2q|=cBLT(HJx{TPrP===3u9!afI7 zD5p75LJuddpvzr8er>1%ec|NkD2WH*bT9Id7%%H;i?p|9qj;-~x2OyNDEeqkh@(a9-F?45-)g06#uG|yy8BS0oou-oC`%ffE(8N&3HGC3A|Ib;hE0nacHrJ3#ChV zKj73p%LQx8B?<-bpVXO>;a=QBOnP za<0M_v_*;z&(yXJa1*rYRZs+78ghK7?s>XbL6wBG5&kYHe$E~+dX`9aXE`+20!3@~ zpW$t@U!4v=Dg@cjO-skQLpn?E%*Bgqc~i=+xt%zH5>Y_-7~(hiAWGLUT%hqyMzU_}R$3K9V1LSrAG~V~(PAcQ^ znwYh(MSXQn;9jpEjyR$o6I9a7R5xA=g05?F$zt{8_C6RV--!Q#r(^@P3 zU)k9PaQkY-Y_Pi4irYD4UbB2d}5zEPQay4xbdw4G8N8<_LJJO{RC7Dp5mz~uI9$Kj&;vUF1 zQo$uhK4(Su8VUbpCx8*Gwi>y*?U-1v$Uj`H=-v390r+SBg6>Aq-OALenhgU<%I9)H zQDQGMsS{M2QX(DQ=z1Y5%NL>*U(AZ3<}^e%QSaK9uqW+}pyg45N-a#!&((&xv^lhukl;chpYI62P| z!e^vJ*>YxBG9Prw%q4VIKQ^0*d68pCH(Kum)7>(L76CIS709uvQ-JPd!Ka>7-evWD zrLI{{|8rM$t9_amr!yi{F27*GM&enz&kpkiS)R5x6Ozdy&`njC9YfKCI&OBF19}!L z1tN#{(xLro#8pzwi5_L%S0XMqbto2nu;%ouQdM+V@g~ng@W@F6A$Djo;dB@3gO?8< z8E@yVH;CVdz;>#C?fqA)T$VF;;LGzxzP7L63oHPdrVjSNtSAxrq+IZPKaJysNmWG} zbaVl?s8Y6rJ}U&Gn)rAPs#;B-j%!U2`*`yi!!v;5lgOuc`1yyeTJwFGOGXvCZF+v{ zoz}hM9;fFfQ_VT=XR;UeX%m=vf9E#tQ8$l@1bH$|d!l%a?N_$rOQ!IJ~_gwYz->8mS;~>7jU$}QJ?A>=5 z16b^i!~GK448wklmfj;emzGSZIc-;1cx_TAX@|UVIbE1lpAl-fi+S_uHOD}7!Q)nd=fs%r~S>$ui#MFVRu&U);(8@R=yhHQgq_xl|b_jo3 zvt}^_f@*qLG6mf8qGXl?UwNP^Y8YYrVT9!7qYXh6!2N~Gzzto_c#$DZr>?IC=^O0s z^2i3jHK}{geyKAP<^XQON(pCw7fqUD9sE;<;n%!pu__lRk#q!CJh3&WI84?j3Z(#* z++C1|FrT&kGwd!>;%Fx>$>re%IUeVT<&Fi3=g*KWdnI%}j!sH<+|hcW?f$Sg{K6;r z)2&5w#M0ICkNHu`?VcyKl?ZVAD_$}C0FFxB!hXU5QgzR&W~c!*Yb6d|xIx}KfYI}EYw|TG^n>O-mde_bUcy=C7oYIM#ba79XYY!0-*`2o=%velHMZvmtG(M7ursb zxlGs4_nhDPn>kFfn)0<`=H~U-QZ7kDylOV~@ZN_(yo?VfieB#OvE0RBJ0iMBy2RNL zc*)AVu_o4#P}+;**2kf*@{rjzWH^=l7=E_|?URbW%;v??4E#QTT~zR@&b|zSfSAHk zQ>V*1)=w#$>Z1BX7@SL%<;%*0ZhcU8Rm^F9&)Njp3sT5p_)wy$kK#Xl8gX`5H3?V< z)Gzoh4P=~K1ha!bgX;mbkoh4&6r0B^RHMMDXZXvA1f{*tl_t|breF1KBaKuhMs^r8 z3KxUvL^2I=Di+HZ!upu~2Fdm~{QATvpanGws5EiJu5#a3Wr|+;{hQflB)9_mI$;w^#O*b*gakXl#6BR~lh~cUvk3_f z1d(h)=d_@O3qx_Ex`6O90RxEgRts^9kniEX&_%w3J~*!+Do@C5CD-U;+qT;MmI&4) zR{Z3<32<+=e>jy|o_ibabKWoEtR~=Jo`pVlW>GZNE<4N?R6cIDB6dRwGW#jaluTdE zXi3%+E&_!sEK&@k8F+zQMc!9Uh(ve^J6v7TH^>NT6U8!J4;=ciZ1-*on?_g-l+}Pj zwUm{Y4V@K=Bu10H-dAFtxDiEzyg~SMt>5q-AT|1dUJWUj!56~JEcQxf4_{TI<0I$0 z5awK3DopSXx(hI#U_PeD*Wr$zsgIR_HL>p5eS_LnfEA04z!7hGQ7nb3J@9_gewK%? zq>}8d{WcuzlkD1k$ZtJG6tb*$8(n$TQEiv&n6wxXrrdL)BL{8UBCH8e&dNMtro7QF zmH-xe5&Q^B&!G!RR1?%zXq1>#_uV|jneueBhW#cZB>9@$=VtQicn&0>Z?JyGZo3Ra ziBQia#$X8=lK|^_=hr>ZAC_`hM9D*sA*HpTnVL=X{U=v$K6#-j8o{gq_t9x!I#DIP;6>N;&@ zgosM(aR{I$#X0b5N9+%=ee$U=47tfB&~hW48*(M0TtU~27B>~b9~q_9;+$eTG+Acb zHBjc1!lL}8T@D%uI?h2-Ixy=3IT!#kP2a%amQDnV=eY7(U&H!c36)(Qzz|i+-OW`{ zN9Y_uUw$@bsn;3B4yq}|BUp=sg*wY+<|6`7bfx*35$`B~^;1JIH?=o)Vaf=P;>Q3& z-n+!mY!erok)uB_1c-1VW%|8Uocl%UOrQN3nbEM}KIpGz5NjfhBI!go>0bECoZVN3 z;Q+9L-&G^6tmTv@QScu41w=3(3bGQU8;AY&QMv^Xia($doUuGIv77xI@mF|l`hw)3 z48V^^q@gAwwa0rpsp>H+C}KcWE|d&p?_v1Hh%K;yt++fSeoI@u5*QKZ>dLs|O9JSL zg`D$HeUv50;z?wRe=}?nu$*c?!wvQ{I7xTyv4lOkz7_Hs(u zy_7BTB7knz+=uxpn!m4)dU=AA?JauP(7g(%a05+UWT)wnI@vlhpn*ACOLkt7rZ__dLEsQOTyQTUVvcdh4UY1>7 z-I=Pz1CN6Z1^wg9`~NrhIx~q;bRmn_Hy3(x4I$4jrWpHIA#BAPz>NhuhL`9oUK9#ihe(2XNAl3Q?cEphx&7kp<4>v>sVOcB zV&269$%D!xZJV0wkUZENEbqC3``sOo=dBKJ>7Ba`#?=;BCDzh0ex1!NiJat37OWpt zKioO__&{ku-4Ql-q+PM=zkGchdaM=+(m<{atZ{ZXI>R$6qWcFbf$OS_bTvUKW~z!` zm!f7S%DZ^FtmQeWnAX6rX>UQ{xY5IB8nKNBDwLUdJR5~4Re%Fp&oq`IOV389APRXW z%ZUVR$ygcaBn;+}^6vML$!HKvG=`zz!wOCeo-hqa`Svj{MkWK|fT^Q77UZCv8shJi zZUL|INR=+&Y(@r3G+u?jyJ&2yTqgQqsPR}>pfH`ZA-T33YIab5U`_&3t!QuAbWkXKl2 zK`Vw<=iS;qlH!>y|GcPPY=*`Z`_QHm#zVCS_w|i0;!$`^cE71JRzafktf6x)d`g0C z?e<}D?_9fLLy?#kt!V6NI;u9}_v1&j|jZ>r#08`;#1luY>^ z4LRR0i5??iJH?>N<(ItW`U}}}X){2#AfwnfyTntiDo)@-QbJ7eQtX)zv6{b$?{|%3 z*_o~MM02gi-?WL9Xj1 z2fAx(gK(`y;HYHENX&#rB(te!?w8M=)r9n&k7q*})_HPXrE6S~1P#68kg6s%nT+vH z8DJ#i3y0?eZY?)38blYL(1n12I6g>rNVK2)ZiLb|Q9t4NJ6|~2hmky z%~i-B?t1!nYYPB68P$m_HZ68|er*dG`8@p`XbhSqAI4@>;`7@S2% zMaHxXl&yQ^!C$8l@0(P^wbIc?4M#WOKEiBwmPiO(wOvWDG^g{IzNz696eX~i4rx9~ z0UQs=GhxVeSy<*y1anRa1ot)qk*Z^!PZ=*_vTH&`FbZpS%z0XP-8IOYp_!~-PEs(c z@62Q$zFVK4epb_~#zJq{Sf&{Nb#JQ_)ActO`U*Z|5oh8nXv95E2SY+>)s;=5B`cCo zOS&u|LUk4J&Xnlj+`nd6=h8aP_tpBEVE@9C^BS$@juEkxlMYv=P-s;OA*n?`}{4 z0G8+ZHWW!H(?9obnhSWj2to{7oaH|IvBXXIodC_x+W`~|I=gfNz`-&pNLrqflYqtt zFtMmfF+JJ8nX0m-1M?*<>Z`LQAysjqrcw8uWs2>1ghMPwEk>TP<6`1q&bHz`KE@S@ z;QK5Z4u@rbP=jFQ2^K{o@XcuvioVTz)&_SUC?VYbv}Gw?Mi{=G63GpWW<7JW$$DQnq74 z1EAu{r8A;h;g>}a@k7w6*sB1QAb&fT*@KQ1Lj3jmtxyl+V$-z6;Nc>NzV~E7)qYc4x!f{{V4xKpURzFzbZdPuJG*{Z9$mtL?BKY00lH^wM*(Hy zy2OjJi9(!F5g(iA_h!Mr(+{{J=4>p|z!s3Qyv5Z;IGVnssPjC}T_=V`C!ZDzkVeVp zioyq=dXg^(@$6l}ylZmhLc7F)9$*i3CZI0V3oNzuKlKI{Hqc^lkkvI%iT+i(z*7?2 zU=f#+o#vtqf<&Z#yzonykA1y*5so2?@_SYMS7`oV>NaafxJky!*vhLA%W2+uvZa`# zQVdnWZ{3mrn}*#EnU~r66^LuG&YV3UZSj*}U6p9Pz^);LB%IDSI6+JZCOdIbBrWU4 z0l%%e?PU*{&jF?hGo5r)I6bHaMgRV^es0KZZl9!#3z&`Z3OJUo_$H$SdqHP0OOBr43lf~V>ZU?CPfgL#0>qzZ7k%}{S6 zMIN;Pnc8pL(wRdLuzXdjG8KMh_U?M&HY2{|#%)-e(7TIg=vz0)HgnG)%E8Uwu?R_5 zgoUa6bB>9`Y<7D<4R8XlTL!b7U@Xh+aXu4Mf|3Ogj9$C_E$DaqS3u*_*iP)ZYx-Rf;e)|FDm zGTx*R|39FOz*uuM?%4i2-si5W+CUnDwQhvnPd^Wy&EF+iRt%Iz{iaE)QbIWwgh)Re zvEF3y?j-P5ZR`?5Ae#D|4bizlbtvb*vH?qMDO574f1!(HRp#}s4QEF8uK zVpylh8;I`RmPEyi&|gY6ZEP8kMKas;=k^hWb)d*+CpUyWQ7Xiab^wN>yff}I^HxWYXXN~S)mB(`{=t=Kqm>QqtB+!Rx{u! zyUPh+xIJqT7JCL^!o{cFm#}n*oz?&yE5vhczaBe1Kd`G!KB$g_zZMkcEwU&&VVEPL z>G3v&6|>-WF0Vk(YYUbtbMZ&&TyKTmwnxmyA%Tt6$RXQrr)0X*e7srXHtJ4+V9Rq7DH*!d< z$YcQetLoN=*DmUDZS*ISpTNvFX9X{*?8w14=!PC$aXM}_9*1IrF_zYB3V5S^yaZxe zQI2K+V;{wEy?MRLlFSTeT&;8e6{a5iK8Zcks)Qvbrket~xQ@(wF`Sk9&|NhoL-!*W znUS^Ua<QINM5q>L0~ng z%)%s#4pl7j;$Ra6qdng2&*{1AOVg^u(=ly)W%vmDf?*<$Suoi%~3jglVF-Z(Z8A2WgiuoHaNc43m*>0;Y8wcTCrepaeVyOGd9y0 z0fHjhUz|g?`Pax#tQuT!w(V!T<&qlpLc{yLVGd8MRR6qRA*P3rHx?`@!O)h_Hb{j? zu>3UFChfzoM}ml|q}vbtK<@<;f&N#13Y^&ntU&RPT`bxN75SkU8J^29R^G4z&5|vP(@V7S%JH zOkUbh{!Q~6KhnHY5|&VqNdZhcF;s3lBv(`^Ls4U>jK+?oV|BOQM2m#II}OAFK|VG| zd)z#&x8qM<1+I|O_RgbbQUl9p87sbsLM)F7TZe>dzp%<_yTelUo*K+zB?Cp8U$++V53&PucBu9q^;*nLU~&) zV_!ZudFybgD@?$kC!*2N^o`g=pv);ISTTBW`xG|DIL?}5BEh?pU{G$$EM#PLe`=N3 zR?>nQ8VQgk$?920o}W=?Vg#%r;^d|x4N6`n-F|pE#rl~*PI%ULgyo`V9Z>j^>^Q== z+Evnne{naj{jQG!>)&e*_!w6Ow`dPI`u)Aj1$5XuoEQKO4{}+X|Im3JlD322uh+85v5h3pW%vyi5)1v#8$x6jP|Njb#bAKqu7tBWXW^OT>WF zey*l4nD3phL1>b9s?h{28C7m{ zbxh8w1%nze2zG>>)%kNuW{=N+{3MHA8cN)b*=}7Vt)eBs?J&ZXYdBcf^-<3q?C(U9 z$v_%=Ua0{SG4nW{`q0h-#>4*UYe8#0xTLaTN&7fvZD^@0xsCV#eu|>KRuj(cCFX?i z`6T;(b1GTFJ(UoGzwXiLa{Row;EG7+0IQwlu>&I8wFGygn5 z0x*+gx7eac7l!368PK?#wzYpJk$L2Dh>|y{~eC5&tkVvpm_EGpf*-*x5fcH)-0NEIf+RLm{+;QF*)DbALRN?l8 z+5pKo21kgH9dfZ5r_c_x)*5K-@I@!FNtyN8yN-3w>GWj_gDY+pqsEIEO2wZqm6ON! zh!Jv2{PP|zF~10H{lb0Wt2s}-n<^8%uLJ7n;9ttlz>nYWx63$pp~tflV?sK7q-V-kqCVKMXw27RNTZ%?3+0i^QbtC#UE&oulv1%xASlkKX&Y4si6rqv za;sxou$Uf^CwG?g7#Qi|P|N3C4G{X|5Zg0d+qIl4G{A4l9xzyxsB!7-ZWedpX zU}zq_HAD6mmn+fVLq*#W`6qiIYu;;f=~d+L5X6J8g#5zDX?^vh7oHA1B&IA)+_10-X zC>6z{eW0Y)&w$|5S8ojymj6D#T3H&g^cPB^D7^FOS|1wuCu=u7GY(U*Ypv<|>NFYp zf5jaTHkh903pr9AukjB=KDn(@UL7Kfw|*0(?Enjodud?a=-qARe=4gd435b{f@44O z6?4#L-ncTZZhOwm8K4~J{q3IxWB$fed-u-L0$^p8*%XlM9*V(?2Ay(fqw~hz>;6Or zR%x>J7dQ`I(FfD2O(@Sc!Ga9yu%R@aEwjWcr4gw?=SUledNR2ALvk~dHj?kW z64VsQI>ZU*?zrix6~b6Tg{qmc$xMbND$2Of!&q#t^4cQAYq7cGhn^dBprCqLH=a6E zeE0<=^sPO5zEp|-h`ztx(tSrm@~q*5v{}1;=&LUj0bL5!oES+1K|R_vWkC0YS6E}I zel(6afM>pTdIvdA6?kB#V0>rRy6`a9nB16=JUX2Vg%HrcXl!#xHcIx!c5qmbOP`9N zAi*`({SmC%{hTkMLh_v82d1FPt2YK!kKc>1*A_@tq1#xlaW^&*$72C#4CzrslK^^( z3q=6#s`P5a{`XJ25E$u4_Han|?6c`VXfwG^2C9m7=IYPb{Y~0v^yMWe&R(lK#F7np zqjAA$NO}b`Y`6u(^aD^- zlmZnh)oRb}PNJfQ)@J9Xu3{t8m-zKkK$=L8BCiX`LvV5uS#$=t#kzO|=O<>0uep~M zPqk`WTww@PtOcQXBCPL^_gR=47Al4t(({*i_lV2GRdrT^v?oD_=E@>v36B zufKv;mNfae3NvT_ezq-Hj)d39O4i8=si}0OP6>0N*<7iMS$dnALD= z0>iAcEHRby1$~_Y7FB&4BA$P4E3~7BOK@)d)h)xC{xdiyns!k{G|jes|LL}7sZwX!e-Vr#>L4NM(I5XY3&jLG@9 z75@}UA1^k+)n@_G_=qAeN1I3!;K3NulZsQ4K>N;zZEr|!WVFgsw4%55Gwep_7**vk zEOVA)*UufB$<-8sIa%>FFt^==VL8Gc9-(spCq&n*;f7(6#|n$;e~<)+?3{nP_=`16^>hkTS7Tj%L>OcG~u5lHi5j zZIhbKujxIY`sjd8<5C~nxA*W>j)aIFxQ_OSQt{<>nA`RzK=(frZ>)@Tu4^;n+bRE4T9 zYJq?knqM14O|6^j*GRpUAZN(_C>5nef@WHxr0N=<1WNE)TKn zJ(p_NT^JrY-Ih!NLZU>>dD(IL-FRkU&*>>q+$1)dlOsIOO$cd^n}5OXE}EmD2uV_h538R?kAsG%8 zT%@)H4!UfV#_w%eOv9&y`F8WNEgG-nE3-Eh$+{hdbED%e4ZvXa5d~ z6QW+^s#X6MutuFCXZ3dod0t;w(wG%$Xi@3mOCt(3UZf`RF&_hV^V^%taBt~{GKjGR zjmAa4TuZgMY$xz4w|de?P1zDJSV;*?-^`rs|{xbOozH!jp3yZuo2|%cl~dK%$O`*L@+M=lOnn|QS$aCi z=f2k`hh2EMHN24-ix<#t=x%di$*EUIZ+|LMZJ7G-%Cr0a2d`meb}ScX3+!c+2-)3` z>IX;ukde^I>F=MfqoOH4QlDxSorc!k;;^sp_)zFe-uZ`p|4k$IEk>+8nGX)?sYP?x zXrzzSR&LJkT?;7~n?_<g$N&7&Hf5%7lp?d8YHUQJ zm8WP;OJVdg591;X+LY`5O_5bAzMW;a;aE7FZYNqi0;lUMer92}~UFq=0F0}!T^Nb@bQviQ++_qiuB9_#%E4$H{;5fH( zbX(P^H%hpCkY2SgY(=vR1&{@XacM#FFLB%)C0VKsDW{F?-;%)qe3?`0{AzqknXP{C z8sRt8lXeDeQjW47k;3 zN>oz3*E!qav3pMILU&70WM-KzuK9fNlVq5mx3UPH@=A4EyU}`}Cn0|k52t_T*W1Oe zOVD8e1^=2 zajxlN8>(P_q06|4O1#0jG?gPUsgp`T_34L>A|3b=5mXoCvWK@e*3Xmp> zKj%3)f&_LNg9fXT0L) zBr-aZ85q|)?mM9KS+)5bf1I5f^tIr*mQ$>r^m1aTnF3_T)CG0a^E}{G%#BA}P?`{Y zpUi8B305hRxm$g9+T)MR8JEkITINx`8yWL&lS&6}_dReZyo9=DP1M*8W>?R?;&P(Q zGJG;(Ts1{=qvD@qow?R5efoum3XD(af|izI)bT)B<#K|3Kw`!Wq!t zQU&zrt^4c}Z{ZY1Imrumvl;jjDM01zz-v_s!V~$g=V}K?&1yV@e7l@c%4;(YC2WZ# zP(}FyfRpUX`@*EtNPZ98%&{4w;+r_!pc{RTP10|EQjydpNI`#wls0r2DoyF8&Ihu;&L;$DMPD!_8 z@g-N#^J@WGo6n?)r>}D~9l0YlYUV!9urV0s4s0^$_!_ve+o=TuGq|&zn=FZU_w6{m_KA^Zu=39BrD*Ty{3YrR483SCbtHaubvIA4_Y%V>o#ug%N z|3Lksi&Vn*1}gS@10pjIO8sHZc|DwWwWR9PaM@NA>{II$4*LBmM1e7V?Va-j1f$4fzC7{E2yj^EhTGO7 zBKfGwb!j5?hnX_W(E`5+q0W33<6u|4AtT_js_)oq5hV{mg%-%-h15U$0hRH|#VB{e z5-Pm(sAZC3*&ZZJh%;PrHt$_ZLP?5sLZYk4=*Gb#LCc!v~^|VRAfc} zW=7VAZ`n2og^}959v*A-qU_k_@EarAHcI*A+&XfuX|FF)_Nuc-41DlpbkVRv4sw#O z7^jy4o5lbPOUNX(*Js(9C$>(ENoYa3v_nzF$`@0GSg#Q0$@#<~iq+2&SN&6qzSjas z)oVvjy;?)eU_8Y9ZarqJp`erb`yIXbe|*CLqH9v*rILc>WyMyxY4M_HJ+!*nS(7Hx z!vE>LYf?cf0NeuK->}{$J%l#E!mMBxh{sLJpowwX3u7LyJmTj=ur4r>?J#7aRvm2m z$kkV&7ovk%+Zizxe5u#E=w`kpwJvk*4pmVWBJkqzVStljWz#X>y(xVgUF*YA*{7_`8G#VK zn7?_=>4Sh|=;C5}a+PrF_`~7pDzV1EMiKqofHkZox@iS#lg01;o(K+FfKSlUam(uA zZV)k@mR5Ju5z1=LvfwUM1RsUQmaq2TRJ0_1{iKeB79(u?9R1Tr?a8%DcX9ZP!lbS! zqlHmLL%Dz{=XbFJqX4qzq!)EwnSbMyIpdL?Rtsjtt(z0%+DD29fV#|M?tO%Rw< zA8S6yrx>uCN}Kq3G{|E>up-S8~u1B&sGf#$}IGFjB(E70P`w?uU$8)kon z3^Q}Lk8n(ql^b$$U2z*IlWYD-w-ev;(+6`4p#Hn|v>EZT5s4WByigOx4a1&(A_oAP zS;4<;a*}dfzIoNg4IZ{d)ANMCTwr8ILbkUDfHv8MsgIthHJ05ru8$`p8@_iC07cU* z(||=~HiAGnCa(RcCNG$5f^rp3L&c4rvJ;QvKjZ4Rbz&GXKkKG5-*)gD_|CHCsez|V zNL27)`YjO`V9VryHsmyL0=Eb2qnJBL`)g!_7Nk@SsNLgcSt&L|;L_p?+h(fB^AnaA z(UA#MZ#=!|h^Zneq9xu`s7APt@5cQk7O+@8N;RFml6fIbp)TLT$GFPG@Z6Ah_0#!B*~oLTny#=67ug=0$n`E9`{Si}%Rmm$BC-%n?Xv36t&Wn)j;r@u>@d zYv)_|?*1n4J4B+@;{452a(akYaFE`-NaSW3N4d-lIno6yU6;mH;Bi>Nc={X^e({rq zyGCNZ#=%M1jZ;K6>^swKpCHT`UpvtNC_WWLZOjFXd42gTTl zaPk=eYfTZJf|K#Opb-Q_6P9WBeuejvS8lGuUVND_K4R=1+F5ci;_Rj5p1%y_nHo8r zIIG}3G!%NXG;=2sP~X3WeA zpHDk0#p#wwKe``u7U;n!QXd)z1)6I*;^j9+MVuTh&L2;Gf zs*dvoCcZ%6x7Jk?h!`>20z%WxcG!FO>M#6@$5wfr0CKd)WnskcKg)((oCwg6&r9jd z#mHJJ=Eykg>J*94@-%DeTGX|YfrdzAsS-1rESm2IrFIEKjs&<_tYducR7La2{*(ga zs<~cDJF}3@CO+uvA%bGnwVab)c<^d5k)2>x!u45TpFCZ9<4nPv8#ih<8_u0wNGUml zC{Wb6t=Axu8Qs+;Bmq}5!GYMZ{2$aK6?448{8aQrzy-LZQVVX)RDQ|A3C@rwmW)`8 z`z86U4KQVO`Rn%Xw%b_LWmHabsFwEoB?ghx9oCwa2+})zQO=>zH{PiQ|Y>kLp;_&AOnv$l2Nb#2UJ?ngqO#uiWZ)TaSyUrIYK3I3@I$o zI@AwU$0wU!Xx4-e;nt$G_W7%&Ev6#8cL18f-!5q;n)90^@W@IfIZ$#($-Q7%d^m@u zZOW)I^w!$%D8W%ivx{{ zeREoEtTPRgi&a_YA?u;Z3RPpy|+(=>I~{B#kB zeT7b`zJIlbWPA+YMWow70li9iFhD|{_%@hNYI|Y!oi42?Iuf63cGhH7OzHD??9G(| zC~(W1hRC<|G(dc=M{OI9$wuaAH{0@7Ir72y`@zI?D-jp3tb_*4aF?oT`*K@5 z_b>!VJf~tTbDD3nHI)QLyQv~^iqF)F;Y8W4$Hz!q2QjFN9Lxe=>9EGPuLkdgNAI9* zd5+LcP%WRszV{22SPhx$ePN(|rl=S9YJjWoWN0#i9ia^E1^0!{F0pe$@9^4Z7VjFHI#dzBq>6~#a6B7TcoYYwT9Q~b06)v~uP(HPYii^| zMdtq<4E?|(Ryiz9ZG+U8M|&-JM2*+F&%oGDvQ7M#V6-0o~F+lk(+% zX4uQZ^iYf^{O`;Y-rf5qhk3c1C^k~PptSaNJYG2ex+zOYGMR1W4Lm#^ z>wi+jd#2%pHa=^j*-`d_LYW=XDS1_{er}7E z4J@tYj-PvgWV$s|CP3aivy$>DDdHeR#aLB&jhYq}!9wFo#5K{%IHn08#slYk>M^&) z%Vo^0{lTUHb4d);-`HzkSTW2q$TyFV_hgChRw>cSsEy~bp2vo>B0Gkg+)Y+)U){WS zV7~fn*yHcC8;CNA*F((4I^esm7Dv z{Im9Sf{pIfS|&I#?lz++2Th3;e^a5KAqMmO!j0F8xI&j@rtFY%9zmBW_7w?@ZB9T= zgkIJ3(3%QGyG-;(6vveH0UGzV>v4r~!ii2v5r;+`xrJ<`IRM|WYE@iLd62Ai-~T_n4D#vJD@+qP@lwF_x4gc z(zqN)W8*K|>Rlrf^b)sHxLk&Lo4a%?0YS>BK6|xN_GovnHRqYgFZ{LAk7Dkhg2T9? z-L*f9Tn;N4L4gU}kd?pB0QmhJ@&E=`$F9efv{;9nk-VW>DQvZ*_XzwSvk{uRw6YKf z0Tvm#U-wDyE9p~qRpf@{Z5U%(Apx~BSzxhcrnUZF|0@$Zk;D2}D8%FVlTY>2ug+{+ z9yr`tDM6zFGYXxI03$X^?E5}Wit&TDP%aK4g3%Pum*FT}7~3(QV&euLa6%O^`CvD5 z8!R!I6-Nxf)Ecp!OaL8>cqftK{@P3p*AmgVlsuj!(g*pDnT1r5Z!EBHnhG>|JQGo^ zTq7kikiBG>^}88*(!MaFAp>8wmP zY`9;-1mBYs6d}>fOQo~4@(}Wt?&ZofF5w5@Q6lmuQscMq(8UIAoTm*1vu~eF$uY?4 zP;vx!Ak>kFP@o4o2@urwt5O~GzPHNlNoUVNwLM5zZt3{^1mbSGQ(VDxQ&9ad^&TH> zCb;<;{@yF6K0sF8{Nt7Fs9CR=-H6u5x&<_;FE?+1@fc$GVSMe`P3P`K1~yv_<#)lmDCK!Kbvn?KX#HTYAPE-|)P4F@^sC{|nK8>n=2uaW zmC!)!1K?j8**i%xH>i&th0GN~0Sft_wF3W4W3gZVnJOb@ z_v;Mu;`T&et-XdH-Cj*B%$rJ8k^*T*8gQvX*hQ=)BozI9L!-N+`&}sYTp1zC2ujy} z1cEUmR6~KXcCGW*A<*voZSePr{_0g3iDKxZ-zz~eWnx@wj`cjqYULn!!gwZ;EIBQs zhet#U)l#}i7~1sM(+u|%cvDZY9l!mct-ixSu&1B97WYrQZ@l3YtRF+Kb>aus%Bif) z+M3t6Q~yM&C^%i~<9yY@%-vkq0WVF7c^oZ-B^L1@Ig})K#CWdObNOmKjm=T+G~JHA zfj_|aN{;CtWlK$e^pDTf)Qe`oc=c_ZD%3VI=9p6J+O95*6!5x&8m2YDCQX<1L!iQr zYgiKc9r+HTrn6dKhPQL4N)i-n#*f$e&%{s*yub;PTzUN`HxC`z)65vV-lYj^NMALsm^)OV^uvs7+y|m_R5PX z5<^}m5row6;+dzd;F={TjWc^VzKY8HX+77{bwjKW-ys6DwHs+j#N%aOw{cV403iGj zd@qklanEa6aa&`?oFf8Bt$ojNgM40>QTcx4ME0N%b|D?Cr`z_*Ia*oT)mU&;iNygd z^l$|ogyEsj93df84K5=np6yIBuB5A)dkn%VB$~fx=E?=+Y_kDr4AJ&ci}+~ z$k|`pTYs5YXdp}2CTF26gaBEQfsE@OLF}kK#kJRDzo4^6t`{f`Z%H)C!hJ}^~(6i>D3zS7N z4dt4?Kvie+RU!5(05UqD1Z%lw)Gjc0vAeKq9pH;7=SS)}7do>#>YUmzW%d1Ph+_S& z#2cAARKKntudsvXvQ)Z_u>}k;_ev*hczN&Hb5&`p-G>I#$=}9o2Sb`+7j$)MZdvfx z-;O`6I*5s4T%xpbnC2`jFG>hStFpZ&ru8d@@n;3Xf z$~hF~v=YB6XyL@tPAX8a5<7~HU9*Ek(<1sZE$OlFc(r|b5J7{xp!@b*3ZKmSpwlmh z&OfqnF4_VfIg=*I1C*D*lqd`yshzNk~nUL{LcTg@2#{9GNJ-@5n$VbjAX5wdf7^F~Fl|wkWn6;^zO&wBC90M?EI0qtM#U{DgUMD2u605?F$zs8DA zGDJC^-)lumx`YNc$Uw`4m1TF_(WO#=<@uuJcP0eI$k8hSS{bcjWHg!ou;S{|CR2NS zuETkCJnSZ^EwC@CLs20e>3?u15+k#HDp|Uxa0kOR*S0|t;!zs((&aGTZY@DX;DBNh zpm(xv#f;SXBBhCIYr)gLh(#?&a~{4e|E1pyN&0H4in*fonkyX#O!TD_AZ}orUO9=I7hw{P7u-p+1^onUA4AMm$(4;jEJe}?zxoWD z^LH1h)oW6osB3%PyStTwvtl@vP$tkg`^ZpGzX__v<15<1cYW|nuVK>`z3rf31Ul>( zP*+4l(iGF+$qs%%tzxU^`F&bqCtJlM z21<+n0)}%Vu&+M6qND8&1wTyGwE+_%_A)kwk(Fam6|&YRN(F?iApk%jxr<&DGP827R&nx83hsu0drV z%!zCFm#C)LMY{f@B3J$0Mc$T4R9hAY*lK z%W!7i0Teauk?Gp77yH${a-*>^w08C=9=Uf!4lWg>i(i(MmmW>Wb-xSh^Ulpijjy4; z(GqK$MB;d0pur_6_5mWQoEB*Uy9}$naMEU?QOq~U*pBXBrChw=mB38ArGB}{?{pA_ zDLtmI{L66ww}J>AnqA)`Lz;7z<^~$pN4-rR}#(6srj#`P3)_*hC1%OS1G-iP@wGnwQ-c6T^0& z5Opw2DQ!p>cmN3_e6U8)P|p&Ka`mm5CZhiZo`j8;i-&G4Mc82F5qMHud)z799+R+v zqn0QSAh$dNJDg?L{A>6`@cP(LnB`Icz5eLsivFk0M_=S^cmSki|?3!Co53h(T= zqI}PA3;M~W*X$6l9Su9(MM8eqp5t2x3JxEeVu9=++(P&vZ&rzmfSSOsx9RIz^OG9* zE+mN%s{1ez^|brzjCcawGic^Edw>ZJnbY%}O7p6?ODSLe=^5X35Bx-?-1T05#M{Co z@CU^^lz84gsBm66SYkiP2@(p6kVcWrk__gF7!2*ke!}qZa?~7ir_SW+*lIeI^Sgt> zp|HZay+0&pe^mYTH|7$D9KJs#DCxOo;P!xb^P!a{IHro zefTtqzY#iV$&31bK_gl&Pn5mU%uGNCSgpakPQfI@!&ZK0FS7?=VSD~n_7zAuSl49E=)8Pd6#Jg zEHVwGXCQpj5l}FqR4kl%?A%8?^W=D3bL6_moX+}HO~Vkn`n5OlFWaks>lZ++JBtWd z(lt1115bp z#bnY7u`2MVu0yFs&@to$@)iXxIgH;_=RkGI?tN2!af2BmKw{!-}l826UZH&qXSF25KD>`)|f|t0K>{heV1=&Y&COHt9U=yYVF8=v;8f*m6+9_=hEXUI8+4 zPt6X}VXT#QV7O5c1~%4)qQ&^&!dgzIgmNI4PgP#<`fH-nJ(ANm2#qlOErVh*tG=0B zt8Eo<4P*qA|6R|wjZ%)MR^h!f5~M$zrU!{xY7geqvz(=)FI-bdOW_T^AMd*Vbm`N! zN_Zk*&Prr(0r!XC&@`pX|S`RAq zh(oJ((1x3=s~rZd#}=mN_5fUl8bI|5paFO2iseF0zdf2ve=Cbnuod#>?EP9(t}aP^ z;_f#1%UXNm!pIzqG9T6%?28q`6Rs|@&P$wC_#@RGzup~!5{duak(?jjnFf)M4+sVC zTJyWvieo_}A8-YGb>9tT#d<9OWm&+l{E<0E^h)D;v!0p8+q{e`t@EmTs25?DblWpg z$W;I{ITPWfx$@75LweV%hMbe;a|BLN99EpAe1(06d)SI*{645Acr+A})p$-H9j}Sl zxS$gOTZXkEsoP3y#j#e4FuABuxgT>p$-IZ=#PW)_Q#G~KX8Ex!L)I2cK4x`c0lX<5=1Pb9=F)xbcXs!gu|H#$%JGr&9GSWs{YzsSMXlPaTcDCKSF-my&$TWAN zPdkSHNd6Q25I6lQg0mRdm+xL01I-CsaH(iYEj<^VkNOyme?ozjtK^rPE&=XqmC{W8 z9V#j`C4L%JQ*P4|=%D&TUnuq`!#d`F*8P2MELGl>a|If-dWr$-YUDV;;wtdwuPB|9 z*T2g!D^L;hnhzYq(=T=HD4j3ReCsDZQ~MsKk~4De6Ss`mn#W9qVY*e0z!0+wsTU#$ zCorKK8$c(MITOw`zXp*GKnA-j%L2fJ^-3n0@YNaUe?WclBrU^Wf!tHaaT3&44f|BM zl^KzCj?8&EqHK`4ent}qBQS<%6$D9UyK?_Yfmt9O)Ai6Q0B7d6O-Xb&1`MED`n(?e zV6jE*ogX;gY_KY~4cQ3Xo@yYmVJX~ZH=4EaSdRg)od2V>yHbsk%0ei6AILa-<5vD7jB1W`Cv2MW1(AkQ%FHcBb0bl`65G787Im^ zsk<32_4Ki^LOO59Ws&MP0a!%&4gVpGJdG;>P49co>`9)Ms}Q*+8uQ^gmP%V1P);Nn zFlfYGkqDBWi8GtyPbgj;QWMoe&BrfXJ1tykL!@8LC2~+l`Km#q zG?vm<{I95N0k}DFC?Mho&&;JN$@t6*Jt#~hLa!8&ayrHC)VeD9b51~olfjv!p=YtY zxroV)L-kaP{fVR@pZA&-weK{n59!|r9==8z>Q>|dFar8wN{>Mt9`6j`!jr~T>FeQ$ z*R-i}?F^GE-!(HAeb3GF4 zo47#tPftlZ4YC?q)8pE&RaU^?JApD;mU-%}6xNVPQTT#iP=3&ew zBfH#cH)5;w_El=06(>G2d`$=M@Z)sM4qdg$R96aOARYKIvS)o(zEoah4h%@Vx1;>> z{}7N>C2-vgze8C3IeJ8u4(T?md@9Img;vuuNjIn#3<+M+rr>Ct9TV<(CFR`kFVVuc z)IU3!2p9_<@{R5xcpjko8+M28aX@s!{3bwVjIb&sGk#~C6JX+|)&T4&H zk)OV&wo+#(+iAN z_pt`+>J@;cv)oLm@@2x~*=>^Q5dEe)%V+Q$Kt(D-!NR@^g0Sr}#iV4RuGp=ai3Oez z1I_swTTW!y>E!Bq+);r`VSCT87D~NzA%w{7^BcW$nN1gncE4!AQ~Z7;%}GoLeqUYv zd;Z;0ZO3GW`=N@QK_dHXgu#4o?TmCW-`*lDAjGnECfn>Ym)v&&x9)o)y;}6EfFSM< z((HAnl&-F$MHBGtbz@}5UNTOXtwcNrNFoY$U1AA#{laA$zgmJJIxB^;k*+AMd$1T$<*lC3#+Np4sp#nO-^lHT; zo+yYfc%P)mQgeq7$B*}{=iN&&w}u2eTr)O^lhSV4z8v1+j>p>P?pf4Ph>0#1uV=G6 z0xHqx5+TD-UZGA#62yPt%U_(CwY_KIjfVepoeORSpy)E>gA;;>+W>*gnh2ZI zpG1v(eMlqUOg#~?AJe1l(UaSL!T$A7tdtAR*#x$ZL#!dq!rcKw3I<#8?C1~imK8?k zvE;092g&=0pfutScm5ceF7~7})HGCX_2(>&<+^Cb>ApNn7C+mabL3NyX2Jx(!ciXb*E|?>WTi zDLLJc;cbMXu|V#V1duM7HU|19hXhgyqdmREy6mre<=>3z!vBolkZg-TX+ps`Ek6~^ zj!O-Sw=%1XX*ko>*z64i&zldUp;a&^IYy>fe%m=$!pw=IckugJhzHNF8vaXI* zVf7-Az*EuFn_Jw`sZ5UYw5w&#>KX*o8y*Cpm#~HzBG;!{tGfubH9n0DTpixo%;qsxcm(l@ zlv6MkX-vwY^tFeW-McRDYI~6>7Ftg^-3E%~=n@&ux9WdGDT*&im(AfypS>h)Mp84U zxbDjr+Q-E~$B;Xop~xUqWK*^Ax3f(;b84l-mPNPt8B}xfGpaB$Z7;)RI$4@I^H!k5! zr5f-qu6`|z!hBIXGi_U4Q!Vi7FDLMU^0EV6e+n(F-4=~ZzeB^qY*ED-clLQ@hTJOH z8Fn0{h@5R$mEoHcipJ@Svh*pP5qp2(zub7KAlRGOfF_Fgql7@J?b~mXt0w`Vhc~6G z)8*$7IKg}f_T)pH|wRCsJQ~aUaj#G34Xg8 zuRCYuSEiJ-q=hgNQF6^-J$M820^OH|&nmZ7mYKZTJF+mikO4;}Gv_ZPtar5V5(u~Y z%}k&H^ZWojWSfay$Fp#;nP*YKfZ){fbjrPEWfQ4i02i;)1#cnldSUEacH9jHySVVVjcu3+rSBgo z5h>|Blytr#bRuld;BJlvV!wWS`D4!oleKS47ZG0HRvLSUF!qQ-3ehR>%F2Dp3;pZu zpD5SvNw>|vIX2ROaQ^bup7uZ*uWH9g=-cEohdmn|Jy(&$1{zv78AAd41b{HcAH&bp zU_tC%Hps`jy2f-XVFZ^e7+P5yb32OygFniTUs#@^Q`;@4%Q;liUks`UmvChxVJut| z5PZ^FFqe`6A3$jT<>j()x~IvNN}3gX^|_ZYm3HZ3d5(63vj{!wd}-i&aF}U~Ow%T? zLRiQNJP8p)+pqDym{={_eJRNS)Nv;`Q`K;jEu}99vuknmyNg?z`xB7!LOu}9Qy0;b z9ds3bTJ5rbbGeGG<(44iU)>s$N;yq)UBq2pf~+jThFU+WbazC5|v#{SNL}-)FdwD7mfoZUxupWfQk~z&1np}^MU=XQJ9ESD*CUiIUhjL6Q zhX~|1Utvkx?`Qo#`-}a!znr&mJ&B96?XstezMw4MB3Z(^@rVo`I@B5QjksWubE-;^ zmkV*?Z^*^Ww8BX*yTpOI&gn~d<7faS=~60Uc6j_GG0Ps4cy`YhT9dlZ`JzI>;{P6?<0k7b0NVTp zzM=@gBXWj-yfBrn#JwhPP6tgl_mRWbCUZEYfO}tu0;_19az?Jd7!LadQ9l8NX}R5KSou^;yqP_(rsvv5pgEfB z6Win#pk7dSs8D&IK|Yt--N|AWSFiVbbo0pvTD*-6wf%xtZ50&7&MEELA zm}(>#8}V;7u&6X7s17mtXN}x@^7&aU#TZR)OmX_i2`G*YD4ji=o*uTY1}4|O7T}lY zwF7K&@wh50(t3rXr}u~uc;_cX=inTmF(YLlFWOAIX}|J340cL|CR#-R0FUtN6*`m+ zR%&*MvoW~QA9y!HIbAm|jE-oc%#}@&6^6PBgKefcQb7>Jc3^}TZde954d7Wb6@jk# z*a?Lff9@)MTIuZq!RDfP20TKDNRP-W^F0J?!3hZ%?1-1Z=Zwm_R#!cJO8$YJD04t83r|JC z*Vgr}FhqT~q5XlY&(>CI#s_Vmrfl zHjq9VAZ_lQoy&~E|x=Vh)UwSXjp9M1wELGG- zEj(s{ofbzi#Qx;_;*V%j8h!}=iVj3@^Vv)FpAL-L*9ZpAy*q9PwwV8|Yo8W}ik>c2x}@64W? zFcafubjyJRMr7reKh#yZZelN{wE*4m0@?=ZqC!~nt99k)yvc4N$O@F?Wdr&sjpe4= zqn=ZuLw(;*`u5n}gy&vuON0wJaIz!UsaOjRnSusnWwGnLt%q-(;!g^NAfMsY$M?%+ zI_uD-Q4=Py1bXt;*-3XyLV!=F*YomIkf#ExHB|gX^)qn#bTiJCjE!wUn+dGZ-Padi z*}{ZVv0^QQ_E&2%Cum7qyGzIzimYsP9dz^57bpe`H3FOvLRPx$>@X5ZLqKQwbcM2c zx<@-lG1hI%`*iJ4$QiVr|EwS<`s7*?uiL5Pp-K1)Sw8T(r->x`&^eDK1(i6PxZ={@ zaeGzj-ZfDvykVJjMZHl32U|bErJObwe2_uVhSZP3%s}Uz5>hA5S!So8T=AS)bRw8a zc@O@=z0@P)*X;v#4kZ%{ zbTWz0C1rpy^Fm%md)-r^3k9pMuP!1UYb5?}fdQVJu+vSqGM=?nr*pR$o^@aOH?mm; z;pt1&aC%98^7xoZk>CfBRPRnC>M>@Iao5ry+Mh3E8&Gi$RCMcnm$$$Bmr*)Xkfs`bCp>|>DMjP`Ru zQzm2V7q;sc2lOT=Z>R*!?RdnrW%Utb!@e~p7#~Y1BI*{rW};5I-r3yNqYU$S&DyLB z7!yPA{I19dnUH{<`!%~}!vz_5!_`Vt{E$yS-46~{8r}Y)%;!jnzED~;AaG_vvOh`3 z1D+yfq16EX$*Qi{u*apWn@nfp5P@jDuAUTnt3NZg`PdKUy2+QAL@{q*2VnzAmQV5u zc9EXXmR{ZoPd3wZ)P3N|vYkt>7A50F5n&i7D%M(#LhJ%*UOK6rcuM&it-nI`Dv{0C zq|*w-51DVC*hzbP^*HpcIL0tR=o}E!d`=!GC7jCnj0cf6DgT;iDDxZh4~no|Hz$5K z95C5|-ADWL!}cssutEqdC&WX4eJ7Z% z4}18BO`YP>ccaiOM+YSN-c1cg3NsoHFxz_)bu))I%$opUC#h|P zdc9p0eM+G<0f~`Cr61iA+$;@cd}Eg%8>U`W1v|(x$yE@%|z*V8zx>Xp$12HN>2%IZm zJ^*xE<4GrS1RhR{j&=$y{9c^RD_GFIsopN$qOeWazZ0XW>EAQbbxlzdr6>q>_|55D z5|2qBNY@||Y@X(H)pbp`MQc>4y@O*=Qp1dq0y1;5%?5(b#_Kel_qLr~2Ne5&{!l^~ zva#ThVJ8{mZRu(zniV7`wjO(C4}r2IukgCk-r=*2-3c8_XAD-TvWrrOy$?Vkccnj^ zbh)SH{dwWVN=)-xa9w;(v!_W#wbty?iwwyrH+hemc z=DrG_{M%aDR%yeNI!cjd_+_d)N9C{}EoAMPYF1Ja_g%5?2}H zavuZBBOYf<>30JCfur+(Lr%MQV^r|#9Xk0}(LLArubjKn@WBy2ve>FnWE#paY|i+X z-_&WswmG|N2KctZCV?$CegXfeFy#Gw)dHFu{uuF?JImce<^O2Z~*1>2C&?&T_+s1x)E4L1=In)mnp*;KY|zrd=WNs z2m%__8iJ0LAFzmdsZJ;R-JJOru6EOu<|k=A5-!B({DxV18bAqFp(@UlLL-{c$X8{z z!=kSX=Vro0Mu|0gh}!l`WTqCPi~91|lvP++jw+ zt~LCeF#v#H296HNwbXLVM(}O@2>Hj$Ol~qKwm>kA`?0QT6Fr+QQbV!$Z+yL1_sAuc zJTq%)uiZL7JQv2!!Jx8NBpzq z<6CJufG(cL)zPct{jab9X`w_$auUswl1f$)|7O=a2v6?N%N8qCYZNPdS||Od@C=T{gNOb6v~< z-g+nbkO|4QhR0eCK7P^a_m)~rIWH)2qM%&laqCv|2iX}PGdZZQHCm#=es6`t_naJR z$s-65M>SeYv2%_lrVcG){DU+7Kf1?Ewsmo!3#$~ z3FpiomEdt}8I)sBuX+HMa<+6(+ z3+!v;_38Zs{-LS4_J5&>nGG-xV|lROGifhXKU5sBkCO*Sn$D+U$+$x-Lv}o5(Cl+CZSg|s@W5Xqw(j8<3*&w!w%&VH7-aZj zfVSC+*WKm31g51Ky!i6dTQ0^>ISaixDkeGi8aoiZjn+}=<2#|f zN5aNkPHXpOH9BW_er4{ZJPB8QDBNmFEcWtx)qs#eLg%`cs|B*Mj~xG8YG34Ca5c_t0?67jViiK0q9Ab=H6i;OP z9J2MxNcD6ium>e1rqovi=c$x_#VQr4WphSFc{CCrORh_dR3g&H_x|Q0DZopDU#gJl z(ZtGw)I{&v^w0d{8vO6DelFx`iOclr^eP(Zd8Ns$_8w5&hKeQ!ddeJt5Z}9kxn7!l zZgv0em${9v1)$lavgvp~*gS8znBHqL9G`B5nhKV65c;a%IY!=Jh`-_50a6AdXCs$P;wW57aMCu-HzP8QF4B5lU|&cVjRrI3Y`)_SLkswVY(YibIA*NvMyo?gar!5}!X85wbAlqgl z)c%4_P3zT%-Pb!x zN!U;dc1?O^!(iBWr7c76z2V&Z#+g?3qfkruysM=IhPSdas?oXOM6Ho2AMn7?V438t+v8V{yx9Id`+<05`lO{kS zn%(IYfg`5)YBn_r-wf26Aag!qMWeY5k*@-}KMs1}SIAe}zh)b6K$RRhdCy zXEDt$!+F7)sXX!GvM6V&+BrL_ziAZEymnONkH8Yh=ekX=OPkixVc>!6cmG->LYB)n zeu7|u5s*bk=84%>RU_{mAfG6UiTsIB5U7Jl4JzA9rcm4%f&YZcOrQ;4HT4i{j0O8p zDVF_7$PG}1)p2WZ&^J{n1aX;Jnk&Lusm#+r5cmX7W5GU9DoI?(+Q(H1G5~;oZ>FsJ zYwO<#jly&Z{YTK zT>)e*WBp}JB>QuN*6gxwh$DwTb$szxFLV!>E^>xE* z1sE>ID_(MoH-2&1;R!U^YXl+yippYkV9e@`=B`f!#sMwbKUJg zxzZduTHsp`pPJ+V`NI zMKnpM57K5@R`R3!6~Tbx_iC-8Xm1n}?*;{*bYh}mPDnOqjg9wgh&%@;mDSyXL4;2H9{$RC5yWv;NRVQ?0 zuQcgKioKX6Q|y=7>Rcg*TxfypmntRO)sTO6m$E<k`xdr z@wtl%+Ah}JL;-%25AtS?c)_wFGah5!_q|xW>{aZq5b*^~-Z8uiP=(bZl{cRfGbE^# zG$nGkF9pg#ihwFI5586QvW;H8Sn)W>Q=@R~d`(@Gba7~47aSGCeI+>@UX!{#1g%2* zLWx!85akRYLVqrp+Nw%D7zZn)8{NhrH}7p9(}+JMr#(-1nOP(jZtEro({WcLQV@)p z?PSLtQ{h3vJ!uCImo`i09snZ@^gQ3)18Aq1}!|$T~+b)q=L(BL-%}5 zc+Wpo`e$s=q}#1WV<}vbB&9L)S~iKxT7K2%xR39T2pAsu(5zphy~cNDDpp{>r->`- zQf`<&jAczkTg6X>Cjls^)SA8H^&T_q(3xFVmI+%qAjGVOZOR?(J2Lz=3R7`Ei+%uA zcCY_(O1bRq1OPkD*;gPH(q3v+%h1$mc-MWM4IFr=JVU+0U0q8CP&0BbkyBs zrW^UnI149KDySPH!?Y7%d-TX=N}>ykZp89kof>b8+HEdn@X5xpthRPJw+bDj=vd$x z7HNcNaMaT&3T6|e&;{cvkXai021<%V{sN8AR2K%1yP!#aM~knWeTxzjb6;~Lp?SI% zkS8qsun&y9@p8@7bv-0BO>dpk97>$!pz#_r4423)x*^=q1ZA){nZV4Ju8-1fSOTA$ ztMwl{_)NkK!thmL%?Bh^969{*$?G8>&t|Z5cI?!zLHdtAhtSs=T+CG1IHqYLNfp9y z*$tw|imb&XhcH#5oTh`boQpjCG|I}=QfI<1ARh76G*>9+*mVFs7Q7_Jcp$Y2emOuU z1r&&>tD{yi+C;c4AjDI zj+3>$vS(jT?5`t9r~K*1o1kb6sEKyPSNw0cMNR5r)ZRy_3BKcj_U6&QHUQ zRZ4CAl?ug*mrPO%z#$*ZNk(N2!@l$u>&5f{FzFR2WJ6Fba0q-%qt+bohe_8+nmr2q zjTM*58Vu1b;wI4aAH9sJ~FNUM#Rphfe8cUlg&_VFg! zQKI>OndQ%PMJuyCO$hsMv()a{&-zpL!UQ4my$_74B3-K!59}}~ z@u39fq^<_7KT31cnf{3$zXvu(yGp=gn=X)U#G+P?#B)4kQFx`)A|KG^c{kkH;XV_& z;cQIuTHUWjiyH-Lh8k?9H1uB;41 zN}nBAt`rvR>=dc|L0Do^^E$aTSut$^QxjlYZ|SYS2>g}6Kwz9kv{@^gV*|4B-Jc`*_dS*SAHhHXafOywk3-z_e<8 z4?_fN0J7#Uja-ekKs-WQX|l8$?`Z)BSTT+1zAET&K2|^_hyl-4vxs&2ghqwr4(((s z+}0dZhur-Y#(2CtwJSOCycN>G#e#Y*A^DWVR~g|LZ0krE#^IE`>6NsnQf*eA%ow3K z^mT-4%RdARuFQx(vCZ3-+KM{=brxs`W|)*tQ;zW)z`o1WI2=Cv)tNFlSKDfi@$z>k zE)U5x;(86XEHjm58#Gmb{TQ$l{>CMDK;Co z=QpeKE9AK!)Saevd-_r&JdR=sQOf5$H44|0x!Vz_rrlecxvSf}uk-NMb%Pgh!bz^i zf_2tO2&RUJo7f^aJ4FNPHABDkB)jvScj|B6Y->Ij>&n5WmC?ISBCHg~>Fb=xr~?ZX zUt^smH+fd$U6XBeqRpHdf3l`l+H1)GBF&Rmc#{lbBv3Ip}TADTknV#)OD^kMh*I|5c$*}6&5EnEyW6cs+_44AEG>7|6?;V&?I7Dm(RjEBx;}b=46*C zte*iGQT8-CQ&WjbbIE?()}qf&=(t$KuCil?-wgSQJikyUkc3KjnQ`sr$R9h0kH%M)1qzGeA&Tr3M;;x~3I3 z9r%aLRQ)BxvD1XpIH8>~)TJruiY2yd-@qjTpDt-g;x_#gNJ$wggmf0W{S#!=?UMfL zlKOtDb3eJtyE!;D@ox#X(ok;(-#9l_Uo6EoME2v!=6id+)j>4E$fH0c%~;=pbH z=ip(0C9c0|>Mjt*l%KLqx_VFrwJ}fC;!%GQxbq4OuxZx*KuHnPJ&=K^6JVH1-{}kd zf3UQu^yHZ{QJB*MTT$9mWSxS6wIPm7sIKrYQ34NyYH-FGV-q?g?qdZH#5^Ih{`*mxo8I4M&mcOo_ zCB|Z%Bc-?Ne~kg0pGzF%_nBPCXXQGFBRMtZibFUbR)?{_kSi!Rxd{`JyX;OuVwdUo z5QLI`a7yD5hEqfEL6DT%=v(J@Ank}3QAX&}yYctuagzP?Au^HIiuD1E%$cof;mj-@ zl$$$m#>%qMhP$G$ylWYTJ)q_=!U!SsbwF!R>htldAGVawzq=6rp%v1srwe6Bv-Eq` zGABLs{^KMc0=;yiUKXAs-da(JWU*5>i2f`nj8@9tYxw;=sH22Mm<)SVBH`Ze)1ymg zx~u)6KxN0%k6&jzu~5zw@eOTH4co+GzYXxy@=YVEXx_ zM6JP9dQuQb@K#$xcnWW6$tlCtfQOdTeW=U5$4tN3@M6^SasS*J$R{}}36=lFu&5$F znXw2$pe8My4ckV70~tIs@25QtK%e2=zc^5Mpmvbbn>BV?07e4mn-V(Ulm}QPsQ`Pr zE_uV)O#cN-#N93!OUUlK{5n28T5yTNyb-|4Cv#P6=Zvs!5ns}6WS!Rnm)mg#ls+?( z!ul#XSM}>}In7uGc^9gSl`bh>PV9O%^?s1B;U19G0$^0LPd{cYN|EY)$ePGjUDEw9 zq2v5pe8LXL0iwzaqzj%Jv&G}yibimLEJZQe zid`d?_8G&hSCQ$D4Fbi@$oSXwJHfji$W_}73$0s%PzBW&d*OQ z9t%k64QZKSJ_r{=D>SOb=0%C0gch0JW;VcXw?g7c?yK7cDoefhekH z^yl)fPHG$Kd8jz}gi|V9L3GOifIRZGrkpx#x)#2}eJ9PPXV^a2e9`WJ_Em5H$;%#+ zp+Ir+9E}_th64psmJ)H#XA*p$mCu!@lNj`e*b4qpaoUw?n`_$tHB^=!g0G^Rqq$0{ z!A15sE&8U(yqi5T%*0_~e!)}4GRI-6a+qA{DjXfkC+y0SsBqznf>6_yd%+z$VAr%j z$cfulIo_@eZflyzpKL{oL7mW#8iYgVaW3!9uug=)z9`)jmNC9 zlbr3Hbi``{Asp`ttqD}b53u)rnEx|BU1A37_3A5Nt+$b7^;`cJ@C?Hm1i;m&fTg~& zt01>k2-FMiPSZ7t&&X^qfhi z@*s^_&Ku@>S#buB5q%8()pfXvy&D2ekxbw)M~Q=~CRyyltnt&*KYdp@hdiNV=df|H|yTYY4_LjA;Z?wE*VPX-tld3 z&CNY=;a&q;)Ha{R`#y$KRm=-vz##XH8)43MafKd6{-!e%*aukMZvY@hAtB-ryKzQz zY2JrAzBxT1EE2N+i**AkunLPa69lcvBBd&Hwy<^}bHQ9{*-fByNJ9Cf?oGsoC!OrT zfoF7%DUeO0Ycd}w#q%7XcDLu!dp^2*VzCG^zX z3vJQGQ!S6K1r{R|{|_DH*8b0$8b`I`xCT2>?_g!CaN2s~QvrGDio9e)EKV?mOo(Vr zg+*Amss{Wk{s_*tGuN(a6CWIi8aLwPT7M%|w^gdf0c{IuP^Y2u^>5|4$FBHn`sy)` za_RL!?495f<0Eg63U|)*roF$WT0H%qIV9>!3dTo!+l(MM^Q?zEp1o2Jexrztd(egx zszcZfDoobwg75b4EX&$%`WM9QDa#y#H8TcZ9hXd^3L2B&dh8&0b^Vs^Ka7X&q@oFe zQQ&_8$>NaYK(NR+=srYQs4i?AaKsVOhZoRIeDp&sxHfjnZL|itrH4M}+z`=L5PPYa zk;ON6Vi&(tU$_Xt#9tE7g@Uy&W^3v4F|qz%-Y>g9cB}lEzp}PImFw$+t{|n@CR@1h z*-t0eZ~mI3Uu&U9Wf>sR#J_a*g7^*e}{PpHQ#u401O*5{)Cz|rH44E zMGm+7ZCGRj0iygAF+qs%L=FkcjMix;>}`Rj<%dqRfT{js6YVF$7;s#dzF}PzGi{mZfRec{{I1 z)Y^t*r0ap7A}vI{R_yu5Y3T zw0yr#ysk97Ta>X_=*PZ+#tbs^*GuZI3I3DB(Wf!|5OeWoQFQO4>%}OI3I9VrOs4G?3}^K~%fN{>HO%eLQ3R;zW4W znzwsGAx9m(pgj|0K(*gVA(S?agV$zzL6_5nEm-$p=^VFIQ zT;AV2YS552_Wf9yV3ZPv8!-bVGOtJelrfvMO+HY;5we%Pzyt^`HJKhK!GcGk!88{v zLL@W;^Vd2nt{sWkN-*84lX>V>Un}ut{i97Vm?(tUA#C#o$lDeYqr#Qw99>bHe9_+` z410#^EazsTZtdnRZCi)3mE=nrY6z8Jv7+ zUhM*KF~H^X@(Sxme9Nh*!X&`qG>Tv~O$yDVP z(t~7Tgo=B_1CU}6U2V|FME>u;`+)JCYiWV6#fNsBkkbClP`e-Y9!Y*QKfh{ zoZ?^qZc4M~sVzN`fjM;pv(?!|F=FktUw$w zk<%b!TB@+TJ_!y@vbF9*lm$Mb){4aF>U3TAe;>}I;Olp;8r;eB52Fet%P{l<{vSKLj?L4d{+{JxtovkifL64rcno+C-=8N8$)7 zOytZq!^=u%CAf$BJ7u*>L4|bTl|Vk9bi<&_`5oS%J^3DJYDiM94XdsY$QeL&e~#sn z(y?bpUcD0EQw&Q}ZtWig`NidiGONz9(s}ugHb^ff(_+D@?t#i>rvn{8VzZS#+{x8F zfk4nq19eF$F5jG;SSzD$0rKevU2v+Wrw$?5lYK=ey6ekCq8hec$luiHPVad;AzW=r zDpkV{cT&gHmFIRkj3@ud8&8Km;D4kbrIiR==Ba-4C4dKHpvjy9YR;Ab5;TtK{ud-s zp#8ndG#6-#8?dcr+Bb>S`7vmG^O`?w&@gm^rk$FHmzT3TPVf;)QhAmdU7w4juHD!( z`w)=C6>1{Xlq8uc5nPWj9PRFJNpnXf!Q47>hq&PUpXm6i`rnLivcMM_I=RZWy3SahOF_7p^{SI3!oqK5CFixw${;0O95UB@1e7kAmj(27@J_ z=Og`f;kB@TtT9@2C5Jl4q`$v({l9-Ti{_n7FP6Usg)BuHD(b1Yg<{hlc$#No^4EG> zci;w?yicF_Y8^@wp`cDor@J-K@ofLzc-O7j7}d$r$Wju~m`s0?TSGi>xdC4uF_d(M zMS0;VSLjFwLIa4Qf-v@$P0;tvkbYCEhBojnCsr~cJ8g@C$2r&n!kRoFC@@?zb3+TMCWo6>4QDja0o>;e4J;ph zdL4h0G9I@39nydNmohGP!QDg9v=WNxZ4REgE1Yypk3cqE^OBMf*v#MKvNRbGDsOAX zwiOapY)fS9v9jv>~ zs!p_)afZ460AqZPl{Z?0mOP@zK6gY;%E4&6^m@bUh?+OL7*$ARF%nz`^#AMIW?zoz z5v=v7UXxHs8UcE>8-!q1_)eruKh})EYt@|01X*KX%1@%sPqkgpP^ly$6`=GMKF`%; zj?nuV8b&kKz0&Dke^xQD8RkxrO8-ynxf|UMMMs4xv#ymb6P-vxJ>j@=di%ncPas)3 zC6ul?Y*xTEt!K&bgt`OvC4OhSx6CX+l&}s_B`=b=)q9M5A!AS3c7IyE|7V z6X*{l%qO%(16#$43<;VN;_pZy&a(CCjgf zg3CsAVPI$*3b0Vu=v-T zofzCvZx0bA@H_a*8f#PEGz8N9TizA-1^VmXe9oO|jfIU17)WQsyNxG;SXMPLUorX< zvMic6?~*7m+G6h2UCb${%gR$%5lC1L;!N=w4q4L7ove5(<2@5XV`m9O+-Cg6=dJ63MvzG z#&WgUuI_SqRcQ{rnyF1Xo>(*!4J0ra<=DT|Iw;)X!&t#mm?63ma>SI*gUZr454{u4 zx9G?^9m+B~d+Co~cGAWlK85wA16yv@K6RF}$+MO|8Yx;)WmC|vdGcdb-lXG6%GvaE z?~v2`YKWx@JVI?emm^UTT9?*C+w_o<7Nj-h0e(>FgpH#pVYj_3JL3^2+f0mbdUi$cQHyH#j6 zeKmSa#K>RfSfmQT*E5LZskqAE7D3JQBz}}y-DUjzG08w3EauqAHk!wIik74^aVy`O zjg*@YcUSP2ZC#hSq8remMC`lEY_M|mqp7F0D4UaR-a=}%0R)uN{F+NQq<*?R{y!8e zg?3lG9FfQfG4#XiAOOOZciJ{U5y!h`NwK#t#EWz4?eFOPrW`NORd1bgf6bv$tvMe< zFIoOkkxPnaO|uMQ8~Xepl2d`)zCU`DR;lBdd^$)j*ZvLHINei7TZJ}Y?C-BCIyvk1 z(R(aod+1Et@)(!fN)k(C2rHxDRLr&k*jEYf&|&icFbMk-O6i#kF#Q~2Ed30ej|IF8wuditM3-#?mo<6P#U znJ{)GqUalX;;Vi|F4xU|Sk$S|m1z}ELbh-)Kgl!7c!4k_~j1Q=_j^Dagw002eh8HjZ%9=|*P0K7bD zX^C2*;F-p;>U7R%;gGR!!NDQkb$O~ly>UkKRE6+k<^KIpd`uE5vp8!uXH$6nLM}Fh zn;xaS5-O8I&Ter5y6?IW?W%(yrkF+vFG&{YnE-wbgy3t&v&~hHzyLn_hvbfqs+)xC zFA*Z$nTVwfyG?Y+B%eKq?x<%P#WCPI(W{5*`XMokq8)jA;hG^qj0}C@!vFk4!UBsc zYbO=oE#4^~kr+~`m4Q{r@7v$xO-*+dQh>qP43K;G)4q%vz`t(t0FVGb>&9?i zK*x6P|9plN53iNbUnM2b!~ufNg9}_fEEy)sjdvo$80K%mi1(f%d`Ku(#^?vnVb$BM4Zr{W_tM%$*)7OtoDz%SQt`a~-DX*`{Pb@~p!f^u zBVscY(7L3EDdDs_AVNsAQDAm`*t%2YT|3J&uYN{`t}*9;E5s=^vs2&k5gYE<*{)wd zclX`W>Wf4TE1d@)-9w@-h1_lrMx;D~+*N8dVkJC;4<_X2u5fM*g~%Owq3eq525TX| z&?t|+k_h!!0F4jD#*Wj->F1R*Q3e}`boDh?gQP*R)<`t>@=XR8+V#+c3@fY(@1J4W zr|d&eP{*KWx9MDSyM<>B2CMEyR~o|H7iU>*$u}u#0)Z+tuLqhQx1R772Td}J>W=9U zRzGhX2taCACQ|T4DUjvGjI?Ky?)afnHaEECbQ z+TnH{qYSC_K;|%30AbrHtSkUl7UO9US-A1OB|LM(sa-QVDu?{)N;c;kfRpJY72b_ADQ(0!Z(BOe(J^MsZ z5Xpl@xl6PEAPq{wB1{w>-K(KMi%W%UMNU7TOxz+8_7q&s4N>~q+|-c>mSLT4t~$=c zE_;;?CEACWsd~18Iuumbi@H<1k)y=>jqtux>+1^PRH2ifLGOBj|b@j*Vr%C zeLtV3s*4bWE?7SScZT)Y!m~6Sw0Sn*wT_YNt(Fe!Q9Av{fW5w+YVrMmQ_I@~s9%l9 zb_$;a$Dt3-aQuhkSangbrP{bR0p8EhUwIid2nKeFqpp)cQoZ{nh_h`PFTd+Pa+BdG zMJR4HGGunzWyF*_(=Wb!=vi+*n0UZLYzER3`6Bps(d25^1xno861~ehy$j403U3ya zbkoWV^c6=y1K7;Wrr7x&ACQG22u*2|G3|0jtAN$55Bc zltY(oN!OEg*q6&GWrzi08%a#YyiI#2ST>7d_k50K@?xhNsuu)rAKze;)Vt}U@@w^q zJZ^7^pc+lQgQsCy22=uR>R}eBBI%VDX{~_puc? zZ(y86(u_Ebc9K`IveDq6~x?ZsI8mUDX{@nT;)EHai^v=AauB1gwwQ8m)f@5>tJwpf;7BbMx8 zG_o6(NDXDc1tUGGe!I55>lvDEghqt1RXo}=YFdu4J-&3}k{SipNFNG~jJ%FQ4oA~Y zH0NZSz=%Vdc<4YeKGv0k*JHRJ( zdvd*xXtX&ML7Arw$>)}cZ9mjM-R89iP*<;M^4DyCqS3ZzA9=bMVzYmoTS;Q=U)qO> zPA~0V30Tp22+*WcX34Q(OlXDk zg4MdN{J-E3>rXOXb1d&_KCB*gn=h$!k(mLYJU($#(`~v5pVmnwF;_sXTj5h=57s8< zyokGXa4l+caQMdrg>>)|ee$_1>hBgP=ruwW0$vjZsjK;=q`@(Cqw@5kv3MlvWhB}&*=0!4_rl^p~`XwXmns$VH zGh$dlkxM90vPUqf){4z;uNZm6R39^bOMUHGoM?1ne6M)a8J)%7 z!K&XA5nVcfq@khS_P5r8A|*iVzmxKYeT!L*ricjM?mWnq1&U;rQo}7qRufks9uT8< zDIB@+2>;_MY`>TlUv&|(rI*OHNCd&hH$l=wIvugnx8id%5`w9Qm6N<;S-V0AhJwIy z+wtd6Ucqt!$Gn*CkeD^(n&`s>@IC7}2Ymux4D!iy6gvTw0l}q9;#_C7H6kMn&Qh2$ zK1e&&Chla_yJq|rm;Io~I0OW>xA=$d$pT2bZbpUhxs1>MT-;lTGEvhhjZ!#hv9ewq zhY=K=nhb?9O|f%`Ucs3|=~<$|L&HE~iY7;3AZoGP$E7p>3TFderJPkuKrK*3ZxP9s zA~o4qtFO`7*j-5;@FSc6IVd6;+Sfi$qd&F)y7>l-ESeG$JY^biXU-(?pm*kFXGy$u zY7Pn~Yq3x){{h2Ck{b0x=S;U4xLmyriT6%0jG33uUO;V=2c(7S(ij z_p6g<^K2cspY?koWnIGHUsNR02{@TN*=E}=@f>SLK3Y80^matL>o2(OU)h?f6O}~Qxyn$n@0A4*=RN{qifXNO3tpv@7pmVGNI>aO z7CEyY&-60?2McLZ*A;$T>Jb%W<^Nrn!f}w0AB1&szmu*h0(xTzX|rr}u=E{RZbPAFd&N8@>!_{d^@)2H3tk?sv-cJoj05T*!alMtE=qVkH3CP)&W@$q z12wZUSena>Iev6yVfIJ)-P2Mh1OsSH!q{e66}fP4VCuj3$o2Hc&&j9D)pp=(JA<^} zU?aK}M*75PG;iHMThEKmRnV+q6|IdW0=eg}a0^ZH*cYs!W{hv`DvQHlH4msB9mLXn zuY*pqN>&i-(h}D>8sVEz&Q-r@-6RZtK3!IUROO~(j6Cso+L{vBhx5-aDQyMy{KDUH``})fbY%q{Esw7(>>-qRYoeA9 zCzJ;)KiqrV#xR~`0s1o7KjdzGrMEN@3C{Rn;u`dMl=GCSwOLaY;;4tH;)}y?mxI1q z@N$yX2!6nbQLCz0@6{|jhUJM!_-2tsO9WZ4m9P-HqQbH0zuHQ>T_=mHFgAID2P~ZV zqCUbBm#Kp~3@b|3y4$c?K)pi?lO7t2qq;Eg{0{A%QaE45@dzQDvp(N^&=(epfjNIQ z-3_P*WpWCs{`hY4xC&10=0^w)IS>TC>V}2^p*=fF9;Ks3Y^h~!*2WrG5kjnXkxe7L zN{mU^_NEV43KxuqyEabBC{_^+iZBE|XSz z8?`Z#{*M0qB{yr#4`P6wC9puuZG$Ey`;zyO3LmiHMdTP$}b0KtxZ4bZ+D`UT&} zr90blEqza@hq%MXHC~RX_`HlDzHH<-8&4cQ(j$mE4zk-uwfWW(W54QL*bdfsaN&kr zI9Jn})PP#H7gXNp|Gx zeRx89vdJ7+aEY_V4NLse_;JY`tLPbG;_;r@8+&sZDAz|f*d)J^!PUKkC=FQY2Dh(} z|Dx5F<#T9z?2W2Vcj;CvJugUFLvdwzhnH8tzL5$6vsE=9%`YJFwg5Y6AIYEd%ZqCqtYZW4)r0n}X}J7`iFPbIb{@2RF8JTcKbOj8u1Ws3wi>C3#~f6fH& z&#OSYl8Rj`vr-K(o@Pb4;ADgZj})nI^tarC|Btv>a;Nq~m9W;^cEg2!`qq(hf<97; zYg_0Sbq#2KJvmgi?!$>up2e@^Slq{nmG|I%Zr2}*@z~a*RhP+-6!3b;%?@dKdE4>? zWa%iTL}q<$HJyAj!LQsNTzpLQac*1*aYsaf@ByXZX3Qqx**VE%1zSQWUY;VRfu==@~MFGI8cco!K)$hwc;fH@`o1 zA+%+wBX{#t(KX4<`ZLUV-6}F0Y`dfz%*nBlmx+JJxzelxf%fF%2>Y{;EB4ps2Hk-d zw_EB<$3NiWl-RRpThSZ75xSYecBy@G8SwAh99ra73Jh2*8=@}*7Qlc4>V?v`j48rI zOe4QA?($vQiv37L-+;me(<69`Koi#*mit_+G8o1cKy|McVDy2^53bA}7!nseuTs>A z&oV+Fe*)$aMuH+p&|BP8Nygq}@N^?0-j_48_U{@$6b>{pRI0eUBd*WT0RlGYu;eIj zTt=tc`(+KinwU~d50kAhraA!j*YIs{|i3^XX_#?-zR zf#|vt(N_G=kU@9)SfiMfqXHP7Wd5`I_wxvXOtkj_gyN$eg(WVPNeO!J>j#>niU1=G z72x9ONr3tvMyWux9+FOc7Zc7n!3_J+VYnl89^v2(V)H)L1t7e8q8`dA{S$c3gE2tW zt|fr3@;%iL3x26DnslUEkTi`DGWXkv9%ZT`D6KkGDzcc;gfk;T?#UkezzMpKw9pIO zaakLF;TXs(0DV%cL-gky%b;@`?KzTvtPkeV(m|*AQXp0!f5Fd7X{3@$>!jFEM=Dd9 znB!|VshP)DC$I04HrEKG-u7wm!f%OUnZDCX5+AorG7Ep9(+n}(7(p(7H6dvDVk81< z8m}m6`ItW)2N0h1z*Fr!kJ5sTY@4NR+N$2_ZjP=dS?{vqy*H9{NYOhQYvdA)P8s!X z>x?YG031Z+I=Q*aro?!?2Z9hE<2|$o3(>W5G@lpo(mJS}Ns81i@fawd@5G&B_h%Nu z*UAnX>&>ADV()~epcbs(JK=O`S}sirp%fcH(0mbC^(oJX(Fv)lh>D9T3dNkQ&+cM) zk1t&GA9In)GFTQACoFH#nqKzMQ=Ra3GzpO6N!~iNptn;^+wOG*P7HA0pBD5X!ig1^ z9@ezYjba7g-jcUsuGxPo-Y)>6CWc#wruWhvC%igXW`bV0b;f?VkOax|f*KZ2Zmz3#i z-06kF5!2QrUat|AN|1~!SXa9(+vzB`hg#Z9k=WyCw${tcu``MJkL#tIqpn7HbrlD7 z6@$1GgeSMPn1c$x-v*Ku=;0{1>UFzKo)oEz=j7`ul>wYTZQ>aNV+Pc=UPL6Xg zr4;sew7JDHqX`DDpdk#;#H4~T!wFl}kOJ&UADY2lxC${B0<0=^AV8IUPj|nrd*Fld z=4FrJiY^l;0=iNHPHkc;J2+p>-C+%Cjzt&;(7ChoEnhYY?q>QzL@w2y0*f}7p*^}Fnre2Nm@c$?6W-b?9Z(^8u zJA(n{oHbW7s*TBeHuK||l|m;cxjwYwAonDV?_J0lN(QxSrr;oBG&mY zO=bKAZ{#inL9f{x^AB*lVHSRy(X_>sWnlm78yD1HTUSyZo5G_Biu+VPn)q%+T)SY= zmk+#+aeoK7U;qep#Nr;DU~{7tuBLMsx@;omM}%|Fp5Bo~Z!M4FtjoEIQ)smTg1tr0 z%HoD>Tjo9NPXfmLc^@B<4+lvC9dE@#N6m^Q9>1`rbH@`1+jHHmP0ev{$8uOK(><=O z-h$7f-^ZI$Dy6)`n4Lnlj_y)^nLBWa(%3dX`N%sv{9-I^&?mS794%3pwFz^SpY2rT z^f!}e2bN9FOy*xuj+$Cw7mmKQ{(3$rAMYzKF^R;BLUAWb_Q=*Hqtg8kp(<}c+X98K zX5$%WQaq2yt7B|`anWLfZ;~Y46-zsRZYh~gX84NXyba>%3@X(6BEzZIdwdC2-z+fC zT1+@B)nVR&UuJXNnOo}KcBjike%Z}wBuR{VH_?Oe!Ihya$TW3J2jEpKZ^NNQ$jIN8 z5_<-XS4 z;Ypr*NYVz}$VCLJdT?Y!I_To%PMfw(4`@`O@I|zB69p6wRbY{+l{7WsHsDm0llPvvtSYg+KjM9hZTZILu533 z?87iaA9zQQdMCd0pwIaabg_uYgB8}O7BSMVpgO#k+?x!3v@-lH z!MX^YOJJPoXH;^0&qjzn=ppF6lTIr}O4bC`3oI<@eBWCiZ-YQZ8ZtQa;iGS92L0Oh zePv$SsE-SLjVYC0O5Cj6t0+_}jw#t&>_^0BKH;8=o$dVfX9<&G>|QxsLHT*l%_3TUS$B89^*?dy6OJ;ca+#Hj~E+mD?ZAAu3Ue65wV;Uoc>Xw<(Ox0|$F@k$FyEBVGzZ z&HDSsEYs4H9Ysz-Uc_sN{6x}sK-lY&{iYZ4uRP@e-9f-j-#yY{SvchF;Q-nJijHqL zfgA%4?6O!ND&iipxO2g-k7yP8!}%g5zU8m0w_26FBF;}npv-yuS#U*kor_WH=-aS( zD?0UkpjwxbC&JtskDUR4k}XR1+XTL1sD)4>$=(e`SBiWXG9Nc7mhx4|{VJ390y%m% zUH;k=AwfH{6mU1v^izu$`@!Qco768i9B~sT6u4WMmywPSmm#+NuR+aIyg0vC+IR9% z-<0*`ITOngq;??obEaoW!M&l-pF7+@)+fF3^~kE06o)E; z^M0-LIo^48dYr%+`;%dEm6=;%xqi88bMjXEG!#0fD#?aUiD9R z8ujDd!y=q5BRAr&GHJ0`bkdbVi#WM|qaAHUYi5Y0E^8gpg{`vp*f2T%kl{mgiL3Z; z6SW6-Ngy*miI%HqHka(b)nMsLdV5970whtUI^;_MvaLkK>xBg{Z`bj5LZG}Ge~ET@ z`eBR)kIIBiQ5zSR*G>1v1)fk7VO&P2B-rC2Q&3WGB}&VWqG zdBiXL&^sBElf)s^F~eBM^Fb3~unp$@E|DeOQL^Vq=i>8cN|JSOhD5B$KUWpL$~-c= z8#1;9-@(bYGtc(TBomNNt1e2p>qRnT91JoUjZ{kd4U_DO1|x>EdnEX@i}&Y@FVfhk z5PUza91B*u9I8Jawo(1|Hi%CHhg$L}oH%Z3efqbcluiRhe`y|vahwfQ&;><)GD;M7 z6M4swiyPs4)UyAFL*K7wN&vpx%95;+MKMT^@SY8RjKkGv=6W3&`q3lHcReWd$eK9F zv=oGuJLf*k>#2UCMm(8@{9CRg8I|bB;rF}TL8vj2zj+7&TRhy+W~Vqi-`Gt$aKShH z`mL^240Gr(@(F9bC>{;F_;orfhk#RsjnU*VD{3;m(b2!dule9idp2kj7(9{r_oB`C zxL$gO*JK3soMu9QM*Tf;HEQ`DZYEL$0*&Dyg|=|aAlwUnY$n2wHW9^mohF%61{5B9J?5>*f*GB%^C+VVi6^onD?AayM1-RLAg@@yU zMn=CDX)x(uCo}eKOFq&yagsYt9Cko1m8+K8l^k3(qbop;l{k<y0zZXyRZ4 z$ox04jB4uOkV=$<(Zs5oLMH*E^sh7!g{m?-Ymcl>u9y3 z6r>48s{ou6cUXP9RxrdPNuhZH7m23&%((m)6`gh9AG%L8%%2AAudfS&OD ziWt;?UCx?|M`H$JEuTJy_(b$_733u$O0im#!n&IiDpbZ@pJH5sdQuhzm~5V~YR zRO03`q&~^GVMS7tSkdOfwvYrXd?5!jZc^@4Q6c(^Os3Bv-Ua|Il9ibVXM2S9)A10d zN9%#PLbpQYG*l8RCW@CxX!9HdKE;nc?Lv~U)0>IIe+uC7*=Oz^=%u+$I-c}+IzM#| zkJ8Gy&s6I;DOvL?pHbZcdee?+RW)a^n3!-Td#d0T(T2%pdFHE8E9A{Uch@I^(*q zrC^J!&zOW*)wa%k?#ilJFp0=~iB>hWF>*=tK{Hg|1*+DM)eF7SkQZ^8qf^cY%TUgF zDYfdn#W(4VE<6!hVpq6{Z5K5!(p8~3bL>v z!C;H{ZIZVAWDd3J2N^gi*qXCRwF>=;2*k;gS|-SO0HD4GKZAI1qW(Gm{v)auGD;^0 zP-iUFxo(?V*GgbZo^j1aOjMPx=JKU2#!|v1#bf^Px2F%cOclxpCah#ZE`=$cOE7fV&xHFQ;-%jv>RFFLD#x(=&)x@`dv=GjoeaOKay5<2f$6d5-{+NU!GEA~eq z@xR|2%YE<-jB|+)K?8K8HXARcN(v%8!|l0B9cQ+IDnpVpHpOt-X%OgyS4U_7dMI>f zURDa7;3uM6zsKX7ysXDp3USsWRN;Z&1Cmb?vAjGaS@`5t1)9@#izYxExwV>r-kr|m znjA^pkI^kM$*4ySsT1k=H?Ak>aIFV8KCi`7Ey&?fhC=2Vb7r^uPmqnfd;`~oP+38f z#9e_W4y(vbNE|qCCBJlc>;i$ae+^^_2QZ_&6eagg;h~87lrJHS8C%T zTBJ{m%m%*Z@=W|J{*;6Hat=|x!qj$w#s$xgYW4Z0Bz9a~lOjHxVgv`U%_(GzyY}=H z9V*Jw31v{T&|DvPA94$zXsEPawl81< zK$WTy*}LKUa^RHJ#FHBYE3-@9ByLaY%m;cW}U)U9TK0k$MisEN1ue(8*VV z!(u={LZvDxoPtS~nE5CN+bF)qHW=JxyU?>fDg#5*Jy$0hz*jUC zC1qVE&0-wr`Y_68oX33^MhgzCo}M#C51F92H+$V9(8sqI<)|e$%NH>ILmvD)5?Sc` z)B=n13G=9m=d1s%m4&mmhci{jj-P+!e7uNJX`#w@&TQF`^PfsEhWY_sPz_6)MsyKu zoDM}Lpfk%_sIVxf3xhM_TommGL6Yy`0Ry)h6fMly$ok#hn+T2iN;b|f^?lW2HY}C4 zaYK6s#eK)reoBay3r`IOq1Gt{SAICYOfkXtrE98`-U3W0*|fez(NERhr-1@eugQ9^ zWl*IO5-pvIR@Cg@v1uIYFn_+kLl2NV53M!qlUo_7wH5~LyTPC4$qpNebq97NZk`!G zi-nlL{3Tpg!a6cU+x`pT1!PZrNbax63GPZB8G*db5=1h0l4H#=k(U%=1h6x8-S6YZ zH?%s~1+LPLDN6H)mk+l-Na3o-0p5N-?6CMd*zO>ATLmgP@EQ!|h6qCU+K`ci7-8PI zF63?(-*!p5Ey^O-U-ns4YU{%3F=?$tVzTbRZXEa_>pX1W8Rfus>+KEcjO0VLrkl1lZ5u6+@zEew3J?r?w5_NmOSh)V;74rDApU zmv|VlQ6G6 zN)c;qMQ)dvWHO2g`l7*jWiBrJos$K#cK1scrK@0^TKQ%(ys0((+j8UEc~8F~Ziog@ z!G8iz3KQN_O)Cfl#Eu{*wehQ|)^>@F@`0(MtrHi9Yvvsb5SP`r52ng5#ZLF(S2hVG zWu6NjvYRn4MM=|H*sI^Qt498C?eMF63L(2^bhgN^*0LjhQ^P6PL&wRd-Wt{7#4W^i zI>o*P%WRAO9&V{hjF$?fo#Z4{<%KBgpj>G&ix~c0*!0RZ)hC$>cglfd#IoW3lRph%M*&T>=1^|n&qop*LmN^>JUv?&7@$Z`0Zl4|cwCfhFqgaz? zhp8Me{F!YAjyRmCIazs?pKJ$)1#G&-qddUpHgR_?zLqi^v|iMzQQ8%p zK1dcp_G{{oJ<2jhb^uK%LTlksi4LSjkbc3%;@4*Sc&;0em{``7M7_@iUmx2={>x`J z_iBflatyK}*C%Rbj#tX2Lh$*@0%U`Npnf?*A4%?7Tt?w(HUTf|SUrpzXIW;lqNUMo zg7gbPl21Vd_nW-Vd2@qPBGg+X1$U+O^d9DclGPZoBMeL~kIk)j<@k`eR&NpJ%pBU# ztRc~Sl>#+3E1&t@$TG3yzsCXp{sfjpdW>gdNdT3%Fd>Nx?ZzKMfrX;zK~u6)zP`2( zOQYJez6+yDvL@~3DB)4pj%YcU4EoOgL>#JvdWEjH_xKg{hO|O3cN5_I6@W5xl~@rj z+Y17png+F~-(99WH0pjFX~7Qm9sJX3j7i<$r!_2DOA6)Oc4sI}Vh7YWS2Qs?)#Zo4O z@*J$8zb!>?3Dc-I(x_fZi%am93x2C*qY{icxn^IybW*)7H|o;h|<_JMBkWRsc>kD>%`qz$x>z5%Lh4`7un6 zx!6@08>w0%__1(Nmskygrns};q?8;9#MXEOGf3Hue99qM%`{aq8u01K;8ghA)S?u~KC ziQ89~Ndp|$ulE#j$hmf7tEuJob#>iEO>mhq+87Kzgj%8bQ zz>%_BnVeLub6Z&#j;XY_gdG9m7cGm)7S`|x=#nVL!MB)tqxmQTF9(OI0qHi6DfXun zKU7c5sMhWJO+JeF2ZI&*7yL9Vp`hxeldR+O{**fHB;sbkUD`8cvo8pIc941Av{I_2 z%YWxD&kGYVmSi(ICMXM=?lbat^L@gbdNu(c2sKXNv_?!qI{&P|2M1{-LbpV`cc z_(}ujw0fGG=8w2zjV7o>P~>OVX6pg0g{<9q($!c$blN{gr8GOtA;OG&V($Jhw&0iK zPVg2a@k%n$`uqQaQ+)f)mXq#A%heDY?vP{gf}eA>Ctx7N9~K?ewY;~lnRH;ON_?4` zDNYyJ^JpIG>9q|<#U<`tKaf2@PCg{D&*?)%fC|WHS#*y1-iR+PR?9CIWJxB z*)0=db5WHZ#J-|#+z=g)<{Gb~b3y2Wg?pM89oPp4a`dhwlB*ka8%PTWj2(XB35mHA zzzYe^#QCFJR*jt(?W$m$sbA9lkB%s>wy`Q2=<8k)00niAQH_akb&Ht;`Ew?h{Dq|0 z9#%_@f@kq-mB;ujo8~CU7V@Brb0c=Ed~wzMt-=njM;-4(fL=S9XF#k4II@Nb>q3_T zN>*6U<=_Reol`$NxIr}F)6DE-3)iXgnlc4lyR7)+WY{b@xgrvoQ`Q>ei4{q8;2ZcJ zC6mp)ii~7JQaP| z!s<8Ey+Owy-{@gR^u#~|4fM0f_Td06PS|3n6^bofZz6UNRPDoCB2qa%427pw-3W{s zsv`UXu#(g6H6zxwsc?>dkEMZgHfZR|Q^wmi>|I4-;*0G4GllVz=*Ipnt@8b0gNS9;q#Ud8s8G~ksG(^MGu?)P8^ z+^XQ#!SNoS7Lq4|`wt&zu^Kf4@RfH&QQHUTGYV~wvy=ZWR3y2n=)AX=0VNg8RZhFq zpz&ykgrP=b3Dr+MrBtU@TcxPM>$S*@4%FRrnRoZY&xCXh4tA9-RdT~+T=xkV_d%E5Qp3wh`Kg=Ps(59LhG(5~p&#(yN)XUERXq)Dsv#e=mG z-ZT4LykkOH#|x1K^yWw=s!n6V@QtwPvE{fla3wF5HE}LuCJjNZi7TFgMmkslz9-<+u-eF5>HYHJ@|CPAh z-t4m*svpUqaQ?H7A9_aXt1qj=;!3g!sfsOD=?y>=AHtRv2R~gFxn!V#m+dH^NDQ<=Tl?mZ}C7YKLY3tm!K4~iu>}AX8 ziCV(JLX5tk4a-S{y^BJ&Xn}6~WdvxN0}mE?aXi(^^>ZtqaeW9blCycx1Ql#v79ycR z{|Cv5N!2%inNrw%Nq_do+Qo+YJ!MK7SSjo6xbt%2W~R0*gj->zfSlWyKUCyj&GA{o zEfcK-eR_avx7ql7tH0+v)vAxzD>OpJ#s1gm_Kk|SH3SUh|B5@Ha@A%M&MrYy-41h@Fs6L~Swa4nK)h3dVwm?LhJ+oOJMMMJTA z8QGT7fdGo_QLkJhx{35XVt~tV?%9YoB!w__qth&!Aco5ro!3p2R4*DW+6iF?jkFTz zoV5j`7B|}`5}UVP4j=_jjq4@^3%Dg6;wNi(ii+Xdyv@GH*U5xJ@hiWEw@}!^{;9O< zRB$5x5L<(-^8%(DgZxWW!pWDvy05jVA4|0+NzM9%`3GKbD}1z$SgYPY-k7c1{J1Q7 zZqA`z^25(fY~o0=?LWwWc7xEN@-ynl~&`%U9{<^*xkL|lyu z_BDS#&Pz4gXMEP{UQ48;tydF%oE)u5;I9PKZE?Bn6XNJ%AngX72G|ni@AFk?oFgDA zoocXx$oqSOj&9l!ZwQeMi$`J)_86Bu!YohcoYFa=|LdEllZBR)2oDIW2J^ z@X~8AZLm;CT2%?hON4IK>hX{@q1RYn7qT9Uy@^BT71+2d1T_0vA|i}_$Im;r*YE(* zaO4?zxIrTBwaA}jh`8wB#LZrA#P9SVIy2;PB zi`Xv`ZduE7vpqN>gVmW`lO_yh05w3$zq)5xDGp8tH$UXtiL*N5(Dd>h5S_lcV(08} z0ZB$LQLpQ7Tf19i<&Z?wwU5=<;^~>bILgBpwrb=uw533Seor48P2Fu!eG`3bH9k&+ zUxY$lAtX!N4>U3#dH4&J=g{5?SS8GuNRg5HD2iYcR!*jEUQT|qOKZxFf2tw|fPtDG)zXMM$`NI*-$#`Q{Pr?{-O zrYN(Vztv9t)U~=p(IDl-nSB{(e56`1nfXoZ3nv3C+Uue%3~#BU%OPcE+x`Wd#ar#m zK-Fgl4=neG>shAa{^i7TeeYWq_IKIL0b4yxe|_7f|6z^Dpf#iCeZXbXw#hL70Io{V*&|f|G(Jf{aJhuferu9_N3Vb_5|-^*^j~wrA0b6245$qby5E-|A*1SC zdYjTaMWjCBi_hNRsZ7ER@+9c>(=+#(^24T$rCol$42MS&a-ECixb2q6uTsu!sF6>w zU!Dv32w>2`)k2L;aIqAfy{%yvwmP~q`9+ZZC=XzP>S*pF-1OwYMkih3TtQqi54U@d z25F<1!rO;ne^AiIeIGg_vVtC=N)ft4uR1$#sso|FFHM&*I67DYf#;+mp_4L$3E8}y z6mO%F{5c_8Rm&t~P|EgHz5X=o=dM~jQUx1~~!{_`N8v1X#*d?&|Ke@5PAQ~^@ z5rT)hF%wxgMP6IVDtElyEGe)LpFRNQ_FNS5OYQA2kyIR#K#?Oe{&v_Zl*Z{at$%k|#6YQ)Jq98&h_R$&mQ61Ex?6Ph^F zI&UQo+vRO=3M=;Jx=&b8%e$1$Qwz3>kD2k_{4yfLoIwXFJyG4kSMrEbv-rLVG1y?%H%Ezw(tnvBQftT(awGOGj}RdQdGCIO;trG zc&SOGPRD=vm@6NT3U=l(p3uUcVrc~hG3Nciyh?DnHnQ_0^fXZ;j;V$vllg=i*PZ2( z$tsKw^Z%IR))cC}=XEktKPy#&TpRS>_ZGn`r|L>M_Py*lpwQCc&?_-4k9JUGdAVXy zw*%s%cZ4G;@<~3U1x18)6i7s}1J4CGsPe=e0gi;awiT46>#E3xM!@=@q1B2ZA3B;0 z0T9%H69r&(C^M0K9T*2r$J8u&!@%s66c$ybIPnlqj;nPuv;rgdA4dWJ603tjX-#vw z1iH+%3sHrZ$8Kp8w3TH=ZA-fLsK_GLp(J~}b<)G^llVrTaj~3XrgX-F4j3BZ_@KfY zxD_IfBEG};-#yjN`0Z>d@D3e4)hLPucTZErhISzr{`7p+e@Nz@3PqxG(jE|yJ1aU**IU3I zk4QA;@j^vU$AUsy*AgI+C_$B0DfmIPa!kx#?x~H2gJBnL8XsK-mca4S9h>m8Yp4wW zZ?;5CviO~yY$aembkkhw0`@xeR>_N6iAy5Dt%c05#BTyq=N+#j?Ji*7BsS+>Dj&;L z`Z1eFA4NvglL3z2-(s->Jj!^rZ;oGlf6W7Xm2~dvGeA|(NeZ;C$F0I&V&-TOq|(v= z8}$W(W#yym7mAtPq2LI1+?;YGI^Wj?+iXn-RCI!ifr*A!3IEDB|9GC$IS=7`gcpke zS%`in3hUD~jPuVSCY~WR>iTw;-DAiMsU4P-T!Ns1Y(NfY4p(u(M#Zi+^76-Ds;d*p z*N&3K#qwWoj;SF}f#h+BySBl@Z+%Hvj>tIMZq!_&qb`*|zWRqSXFXSMbZ}9z1o5ds znV0s2;&|b%d6?x3&a$`j!75t{gV{DMotE`xlLi7zBgmTc-{|7_FF|9gaH)zUn4G#Z ztQNWsEzhxs2sK*~R5Fx^(lF@PR*MC`Qx~$Zkwb48OMu|76nJ5WNMlbZ;*!5|Pi$FF zorFg&pB=KVcpoKwj@D1vlNkwUncyPN47a2Ct4G)`sk=gN8W91#*ZcGt4v8Z5MGqrT zi5C-#EUv#^5(aaLY^$kt2Fj^hpy&;4YKDW+hvm;ulj`ZI2V`Tp9fKD)_}tz?NwS4u zso?t7G91>5F}#@lK5ab|Q zYQ0}BwJ3`olWvf(n>aQU97z%nnyw2yq}P~6L18o?OC~M~68;w?q$?ITFn(&u{#$y% zqRtfxX`A>#hOQO-9P>dDGE|skiAcx7Ut!`{!&P2lj_+pJHipeLhA-JM2v-H(%f0rA zm2`XX796U#1Z5*1a>xy&6al>THMf}XniGgrVEwkp(jlvVruUTP-Y8aEs6U}l(Z-Z# z%KHv2K0kQD%g|X%SJkdqJh@*?m$S(d_K99^P3QC>Tf+u8Q-z^#1ZsV>UQT0uY3kwS zZvkuQqj#cBl2OB{VhbIX#z$mDx*AAtl=uQz(YoSO`&(jba8T3%!~d+uszb{`2Lc(N zBBQ?=La@hAauj86kVUpcsnwq;pXAa>QAUYIE77Q4(J4WC@RX^RAAViS3=19}TylnR zdvfSv^}LMi#%sg>?8`7$)iN1c3G!SoG>miUXa=M?kJ~#J=j;BU<4n7CZO_}`Wg@PX zY8yZHUtkBnEqe2`dxM-dWEVPeb-HmZP;QyCfXJN?WtWZtk2`d(FM&sv&g&PX7YXsH z@p(Lpxh?Up>zPs;4J1`MuQIuco}*K!0DPmX5~8)=(;v+}AXMjw^zO z$~h{%WSIy3y?sK$1T&nR@L7XDL#)=(;wpe$I0&lfqGNO^oufq@dO5MH1yKPGwMP{kovwoskUiB*V^))KOhyt2GE(+vp$uzNQx;)a6CXh3~2 za(`2v_bw4PlH}ZYvr)6nTu7)Zmle|Pgu5agcXUK9$zU4=`*-pFp}z0MiocB^%V6G+ z;*}r1=0=LxaifHTeG7N!sS166rJa%)#0(^BG*+L$uLDZ6h${285QjpzcX(hu@d|-e z&)2>se*o$OGf-HS z&Z9%(yxqODffnlVcABBb&J6QE6VqR=x+e)A>1>_bmqM?3yqYJzXK5bNH(YHo39((*)R=@-}2 z`60+0eY4VBh&=@7+aSa9jg|wcXrTE!;`q?$$d~Kq)dgR)7f%RfM77o6@RCMu?-mF_5@84L;i=%?o1jUPeiU> zcliYNiNnn+1H-*WvZD>SO-e$l*AV5#R5sHpc17P&Si$J{C3!-r0cr=wp4(PE1os?V zW2>E_wmuMQ&-7@L=-PBjCqMqZRwQ4au6&?5PVu{xQ2PKx#c@d9A094$4tWxS4P}j!kt~}h#{6aq)-F7Ehc59H_B{SvDui}5v{!(@E%El;ReQn z>MsRngFS6}`q^_s&#bYtiu_dzOJfodTPuuq&^%PHutzypB9E*M@ieqos(=#PL??*XkjBOgqdZir z{dOfg0v?E0w6p0+145yug?q1BHI-{;`(rU}Us?vsQ3*)LaAef_;|>;@Jd&z&(A!Vc zVDeor)1OkrP)2Ne=1JFxVQ2VPTnUTj=CJs;*WIUTxp>s4DtkpkW;;;rr;!1xv-LnJ ze-kIV`@R3mvN%U>n9%>6@ab2Q?2LEL?PQeQbLX2i`6dC=q63*ZBmw#!e?1)++7_v!9YIF-}`1j}Q*l%3)w-cm2<~w@D1!TOtB0r#b0TtSJ=fI_=VlCu3pnl4;RIXdZh0L{) zKwDt^13XyE>p}lcz1~Q?rdq0L*Qn?Wl$@>;**Uu5A8Bqo*8xTfmo5b8Jmay_t8R@5 za%TMEY{`*S>yO?jt+TO+{_`WA>AK4Vd-b&)g2z}L%t($Ky1QD?yzmV8NQ|&bhErGH zxIvUA-lx!dMj3;-dhG!I*HSPC|NU!>_P>glEd22988_41%~G7fpkgAK#9|saZNC+B z#@Ji8;w1eRpK+{k{v46rO0q8W**bU^n-qTNSn$s8-0z7WR9)%DOD`a8ADxaiSgkOy zu?g7wwz$R?Xq=l7I8_$*r&_S;zYc1aXP+#2YhLD#xw;P9ACLrS9`U*qjWrNJ?)k|M z#=gM~G!?}UAWwg9r>dtu{qeRbQC%7R;b9nxNdNYze>&Q~)h`v4WTM@_4j+7XIi&wk zmQ@w`%xW=itv9*6;=cb&L9*dbXD492Yrw*R8XSl{9kBy0A^j6dv4xIQYu|WYHz^4f z#Nqew-bc@)m`02wObx*74&%I>Gl#veD}L#&bQi2gbTCp;<~w-1L?Tm6?RAnn-_CWf zXk+kxWw{`#*Y$5z-Ymx6u=T|(h)$D~sD6RupD~o_$#S3lgeKOW=;Y*9OM;GJ+7y03)X$KR)67%>Wu z_+@@D#g#G4;qssEsh=RLzb#WKrK4^uVtYXaRt(=FNGNt7T_;oia zAY__SXI-z^|K@ZdFrUurpwI@g8V3H)(#7OzK#Gqg1y6G8Wc>h}`8Cs2P1=_54#ICV zKoR6wkrq6LyKKs&60nr=9UQygS#&WIJymy{BKh?Gh|lUTBaDwF!J1DNDi-m5>i1(n zX*n{*v^UK}zu9z^e1ZLGn-j_3mu?jM{E4y?sY+x>R%{<0(9j|7!R z=r0&<%s8|1bnjo*TIynf3yW=a)uz`N#U`5%QylSNrh(YLL6!JX7cH!0*W1D4mUZFF zLW6I_AB3UzUmBzgWvxH`d@!X?xt&-QL)DQ>(uI{1-RH)lX}LuMaTiM=W|S>gtk%;) z0Lff1PQ^YTH5UtYYZ+o(m`4*hhM*^nj?1L0YNu3l;HB~XM3(VXbAc?MVo;Ov$(C7T zUum+%De$cBT*Jld+jLRzGP-#L-IAsz&sa{qASMG^IBl+hGQ(OWYV&(~xs;Qeg?F(e zx!y}Hm*>XkKrr<*&_s4AAOouOC#Zt$46`is%Z`I=5@a8T_9ggu>mN%#@)$^%gNM;G zmQgM0ej{&S~S^yTDRatY73l@Q+$CC<*vdcOLC7lGj3f1zoEcx z`$qFZM@yQXKwUJwicSCZ2dN2#Kd_RYxzsnVQ}rp0wi4AQY3NdBM4c|s#^$$q%*88y zqZ#0#jG0IPCRKT;g6~tG-#nRs*bJL{R9ilVwxBHj)K0Q027KX++nR@?uM|Ie^^|T! z!-|sL=Iz+>tpYfJcyBSp(cZ*sjIh3BtwYdp-DN_Pr)PWi}O&h37;jq6ju9Uyf1CKyabZ;FemTmc57KweK8c}(o6RY z%37u9hh%JcXi~r?Tp&N<66DxKd_LF@{$u7cZxdy95?-WNKCuaGlF8OgcAO=L{$%O#nATGse$c8up>fcFl1|n*di8(tB$Nh@GQVJ8NR#ON|GNYsKHO0?@!^f10|oOPu{&b;TG<@le@1Vz-zT9J9L5)%}qgffAdr! zq*+ckHvv9y`+4GtU8MWCWA~~-k{Jin_)LG?3@%%zB6uxDB~EY|bi+I$u6T5@uEdnr zw~)q@E$!kSYf;BXl{R^5*n1x|cS5#iZRQfpuLMNeM4#YbmgOKR9}WMeujci8&e{vp z6&9Ad?MPb9028ETuM;Xy5&xF+U23;0VQE$dl{qnMix!aNLDxs)!oIYwEi)FG8<$!2 zEF|KQ4oz4Mv$U~|TB_CcZC4TklgWlc@*i;B01$w;dRFzip6f_t0143`PP!T2zsfedUEmd#*&4 zDp)%*Tk+T-b*8K&@n;s}>%phCb~SwFavzr?(=fjWxxb%He7Mkm%4jG}Uf@f2KECZr z|6jpq%t&mjRj&_S;Gd`k>wvEl@BU=C<>9B0$|ywk-$MufsD0=3iiahp!@?TMws(@r z*s1$}%R0c(5xcD@aFXcC_{HpG`;&bTrIvXD#|eAJh~S?MTHJz<#W~pR)~8(u)*o#! zh5Tdy)M1P8KC^g&OFu#?woIpCKSx`IOO5kYiffzERkrj>Qw%dF1EIlmH05{#t&J`5 z;5YP9+W`FQ%NKm1Rmw!a<2ocyyC?$n`hkhMKhD&Qy8b`_5#z{ea5u>N9Bcs=^Ij)< zN!{jgf|w$pq-h^>{-KBCYx~!y?Vbz>G0B^}i-UFN%Y=4Y<~*`m-ybtWicSDPD-1B5 zj=!6<)aWM|k*LUbhm++NeHg9R&f3pJHZM1sGTmKd3wfE+zEHB$ zp45Fz6;h(OyScIi<~(e|Z1AdNTl$>m@ji)p5eR;>tPM$yI@?>fKw05!IJ(G(lhSqP zO5YFnW>gM+jV9IUeye9rrH$I+(7}yVB@q=6fc@rLES6D6s1gdPYY|UAgnt-!h zx`UFwiHBvbuc{1`1t8IhrBCTHDq)kwH|Sso`?hY&5}Q;_os>j|t9BtZlIznkjQ|=T zK&Xc=vM($~E!FnyF90bgvZ7ZF2K3seN86O z;f!eT+!@Ljvcug{Yjxx6_H4;pjfdJRcT_%yhBj3TIz15G`5oqXzXF4A(O#al7P(Wn zOs<+&#Vfc9b)_Cg4sd_Nv)xfdCrEHy|viQR_Tm`;bmC#JA2 zagsHqJ-b@MEbY6)pv6>bM3B*ONrwGOG?8`4gCr9tuqd9YrxS`Vu407|mi`Etz+jjN z`1cvqDK6@0bfTAj%Qa9OOXl}EP_Kd*JBLTg#8AkxTN0g-gU;MEzHmv=+?*95&AzM0 zRZO$raNmtA6)#+@?w4GOG-f^ji1CsImTprJ2;~zJtOyk3Xp5i$2gIA#zD4wpUA`q- zQ&VA~L@7E%1}aazey5N#31O5r;D`iIMU$MFiGRC#B69HxW3Mtl-r=d%^wHU*Jd<(O zeA069M|ZVPVKaBxFPgOnOYqqE+0sxl$$I|s$U>B5EaRoSOr&1)Q$7>FWuq zlF36x)^L(X0-T4X80p1}m6+jy(7qmRfv$sc+GhY~Pg>-b29gCUkKHn1hZ_Rv-u=4{ z>)9UqX&syS(IWEN9Kuz(bp~*Al}i1q7#J{}gRb>^!o8Bs*NxOKDL}N87mY9$_A1)a zreU=0??`xF>F3jfhwN8@K2NMV-Z=}&>XOoVdNzy{spqa8Zt-zAn1tk-uu~aAlG=O?r0QCCe-VsFh_=D($PweP9iMsPlo*!jxk?Ke}F-3%&x0dO1Zvkj>BcX2;`bioRz z3vG^Loe=ek-~jru+OnYAHj+$?Sok5M#dB|Mkw~3QKNh(6v~qyppI(2GfC=-|gl`PH zS%mTV)gcLdqF@ekz^0~fDtq2zZ_4N<+-@u;tkY*fvS310+mzNWH=?TwZD5vi(l^nV zKz4=dC~B~}bXZ;2$|KEuceIiuFIEf5P^GkcCH$kWI4~Ow=Oz9JPt*NvU23WtmAlb? zWP2IEWtkAz}$b*x4+`udd&@A=CzpW~Vh2JrJbKd~;l z$tA9TUURE0@IoO3M z;rNPEd4ur>KLpuUEwgCJ`R({ow+l6ZtR|$%Y%vekj}tF@)G|GDAE)^P(H7l_Lu_D+ zR{T&xylCl!x3!y5Jwin3lsyK(0q;L!21nJNs_-yUV17y_1FdI_OU1>-k`5RExCyuM z3MVWI^$cg)6=GdasPqlTeqlGOV_kyi4GyhRsk-=ex7D7jyK)+3Ag`ck|MX1E6(S?i zBU?Wd?G?590lv>7?M08v6LrzpunL$gW}O$-*fae7{OgNJg1+icFx}RxzTBQ6lg6ph z&$^YGi3#N$@dhariL_D;KHJ3uNkXzNvT9@h+1Skd@|LKnPIfw^EJzZ*uCty46(fYN z!^0gR)t)Lxo-m}c`1V2#j$lr|2HH$g>fpG+!~NiVXjv%Hrsmkid;f4sr zY0ZB0Vf+A|Rt#7}J>}SoQ7U@t7ok*(zJT^RP9e2W3-x_(NALoOp3D;Wnk``yJkAgY z#DjXBo_KJ#@6Ys}_C5l3kOa$Lv!ADp*@W|}eNl8m%44^6c?d5%g?|Z2nVnM8QPJ9v z!qrj(Vr{mxh2Xi-EKTbQXL;rF`;OoXv+Na4T zMj7l4>x6_X66f3K9}1howT>ir-1>Wj;ZJEcNPL$Y36ups6r%wj(h-a!Mg%uhMIK#f zW_ik|^U+if?8|Ir%rFIh+81WJ|FtRuJHz8zcLg$An8e4<3GU{)^Gn(rZBO51RMqBU zsK8*>cTFqI7)@iGn`HLx)<)8ww*SLh;-Bf}S=%mA0~zh2@Rl8CFBU_Pq@#K_H{t$72bK+`a&6g9LT=1a;Z6Z2Sofkr zq?t;pW$|NOi6v8LWpN}4P3hM7okiCAT|oY%mCKhoGdc&#NIW1C`;&L}4*d|+O6?`5srxnE!ka04$fZU@$C4XC`;OVXIJMfva&*SU^Xm=Yt6Mf*=s56%t; z$6k2MQrorQ01zM`Px7c(B9NSMVT}dF+{Pbbg9o3xKY`Qp5pe&a$cn!)OeWv$)p@Rk zY@5KL4KeTeE;nBK*eR`!81u~obm2G5m_Xh3RwWp>7LZE$+Jc&!A@X&Eg8}Na;8Ma4 z9hqV$#+}8~X?MQsJSLv#f40eGrg~^2tDU6QHwJ$(ei7VL2`GghyQxxP-W-ok-gU-* zUd_Lr{oLBT#`t`r6){Pa|1E)OXU$~lKX!52j*&tVz{P2Rd*c(~Tr8C{t1D-r43O>z z(pfmQ%V@Olz!95ZWxGc3{J5)a<`L@T0IwzshdP>N2?_$fdpT+}v)1X;WQA>Vpz`;1 z+#0f?uHmG!C!*v;{Z^)HfgM3TpauYLCo}GDuFBi+tO~UbFUH>(UMuvYD<|Y)_mMxJKn}YdPI(7m`f?U9W95)KqYkqKg7rtJdK7+Wo8RK5v{phCAu$4ww?mAo ze4k-yrfNweT%Ae^M>X)CCo<|gl{Ta)S7%E$TDjwG>i2BHlIW%fAVz*L?Xp{r`Yw9n4P?_RCKYSm8e zJ829|APfZBr-tsnXq}&Pt!WI&P&qoG7v^#w?dU={qspz5ph&Y4Xc~YosiaAJqNvO( zUp2Yy5S{~L`S2AdDOqy|-^vHWoEe@QZ0NNu(l7?pSsLo7{0ICCcA4=r1NoPy_-Jq}_8WbK>jB{D@g!c{u` zLo9Ww`wO~NXmO_)HqeBlIVS5dmVA!G`7Lm1p2f%*!*Q1+T_Z|bmzcW&(p0*U>tOIV zNX89|JykuwTX7f%G+M|YNk{#Wk@ij`68ytRLv*^k?T`(k?S*kO0XU%|Ki&il^Y~(4 zI`wCPJU{sjg>@v2rBf+LX9lotln%~#%;O@-ZIsCED~yZNZzl z$TX(L5VbmPo6z{TPECtaq1Cg}Qmhe7#Jg=xxk zl*r$~E?}(T(|oPVF2hInH30BYW$OhJfbfiAO=0=lLOtOhUScP9X(BMC*=H!dktzho z;g}%Lsg1C;u>Nv)>T-{E<2nYY>W$3*tnwy_dB(jry6wt*pH3?NH{w<7{p40>SJo}1 zU71_uYMT?+enM2^JdFHO70qoNF^2bBd;%Ud^Kht7kPiF)qs&9Bl$!24;5SZ@$%cbZ%j0)D_vSLYr@DOY++KfhjU4 zSp75son$($D?xHYBr2XugI~PYolHXVr|D6Xbz0OzGrQM3S5ieXH;t+<+GFZ3*DPm% z?V)@{7);22%mQvU{_2{Qak$F3a>O@`g2C5Pp55?iSt;t-}tjZ&tumJhf^ztZ9OPAs#d~J<1xf zdaP(4N{U>sX_~D3M3j5}ndiXX!Om7FsfF6vo!b|zW%{)L;cV(DSBaEnS)REFJ^9oe zJC?liVo*U%WD)C8RKQPQ zB$f4}K^>z9woSKV1s7GvJKxS&jIWng09CYp9tuA@mEK+>u}D%ljqf=e@A%8++tk`7 zP!WIaVZAD`eNxvkyQ|K$Hyw&66mC9f#~qMK4Q`fDGb+i(oVk3T0RDgI##K)oly+; zz}zkS3m20q%MMY3A-R*jcLSa%f`p%&G)@&KRPY$+8WA^>P!>X=J?3hiYG0rb&fL7 zPDeHD=Z$zHBFN? zpb8@awA3#+Y?AoZp|IK_CL6hnfrWHNI z0%r^h^E=(P+1LlnzD|UR)n*IIZRENX8S3YdNiPkt2XP~ z%aa#*Yyv|41Z?dt>Gdl2b;tXd^#72wNJ80V6Av5ZsQB)Z$Ul_N6aUklj9_jh1eIUK z@%(ihjL*o8q))oknjp~dX4gvVxviZK7YzX^@yrrSpm&u(WXgD`J|@TGYZiEjMcwec zfZ|7dWhUHMv&*ofB%VCxE}l7g&>jY3LJuf4%2EtyzcENu`v%=siE^GGprDvt*r;O zbEg-Wuf%RBxBT>WayNEU1}-7z_c-2{-+oMKu{!Ihz|Ye;G?j5L>E8d9wuF>vyMJpfIX4$^Tumo@jK{%6aHGSJ*5+k=;}SxpFwbr z9|T{BnG&V{YVnPlU(;L$(0On^q?VvhlyW2}WkRJ;Y5Ut~(i|eI+rVzaYprE*h$(r0 z1~Ast^zn;rN^5{Ff0UoFG;}vwm?z2|?-Gs{h@gi`@h%g;u*?I$Ps8-BA~E|atc z(hPN2rwNI1Z29?G#r_`EowDl?ez|(oqC5_3aA+}L;nF4$^BWP%^OJIt$InDmCT4(p znEPtPx;n~jBKycLw~#cC7>LZ(=6sw{x4~BO=gZ5FW_2a8s6~lh1i@mwCrkv#m7Iq@ zuB>c({-xRq2L$BY=ft9BQ`%hmU8WPC^hR;{-kcJ8s@GI-;iCnG=tq*pugpbgPuRRh zFD~o1WC!0MD+-p+ia?;T#%r7fNurKpQW=oH22pItwsm=%W9UlZl7?>dGxVgsQ1C@( zBrPKSmGn0}>Vx_dkI3p7&P0wcQXl~l7JCMGN3M0rYDe259Vx=+qfgvx&@pu4h4WwDLE?1T2Y9y`)>pc#b8n80S`fDEV>2e?kmeRU_b z3X<3iO^6zraLk-!VNw`+QIpQ^IFkJm-%RYyc*p$e)G~(+2=(^udx)GeK17l#XKvN( z4U%`8r}@;b6lX_2-|*SN?V=N91dKYP-rgN5NP<#?x@mO7JI5=wIBb`MQ7mupIBX@b zc}kzBU*6Qr3qWl81+kch3j=?9;MVIi>5lSRqtYC>KTvfUtdz?n1Z+z8zB3m)?pJGB+QL&Pi@<`vSv}_z3_%UH z?SD=*ra^?b)x7X(a}5#*mC}sT=yb|wvw)v%aN6Zz{V`ux^}e6tjtd6t>vnCxA_72G zac49cv9j3Bb)_9g#E~Vo$&ffQE`XN+Fdw;cA%~(a#^rj{RE(afVZo^HS>pB2z{LMB ztw{Bcc~nKIW9ZUl=Qi0F*ONTE#_Wlvq2EEPBOx%dNs5Qt#Y9ma#jy5EK;wN)#KG(F zX?YKY9SFwlP1|=nPS}qSfdh)M4&gpl*p)Ro+L(p20D1&|ew9pqPRIPSZ|;Hh)P8+c zNA^7_2tV#Zt6?zr#43C=6ZC5A0*GMk{K~D=ffV=_BT}Hp3^G39I?UB$aN$ASCNQYF zq@6}&QEJsne`vpVFMV4ML=494M7q7XiW}ss%VpzxGar@3;OQ1BL}6fFDzn3xl6BQ9 zr$;+A>JPH$%@OlG4x=@)j7bV6WJz`@4cf#U+$T;*09t=e0HFTGG#)}NWEx;L1JCC` zJt+7QP4o4f%=ObFr!#nGAlPl`@zT2KRsbJC5X*j*9<#xNxkOOm&#^c6n#hVKF@_o# z4qsx%g+QWKmTEc4Rp+vqzAA!sgar^>jTnk$r2n|>!QUwBsWt80nACs6!oUEaEi|XQ zH33N6ft(jBEv`9h$Am-^+aO?Q%E8G&oLAT>MF5MnaP>t9@NpmPmTEHY_zs-X=0z0N zk}H{x`^;phrzH=#in1Y)#Y^w3)s32s8`)M`T1o~+)T-?%0?e2I0HO-Kk02I62fJHm1QsOZ{+iNal*?%OT2A?FA$a;e3zu$DNPQdj`Io zrPwc@eor{0&#+3u=1a66lPR8@X|CMNUO|W&oVa6n#j_?z0XwkaS#KL$B9QU)db);vn_jJU6lpR|{ohi|uMo&4h zCVbDuC6B0k)41=Q@lkjEXHVdl?U6<)mC)EJBhepkKrh>g85JZ{!EoyY*|rQU&YG=Q zYGeg0a7<*+Mhr!iKbp66^DTMDJxPUOn}tWeMSf=3#ohlVBc)~PFizH{K;UZIqu#)F6)i6_94@b!tU=p#kMiLf$8~yKRF%GCV?D z809f#>NV`Lzg%lxLl~-ROA49<2BNXK29X;wFESGZsJWOo0AQNSV0jZcoY4+W0WY}P zK{6XFjw`@cP(Mg6mpn9Sp8IywHPB-A)y(!8Q&iu+l97v)MdF9Ztkl-UyQ_S$MykHy zK-G;G%RDB-!iUpWrwqa#^Ks?Atn#J5J1IG@9S>AYPcr+zB9)Im+P32sD11tlU)gIl zWAsfQm57j=j?L$yVzBByfcc>fYdt{PCY2Vc68x(p%-L8(6Qz_8e{IEmlrv$E#s+1* zY&oaQ^`w=#)28>zFD>XzzKs$V4^L0k%**t*v$$j;6|V>Y9TBocB`KDP)%t$~3udDr ztWH2pdt$yf`9mlx%HX zdR-tb_;D&F`nWr|2vv$&9Ql4(`$D3dolp&l03aQ<*7hc+VOVB2R6s@YHk~J1#(RMg z-s}UvmKu1j#^bmt)sf62JYEzLoU=lBJ}uIc1EKUMMX5V$jxol;j;e`BKEar*V|`c+ z6ne$+WxKDyk4IbwC{)O%3drdqZ(p=F`vEt`R_JuBt#?>8vruHVLuI(!S zeA$%>U^ZXE6afOoc**J`wz=dYr&w2@Z9ukll^5(#@gy`t`=jPv9(JtfL>~Afd&uWE z4gPGnx8!FlkVDfMS~WGdDg-ZVD$c9h>{!|>@nIyF@xu1cU7wja_Ub?8*nJ3#J#azf z$DFuPhc-Ya56SgDw$b5KZ&|oKgEf>q0X;5VGn2P(xK1(V^laP1=F^;dQ?Apd-F(q^ zMZ!h7rTTaJJPEW%_Xrx_1aURe73&MH{tfaJ(NBO&x*3B^8jH-v(#e_0!nPWu+pS~2 zF?#b^=jvn+iqNbTf|`KDBq5hfVv(MD-sUVidqO(F26kP^-Kb{N!(dkO#0Ythk) z_b%j`ITs^srZuoYnCIL4t^-lX&)KS?y|~rXn-dvZ_mHC-pff3}lM_6iz-ZHB3$ARc z+%YUZhU}(#&m|zq#~L>~9;t!prse*0aAuZ`5+Y*R1b}|)$_TZ%=q=gG!6K^ludJ*$f&=@!SY{__j%e^_eOi7C#Z?7_&fu`>_7*Rm8a>1w>VbsS|gNNuNYG%x)uZfa(FkgHw6GR-h0;FK!Ixq zOeC-?B!SDn9{_8V?aAsHVh-+C)k|VzN;d7r1ZZ9p!l3bUQ+)**XZdriX{zJYt@=)* zMve1VSa;0Nsfl-E+jX>GY6NaM(|}B0O#nZ9wwMHrJ`wj|i5jfD22NHJQ&PYjmwV{R z1(Dc91r$so=$kms@cLq1;B%_!xlcmS%7kprYw3O3P0?vggcic|Q<%0_o#ZRO7uKSG z0zuFBYfGIJYK!o1(E~59Km`a^22EXV#}rPeXuhWyqIe}LUq!fK`2)h(iH@gPyUIdS z65ZFlMw5iXo+3L&v_9n~&w?`Arxrvq!qa|aYqWiH_IZCy)T+!0*-V_DnmH=3MXLx&zgv#vA=l3>O z#ox>XIrNeGB9Z%=ajw$+6DYocAgTCrc(2gUdF22%K*+yCeKZPMeJ_jJ<7&majLqNT zfF{J0gQOVHOwPFwhZonACwzd8G|{St7jxga_+vv5E`v0=V!WIGguhyQix#?BTBezT zl?7zfkaI87wcm>@+O*Q)4?>V8{Q26@HKFj}V2VaRXyf)2?QF-LzQNI1H1Zr8d0?7^ zyP85=$7G??}uCljTtypu6HP#9@XvRyElPwqmZC{ z0P+?1445g8NBBl_i=U`oUKiPVJYVP3CBwU3A*;{X8i!E3&kjC``9h7lKhK5!4&u61 zPhFz-f0d|H)jvzx(e6YbQ*@l_WjfF&l=LhT#AII<7bhIY31DqD|X72G-V3{a& zHd$)JpBJOAuiMe$u_!!}YJg^(`zT~|hX^VjIIH8 zPhTAgt?Mg0M@xwjSzrZpA$Y0`!_6c;ZDq)Vh?xB=Nk{qw#~NQc|04jx9Zn(TVbA=I z!LyZ9E@Yj6&=~Jl>`ZkPn%AI#nkS=<&4!(?fiGz<1-EELAd7Ed&u0H4Y<)?(sVZ7e z(bRH-s}Q1ehQLZHVMqLb+5Q}W?tA!rU3ho-{3mJq2azkz#A0-!d@*rOY zXRe7I7o<(K+2;KnAPqCzqEyURZ#x5*N&1p|I<3x$>GVGpltTRXJdI?PuPXo`>H0Br z^$mp7m{{)a^MXcUv7=60g$Q+)P0@4ey6&4Dnax5I!<{1ovCcx?gBhC{k3}M?JtK#f zz-Fi=UPx*jT!+q?H%D}A;dLmZl-k~E?Q3?SJ=d0)&J8vGsAR5nIbr|Qm+|1RqBFm7 zan`qX)S7Th?!oA2XN8-ws!W zRAj3<7igjXA*O@CPnc3@&aCv@^}&R%=>f?29w{?O(AB%wf*8raJ?}RL>9pHVsra0Y z2W<4G2 zhMH{1L9rQ9uz)?qU{kREQAH|q8kz4e>_R?;Eh!ZNI?qC4bn8CDA)W(Fg~=Bo(iGjy zKGjnEUE(Jv3XT88fICab(GU?^3|H~kPQCowpTEU5a87P5q6!Z-+sZ3X1c&9oAY zV}4E6laXkM;)^D_CHVQ6S%yTedPulRH6nR9Y^41%CbxgDFmJPEZY9)%%#&8TJ*eiH zw)cPe;aIx5NY9=aqnod%`Zd?u;kRY`a-HuY9F#VncbC211Cfv*uE}qW|hbJ-1TYnmPu49#% z6q9=8zrPTTX}4RYf!M^5zn^laq=PfNtE+4^Bd^{mdF%^-CxXK=++U@@r-WO?U^LsM z_{NSKBD{o3WaUkEXKS|&yWN;?oou_~pFy0{P<&qhVbFB|McB%0g-E{x z+v1fl^F6E-*Z=miz|1Y`+#*BW8&7>&-S2riKLb~ z2L2PVo+I+3yU!Uum1%4b4c7k}PtR>VlG?p2XukLwuD5gq1k6}2#{|6o9pv)Oyc&xs zO9P+;cB$0?N0m+y4YBlJ9z8p1{ZIUh?Ku~kSzN&8bY}z=1AoG}X-+JPY7Oozm4b)V zn8S~6#t$ltA!d$LNq0= z-G^{$Z|FA?e2EjK@&l`3&&G}uKmnow+|(V~?7Ty|#&gu}jJOeepF->29@m|&G@n)8 zm$I*XnYf~`1VoT&7oKX6os*0gAg-8?4`|DQ<|DmQeM|1@oR3+=>jf0b0`7 z>%_Np` zEZc$)+!A?=pZ>|oq^}Oa8V{3+WK|_PCZs1*@4*`Eagd zxytjt;?^3+ynt{R4E1^EdE$r9psr&Yw#kHr%)jHz2IYZuP6{n|8BzyJ4w^G;9E2P= z@KpoJQXuWrqz$nq;o_Pr+;f$QY{T#1b<=m^dCA^$Z0?LihsCzN9f#j#Q8@sprI8@cEKr)<@uavo?P&n%n*|5R0$Y+%}}emImZjiIJ$Cin z3lq{fekm1jltn=Jw)S6wf^5>Mn z2_8b3=N097(Nzt~^_YXN*8e*vSMXcRHphc>(r&7Ngm%=FcC5L0!s$R=Ic-&`y^I}f z0_m1r&kx{X@&0ej=FUk4N0fEV&2J;C)`s| z3kk`*y|`SBz$=@DGblta9{Dxu6K$q$MwcQCGuBu+3--z9F-e8b2U*582DhYnheP3mknB==~4}(33(Y=Kzub;X6xIJE5*Gx0eo8b%- z+sRw4QyWQ6c#rWP{kXvQV{gTWSKC(u@RAW?Xp0}>Z_;NE2U{SyWlz6go>#a9!xaIJ zwm=m%9$T%)YTmRv$Pu+%RQq%ToHT>O%V)jzYbu;LUR2q_t6KjtV{Owjex=N2jGFu7 zUW^uX75Jf%I*uF7K;EwF$7Aa#i?(Y6H%eW2LF5`uQl`Bkru4?g5M6CfD-+?b4Y|QT z5HYdvW}WX2cKv`eqEv)3E_rNzkBcup%cEJ3f`s*_0wi-7y^qZ!&LYj1nw1NJziX=u z!O2JLw9+mAu%m2Wj63}BQ?dmLB&6XU@LI9 zpb3YFtG2l#{PJNJTB)je1y{_D-0a7tX<6_W0zYV}MT(Wad*eK?#PCK*dH9nYN8Lw5 zrt!&KTr^0I-irij2n(jsdZcrimu(NZ@l9}huJTrRg*fhI9!;a)-!5QjN_Mhb>w__2 zGBW16!o_BUMX;U{v7z8ri~a+NqwhN%X)rZ#c>g)r%EARILi&oFuZmwRgvW1AnCh@X zR97lm>!q2a-cN+F00DUwbW*3V(mh^Rn15Ze*rc4f5Efb=812WTvwr@ygyfwc1ZZ*l zHFiE z$ajJKwI0o%X0~;pJ$zYqs$M#2HLEcPN(Umz^kl=Z)Q2+BmW!0!j%hL@p-8u++}|;R zQ2+tI&*OgaXxgP(4owb0x}ei4=<6Obl)VOyB!BRV^gu#@R%a6{S;;YqkU0g)#otb zmI=C^d%T2kZ9Byu1N2fqiufClILbk|0K}?=Bvq)E&{dYX6^octtyA*U>?cdS0%fk` z+OIJ65AjX7>HN2K>|E21w(%tDzoga}8I@+ybp1||9y$@=imT?MTurzgtQ%;cl0G&U zQ-<5ZkE7dF9?I1Hp!HfA6=gJSslA-=mshY#L`6BIs;=d$G*MZamJpYArY}+cSn11} z69YF9`cczRP?Sg%=RxX0=%R-r7d~j7IF4yl@k)Pn`%}} z7~St^5ALxW`{s@!qN&Ap;xe)UXn>}vP(Pj&ZSEENe$%;65A-#2SIp9OqM9J2M!RIm zLO=|;(4u+uUq_&SPd9?B+Yb&~`zzaMI!x0|5X&(yq4P(Fqk%eRZo>nfjBagRqDXKS zacbWp94LH?Z@Dnb3eBDHd@xrXoj|RNwo#}S4w^CqU$DANn_hHq0|ATvtLh(wgU)DY z*bclp-WN&)`T|P@vMggwKFla^a$CTg{5q1|Ii#JWjdN}&S#*2*Ht5IgUZ5?~;u~OB zU-c>VEE4|Q0LQS{cWv}Al_?RciDTQhCwmp%w@X;H(Jxd@sCm@IgTFGE!M}i|3}9we zuA)rpoA&a;8Km?x&>Mvi21PBa%r|WGNLMFVB|l85&e$Hnq|wCU89iZrH6mSnlN*=} zCWJf6&_Y}lY5Le@IS?Fhvp*RCQ_E+j+TO@&7eEn}_3S z&q&AZpiJ?kk|zwecyH;~^{f{A=dTtQ4H;0k5qzsije=G%1u@77SY)igV2qPXAdUA= zx-e2fCqn4A@v4@;;9a^|GYjD6nWL;cn|_Y?X;%(2)UUuThJ|T22r<>qbPiHTPZF>O0ej>{jlDm?SbX#{r{4mbM%o| zj@3+YqmibYXPdIMoxt!Td=gNLko4KYD7edf(-1ZZT{n0m8+M8{lc{Vme6IH?OzZe& zzhQ2g>8Ni8e2aE13@nI|H{US2wCd?G_^=+IZ({njVJ?9b3#~6?Dw-IL;hzaq4ez)> z*37Wph#-Rg9z;vi+JYy|*s>Bd_CgV0qllRkzsP?FO2pA529Lt>V@w?@Ah%gIsBJyc6y=PLB;jZ$n%J$>I zKeb3Z&n(W5vZF ztzw+A{Sdg3?MJ$?qihrKcFLJPiwbg3Q9Va{Rxn^(+a6Uk#B7Y{#O}cWCPQ9f=MT7NYj%nW6XrCM}WaM5UhTBY% zlnaLx)4;?^259RZX*5-#^vE_r2ZKYECZIIv-|eti@d5^83}y&Cwm5Xq5% zQ%gOf50tX14WPGm5KeVrIdib5&K1N#@mc+|D#lBPkt}o&$RY}c$v0L%+kVwz;@~am z|C5yCVcpyat9FUDp{~rYIYaNVWWv+)wZW5CbMxN4G%JKzxdM~^l)R|E*_9av7_D(K zW#_nXt@IGT;h{tT)7unaq}~E?YwAT0WD{9#syrRrMrCRVFu)u!2QRf-;BriV1`&-b z)R#KhbTq4LKr|giqfJzVe|FxHzwJKaJB25=WHbPryd*d}zq2GZvU`~h-4Y^UvOip~ z+j_n}f$)OrVsj0(?&Qdv(oJ-pPWIit`b)9rgPM}g(K0YbQS1SK-pu>SB!~Y^5bLdi z6vaX;pRT>*JH5WUq4(Z-J~H-C<={ZNn_eV^zoJcq=PAiLA4S4%R0z3-Hm=$Q)!XBS z-!;;im)9i+Xi65Q012$)P6_fkD^?G=>Ro;-RwlZi$JF>E7kvrMU8v*o&@_2{+`r!& zW(1%m{|Qf1wgdLeb$!jYTOO-VPwSqOgoPLSK1w^pHB#WZLL6kRbXCOoQe?MAYW3WEbcVEEz_# zfHbVCRgk3YhUd`Yg?s681{i?|(Fn2f2b?dRn06@-0tF z<_=J}v1?>yxS$8N0+}Ct0UqHhH;(w!bj-tBiT-A_vvRpQJy7MUnEt@ z;iBPfnP$YEsM!9bjKJI-Kdb|!&1&}6kX|!)2x&YU{y7r3Eew#pH(2;A%}P3x`I|YW z+3!JX+nkMo5gh~t#Ui@)KQe8mlnU$G^;lH;x=AngSc4N586PP7GjX5lZ;BYGoVgc8 zo4*_DFduresfZh1&m}ss-t8?=u0G)QOGUWbtYmXZq=}E+xJ*L@8@tt;`r@P{)R5Fb z1tp|yM5<|ySJinv)pb5Q(|;Qwac@?;;fv3$Zt(QfQrAr7-X4T!ZNVKIYIgv?OGRBq zbftuYprl-@)_w3xu(>ZjGiU{K=a$JiF|d0bQPg&N(Yn8x@N*LjdpX8JMEWUnIl+Ji zc4u?K$Tnmohb=F$0SGLjw@EY7t?^t9c;Dhi0&Rx?pGurSlB7HMjf5OPctSwB5e&bN zUdKVEKJpYI-|4!_2Wm$f@)& z;2RQ=OA#e~_S<4GB2(TJh%atZt7&{19`7Y&9nmcGF-AQvR67>mdGt?-lGnbnW|4@e z;elC09`xx-B1#rgJvOvXh3%2{8nYb(;Tk<3PAZ9*D`-1dj06Jd+?s%Xy?+>WcTc$G zT9Fs15epZsgu^Zi7&q|50O+DkSIwGwV1f-=E-+>7(6T_vvllWya$^!2oJX(AVK&^f zo|}4N`pty3`eH+2Q_?`7jAsFn5IW@K(|3%YfP9fNV<7bDFO#zDjoGsNn7xrAx-CB|21vmFXxtBDuOh_*l#$qVu(pt45ytTDw4m)&dpESkCW zR(Lx?rm?7c*1S9o;pz8nr*}oPwA?AOU;82e$PjSVEy>!wQnL#d0Hxyw0yh!HDrBb^ z2C8zOt|XmbYdd}QHO)KR+S2(FM&lBCR-_;kf9%K<_yYX`3c!eXkCR^LYKa~S(L3A= z%(-utJ5cqVhQmg*VBYobyI|5u-WHwYjnNHX?<1_kWS9Y!8-f96CGfQbMy`9AgNRv# zfQf;kx6KRuJvJ-ZjkxlgGu%nrtX;2=7dOpBe(HOp`uBOUSvAsVTDQSe42t@8W4)fZ&&xd|gUJq*M)8quH5y9AI; zH0c=S<+Vt%l>28VD!&mD0$1-YaR~FJlE6!taCHzDr61rOW%F2(*J-VkiAo?H+#LYi zJL(!kk?|>itlpK6X4*%{&x0w(Qy9!`)i6^E5i5YJZ|Z#uzndcH8|*O{WY9(rm~klJ z+r3s)96rL>e(t^xRe}% z?2pW4zuF4vq&+kO6ds=d`Z8ui15Tui_HiUBD1)~Z%_*!OZ$BZc4Ut`D;GH;O*b$oX z$hacH)6}1%zhX+-%Q<<;E5vZjF!$}+7J{0w(pd6PY9S9s!RRsdqsU1z+@sE1wam-O zZB$f(2?{w$Xl|7%ThP#dpLY^5Xt;ic1x}59{=<3r5NC^u_CqhYm-eaQ=Z#4&K*FlH z{6Av8YZriJvcKH5R{3xL2%(nG<5rwCmbG3{i~0~I^FHP!tw|wd;LNR@x-w1x85q2i z#hgYnTaQeP4uB2jvJBg~F=5%TV%Incc2?c*0hBvrv2OK8_($Pf7UP=43) z*5(vTOwPM4+hc*gIfnu}I+3 z&{Rls?N1fxltAi9uF=wT@;$L&coyRG$|#2_E5IiR1^zO_-<&K~I*-XszrxcvKe*1r zmU;%WZj@i1VAuDe?v=>Pg|i^@{iz2RLH!M0H{96T9o{fRqPI$&XoYEnhOMA?3EN$J zPfOgwg=cmRMS1458sAn%?UtsO#>4@4xwO7x@PaCPeUQ{R7Xqw(5r#oX?$)1LpdN@n zK-tTOig%aJ|00xWqfebay2M2B1$I;2=xXs&PMvHgXAOYCi}gT5D~dRc01#lQ4kyfc z2B+%Kx(0x-&&^kj!Q6Ps6yrLaN7d&2Gia`H7_i(ISh`sQhwMydq+*a}70qaxzV`t# z%A}OpLP7>P=i$agZnYTR_pqJaNU2PLQ35FZR2xy6)1q9D(4)=6cR{tSceiB_6`9^F z_-=R|KPd^0s?g8BwCZ;YV>0a{R?JY+;CBPi)RUQ6uO(pkQ5?_h^|#=$h|l3+=ki=` z(})d!$z0Roso=Nw(}C=MCfFtWB zujU-p4%PCd%sS^Ktt>%H3C!dvSID&?Tne9HrPd;zNG016RJkvTxEkDl#)VyWdT=qM zJ5wc0Wx{7FJnG(|tV}I8_40&0v1Oz1yyf7HafN8{kvJl(sGonp=52fKa@%KW zN#k*!Zf=6+dq2-D4N(g=Ln`u{$BZ~h%I;&sK68@BHQ0FJ2tDbLhS$k4PaEy_{l&VO zTEZ`a5u!917j3Irmwnp@SXwTf$SbNrW+rOnvBch$7Mq9K+jLuIZ&Bzo4_zu=NZ!P=ktiLsLOtwW?l1%_DxiM}k*?okv%#xC;$!V`=^%43 zle&QO!1=8W6Xk+s?fGmGd&YU8YWk{C$)RB!4_R)>O_y%oFB z@&KO&gB~<&x&LY_cc`;UXz3Uxj7mLfBVIjI2A&E(wSz5aLE%P1$xs|}X|C2*8Q1H5?JoT)^Ak?EXt z2@)Wg28>5|)Xx{=br->Ow~CO-pyWWg{Ynzo-h)^2iQrk}fx1OdGeWQ^4-hRo6*HQ7 z{wRCgSWm!rdnG9bDi_i2>b>W*b1|r=9-M1*g4sF=oL3U8T7y)dNp6TQ&PE}hk(-4w z0Se-3It#G6CUd4z;?}Sv%wt~&G2@dlX&Ma8%QA| zt%l6E1|@ij;gpUb{ihhd3n8`0`29*!O99nj#^i4Xj0*8Kw@S|46FME^yCpYE(m3OJ zb>=!edt}(2v9|3aLA)NSb1A-VzXYemm#xnz!RS!>7BnsJ2ZRM`*eF=39GVUK+IP&b z17?%fs@v>KW{21S*|~XWWho-VmxJlfh19fxZTN$);xL=Ng_168G^GEbN%onP&$R%* z0<5VU%{%fDmil&BbM!=QI~0+f$gdrqHC$P@$MNXSDn*E`^#e-ghk%I+(MzH)Qv7zh~K!X7fk>*+d=-m z0zk_|wJy~ek`z+eqvAt=!ho9@rJN7J@Y({Qm*L64Ako+=P${yi0?_Q{w+2|C{?h=+ z?_Xo{H!bH~k#{(ap_c@A#6aX_e4OqthQms)6$9GR|4LRcj1&~%t(fOKtIBGwB!&4I z+&b35!v`X1mQTt0nw+<`k^lmVq@D(3{L)ieI1IDEk8x2!cP*|a8%XqQvCas2HQsal zBl@)`du>}W<2Ww4(`yxh4+0J=&<>3P>7F?fR5UTmm9A$6=LTr+7 zl5Zm~q~tNxLD(52bQHzb{qyzhe#>Z}Hkz8dT7XqPkJN7m!`;Gkp+15fh;K$_gZ?8t zZn@bO57JawGmNliixtT|s<t)Y2?Ly1i zUqdMWvB6Uupvqk|yqpH zw{7;`VPIm$86$Qqga+=A?eX*>XKJkxN3|D`cJ=rBkft}@&cOzj+ldc1a%1H|g^4NU zn(u0>0r?ln|S(+ud z)g{v*kk=a&f9!}0eJE(Tzcc0pnm5gLh3EK-qt|SPICs!(KK<*R`FzJs$C+Sx1E9G< zz0-kb6o#2QKtWTXn>dyMHT+j4?b?tys7kR*ZzjWYZY1OB5qU)8_OpHrvcIh+l^S&wb( zD)j!ZZnL?*scx`Dzrrq31MBKkKi~vhp0BHwTX zZm&<6GK@IC1P%`sqNh;?#Vvx|66D>{`pxtUmX)l;dU>X zy#IsBOIev@^}67>Sct0M+iZHfYQk^Wm|}Y9sc$irmv&dtR9=9A!d`-I$aA_qv?&}f zOXb39CyOo?0#tnsmk;NVmR{`-?h6{#1h^J7QYa6k1a}eFn%2`_qxj$Tbr`C+q~bQn z!W;@elyvsJoOl=z#$C@bdYI#6(~!Ud4o>Br-B>q@^Y({?td_Sh_V<#EVMD*%;O*B8CNFj=`CV_ zuXCsPmA(BC6ZGDL05rOS>OFU|YW%s@*=H6=wo_c#@4^3YuK*sZF^12&-pED zEHJ`K5kgyl?$;wES-=+ti;-le^TDIze@|bcNjR8|%Q`4L{y@)zXiiv=eg%8s3)$WP z0d*6;j(3!46q9qh<5B?oUAkCO(Ve`Q^BIkHwdU%lq z7gxWNdYLx8BWqvH-Yz;v_X;g<00T#&#{>@CuvR<38peD&RDU-E`s;uSpTc01U#h~o zZd!`Abnhx}8)?Vk(;?wk16Pqz!_AQ}cwltyUroI^vmWNd1)?FF6v!}xAqvoB#^N4ya zE~)uuX-W{bWI}!2Avi~YSy`{VUVidrENEi`1`f8{fbIabd4j=h4R3kPGXw#Ig>pwt zLCt^U_!-c#>e-K>F>sz`6dO5_i90JwAh(iqgUH%|YOZ#$*wd1Er#}%*ypT=e;wd^x z*ZwHCeHP=J+%rwdk}d7_r^PIZ@z9xkwI388a2Owzu}g9_RbRCf~9asJO|aHnJ6hlzLNTG zd+QZyy>k4k{w9cFvnyMilpQ=V(bF%#ou#Y}Z8o}F`)Xz;KVul8j&7BX4owZyke=!0>9u$y3BGf| z+2+OEjf!HCxunW!GrK}dTc?7CVjX*IPI^wsfl(O>A;=7a3s2r?DevbEvnnI{kaP*( zCaTZP5~m4^k|H6D)n`-&3RF^Z*RXoWOU-H=%;aG`0oE-ec^K)f z8lhQdcdOnjc4DcXvz#V>;V-fdUE_|?K`{7$#fAbCcNw*?_*ksp#=>Ex z>@ZdfcGY^+MiJUbYibn`9e*-2_iQze5Brj>;=5qW{;iB|)JABtf1J{yd;6VH2>{r{ zL6M@a8w3cX-SZ{oyTjSTqsRC6DSNIu6ePEDi9@<#U>^br$_s|t{eL!n(PVZ9>A!`SX@%l42CAZd%> z1k#Xu+dE?a9_i`H?^DiqF(wN0ZG_;s*Rn1S7Zjc6T-nXtrh;GjmSe(pz-U`Mn#T`Q z7_xb$j8yORfwMpsU#y_l0e6zr_AuWGH@VE?mI-%Ek~^O8Qc=&e7RKY2OlIB`wCC=t zl&8qN2%U^se&MT1w5IU7QksJh1^k1ZeavXLS|pV4&$A_6b^&^3*1LPI%0kWi4m?J# zk9{%}cuq*j+1lt9f78IH@47+5?n46T1CRoE2_V~ClpzI%{AP`FR%lQE3gD5pdplUi z3-8VzDdu;`YRS%L+J~nPL*+)mIyjW^9>c)nGWEfFnacS zydB4C7(MDGn=JTh&0F(g=6i@i2+PAiQ0wc|gBhv?p$Q8-9H7USg>{m*!`dLemNM(_ z3{>vL%HRB<4cMPTQ5YoOz)#t;T)<#mPQSe1)XoIyldL32wTQCgYsh{JP;;E4CO3YH)Yb~EECVubja=32qO(KY2L=iq6UPP)>K|;g z-|Ud9P!tqwKgieQ;GPOCu_ZFIwzOTTA`;aJACn{OhZa+82z`2GA>TuKi4`@M8@zr* z?H^RO&j&Ll&czrykoTM8P16Pu!Y*z6YDy^vQPH0Mbr|e=zgqf(P@i6!(@q>N=$W?s zVS^-P#A&|!FjP;6(e6PqeGt9o*a;*@&%RCgNSad zl3|TC8!^Z8LCMz_2V4-6TtHC-zW)4>K&9JN%*AEVRobc>7_7CG099579EXVEJ0}x^ z`YK_C2d|URkQbH!0{a9K?&JWGl6`#gJGb|1m&k0 zn1U#uXD%x%#iN4(Ivg;LY1pBK2{^|K?9{7+1mcE#fXc&NDsV7$q z0X3Pu=OjXp>FPxSqb1eK<`O&jFrR31N>~VYTg3ys{IBpOUEEyqIq@Zu$L8fCW|Jco zG*eE!cKOPrmT$vSk5}Wg+sewuC1?p5f$Kz9g*m(m2HjYb;Q&Y23OsAgw}e7*x-Fv; zk0t)qbndqn`Z+9TAOI?7@vwLDPFQ>Pvg!#GrNZUzgXjpu!I$B?BBf3y1qwFtUU@it z{xO6Yl)=radM0>nv@bD5=qX2d1^J11mQbdiUN(O}VOS{z6F(t*R=0Oq!_7htGPL&u zm-HiU=8p{?wiQd>gm=}5QeJ>YTHjeZO^-y4y$QZC;m|ObLsZ5YV5z;sO`2*_dPQ5c} z?ISZlkUh2Hz!PnQ!xg7YY7~v;Rg?YG8JE}warS=p@WS}8Ac3NELEuJHLHgaa+8qq! zj{ku2lpL)QH?Zr%jWmxarxz+?082<@tow-gcC~j)kui35T(R?L)=wq0|n-X z=0qxr!j`Nk1I{smOrw>!)eYMf|KGYpmA_;cevdB@Ri8T$RKqm~Be<;|Jw5SgWxFG- z`g?y=n3$iYritxCrzTXI8h}L5U4EzfxOD`8m~b6Z;x~2JH1w`J^DeW^xumdTg|T#Q zCONasioJ2jV3A0S;C`ABBxHN7YL;^LKqQYPc8^NDvxF*W0g*bC@NfkZL3MK@$77E> zt;3WB1ZYfc3nUmvh~>oOYo6yqSEW{)F$=&hOnpGTm)8gFnF~{#IK*0eEcW$s6;0V6o4rW- z7%Ud`yvyX^COwNM#ld;WQoy89sGuF&C==H1W2|*vw})x6F279~D$xYUowhTIpo|N! z=}~bv${skr+p^k6a?0(CfN$M<2Tc4n;xnmd^&i11x=Cl~>|l&$Og%Iw+>BDQ&mtE81q~?b3m8MIN2D zBrq3VI1@D%dBM2-+=+fgFdDsE`%^Jlo_x1ueqZ^BytM*P7HL|Y7(d$GF<(5|QH7qq zaEUyy&#z4_o^Jp6F2a$1=B{YgSFX-@!a}IPKE9+72B@(-Uu)z8e4*6CGF@OKSZqYO z-q;E?0`; z&<5@~Dc7VJfiGBB}i&1uzSiP<4q;jUyX$Y~rpa|h? z*~SKZa}4q)#))h)U6eJWk{-E^!jj^F>@^2U;I-~on%76%8;e`TspC`NenC=2S|VPl zj-@`KIn`5U7EI&V6@-oZN(-wO6U5wkEdYRgB%#I*f-3CrbW#eGLOu$)R6uSAhvY%G zY;)^RU(JbgR<4Ev&VUALu<`(JZjfkfMS5s;%OIGl9P*my;9LH~wVZxXkeKP+B36KN zb3`2At$F=e;qwMl70tUi8byLbCtDcBqdnz*Vza~U1jO@A3`Y1hoLui7vl7-KyR$Is~z!ec4z(ePqpOApI}3fzH~LZH&uSb4pP z98ZZLov}zeU}IVQDh(*(Lka89ps4Dgl!|MaJi~`}S$iDbCnqcCAQ!xZilX&Y_u%{q zDNz&C`4E37Ow*E0A!GL_V815^(u!=4o6moW_%T6}YPg(w4LFHolWVr#h>lKfQnVOF zj9g+5{AYYeS`Ijgv&4-wVq!?jVnf4QmwgAEIw&{5Kut}<>&XvViuqr%5f~k%%_ zdNgd$sZzOR6JE?kmL^&LE7Qa8L;h@~4O66-tyjdvPi7dhbdVUH z$9=5L=i_836^T_K0ef`(31?=gu;!CG+IWO-l~9;=YdvGxx_c0Qco1g9B*iww79gl- z``D!7Ck6^r%&9D$Si0`aTRUbYayWbm-a58=j4|`Q{fG05jVQ7=+pG?ltuPy@vyJK<50n+9w2COb9NO z1Qgo+Iqb=|4Z;-Bi_rf$+WJpp{@m)%%Pt-`K-JFF5)`9JqTG`&VaB4s8@~L;>s={N z$p&&L$5k~v{FpG#ikH(ZxP01AsKfJq008qF(oeF8Lno=*0U?@q~-*PuQEHM7^G3ulb&AjW9bnw9{R z3d#~w&ITp-B*N{V$1h(XkUag?@*eIYv-4D~zk$6iW|HMF7aujvbdELbVf2It{NP+o zx2}?WZ;!6iaSI;?!EYK6dhqW&SbO^txP{Xx#a!NciXE9Viy_ZO;+p!$M7Hf~maazJlX!T|vgcBi#c_k%|FU)4b{-iR(MnSN`4npKyL z35Qt_#4v;>Efbof4aUHY9MzEJ3ZQ9%J>UwR1STY2;u_Wzik*HnIV*|Bm?19=xwwTy zKnm1J?eu&WGWHvk&n`dp`HJ3T*(vuT5~EbeOrqzSt%@1p66|Qwa@L&x(O(J9;1lgv zWND;rR&Ng=4vah1FjbelQQ4ONc|J$So?ehbg6YFIVg;RWI{}}*- zwhjN)4J4t_%^k!60U4fv46fD9l`kVgSnw!IN>a*B5;!GLJpTb~3X}D?%F-NS zr4;+?+wQuI{5dF04l5;HAIuFciiv58GsW4#ME$tQR_uD_*Rb?}xuZ}Q=RghRw`%0J ztEXYVwtY4VBEX)35u2slGdUCwkN?20v|Gl_f6l$>*~ne*-r_0W&56R*`eaN4w?T{P zN$*Zcxt^$_Q#DW&gbnLhTzl_N@0EWEvGglRP}VOjW?y6r9YJTQ9BvQ*JDkBvXM$(A zNiB$LZ&G;08!1V&qbg|F38t@q*F;0D*IkDKjZDmSlP^`!mdQ$S{8@Kw`4{qJ+ohSY z=L(BUiGCIs1#MMhguF`V@bjw11r|bRNDKH@_y9=&sr;YPcxBU-eXsYg{6`k=7-{H4 zsWq|@9Dx?h+bBRJUi{~qIC81Na4P|BWGYPvvVKx`;V2&J$r5-9kWNwN*DWcM!sN{r zw4%S$R&$*-!FYs^-YY&ZfhtK8nuW~wC~xcUjumJ&t2y!HV%`{|ClK2 zE&RuGm#pQ%&oAI$(@-g$Xoz}|{v=~AE_>~3BxwSkST?Zb zH!kze8WtRCqU zH*fF5=8*P~aO&KYMriM_RdR$83rRAS$b~`Ya!h&O={Fx(6|>p0)wO)d1G|UZOQ3-I zfa&$6?P)KTDs9+IB2t*~SHFX&s&%#K)qM0)=s+1x&Ho6^^Z}d``ES!> znQBJCibae@_x`&(JuwCv(Wv8p7e_P(C@pPx{f^|+l=pp71`wg&1$sA)#5&2Zx?dR? zia39d?T4~T-ySp!*3a6(5^)O^6ny=lu^}&I2BUWmpl>%r0+Ai&xd07^-?OU`3?UA0 z{x6)xY|sdw0u5b24qMg<6TcsLe(n18G%-Y86O`O*`jYOM#(*AD3Lw=}@KLDrXkr-3iR&N}Z#pCrux%036)BM5I}=#?Q_<-JbyEU*+<6PPAmfRtQIT!V9Hxmls>= z<@^8$+1ffHBm5lKdL_m{Lui=r52}Do6y) zOz;A}$vp?1-1~yE&@f~i&jR>~zA`4q(f?yb)@}mw5t3=?7;4r5#CUI@HDmB(ciE&S z#?L{TY$%cjv!5?+?Gr_2CsngzBe-Cd^ARu1~H#soEk;ofklyV!j-4F$FpMIBi zrNR@ci2`JT!nO*KH2S!By6Nv7skkFGEO+cmoH026&y(yMURDmcUn5$!!Gp{!rx&QZ zEl5EP%+ECGo&irO%JFY>?nrdTLwa4TJI)*r3($Ld8e;Ra76?&I02pTdoDBJM-ISws zOmWb`ac6rN8J2y$kaRo@QdtLU(z7~`lyJzu%`Jr z1_Ge||zn=XDDp3XKtP!INuL zS&{XcL)w&4aWRs?8DdtD6!p6Mm-GHH7xp)wyd(S38^`>t)@Mz$_SI$@1p^Q-m#`n! z*SR_5!mao}dByPsv&PE^S}-W!#;qWXmq1gvJsb#+4T`g!VgBWq+nH0t;J_7X1q7IT z#F{D?18D972J(Oi87vXXQ4iJpae5g;1~CC`%`^byWU|hc(+ofeXy`6S zVCM`EyCRWBPUiOE#xw+kis zDD-YVZn7Z@a7QKJr)u+A@WI;UY3pC|AT*$qb+{RTUsajwC;ON?JiWFZ1W_XgBw-*6 zQB(Btvzsip(aYX#2&|*vO6shss7k<-QgjlQI1|ZPy!pF|txMu}ma%*wzlJBGtOBkl zO@860@JrlYN}@a$193i3`x@7UNsYs;s|U9#bZk$2W(##*uzzco;Pk!3t7QD-gqTR_ z3|In2DFJWM4@f~4(HGWu?ajH@Yk3tY!ar~$q1G$@3U+@9HKI|YFM6E3df+m&HHGd+ z8b9qUm>sBo(sb^;i{ESTO_yDv4x&QrgI+q1sYOkuGHr%_<7A5IaO33Ytm>eQTPmy- zPFeVHV7Q%lLBKG|$TNmonNDoXQFfAhVRk%R-#q8R8NhiUU##nx*QSHC7<3Bh^+&z-OqkNECIeE?)O-T4r*dqR3%YUnsQ_-h`{j5*hnc zEgSQu6F4xjJ%|9%LQ5^4>WK8L7I7T-HE1GV>mf22$pEqrs4kE6JtCsaR42#sh;yG4 z^|d}$ySHC#M)*yLBbk+188_yL%SbB@P7vK12(OK``494G5|KaDxKUy%)YDVmv3 zrgxdh%XMwZKSh~M(~|46DRc676pISQLFxZUh4jbH6i^Z)C3wUJpq_(8o<$jJuv}+? z>MF>3>OFoOd(9Asnq=j4w~8TwL7^`j!DDu-Sev{a^FyZ8jirgi0hx^(nh4>}tVBg< zPsd+_lR8V0Sn6NRxgVFs_G|_rDf$K*2vF#p5o%PK`T8440gu3Sp07!W0VP$`Dw^{M z6F%cV44vS9c3Of#f>siIssDS?n=b$ZH)DPRa5;`E2~TmHgiyOVECo}PZ>Us2E}@hF zas9(U0KvF}E;a8YbkD&PHXD@2uCTqyo2xXf(OsAiu%|Fqv5Nj7c_P-c7T?gdJG;;# z6t)VdpD8Cq6n(~a!SSU*5DFC`|3lrL0t}|5;_4sJua4cJO+go_QB{@Fv9tPZ>Ifwa z4JIEy9nNFS=^{MM?yUxZK}bl_hU!bw67vsM^=HKfTiNzG@&JSK+B4;zQnHoqf1A-W z^xJKbA{z&@2zXMB+gMUY;=f7zvqqf^BVkWOJ)DSXWc*L1M(YDTl&a~p_@!&MkTfMI z-qct#hdz8b`E&TXai86n?c16Eeiq=m3SziPOZRe3zE8(~xDcX2&^pDGr|7&`zAAO=fq1yXeeO>h7Y zCkI7j1XzCqU4#ldTmULv03AXB88-!0djK9s17eE;BT4`xQ37X?0&JQAF<${5J^?3H z3N=^(bfX3=N>={Y4LDp){?u@Aa6$gm!Ts+CLT7pY;RiKKg8t)V{@h#s+iw2fcKq9o z{NNh<#Ebssf&1KI`_(c2(4znA20&MB{MfSn>zn@Qx&H1t`^*(YTnJ8htN!c=F;}|y z(~@hY`E*-4v2*_@x*P{ z%O!P%4QhQVxviw;+cvAE6FE{jxU)^VwjZXUKaiabdVC+~x(r)t9btEuxV}oqzDVrI z9a?b>>a!GrkrCXlGFWVVnWss3i(%x;ddJFuo@=73t;Wh4oSEY7?Ih2!HRZs{;^A0!hL6X{AF-ubnxa>% zu1MR&5UQ5d>E{j3r$nZ&mDbU%+1DzZsbR*&CQD%+kB`6A*J-b{IL5IpsJ3Igxes!Z z6Mmo~Q)^h;*aL{T1*+W!nav2m-P$U7D+@wRCwC#oy)G%Fc5}&Kq(899hixN zxtI`HJO)z4ld$9kShC~^Sn(LJLP#Lqiy2=s83yB)%OR=qC$Yzqv`rV^`EAFP!?O$2 z;WBl#5z(z4E=*U?3VL<_{==-@d&3_ANwD z>sq||qo<#~e0cZzL&)X3q4Lek4~I*o;u3Tvx0FD|;4klg_y)XX)??!`pf7%ju3gHp zzyoR;d&qKl^gZ;Wv7ptX>;u%3Cr>7C#(PhwVS5$;%r!+?;2{?DX+hyb2MjG&MM@R)lL_^WT=YQ7oZllEUz_Bl7AFMo=MKAN2fgx|#OrSk%X1}b(zZMCWzga1k);TFV;zbd zm8b_hS{sPQcnP|NV6?0oaDzZkUH03gL^^>UFGzIL{&TsA`WOQO-38G5w%tKruNV|X zYNnB?EX2%h(l1$8SMqz+V#Zt!+H-)~Cf`-iubfAyy>$Wf!I6#@Ab=mhMABzFkB93a zl70hN2>O$Ij}zx+K{^|HiEkGRx5&A00{|^K`VRUEC8yu!R;)YYyc#*Q7)3 z%z;DZp0k_n#Qy?xw*cMIth+yAE^8*8t=RQc(F{Fg**F!K{M;kz0(3!l1e?L9M5f-g zvk-SGchFa*>~S+2?3n4$18;r)`lJnA+zqGHt2R0-R|MS-NK`Xs!~)8`an$CFX#$%W z^czl<8%fcfc9~zX?lzzKR3?}cisCw+4ew? z8+LXUI74q2<_1`92YrPStg01jB|{f4t5aRzynMDUY3PFY!ry^y&dL=*N35ZBMDUBP;jbvtSNH-~Guv6=yQ&G#Qs2eZU&*Wc6e8z8N7d#hX8Gcl6aMo1Jv@`tbWK z#`UK5;&zG-DpvG!D8}j=Sxr+0Jpw%pj+F5N>FJIV8gu}T`X0yViFBHb6SMAgfo%fX z=BKiB$%tk4a>t;D@a_)!|5F~bDo1Cx)`rflvpLrgmrpO-73rt@+R#Ou_bvH)x$EAw z;qXW4;l5EbatykR;wU>lM^B#4oIfDWf#fL|M=X`>NT3fegU;GPUyBI!zyVi(KBKOg zEi~k&nlf{utLqWEjOX@-I`Ku&lUnB^QC&ux!0u%49U$W^MFjV~O)Fq;qnLA^bDtU{ z+UNI>ynE}Ylg=}OKY-rtp#M*a;caPzuGhj9b&65Tkav+1m!8v`m(%c^lre1^`-zGQ z=qMsPg5FUTJiTRw$#@1`DUeQ!VC7PG0zF?bdNVjk6pLqR;0CkoLhO#dKFJQQTu-WP z{&1#`%LGSy`&m_Hy`qYx^OxPsgSI;8tutfn#Rt%nrwh_4fvKN^4lw_-7UtYqCLLxQ zZ?))r+wP#RQp&+KC8ywBm}S!`EQ0h!vB+1%;1^C`CB8aH&*5g4(9}UE(yeP_*!|=w zTST8SkYSaelc$rZlL&K^_mgSV$X+5Y8!ja}I(E?ir?QKMUDB>C{lKj{=NfR4xuLK4 zg?SgMGdl4FOXu&Ja$8sxTLzuw07>uKKsuJ%(=jndWI4ymq(6zm$PQxGmTvA=cZ)3@tO4?L zVm!^zO)g0G-ho}&W$>x$QI=SbnA`YoS)I2*qg8UPkaL3dZU_CpY6*azZ{u8LWoy}$ zOU91J$Bs5xjLLFr( zGQrG^FUVLYITy`t2mODDAhWHmd9jZ$poBB){bB0_<^jX$! z=qG_r%#Pl*$w>@}Nzh`PGP>1I}&Y=P7J- z;^Z8PEK7A83Pb$5xpR?oskPc)jV(GlVtxw?rt65F&Mi9<=v0zo=pdA>mR$c%t!9|7 zRIeb=)eIaf=E%|8TRZ62E=8~$x2EF!$7q}rVurbO6OYp@y~KQd5gl!kVfkGp_w4j* zHT5d!tif~KLO|Ryk&r=eed4AJETt(63L~$n~$LqE4ECD?qg@m`p{d zY#{3usn@HNx)s%yIuyF?($H7B1ilD5K`zz$p=)#8d57{YA{gj&&QhWPIy7NOF0y6N z)kO=$V_=Q#_OAOi=?eE8u%{J9sV6foS^t@CYQoRg6<3@}?fT<7mhQHT#qKMA%vJ`S zQD?Xt*QPAZyEg{1)gqmkIC}M9EZg@qSv!p6<4QWkpr=*0PJoWW0?gR$pkIrwLQ(%c z#WL=Nhh~y-Vdw?&^IkYxk(icOtm~=x3hEs2*I(;uxtgO((oN%XjE4&t&ZKwFN@!ga z)pnrDoB?{r)W9dfG&m^SITq zXVlMI7Oz(gSYqdtcw$8sg(@xQj+!2@?o(_;-MgQ6cop?&w>JeH>hS2k-l+0xNnT zonB`=o=fU6Igj(hH;rg z@Hh2&2YoBSkc4O^UzG7f!U#1sMxu}+qVNf&=Ijh5m5aLpdXEFaPF@d)M4n``Bm-Q8R+xtoBFtezLgkp-3Km- zI2RQgC&0O%9i2&KVoe~LtzFm5(Fsw5o_00m67eM2^&OTTIe$_|zhOKCI6519to|P8 z_Im|%T4g)>ZPq{c=KviS8@f;anX`kwfpQB$nDufrMziFH)}{TH_79o;K?V&xL519YsBFNo57Bf#!kqu$95=Y9^IHK0p< z2YnmiX)8$|S6Ejw=M5xe*@|+ni`t=WVb?JQ*v18+NPflKp=jxLdAOz5L8sX@h#eiM z1N4fnxBC$s^w3@W{3a@B`@!1m=&`kfzL}EICdkhx+EF+pR03dX6YI^Rq=Si)tK(-d zsw*>Y2Vwjvg7gaL>Z0oC;C-;%Io)&-O6!tOjVtI~=!tU+&pW3Y+8uo>!O_{sQL8Jt_tet=rzzeMUNn!(syLhUA!HB z-r05R-&+%dh#mATg#o3w5vrj6tp?~gmB?hX2py-vy*IzpqT z2fgIzVK5%i=iSV!1Oi=P&My}W&oy*VecD0aV$?2(XG3#-2vwY%NQW5sf2g-*B_DSK z;nmgesMFWjJXMUYq!X{W0`pjZ{tfK*my)du*yQT`gQeC7C3TR-irwG020bL}>gaY+N5}0BH}?ei`IYo( zvX&^^Yu_{OuPPar!ZZ+?*mZ}R&FZ!{AzRja*iDm9A2~~9koII#P95sJk@0u0FJLV$Y ziP9ykv#`bp9ezHkqeCaywYu|wzJjA`j2=11$3RapfNSXIqi6^H@1@~@xPB`Elz!4& zMR2TebVoRfd>mM#_3jUyo_Czg3{7-3ynI=tqOKA8o2N>~-I0Ex(;z%VY+=rS=;&iu zX5I^~AC9WFANCl113Bw8!j*H5%w;}JO%e5#0^sqW4}`TQ|F%AU7f;8blzU5h%Je-D z^wSBHqksbX%8p(R(#O6f`%Z&CEy)4Ko^@X*+V4mZ_ZCS!EI(@w)r`E(v-&u7p-KI} zFJW9nw$5APS%(O*x_EqH{Qh9jf%?e{!k@%^s-tHj?>6{-v7=Ma$G+XC*S3?S*Ab-` z8TSdc3gwL-za)U`2m1d^Jixc3E(y)i6iEWUu*Dwd<(?u&KO9dd!x{qOXcRyPzTAhJ zss9@2z7WQH(1Sg&COW-n2Yo%_;_0$d+a%9(z|95qCI!duM&;C4CI86y9Wmb!8u)p2 zms?V9#BA2CY9R+F zE3-wEee&^ZS;%n-%G`7!H&eq0sLs)zU(pBA!dRi5M@L%So^b_T1<-+^U{0W8&WsKQ z-rGwEVY=a5WAre$;W+dY;r&{kjyve<%{*4>yFTj3y_#qT74;bWrm9%pt?ilx-89){_{rWG^fpdJaa^=B zQwZ7%hinQ2*le(J2=X_Gq5wHW>Pk@}wLum)M5|H><&s;KYb6e$f|V-NQ})tEJydF? zsE7kXd@tU!J$-(CwkItRIQZ{mJRW-;IQ;sbc{6_Y9(3G2U4yc4^nXB>XySXtJkHk9 zrGHxkbcNcvmAat)`^uetQ_2~C&k4(6!8-7Qb7APb=bjRjTOa5Yj-!o{D;&KR9N$uSWhrQ_%S;^@rM zQRL`&K&oLQ5zQcHBG{$y5tjd|>k;V1Ym@(xWiryztpwqjX2!vHN7(a7 z%(8F2x*pLVpw4@qzl+H=tk5ANp3|Of>FY}z>A1h8n}mlJTmd>~nV7~sZG^9_;a7rA zllJy^!-~OHD6w9E{xpf_EZ7J#A=ew!Hu0MdIqdUk&juZ94!(K26KMB%pRJ)ybMsKr z(bFUFY1wB|N9m5I)7J)iiltlPIOtfdK|yrmmoVodzrGLUTO~k;B0fcq9tZJs?d_uL z{#40Cufl0;wkUTNXQ?c}1y8GxsDj!t#Fu4XGxaGK+K*v5p7 zoO|i$^v`vpM&tHc*()AXeJblL%1%4MIqg>Ivv0WP(m3Oiz^o>vj8^_^xuNq!sFSDj zyXH0#>n=I+QiwdUOf*X8FH&E-qswKm`HpT+Y;D_C@?q)z{BB4;7lQt_baWh^*XiOd zuMB@z-YGzT66MlXWEf7oF42)$0%a>Q&YnHvm2Km!O&kB@depzTglfJ3aliBW=JiZO z@5OIWm9tL&<#~g3(Wl`nVns%Z{KfCIq1FyMQR;Sd1U+t*j!t#%=-b@Tk>^U>#2p=T zNl{y)fxiA(0s0d}19aT;MC8)U$klVaOY`K8uY$a zi(q&=oJ@GTvLQWNfxWPGz4}5L@vw_voyX*ymi{l|?!IRNX6fiet%bevz|qCt(w^by zfNwaxCW3wCM)V_i5XIz*hgzPP1yV-YU2!$XbasCTKEKjhyStz5ZBD zulpB>vWYfCooy(i)Fhbe(_!;>EyP9(Y5zH5*_;D%z7~*SE~0XzdNpbOx?Zo1LIb_EncHh$ zE!i6vN_bf{ncY|>~@698`+M~Z6W#GAP<8!)-d7tasMb5Yc_!2Wxh8lHiqjIuT8)(V6$hdvOOAr6b|vDsbGm{|OyB zc{x7T_z%76J?Nhopg+>gUAB^+vRwk7>E&2ti-o?GHXE03?taw9l1Kno?*=2-E;>f< zV=%6N(TPrCVRT(7<7veZ_O%Hli9>Snw-Vq`RV@J==PQ+IqhL z{Sh%Z8lTfgz)#(F@%ojtOi>w_^B9~Klrv}F)cX6~-QB%6n@N9;bT$}VpWlrJV`coy zi~i`gLw$O@W_)}1X5TjmeNz6>{;8tB2)IF8FW6Vp(_ZsbsncHfRQYV8jkuR$L3c+E z8myAC2kJCRrw>~@vlh@v($V>mF!}9h&$yOuPIcz!x)MH!e&1%ovTV;{Qc17+%&K_z z+UjEYA4!sP+N>jG`FW0`yWE_rN^hFkSG(62yL)3J2Gb{l(ag+Xcn;T=PDcH4#BD4M z;q_i`c!H#}KN6;44DJ~GSIzGEhMITVR~pL0^9`Rx{hVvzvur)@;nlZf@7?_xMxuP3 zpc9l#1gZeGCLCRKniLPy z>EM2cpet|QYR{|FdG~l;H3D=&I^#}U~4$;xQ#Hial4~F~G@i3UHgrSEr_Cv7yS6y&) z4URrn=;-sX>n>Q&2OW1(uJ_Z(8FgvtVv*e19G&{^@4*&{Dw5ztdo87hF5Eob(d}*Q zk1~ZXctFmpAAzbLd|H6+RwAdf@N^l5Bg&aJ|HgzG@f}x%mU}ox-#fE0H6peGaKn1I z+mPCm{s@r1m#YXG0pb~sz-cba3xJS@EGjF%9l$p17=9_r#L1do8Ef&%U z%w-;UsRy2pg-j1iHOerxHqgn^-@D&H(#plWTZFH`+_4Ux+63g$Pj6um4mp4M9Q5+# zZw2TsQP!w5OLvXWaVHZ@ISHgB%xH|_S#980JlDP{bu%rsFUL06g2w9#yvfL^|J zl}lPY@wqn80WrN75#TcGf^*j-==E>+Em(qgQSO?zbKbqJZ}`wqpu7*VNQm= z(%Tu*GQv4AC#I!2Nz~povP7pZH`W)6XXK-tz^GG zv63cbYRxh$9zv!2@vZ}vQOi~g`bW>p<#*pLm&<<^ps(cU+$tPiE?%}I)^ul1oWmze zT0l-fgB2r>bxSvSGPEP;-x{3;x?k!{hQr})r>y~{wJen*!;67+i6ZwA9z+59l+4TH(DbT5wjL9g39JpUDS{_Ff#tJkg8;dkM$8Yk(n zhwWZjrcRRG`1H;@Gj#q?;eX`l5^_R4Y~$G07`XPzuo9CG+v6PFg8ux)yP*ouSEXFO zlt0ta9gE={tKdXzJ?6vLE6Q+{Ob9x4-E|yrGo`2m!VjjMNw3%JHE*#IzbIQn=wV{* z#whI>Poq4PQp!6xY_)#ZDC6sgL3pk6l1aNAUFuP$POz^FbR=Cmx-@kvE$NI~RwW1D zVt3>AE^cWZ{l&Y5qd&sQc_dC6I(KvGwXgn`b0tJ}n8@`)W=@W7lE=mkI$Tlr5B5(x zC8+5T^zyhjJU)k@+!b8U75?Z3^xWsV^zi0dtK(LCrtLahMR7pa`geLcE0?r#>cD4a zeNQ#?L=<(6pl>lpCt?|nqT5}=Du>&)*bzH{ZsYkh$WPqeipJ<6=+FNsKzE8ysZpJj zySm*KZC9xet|UOmV*{lRJA`XjAP%Rz<1bqCE0odNH`#fq#2mXT1H2TpThys=nfjU>os{M1 zARd=v=eLIh_Rmv&+}dgFfDX@!dzPn%QvMDep&~rqTuMu>o zqeDrA(}LDq;hWcLQ0B@4Ua5tL&#>UGpc z6wI1xRlU4>1)Kxru!99~uU=iBd3ds;vAe7c3ghHiUH3kMjy1Cw4aacReAquNt03r> zN`B<%>I8#v!*kCt`Nnh>H=x62d#;bV-EXxnRXYcz4Wsny!<(vqywke*3H2?7{E6}q zbN?^)&gi9)BaY+L?oQj;a1R!SE5$;~67mp8U!+1q8Wt}(got7!Nt?I!&cqn7_fX*q z5{eHMd{EJVv_2_|NS?A#5N5K5{;OgWAy6>B})kQ#k*YTZ`GcMP#aXu7u z)OC3o#^`phTt4j_mdhAq_dvHhy?Q#`U|j!l14`p8opV9IXQp!vz4k6+=25Oc!?#|+6a`pJ$iZ|F|HaCV*K{fU}B{laup zzhVKJY*QWar>j{ywW;48h{Nf>n{*o=g+pC=I+-#v25{nBE`lzvW(?#A>^L^iuAwhsr7`WqsUxrTNmjBj?x2(I0o_V;o&sGozl_6gN;O@W%%_ zL91Q=!O_D&m#rY^a0?}Zj-*d=#*N%@Cv(tgr=cQ+-$GAGjawo^09+vDVB(t$N4MwD zX&)Fk6were1kKKEwQ^tr#F^b{Wd&X1#eE^pdQu=eiFO@j;R=xE(r z>AJ%oD`j*G`7{LMJkWD%yx+Pu?6p$sF6lY?{Cp`l59%6Pb{nI;zFt~i+UUdt;G?#> z(4Yq!nxZqllc{@;2-UHGj-B@**V7z*8XllDM+bU+&ZmQ*e`8Rh86tg>?s-%{k6+?W zeom;%YxIn8NdW!*<>lphhXLMKi$7j|%uIA~!?aDw5NG=rmFfU7zQ4V_ynPR#N5$gM z(aXz@%~8;IUs&Evz*8e*Fpd%kIyM0GwZQUMuVKTK+ExLS&z~^_-N}`DCQBZ*`pKgb zf?fjBVRo%`WMJq{b@ZdvjY04GrKU-Wj=zVA7$@ju%v0^XM~b)~2k1mPO@UJjuhAUc z6M{<=YeRG&boYZE+a{!>coevs9N-WoH5-|d$s~$%O&>pP(W{AaK|gu^<7Z*8+3TZjx4WG$c%a|s^T(W-ccpF92R#SxgB+VF3-+MJHrBf{&JW+Wqt1i#8`?TK zZS$(GsD(3X*j)|0Lp3JQCCNlL7Wj&Hbh@VHLjqkTU1^Y~r;?ue2SK+cpbyVBL5f@_ zqik0rk)y|ePMAy34Fm=Wc)xhzVsi;`b`SH1S+9%AWs-3dXcp#G)dSr{o9q~0EQ`C> zmTx-TlLbp-%GX(e?|FiSuLQc_@KVGw_zfZ11c08;-R~80hm0kA^DQAk-#1Ym#<=WL z_+5|WyVa%ctEF`(6Mopb-^dKvzN$q#B;}(|s1x{Oa&t$-_W*Bu>1?Z;n2ppPQR z7e}8t0sRS!+-0Gn9<%*C($NhN6``PG7!F|{Kp+m@V*5CKem*eKGo@GGRH_co09!UQ z8$Rd+xdh!M_Y}6ApF13ceD8v%_p?ND(9spj=|2l*-1(pnnuUBD9vtR!YZCOP+$BR@ z3fX<%SXy0Qx-*y=s?pL%Mx~fimMLjyOTOs|XiScdnkG(l*2${njbhqlKt0$9**vJwp6>iz|N)L1t*x=QV)hpA6)a(#@#{@FTB*4SbJCn~L9ps55 z>k7u`g0n}63M1p#4C~X{a`|S4bDM)qAichYNaI%rL5ImRjdQ_|-KcxsonFf~A@bhl zCU2K=ziyQJ%-(3tzw7OQxXnzi#sb0SKYYJ_{2UkZX^hiW*0`~HIOWK{dcs|(lbE)fOJa;F1FGE z&;yPx;L38N>*kZmMC%)a%w7|Kax|gVOq!=ZgIg*%M0~$8Edmok#r_p znk^WkE1(Pi@?>^fuz>}yGsccv1;{LeUG@7yYrfTY7~7a%Z6W8Y-I*}VDeD|5j&waG zMf4sQ33*J?_wkL^;ywbK9 zvy$D~a@}QHh$4RhN2}E{LSSz@vy~>$`6EZita3-!u#$Z{U+Tf9XWQ>wM;A_^b<|{* z$gIzoj$Hlrf}2cLhh;ca2BY2b8tI{^8^*{9VwI36Ni%d>5IZi=C58wxK`kFaL5xbP zCy|>gplkE!s83=*M@P4RpMXBp(PI<7Mqo~B;i5q|;z5V~E4To5-w&zt!Tw>R(KtLA z3>chtzG;MjE>UIY_2qrhz!MwXr)SI1d-py`2&a}#X9eH8-rBmUW8#Ud1YLD>Eb&`I zYDicoli?|LSSaiX#zn5w#rgdEQYm*YXY<0k?#T8Fa&!guXiwJ_dNLc4jyJlc6XbF# z9jY;aPNe(htwIeVGuF8##ex3JVc$ETE&*$eQlj6kyk%MuV>3BGzQz1UhH}dW1fr ztt^q2u0(*yTQo->`A9W0pkxsA{l&$L^MMhve%-VWE7b;9K^H;P?FO8jdIRD12J@J; zyFmJlEue{@*Dcmxs22(g>3XNgEL&{lw;awbi*tgUynQ=2-^#fH&_NJ{TZNJN>Uyrt zc`WME$n`G#$gAWi&>1>TmlNwD6K&KSPcV)wU5AMguDo1tSVGwN7$b5w zmhb54#M(l7+YvYE^1a}~-cP5u);P$5OLKH9ab3XpI%JseKtIB}RVa?JyCMIEofmF} zr7B`vk-nWRI`ey+qwAq}h48?!_G1DaN&n9ZTJ`e?$WcDzSmV|;sDhe+9`5LhlT$;F zxO+**HO@mF-B37}d}A}p(V6kSQsbTtd04i93fIXo3MrM=$2^TbwPd4(d~B-XJ49MJ-auvS^gl2 zdMCuw=|DFXcpvomh+I_4nV3_RRq|t~jRAeEr0ckU6ZVe89GkAd(WfV%$4MQLletGE zgA7$znh7mqyXNN%FCTP2Jya6bx;sdEMg!e?Uwv;fI{><*Qr4oFVUCSk&ufRw7M9J6 zNcuK*&jIu;C%aWBUw6|mG%xZdW`)}qT-;yhwt7}kcnSvKQ=EYg3m)h@tx_KnM6U0p z?4-#EbVio0yE+!6Fwji`T{HEFrLe<}Rn}KF;*WStprb~HSbwZs-V@lyqKe#{tQ{SF z0(zVXXlm-c_Tij=96Nn3OQx| zT!#R<5LpDh&iP@vkcJ_8kBepj483h}tLZL}xk*?YK21Er-GX)-=FwAjHq|b*pd$~1 z^qv0pcKP)Ly65PL1n+IZV%W2cXm90#b8ZzK+>)Bk7rI=avhIa|dw(2v zL2tud>VCDg+SuzsL~b37z4VSVp7KF|L8EmjwElajsY{vBr#4NElt&JW2t!wry^jm@ zK+s)lf<;4%&{o+;hN19ki8po1u_2o(>C+PQfalwjm2s+~hvvGYM=gkpnmwmqV*)+K z(N7jD2T2ukPv$|ja>9KHE$S{vKza>MQat3AW$_^BpCifbc0vfZ-`>XO?R37+_geY1 zW0AP^wGb)5oIn>M)8GCX9sud*H z!hM^=_#DzoEVSV+Q9P{Y&t4e)LVC>#h-dfOKbykuy5|6Ni>I>qtx{(D*R0Lgw$~w- z%!XqV=tO!b>nOtqJ>ynE(#>J0hv{R4n|?H!q03^dj!vYLPy5I}pzX4+jLGSSNcZKs z_;U^LDA1>DObI(ZJpnz`(W9jwm>QY{QU^WAH?5X|D@4g4=oxlfxj@1nb1qrPbdV8yJmM^pt7D*g$ z!}R%e+ZMt~HFKplekZu$7}kSkR`B-q(Z>4vk>cnti1mlzJ15rpFFVJMVfvKI4o>`x zIyt&*kw#1nI~GSLd64Q|au>l+KS!zV9>Dkuw~Pn23JsFNC^3|*bZ(IgFB zSCDgB_o)ugJczm?18gJ`@4YdxcyJPieTzriwOz{)^y`Byd4{x72D=EKp5! zBWDOX#~KQAZ+7|^+eDpvT}Vx+n7K{NmwP*8t+&rkb{Dp zlb{WucWA_e{>jv*xO&E%fUZdP=;9omFo%LHUDlBScREPT84UIZgHEWSGX-?0t}2TM zFHDhP>RuCjbESHMEqF=Lh0@h_G|M&y&_U09xQTND*=y0g*Q7gWDxhNoZzRw^O(N)@7!%NSM^`*u8=#ML zP=g+DSLcbH!HKJVK;09gOL24*A5<#0NoMm$VwWo`j*!7O7xbb3y6 zo<6id@btHB=*6G)8hf}7*7B9$utsK2*n550$C54niJ%J$L^E#|7q-aqjUQfViKq~8 z+jN>u$4r{cxo`fm^X|jjSFdO1Ucde2!|dz_H_a4-Lc$&5=|(i>M0rHtT{m=Wj>pkS z2rUt<_S5OY{8V@>F5CDK;0=gy%wx%g<4g7y9GwDkWYg3lxbcx!sb)_Nh2(GGBR&}SX^NStFD6mZWLQs;DsZ(>2v2gv#6 z=I*9@1hONY+$|*N;OMlmn;W;4%7L9Ba4gPnI?Z1e_Y}|*iEz*n?9UTx+xdK24$f8m znJ@KOQCMsJQofgK{Q{t)gcSF*v3?{Co!0Jf;r#G=>B9$r9qjw{tG6HC?X;Wsf6mU$ zg2B(dnw$M;Ztm@lgB~u=WF*jsMC77V&JgRvM(N{mbRylSTaCB5GwQD7o2s!#fj<33 zEa=#nvZtR-K#w~f8SC)%?fBB|JQ8#w-PuRZ-~agI#rfu&#>pVCO`8Ql2Z;S!A;8iX z56IFjc-^m5YbSzajR1Oi80gU2%WdVL-;SRIFHizyp}4qfwT^`)itX0we5uez&`DW| z`^HjN*$Cs6c5~_Fhv(0aUjMN3>l#naBu&%iS@Z47m*0N-?O!v;Z(q-TJ2yA?#~NBX zl8!a~n-S(S)a4(WxO8SBz;MfhG)IwH6U)1&I5haD6+y56Cc#y4S_EhAtFf!SWkfEJUKKB+u`|#vbo9!t*5u|7ExWD0 zAxV%VJ;fGGB>jG$zjjCzEJIh|+~e`jCn_r~Kg4O#I_P{)xl*8;v*{mndJFUPPX#(i zFDz(dgkvxLHVkxDwhHJZ3%4A;#c0#c|T{FHS-3Gn?M(Ey?r^nzS65H ztAlRD-OZp|(AU`jU6qCpV(0=Z8cX4glUOJ9c~l88$SFEE0J`I&L09rXmlo!#Ujuyy zo(m-_>P}) zP$LxFn)mpj`T3do<6xenn>a7MYn366+q{(x4{V-heuq9O=pbDum?h8{*8FBuyq6U_ zKoL**r>Ixq6BWfF08r_~NV)}`P22Umtes={h<>kJIh-I6=$AU^2LyWVkn{Y*5efAD ziHV6f0z2q%9ql2zSJpIs)+Xpg#UWKS7~9y}+uM&sdqMz`v7Ijy=+z7OW=IEJ)@e_HN=l!0L8aQhnUB#0skByE6ibwWmQ(aqji@yix0&IQ!BJQd2(CaxC=?rv|`EZACidZ^dAIDHHqfa_@Clw;h3|+rD zf2XISS7}vh^)$}!-cH?l$ocK5^y1yT^u)bLlL;Pla`ZQbqhIDbdcF{K?qE2e24tsPMCwRKM}{rYHPoGnlrQDmUE59i28Tq5lIm0(x9j&!3&LQne-Iz_ zM0#)u)At#_$G@Wi&Q;Fb&|He`SuK%IFIoEj2|)*aeEy8byu6{8|MvF6b^D8^bm7&4)lix=$vGbt{irau7&~uOakzgqM&bebPQbw;l+R~ zO7d~a*ZSZnU>QtyOggWt{ajL})t8(V#P@NJFKuKUEL(ms{yy%V(X;CL#f0U0!!q%W zC<=D0^7@pH&ODu;ofk(PuaZ}>@8Jcwe?$*e-@gUDRBDYsl&@Hky9sk9u1=;P4yS)u zTthT&B8!%fD=>6K-MJ3QlCZL`Y52MvZV?^ahmlg-&mP2jPR^%f=Xz()zM#`3J71rj z?H}$R9)@JM@|25m_dbEWL#8IV1bV5%Ryo?u zj?SN;%!W_x=#UX>XHL$yk0`I348IEV28-(!^y11Ry!5s;-qG3i=yIU5lUR^n8(!Ch z?3x~QFOX{a8GaJ)dMXW?igovl;N4Xf8Vdc)sj8v(osrm~Bx@q5oKC#ERas$_e=O+g zLUn1-A=$b)+MJ>68(d$u7IXAX>|P_#HHj`UlS7KbCg%0|iqK7bY3#o5&HF6M($eFSK^o|ZOMh(Sc3wo(G?o7p2=-R8eZf+-o_nh8R3GCDqYG7Y#p*UU=lS*u? z$FM+;O;rIHRizTPsyEq4PiN0W2#_n(I7U%8!a2~ZuR%X2R%ZdaRdHD5y%vCORjkgj z76RQYx7&T~>7hS&HE+vGsKDr%tjMlOc&9&=bM<)p0>!SWF7HCdHUfk=;nWgTokTU0-YS4Tc4K!oj;%fMX)d5pwj*20ER$$zuTZt4s=z*qtm&z zkAQm?k+{o^ph~C@ZJc-aV5#02+1GZMYq)(l=xqSq;Q;6lbD%d?zjGU)TOAI!nRLg{ z&``*#ki_j&`)Wuj1-g>yN^Rx5T}(XC3gLFQBs7&pZ5&G5ha^oRLhV=LNUd+)B0@ow z#kKmOxPS0C1I#^MuXo3Ck2Rwpm8*Mad35cKrP=t|JJ%9c#}~t=0-g1ofmmGqqc7(E zoXg|B1-*0|ch1m@pOAd>`rM2;o3=A8zFmXfVreRt^>Z=gtR&12I%9joI8iayeZDuf zj#JS)Bi;LObeybqIOLK+DQD=s06Jkzpf5N~ptlg{4y#I_+Z+z7gGsTz9~mDXA6Xqz zpirW5^f;tA@K}`g<*tqqhsL=>Om!r2E_(e|m8_x}v^-nus|{*v3|Da8=SzrzR=)&6wX#-OJX# z;<|AYdRa#|R_AQHr5SXQ?#0DL5pPntX*)&VPGaEu68qOSKzygJNM~&S%SCJ>dN0a%Ef@k43Vw`Z^nlP-^Xb1EiOP-9a*e zt{9+$Z}}ZLI(|YX(D8s*PT+ilq6r8aZj7%w5P(a!cfAAeya;u5EOnu(8uiYrWb0(r zP?v90MO8ai8#}GOT`@RvmBe)3kGmF?GLs?+F}Q@F_~Y}m7k@tPT^2B#_tyBatYM&& zgp!B2@C~59&4=E3z!~?S-q9gY`5)IU=pPh}!&UxF7ThH+9k=(^GN4#NRjT1pGRU5Xhl^VMHx5bzMi`Gk)tc_ zAso6k;C9&LcXKV@q*X;v`7L*J?Ed-Khh1x%i7FxUAn8L9tWT4_b_=fVk8449v2tf7zHIq(t#=w} z#-PXRU%e*^1Kyg3rOJCQErSk~IY6Cf-FRZnBDt~|?qm*JpB!{d-)}JMk8yPT@Ga=& zAun#|yrY-w5yykh%kXjIovyOe_K}kq)x74;UweF%%@TgS(B<@9~;otoEs(XNTTDT_J!S@_o^suCH;fI zoUYYcRJg|}fi5mi`jJ8x1x(?+HQuK#c9IPbI=^X^!CroQ3xCQg7mzCl`e(CHF97@2 z(f_R#I^S!8ck~bBoJr&i-B}3wTDH5hd&6n8ZwXCxcb=~x$+o|Df^!q6P9i<4svJ-p zl=X71L~)SfN{E`+fae(Zz>NCk^Rcu?;EnE1yLpd5foW(nyG16;f z13PN>4|>Cw6MIv`<3mWo$?ry+w`O?_4GgWOQdg>C6yT?Qw3B6Xx3#sFU>{5$qj%Kq z)wn29Wi0lh`>O&XW!0DcC~xHDLEt~Xf7)OtGJI7m!duIWX{S>0d%L+W`$tXvK2Wro0JLQ z2Zr2^*#V~;V7rw7exsFMI_^U-^of;NFSL(x-HlazOC4+OYG)dSx=8h1_4TO-gH9!+ zbMBM}Ta#A#j>skYT6)eb}KIAVa@*?638Ja|*#rmD64_cJg#- z2FvsTi)hhEFAln;tfPOsF!B!NM()`3k8$+R#?)^e{UgPPXo{xQc-<*$=z{THD3Qnf z5n;}<=c;q1n?y}sA(^y>Lig^@o|W;e7I@DZ?ICch=h43LShVjwvfjna+^}fl%|XBu z+4>iF;fq?xoj~93C1q*uzKY+qTHn=<%TPVk)j>ikE!CO|8M(HQbFj4|jgM85vWgCn z7k^$x!>H>cT1Aw&4-xSf9!%VaR{b6=G#AGboj7-)o&8LWXJ&d;G_vWuct>Z&ilxhD z7~ovvuulm(AASFLyS@eeU)jr2Hsnq(J`l(3ybS0jz>9M3T%^&|ssW8AvQypNZ}2(S zn@v$)pR7#5)S=7a^OIp2oa;jKnCP7RIubo|l$f;>xH`VMfywwD4eT>s2IwpLMmr63 znhkVyZ6BA={DV{|R|c=B>MlT6WX$E2WeT;o;$~Y|U4tIqYCsBJ`90qGpd?=S<7hUA zC3-wSejX59>6x0k2Cw(MkW&%RZ#q-i=mb=HzbflMz>;jA8gv$YXO!Q9{+|}j+sjxw zW4wf;-yDcDVQx&_%V*X^baD>A-01EeMnYPb8d5}5v%jQ*xU7V7cR7D&*&*dob5;!(pF?-zHaA~V zMO4Adks*G(i<;ll4r4-D3N zJxlnA)%kIa_qpr2*Pjr@MbY)c;0v#JNsyj<8-g6@CELh-V|H{t$;OMd8~G?l|AJ7z z1)X#BQZ;dO6|a=xI4_4e!<}288<^)tCko--OCzz^o@{WA+ksm@VNEZNF;GYI*0$)G z>Ik;9^$l4S20FX`RblQ9(n)($6pI7ss|31xYk4_Czm_G~tRY8;Kvx*(wQzLnkq^2W zQe9Vm-wx{*OS@C2sd>~fs`-iZ_sFN6#=UV_kcJYc&%KmLcL~#8Z_qe}u{h`!Zkf&* zJAY(4zo$Vr^iS^SOyB!&>D#xU8;;H;RvG17w~6vcBbyFCr`tq2qnxjNWtiiBkN$9~ zv=-~0pkmT%VYTaAL4`Z2;7qd4Mxzvi--|{zg8AV$8*NpAc-ZxCYobS+SyoN-meC4k z_V%bWZe-R9W$FPs>f`EDD6Mgj80Y};dgwsSC0aUOnY2Z zVawx3)2Q0`=}!atjyBh6yauet)7Yh2Q-f_S4K+2-!22{iOFi%Z*3#!Mvbolc(|3Ns z_HmB>WtjS{qvvL8+@!sXpPRQy6$c$MgRb*^twJ}4OpXrip|Twv)Y0eF1Ue*9$?G<} z+ywe+1fVO9jh-kf+)U9ohrPs%9f5b78z`5M5s@FFQg_->JB-3OhYosu8WFkC=1qLN zH0byvV#eNDix}Mep9L98`#giPZs|^SC`q7u{X1mmI_W`xj*V-8d}aoI{`{)Jiz}WH zDzB$Q%+SqAHr(9#EnGUiw5{u6wNoDl`sZE}dkZ?7xO0ZiIOh!g!|G%h=-eXRg!y8A zTFn?jJ>6L+%EM#b`>GjqWhBxWiOjJDx=!_ozB-@n={uJlrj@x30CJo&!zS}K`#AfA z&vdVjk8eQZh?!g6yd02Q!FRNE`f6rmWR);ypwoOn zG0znU^wW+crPoLy`O24EmUfcKrK()9+=;Bc18S6WREEZ079wXXEJqo%#DWqW=E%f^MpoVechR zRK!?*FALuyj&pON+&&h=)e&=_Qe~AuXH$1Q`gYdSbB3;vAlw9c4-PrDJKXR3+D7tO zR*vT87Pi#lGjqr6_Ck4$9b-y>%OnCFP~)!^&C>JaR(7sImqU-Pd>yOqSzq$Psi@@e zv&mY2?e5ZG5*an^tcDTP%3M2om2e#QZdD@Ie(Dk-q|#>^8de$P40JQ;ALi&e#>-^W zKMO}^sQ(v07pmsJO3=;ebAE-+w|Bm|6J$X!&QuF0VYnw^Ah=68I&M{*!2T7a)3A@_ zQkDbE&r#Q&jhrCve&h~V6-JnKBN9Er4Bb867Cp*`@~Ew?uexu`T2$$UmX?PtxN$Y+ zsU5=5IEwn#u~i=QuGG~|30-s~gKi@hcZ|I(%Qzc(erGF@o?IQc@X@&}7(7Y;?nRA9 z<9#)NHdCa4_zbJ_`3A3NQiKA6__X2v6G`Mk976dV8&v!}W&V)WqJek3J{I)P9QAMO zpi{p{>8k|YM0?2a3%^<8FlPl zsWTvppd3Iic@cq|si{F_oJ*nxnGFpM%WT@tXtxxq^gcGe64vbLJn@X@j{*I&Mx7j; z-q7UT@>POvs*J)=woe?d8(K%va8G+*n_9U%IXF2uIJumdS`%F4=l4v0u9_WPf|F}& zbAVo@gKm_%ZuCTs%M?G|Fo|bPhVsMn-{5Xw~l@ z3_7b+B{1))@p=8JV3h!0f|)0th6e8v3grZ+IN6|YSQYaaFF|@aN58`y9aniv z{|wgX4E6uCqYK~vT0w`*sMEU%bbfm(b9An!4A(r$A_|tZb?ouXDNL{M!9HLYRf%z~ z7Mwn+zK?aTI0^J@=Sh`Jx-NA+j6~kc6Oo>AbAjvYXk^?9p`-Jhm8rk4`fU8Hx@{Kd z66nT@WPSH=!G=g3Zcf(0Q7B`zt&WZ^Hx*6=1~R3zn)&@eDA%8cigtsE<;2d;@$tp{ z;NUK1>c>g~8(0)aKyfB8Z%^QBOQMQdyHk1Q^)xI6rNGjRmm=}N``P{bOTe5!rx#r& z(m9z^>Aa)!Cn=kwa-Wf-Gt~ccj&8v(bbXbevk*+_BsvE3#GR=;=yWf>$@X6BZXF>ov%9Z?`PC@n&$JyY;WU1LtRUo4tRueaA3)Lx$_)OPm}Ou@5;Q#i3G0J)|n` zpZ)BS>~uPQ->My#TDy!%ieg2diHk}!m++(mk}4?|{u=KyQNrVr2OVG$&zNnfX+Q@W z19X$P3b{Kp-0bLlIF8?+{28p#uk{~aql;f9=%y4p7j!pWoQkK`jP%bPtjS%A6NEM? zI|a@cgW=xT@JSfz#vv{1cT{Jt8jeA(jdBP<_YEn4W$s*OPbdU*Af2(1p2&(@8Sjj& zpe}~Z$CYjDQYB91GtAQ&fHrwM8f|;W05@Kw?ZyQwShiW+vKztx8BZZMeG87>ykoVo zzXNWRgQ>8iTwodPoS{=y4hc&@epU%gdr;plOU`+(zhOzzq)cO}{=!m;4iwW8fG6?s zI34io2i9Y|80*)le^8fq=(BKiP2c~|IXX^Fc5Z%!pc_STT#kit&ez11i^LhA=Mq3u z#*}^krEuBX{Z1WPssT-Z0 zQ*Ox`?KyXTtt%Gi}yAxO2MRlC9+3m9i z=n#Hg8NiLpq&&`*kVD2}TMeQpV>j`$m zvOp3U?0Wck zg0=Yx5I>-MxU%YPIJ{pa+^f6{_r7&nCSu`mjGTS{z2uZdF@KE?&H?&LcWi~Wk?i$# zc4Exac_ea(x3)-R&!o9%WLy7%Xf$JEoP%Uow!H!mI^Krt2I%i7v-a>|BdHnEMy+p5 zbHc_z$6(S&+qxVA-%csGTkB7#4jB6rt*_iE>fpBE;oc>u!FfP_-cWPDuBO5JXCr2L zO*qSAa49w}?#`9Q(TlH|;m%>h>vNg9|7UFd8wPsiR|vZ1=$xgO?(od#*7yawQP1)7 z^)L{Bi9^t1Z-slVsc`ShAZopa!@X~+EW#Q=K8&+cW5X6z)g9e1blQz70rc*kSSNLk z>cJ4XR*ybJsX<&%PxM-CebK(QwkVS3cuEt62AvfSOwS$00^F{@Xvl${KnlIb zPb^Xz&8e@{Rfv*l(l9zk@nWb$QG@+9D4M%zjLr=iZq6C8UB?{Cf$p|$kB;>N+pWgN zG3`5<=H><6vJ+ih30prvw-t)b7s_mkDMVd%=&;YmLApmJac=|C=}#1WUyH6) zEYFqM=M6JK$pD=t)Acf>d)K5h6{&>f#HZMpq2mR8hK~Mk`uaEGR}DJ91D&y+AE4z9 zVdA<+6=r#YS;XGfT9ixd;n)h_+65Xyc#H1 zaY;k(b6PbkbF)_MfUh8oZbses3Ps(mXYJc=c>Xd#w?np9?Eu}b#r4M8mu-&$y4}Xa zX8+ga^2JR(k96~CF46@$$za#otgPabG9C9kd-0^fH?BAXI_MZPFmsm+6Z!#>NubLD z#L3c$^|GKt4D%a{X81Mw=jP}+*8e5gzx{QC&H}ABABE1R)=HATD137Q!WV^?;a*UU zStiZcUrO59OK*C^FYzW-t@XxYK@q#o!&y}o>1br#&vhvw(>7Fya4}`hOUkA=bRHSt zb~fS-!gCIE_gUkW%>doTK!@A}dOri*Zi3zF#t*ykPqW=<=?X)gmgy*zsr7i{^QTXj zrk_7Sk&IN8sPS$&B~i|;(Vau_nTDs&8obZ*4~{q1=P#a5Cw?MTa@lkq^rDW=Wz#uB zzfPuqexQ^71xNqNr|t~%QYW*N=r)0?M000~xxYmSZe{VFc+wkQBl{6t6TQ8Nxl@dN ztv9@0MGww5-QmkBDTt$3E`uReGI34<*F``#8)q>wAo)4F`Aal)x7p4b$KF3{2P6 z)&1@%fG-Jp`N=znIp2cfb6lhU8?fV`bm#ZKUeL7-sw~cW<6&#&Em3)9UQp$J=Rosw z%Q{x+7Hv_9VEc)vNeW`F?h+|~F6^Vj>*AY}Zt4w30XT-a@uZPKVzl!rg1Zqj!(~R@ zPNGr!TmX8rY3dHSxv4vw`$Ja`I=TiYQ9DyQ-^PIs$!QeBsiC;1q1NlKsi~pctMQ%) zQ2s%eYx@5Eh9^4`fo>*!r|$k=pZ}^ZVKG2wBY$jqG0gEYF3!-u2uJ@|+Qsi6gI<;U zaOdj-oilPKgK&C;cuR+PLl?BMa=#Y}55EQOmAS1L2fY&(YHNB=60+fJWtB>6(Zz#s z=K(Me1UnC~~Jl_zHH=qd4gBUxO=>ZYEi^5D@O@fx|O#1T=pWT0cZ?WO-lKGHT zsMO@$4EvpO%XAa)pS7bi-P|8qpy5Z~pl|$lzxAzOeqErmyeSR5q3f?4?Odr0pIN&` zT{H45rO$;(RGka3T)X>Tn0)LmE)A6Nu3boJ$B449t`zdb>gjT~K8i*e4IIhz{u zPR`8SAd-P@;NJgkp|NEFpUBYxIyw3g_N0z(|JZJTZijM?-VV_1$Ej3mJ59gC_}gSK zIhZ+3wfz8~ zFHX@xcZ&J?$+KVmB{*>RS5Jv|cwBB^p3d!7VT{|C8sLS&Lf*X zlLqOC%1O_>4bL>_#X|k9uI|Z`U)4Zx?|0|Q-MhoDtKj9s-LZX( zT+Yx-+Bw(fjd89Gc70&79alG2k0sph3I;j~jsX*Tmzu)kCNeA6*>Vhy`clXzIPb=?TlTJ5N8AoU07%2juwk9N5&IH+1&;%%kqNpnv3^HD=|- zLhj`PaTc?q3xZbRe$pEY3ivRLe7VIEd^541JqQ;nyN@}&fDrR~xcjAq9QylMEGy7z zw%oKmfB9>3uWs!fo}bUnogc04VGH`%nz~yID=PM9ago$98*vdJ8OHG>!rE zW&*wW;mG5$_U3ej{nhB`f*qt+0Q9b9dv1-MPCUl1J?`q7x9OaN^sCLyWdLr-iXXV< zJds56SO4I}(&_2a{H`CUU(M=`GFUyVJ}{}y%=3o(_op*YpH1Jdd*+gKiG^CK`_-gr z8yj>})l8{oGx^(9IqYv8{R@{KY2huMJG6nRbmLbxZO1i%iMxf|Cf1w%IR)A=Ww*RN z0NQZ#lWZ0ZX0qYV%11v7g(NNTj$W^lQx>R;J5Alog6<%K90+*JsFYybdo_w) zqx##z_SxaltUE_|g&PB>{%SID`&N5(%Z|0;DhImTw%M3!1mF~*OSE?_xbr~oB1g|O zrkXc@Z0o=1=t}1n1WmXpFK1^PU_;sl8U=q1y=X8c_Iv2U&JeuH=BevLQblnfOCo#JK9>X<&vfk4ydJV?S4D-536%HmT+Lq` zzwklJ1GxknCK{_3SgX%tJ)f6RXWg+3sa2}{!rO4ap=RcO!?IKybU`pXy0I;tsf?rZ z$@Ef&{;gX_|G!Xve_yj0K=Qe%&8%k>?JCVNh% z`WeWYxvKS%I%;d<379#(pb)-wg;KZ|8sr$vVs%1Tt&67(_n*xs?l(-k@=3>S0^P+N z9nwH&D(dJgi_V4PiZ9V`9sP?J^KVw_ZtCRDmgwao?v{d>`!~WR!s`}XuKn-<{Wqy4 z&xMh2_$BKEi5;tJA`uV1rT5sr_A^Nigx+HVx>Myi=i+egacogNKRBGK0z`pGOnuP+NiI$dxZELh;)7AtlHbB2}|G-w!`=fr~tyG{Mcjw>bS=z9LHf`R_~R6N=0sjWpb$Yer0`axf{+hoP+p7(pvBI+N@ zR*rPzR5>|-;Jb_#&*?yL#`{+n3S!l3OUD=>seu52E;6s@&vUrR(SbQ1k243|Z`bJm z7h>Tz8cXNU`9^W27U%*-%ArquW1(>HL3qVNyQekK-yn6)K*y(-k|8zQefdaHB}F{! z>OuT41AQ2)at8WK%+`$#3$nP^-J@zpv5d#MhgJQ*d%miYcZvp#g6svVDFbB2tEg&GyygcW?XAj3J64fos1hNVAAL9#6O<13|kEQ8b_*vYB zqZ8%0`;)0-_`S(V(bw!(;<6nH*U zhwUsgX*^20gr{|Nzxow!=q@f85cLE&v^r(?ov>cD`(@xK817^|>#ezBrwZ zH=yst3so6VWT!*1$p!H^MOHIUybW~?FT{W|_xE97w(kC~e$}9HUI*NNZRqY@8h-KX zzgFI}m~}&W9M@B-w55Oh7W6MvdL1*yQijehzI_zW`5k5^u7lncixwD~=nV(S&~?x! z!UqQC&|7qJ7OxvYB7;69$)s#7`-p+gZ%~u>K=*6Ks-WNFGEP*qyXUCiE2Xv0tE;0A zs~cNLb9g}uEttRW=#%UULx{y4+ut=VROC6jT?c)6p?$Ow0rxcAe`jGdH3ru=*661M zI@$bAN21?h1S~YnA?J&_`}bd@(~~5O`TEo3)4O+nH(()^shZ)%$K!ZIzjgF4T+GdH zsEWf-*3m5{A*c36uin`2jghsBnuL|`iA%Z$JxHJ%mM(`5Vk=s-UEQMun?mBJ?NkA` z*Awa6lK~NqF&#%hZe#=C57}(5#$O|ieu%i2P942bl>88;{_e*%9Z|Y;&~5J;#|-cp z=v$3A?YzD#^~zq+bV}#1)Hg5N-MavtIXZ#f`uOo!$BP2cZI4d}5pA!nO-|}Tc)Cf2 zsB2RbM>?*Zi#&r)uS)}}L*er2Rg&`&B5%&?9&%4%<_?#T3%b*dm&K)Zo2ks){l(qi zYcZ00CeShPb+D|ONVS{b%VNB5Yw0wn(MGMv@joY)pt zVp$hG=?ccW*PYj(S7p2RWliIqog+odWN@KQJcD+S zFytKllt7208|!|2js9UTQps9i4$xWQtp+&;1Pkd*2NUrLe*-}G))4RmMk^iJELCME zRRsWgT~P1v?vy0K1=ItA1<6YR%G|wscNN3~bWP%&dzMOO>192g-^IplXS;RuFH_&f zUCz(h#dmTq3pz;0-jzEdf&TN8@V=mf4hj3QY^BIRKhQzXHJZ5W%u4jG1S6fpG7q{@ z_j(xV8?svobV6u6($@x8ADV4J?^mqPC-tJ#tpuU2t-xW;uSSt*voS-jK$8iq(H}m9 zZ<|3!*scA?O;_me+*I*6)jrtNv|Hct*gjTIYxMdZ<1dD*j$Qk!jt8&uVtlqH8;(q^ zZS^JNKHqW9&A2<1S(EDk!MzTcU}NIL18M@S@`Qn-?62$j* zOCN&mIRdFjH8YC5IU#OnvvIT{{jhPTf&;z2`BnemXnPu+(Mhl&*-EFUz5wNT z6(t=~bp-1NF(Ln$jW{cZaz3xjvSOyWcwcm#k8b z&QINM48@sS#o<2b&E`Nq2(RmnCP>01fzD1@JL#SfReF#~E8Qnm3O)QR+qoi}IcN27 zM`+i!>8=!TA2 z`DshbtBS7Cg+kDqHrwl)yPCUF_KF_^`qA3fdQ9EjWA!G`k2|o2A8egB<_KCW;LkV~ zIr)2vy_+`EEcuX8AF#QpjAjt!ucrddd} zCIbNcFL&?$&Em=f-6FiW`wLCKd3p`HrL?1iaz4XmGJso0|FXroIbQtKy_h)48RK*b z2fYUqy29bh38sVC$@&CSESCRG6uGbLX*%gVmy1l>9q@7aTt6$-=(uV-Iy>9~_V&5z zzJB*kUleQcbhM>7skp7;YwonN#Q=0X#l{ZD^Ou?lU-V z4XZcB9goK<^D-pN?Ei-8sd&V-N;N{{(f$ zI_HW7;%pA!-4}fwVl5sr=TYde!tOY0Yg^cEsqWiCm3m+Gw%s-g(jVrj<>3N|HsszM zUHCD!p*Nl-`YSZhE8yq=z2epOSZZ(!m9We8^SRN_bIC$C#vwC-m$mw+ohND|7!e4&zi!^d@fBnTT2JRIBy||^btC*!TQ@;iMOVdBuJNuue zzB`Ub*L#ybByH2uEe&Q5X?O?}e1HK{o?1tRMPgG(+d-Jl%A+FU15+FzV^feUfXZN~ z%_4K6L)QSJ;J%LiWqW__>Fw$5X^U>Z$kO+{_uSU$zMEfPzR&r5KIe0?CDAkXa>tk% zble0xf!_L(3825^0^rI!#J&RUmt28hC>B-}0O;cSoGks;3%|ctIp7$B89q)YFqL+` z(3=Ng>8Or*gRYKGI!hm$yXqQBp#~Ct0O(CC7Ott0rng?}E1?*naPev zZu9+$jwKYkKdN3anQbY_HobACwrmE?rPS(y{Xi$S@dMx+%wuGrA1|JzULOw&Z%aDq z2fxY$dZtynHXbLg(sL!S2UOywwJcNj{T&^W80;~e%QB(pW$W++*A7l&@g(Gkg@)J% zALN)0#dU7<}iZuyzAP!lbk#m^1M+UtXJAI{K967qv z5axP)M*>gWLd+V2Yod`M(SYzq6BM8Pxe2?X7>rR3C1vV9XIVACt&(%=OfPM@< zpKC4r#$-#=H_AO6_HR8<;&9lyMrvu>HM_T3+8$V;D+^I%-C(M_rKBbV&aDhYyu9Aa zO#{{>&G}zryeY3KoGK$mJpI3AoJyG59C}atfIQo%K&BgbRh@3-$ zr9;|aoapKpFy=r%SUSWn0rxDYvtSUX?pn|#8c@@&laS&`}vjh%?YV;lR8C zy20o}IrmOr$Dx?nY#0SI+Bh7^=6>dHTTn{NtPvPG?N~tzDcJf5ieVCBraEXf6SE9z z5?*z6i&g?%MSTj*UN1lQIuUo@J{+#C6VOl-st1Xeq`2^zW|a!zk8;-N-2>|DSz+!xU@8= zSF-EwyGtt(0IxYz-3&wT8b@|*#-+Dq0i8R#+I~xtdS)+9sO@FSr3cu=cC}po*j!xL z{O;Ya-rX*4rZqa#evU58+|#DgbB=z%#Qj|86(=@-Wi~FX(1q4AQbHf-Ko&_>HnjvN zc<&>`+a&JjJ%Qk`i%s2qMGM2>(DaqzK6F=^4#W_Za;G}Fd!fa@jyGl1Flm3`R$4YK zqQ?wAaG-IMtdXfyZ8nr5z|Jl^v--DVrw7L1=_2S;n6XFiw6{N9a^`W+0smxofB$xr zFxSd@sk^G`uKG_;QrBd}>eaKldm7pCF4^5<@|&1v@X-RALXcI!H_=%o!@*t%NR&@`36lX#{|xyUw_;wtFd3g{N+ z<&*7G7H59{x%!j!_{65}Q3Z5M*LUx`ydvj_?sZpnKOIzyV!ero+@p@_C94&rqXTuM zW65{XOx3VluIo1{FmWF>p>8Inx0C?-M}R&v&U9})MKi_6n|rWWM|muYWpKZ`rJMLd zM;G(xqNV35;|`os{uPQV^bFfm#p2h5muoEjFd9#IC`MFHy&+{1YdjGQIk=xy2qk)hKZ_{p2CH|u(0J7w-ZI)c=BO1(^7 zUBe16W}dE!F~5D5KtK5%Vsj9I{@!W1(_UYXs9XKp$(8FNx+z22DYVu?gOv2HVsMjkkT zxiEXbh_hJ=p&AIg&+)04Zh9;F6^?Bw$25N8Q*raCN7nY5*>4Y9mP9(A^W_}(L@uBsX!+0~$<=$Y>!-MY2!Qd78!)yTHA(%N~9>Wt3E$Gu6 zy7a`~685(#YAGGDQI`5e8?|&_gk7ove>9ACqCRG1OsmTSb#EfQRb8$H}BpI++duIxanoq`0`rD`;SJbHB03~|sQ)A*xF z^p4mlt3?L=B0_R*8$|dUuRcNxsd1ChmzrZcB`*~oPf}sB&|&5wHGh_MxPMz*e2qAl zGB%aMCDg!X+?CFhTcaN^sYhgLcOhn+w~M5UptDic1QzK-6n8gBvCyg{yW*e&?Ky)g z<7=gPl*4C;g+s%h7%JkzFFae}fWOUXwF!<6s!;{k!qmA->+)+YrPRf}bINLF)Asmw zaXZT78cKJ~UN5(Kbch^m0O$hab^&xdKtH);&+o5)>UK_|kNdgH=MtbZ%ZCI<2kA(X zS5?u$OGm3W*grd3U4icvW!12FE$G+EDrzfmm||^N#UrDVaI+~w7rC1?$I5DNpv#oc zsFA^Ca(%o|10$!tkj2F{ik}x_AT<~N>ME7PmeXdU-xfDt`yTK8EQ0=Zi*#n@@vwUi z^n)aob0O@m-JY6#=C0$?fG+FM<#|Gh5Dm44Jd#U_cZmxIlls061AE_(m#=ts0--C8 zJ)Fok)Z^X?`@^ff3no}O_j8hIx2=tDQ@KTR;l<95C~pT zFxTi=n}61<$cD%@2_0ZZt+zT$N301mw{%6jx;9bS{#ad1bM4kOjM_nR)~UN9TBF;Y z=iA%s!TF@s>%QDipx1A(2>=%-Jua)|{PXwEBiOKZhrUgAbVQ=nWrH7Yltp0aD(HwQ z-YcsuyZ-3rqw5=2O?o9bhlfZv`64%Ju9>daJQBCJxM;&F#LDDO8D8rhbMynILz-Bsz`1e@9Veh{6Za(0<%v*hOef1m9Ldx=jXV(I zL+EYhK$rK|!}Ig=!@aNjJa4dK_h4ZT(!+b6Y3yI~KXdhkX_X$shIF^q(QN?z0dLwu zO;#5>TQ(momX>Y;X$0QQPOp0%2Qv+fG?bpEG`c-ux^!DF*VW}&5~$h}Bgj@kx3cdP zb+_kP&eym1w||HKn9^&`y!LYdT`P^JV^FnH-Q97=+GYJ?)slO(qpZt1x?xQM-E5jE z>nOWUnZ+N||3x$3-h$^rP0fX0)m)fy*ix$AT_y+kR%J5ClZD6oqymQ_A^NRP1|6SB z?4?>cli=vW#umZS>07SW{Xj`PSut}ncTLbn;9OXt$6aL@r!UFnI)k(Z;vJICvoOE9 z6Jr}ufiwibL#x0Wq_55O(h}Vt4zKm~Vc&}1AD-omC)VixKC9+5-zRet+fo{*U=S3n>`Z)^dg8*Iq#Ukw~ zx`P_H>w`)wIe16n7u5sjaDbbj!q~ar7e7ty{P5 z<54tISp4yR5Bq!(I#r-?bV z&A(t}h_lh_e$zyCurI6%N6;uRA4AXA5frpUqZarxF&<}nvus`O99<^>7*8P z5T2KxM?a6{(a&W99THG~)Qzm#60Ihp0NrAR)HQmf40-dTkFJ`vMeh(mhpGE~OV{5` zcKJ+>RM1Vvr3;9zt2(YcsE$$faLJ(sT_=DpI=X1-xH3fDapk7&2SiCl?ij?>nlN?G z=;*qPj;_Pp{dH?-T8DMxPu4=Mt?+Pkn}|)r-aSRGpUW?THte2_X8%?&oggmVG{jZtd_wA(YItv$DY*IUWYWd z8Uo_MwV~G^kU}2{zGM|MIFiK^3ir+i{ekebZvo5nJ_G7rdjpuXo3=1@-!ke`K%Z@E zduDVe$+3x9Iv9E~Li6<(@c3P%+PqQB+(8x2VHsGrTNSp^+ z-%Ubq&z(GZXELF;Z=ZIb8?;j}&d&5j^Y@Kvc=}T-IXaGeVVIr&ItUMm5t81E40Tn>cQ$Ww(9B5}k>bxC|q}tCZ!mEDLX^&FWJtOGa5*qqs z9lcZ5(a~iYRYAAJr|#}9D<#tf(AC|zq|{ni7uox%zZM>^c{ejN^C^P>`aOw*o{n_f zf~9AeyNiaN3&tHVY3atdY~?apdpt|*k`8}(J$INO=Z;P(Hqs7Y?~OMs-lM8lbbf_L zIi3Y#K7(0~1wym)A?!f+p^=h%+8=(g=Wh#mtlIRNnf9Q!QNc_pZYI28==QgrO?9J- zrA?b&OCHweD2eMFg^7DNQ1*JZ4vnOm>PXDbIpG56SdCAfRMdVJjQ2zJBy^y-6tR2r z@|dHipl;8QXTwMATi-?(xo9fr)NVe3x#;Ku=#CsxpbPOhp+#hh6XZ%@5028ExCZTe;u@3|batb-y>-olD(i%!(@|?M^XWM{ zdTC7p<&Y~D7;=Gl%tj4s*t2Tf+rg}T??uS}Mphghh3*S`VSj9^&!mSmps$yr*Muqs z^OmKt_|70Y>(9o@{L?pBUDX&*;NGY;d=jn~HCJ*3#%g^NXoQ!K}3U`iZj zC&>W)R~ko`zS_~%E9dC{4@zkkm!X9ehHQ?`7w8g@ZhZ;TB^jVIQT2uk%kE zMYW(qBIrq^TjMVQbk_Ktv%asR+h!x5F4WM0smHJ;zhwf`E(ot=)FnI*XE z=ai>Au_fo|2T(~qF09hEZl3DsX)ssxJ?Rj(r5?dXmN^f8C~B=%M_v4XEyQM~!Rf90H$ zn(n=))?cKTvl^oy!Php)2&O5%ZSg+|Ry#Ow`G2cWf8&MvGPs#_fa9EA$@oXB>YpTm z5@Y>Xti^RL#+(+a(%d$IdpLn^Z$5@k}wS3xv znP&&6nx6YUj`D9g_34h3JKJqFC${_r&?2b+XqUKWY|qrGe=$ z)$Z%CiT$q$a+}Y_B<{%K8K~b%gb_F&(N0X=b6kP1iRYVZkv3|7pYBl#?LXwb&*4kv z*PSAnAkPs>-{~D#ew2JMm>u05Q9g>?-Y-P5caD-#dOYKq4`U8{a`u43!j213<#Wq% z$7ts3m8lUe4VorrTdwV0g=l3?TC5)^VAF?e_s3V%b$KseUob_U)TE9=lqIA&}y)lb- z<{4RYmY-bm`K+a8BwadP8~|3yZ$bRrDIEBazTVdLG}3b&nnzA7U+uE{xg1*BYQ4&= zKKd8uAZSA%klqplpzGG#5Ea)gEQ(2MG~H^=wMu~BuXF0u)PRf=F?JG^5)-}0$fp?c z+i{nj?o1>2K3pkOA;`I{|59+F0a6se`*sdxASBiRZl7JWSy;45VUTh{+P;3XFthrS zAE>RNAb|v`TD*Tcy!c2E#g2FvDpv3hw|3=ksItXX*lHfYKBl-U)gk>SLNVStY&=-w z0rxM@DZ>tdKR!ts<-XAJ#KNpSD5(6V)2Pm@a=Yj5 z;p*Cu0s|n{TV&$skxdgUE7bPa?T;Z5LX}^Z4eYMN@xyFqn_UqUp;jUyyqB#F+bZh) zAh(+Dj?^zG?78zRh9|g{4sSp%h{I%xh+FZclVJ8_OLP`jQ>wbe8~Bh6KTAG4Si1&d zE2A0XhEqf{MaoP{eMgkXU1@Pn7xq+Y_l*F;qGN2K@OchxroRc5)Z{-&vMwP&>Mn}# zblHUt^){9*fordVUZ3pGX7Y7*<6rg=Zp|cf>SQ|wmmw~6`}F&h_(wB9oXLDiF5t=-w9#uFKmhZ;8X} zQ#cyEcyD(G6&tjTg7>XKYM>CzCed2v{sX>S^y89nJvu4}$>7+|AbSsD;tD%$G4Vu< z!D9u*_OW|5RSXtiQu>OhjinX+Y}$%w@#^c7V+&Z;Yg>FaIXsA8%^CSgmOD_>eDQ&G zg#XRpz(P}K!*xrmG4H>O{6b~s=i;njTI(bc$k53MC>r><%w;wwXROH&!F-`%?IeKta`FVEF zozSTb!6rp)lOnF^$4jN@HkUx6rw+?A)+(pX9MhI*ow%2e{09rM2_oQLq*?Zy{z?8~ zgZqR2?kh(Pf33qVb5;WG;?&@3c|>)wvfbcES+ucfls-V3L={O?>n&*bpG>By&Sr?s?` z)mfLhZB0%{87c3hDCMxWqw(@`tMc(~dOvQyLjKQ$ye0WC&#>iMTRTH(>Lne)X4jVR z3g=H;09m{TBqzsl_v~mSScw!7I;3j+0ypZ~(&>jai?B6F8lrwk46CU|J~Dpljmt zD?fckl@WJAfVC;8|Ldbl`&!p|#6lXITRlE~hxarB&{|$JTIsKU zR4UgzKK^LW;$c0$c3<#r(;VTmf`|Uou%C5e+gb@BLo+Nu`~*87`4c0#-$EJXpRBw6 zo;))LS3A2md{?i-&O;)Wbd5|OJgT<#zOka7+jDKDkd---_?nYU3-ql-GE(JJ=Ubd@pGKg0L+N5wm5V~6^R;mnYu~?HpLBR0Wgv| z24FDtReE)61y^yp>I-rGPTks?PebRb1{qAJl;$7W?r3lQp~zG0cWm@k>whrSwbFG9 z)S-M_*Jk%u-mXNf9tBFC0LRG}Z=4_I2Dc+0#d$Wk1b}Ii2p@xr>IgUCyuIu%fBF^QP)rNkc&}FF zq13OyH@Wqt(JMf$ex_VWupSDvk&I@Z1&i(4N+1?h!V&(~OeiBg^ZKHmwNMx1&TUz^ zI~O1^#+Dq-TZu@5lc3D-BK@vSQnjy{)uMe!oF0y=y2v}x)#AVv^{uG4@Mig0+-lL3 z+0gZbxH&i^kIi62Ej6c~FJvKH+Tv>dmvvHHD^r$DijqpO6uy zK-Xr@lSMg;lcR@UYEebAb%{bIhb(q)j~35&bR25gp934gaL?d?*1)?)0j4{izUm9; ztiUJGXW#ajfc!vUI=<%~ep3;2z1eH%9@tdxI*;yBWYfTY8fTkSV=z<%+)LfYiVpPM z+p-SrmQ5qlSnC$}p%0o{~&|M7zY9*L{| zj9kPq+UEhlUDmX{l6USt=JL|9x|e*15K`}r5}Rw+^8t+-(NywMypEnn0bmAV!fPq> z`PUcAI@!LQcOXBv7~)|uo|2At#b60Bo?~{xJqwpA>Jg`pYO^_$)L6Yi5)!}PZf(So zXYdSA0)=dH+gu;}k49a2?A%sa7|ZIp6rBYCR^hc8@`u50(aKIiZWhJ-qiHU9U$6UA zKJ;e{XsY2_Ao3oBSq@?M{EZUi(^Ehq4kvoMSS*%h%Vf`*wb=3ij9Jdw4Nb0wfQ4>7 z9nT^_^+>*V^q1t~X0Vxc;Id15WKB!iro($B32}p10Pf89;*>c69K2pWEiR=rx{8xQ z{}Hokg`IwOB z+joVcJtq``rEnYq=o($wM+rw-(n*aZG|tAjj0fw3l60eAoyc*_bJ)xEI`(_F0mI?C z2#eb$Dk?3ABjC&-eTsOtu6g0wOOe>4&46MjX3aKu>u*BlX5W0r5H&K;mYzC1!L^VC zde2W6$cTU?oBN7ZtiQF0Fg>yBU17l8QWOO&byg8>GkSh|@+Y znsnH~-;zSJ&(vIHO1=+H3*}HawjfL~qrB>ZALg_KzOkb3r3+O|jqA86br%UNX4kC< z@Wuu8Mdsc_x*^;y?8YUR7#?Y^d2Kvo_0v{$I5T?T==WmH?yy?)C@87x2Us`+t6)fS zx2EwkVjM_wEQ%I;T+Oa7WR`zpv-7-p^6$|3))e|bDY-nFcyPrbZhyJ z4_&A@rLhYGUq3b7Q0z~?sC%x~1EV^>yHLM+ZT{1lUkEl-_9Zds9w`2ON$PlxvBAil zv?v|Z7p{t*FFJ&qUCwLv&7rqr#*_tt)eoFHA&&6^Ywog zx|Ru8x!rx7%xnGANxll>IPM2f6f-rk!d_g-Y4l+GT{Ys5{lCW`eXlJ#VcS<(Z|8QK z-Eij%DB1Mu&&yo_NUgcI+AYu7)Jf;$?q?PS(g%ZIYtz{7n{Gtnm^)}tO-@cOqZL^# z^7ntzG>huUrOl7CK_c5*DzAmG%67KQlci7ryI7&Rq?}vm%p7cw_ zZaa}ag_NzEWX=2u@s;l^Mx}zV%o@|5#r*@Zsc_%(KTO~aLw|J8Lz7t-YVfqLF!E>uDmR3&FFr&A6yuS{gofLUVbF-8~VG_3;_gU%B!HjW#{S zGE2MPz7=#KIx0yQKGN|n{c4}wbEovBSm<=Bc*aPek;ZgHGeYw_!u{@kzm@H82RsK_ z8XUH16oB92`2W}ZxkklA_|w`4Ka>_atEjpjAD9szFF?778P9@hcr99F4Bh(gKi?Ej zF$I3k^Y;2`Kk$EG_s2L}Jnq63?UUfIC+`aU8c=zld4Eg_MZ^0rQ}REQxHvCY@#7=} z3GjqMqk4wyeKP6c#rwou22!3lT8Zfhx(O9u$N11^klIQ&hn=}1ujVvga4oLzP7GAM z((gaW4oDx;qV0S`zRV{LivRTKh_th-EgB)RK&v$Mi)O*cy%%meaIe>fCk38P;=lC^ zz9!wSr|IQ(X-m_~X-^&3pZP#+^y|3h%%3WHt zKLF3~kVl1I`OzCqpPlsu3c|2Mgk5;K?^*bDW8Y?9Y6m-ie?CWq>DXCu3LCxoZkQU{ zXvmIqBPR??)=%(OnLj2}7J~*ZfIxY|&0VKiMP1D4`BY8(OP35GFbhv~Wj7TS^-Uoe z%6J>QO`My!<2!?i*)6*#cd+C*TaiDuZ$)A`o@dKhyy3#;R%v%{A;*jLnXF&ds=5wd zyx~4*^=oLma#8^^^MN;Y10pWu9Fgs1zqW_7%F3j#1zQ!jo4mLMOaNjn+;*kc2YvzU zrAWkD*up4TF#A@Y8cdIu`?U3>>z3h&iZsO~{%5vhzq>a0OwsHOZ$p)6=bne%3hqY| z6DO%Zb=C0Gr_{c*aGj@`Lnl1!LUq5s8nGh6r6RyBB!21(k0^i12Ejs6nYdX}0&1S? zimX50;QG9*F;XtWMZY4WWnp>i725+YHrhTSSDb_&7#z3T`Q}{{Tr3jBue*5c=!RcV zFY&UhFHk6Qo_7Jrc>DjYu(8hZn`-oV@ZUh%9;KpE$W1n=_1qsogxmkVLtCus9*dM;!KXC-1A+3{6TYu*!s2ye{wjF-YzB zA~jXNLrvha?dPG1G$}aL$m6g+3@*yFcz6i5}=-x zy64uIURufscXwY93j*s&u@h&$g8%50l1g66m3}K`OC+ zlbG@pQ#*a#MRpvCvaru$bebq-=t1*pDT^7dR;}n?5eJswHu=O>^~1Rpjx_JqjQ< z&cPAU|6m~4JZMdX#V{w&PR3S7`?Vs?rTY}_dw%%`Qh2p81!~4qYK~56(J^_6#ON2; ziDyQ5J-ln3M{6H3=PNpaJOI`hJzyP){l~JQnuwIv#I~)5GUhDdd|L=aN^Ib;0{`Tw zInUvvZ}d?LI&6CH9;~2?hB(j`FLz$pa1q9jTcdPU8wosfM@MF;ybm)yHa_4J)$wZ? zRfGOkm_l_lqL^~!CgE``Br?w%{89bd)nPllZuc3xZcD>z@WbM&sXnol+j^uiutF@`wS?wGpkP)`afo{2t*vf88R%8Xhpr5^t_@LQsqS| zFiauZwa-TZRv7K{|6u3Ur5HF&LY2TXt)Z#-NAG}FGX;f2gXQ@}mAzY~d-^OjM zLe@eb){bhBS~4v%@lQTN^8f^^h~VSK<*>)imj;3_d={#gu4eZ-{%xJ?U2Jd>NQ@7L z{)D(ZE0Vq?FkK*}%J51S3KanQ!>hTkxc{v}+`Wrs61=?Ex+e`>^EZl)Ka>k4-_4|_m1}% z!1tR|?RWNRwz0(g>(^9mFEN=4L?+VtD1~d*KO;hJqfEEi0hfVXgf3&&Vg;b5)pPVl z$C7xbVEtxCZCndZ-Dgm}Dtw1Y&HZ2RXgDzduup8QMkLkybkUI*!OKmQpwS6&=nZU% zD;;vjp>wjEmhf9+N|^bSEV;Lsj3_f{*21;nzVg=WVY7=aUEeWPmsx1il{XWk!&t3( zE;V7SL1?6pyqSjn;*i54 zWc*}QtqY`Vl5Xn=)l-X(ia{~7?cyL64V%s|u63cwu0%MG5J+fB z!n*18*#SA>32&t5c+cW^nx{hS^1Rtd#KlFm?WGMh)+|EDqG3MbqD@aO-R_cby8Q3l zwehDR%_@UGEyb3OW#^B;V3MWn?Wpa-l8lT^xca3}m)P<6VR4F<@!L>2k8;Lp-Dl>V z=C5A|(9#Mwh_a>yy4^mv7t@iJ+;x7A(Ug2iG4xXBiD>sO&(1|lM_b6r?kvqbOZH`+ z->A*^;={ub)c!xzzUbPI^u1R$5_(iNZd#*`c9Vzowh*(|YGryytA8!A3=I=2BIeZr z;a%?F=DZNdfL8%Jb}4HL@E(hk!u2X8ro0w+6WxNZPC5uH;@EQjpfK8BLd*|(bQdFJ zPA5tnHTRCs-D*5oQ&$8<;qsn0Hg3nD2t!a?%f~UcDEN_Dih3iZ{xWRzC0%e2raVt` zAoncyl=;i8@@Vl75KQS_RJT>*+yDx@L`T`Z;{D`yYhrb1_UQC|M%%vg`jT=`m`WC+ zd~%U$*#cz8Y$$32YmrFdRlbGeU|wB1os+riCl9rYNM$Ua4F)QWR`H@C^KSFUz&>S0 z0@uZzg{%c+ZNCOvrGPK^{ZE5lztW|i_()||T!$xqIeN*hq+teJCF-5NDHSL=!+2gcbk`{5Fy5( z;zVtTbiovw>Zpu&5|dQg*umcTJc8^_0AyQSA)sw;g6TI|^UdJvXYaq7(cU>#y$>uN z*v+>yp$(=N8!q@N3154IHq+{b(}RR0aZkVul)(k7xFVX-rf1Dr>2ReZ1BR6J?&SY2 zk9^0Jfa^XVD2iLZb%iKbQ<57|+vWPq+L3m_T826+uKZ0>1^n=QT#s-xmRfWAU2AYE z@POV^iyUGa3=&E)zhW`nRLDuJYhumJYxTSjPA&4&K4q+j_%}p_YC^zEc7BAska_$c zBLIx5P;Kx_Bc|hh$t#9n)g8s*g-4s=hI%{^S0%a68&K(V-nZK_H&HGd(`AY zMx8!3F?uO9G2NC~kZ)p@h}u;1;1gxy_e^haXN%!|WN?!A8+nMdmT%>&I;Je~Zo5dkq5Qp51CIxL*w$-2nD70VM zU*W1Zg5UxFgwLL{J%m5HPa_aOFr)&LzD#lH)4H$rXFrC@ z6WYpxbNM9_W$P8iD2WrV$JdH7y3Md~o3tg4?fHq6?LBXDL!nQum>KOx#sMpTZ+zw^ zX@mI5jTn|0)(qZypRl9p0th?QA$nLyf(0;umyEzH;$bbMKPAQ|XU}k+m4#3Z1Q}5I z_Z?J}N}s*6<8VSx=6s{X$tUu5h4pdQ`w=9&{%P+e|D=TK2{-4+r6J^x*-jwuXX|d} zT|L+Ry?9GFLhCm4)@5k>`bJkhFFSY=Q@{_`l!Q;K>k@6a@{*(V(u>J##AMRqT?DxJ z^`J+2mYj^7!gwxPh1TLOvWO>7V6_g)ffnN)X;%ktU+FL*w{Z>{Y-U=IIU#n#sR+DJ zU4P#~rq~4L?e7oh69~4FRpbVe-MwSUf2ZB5p!!_uM~N|d7R?Hh$UFbYw$*zpUlC@2 zT7K_V(%ag$GJh$B%gAq8exO#&9@1&vpCNXmEh4`@0R?XvZG6VO3iM1(kzQZ;kP1Xn z1-~@}WftRqx4yeX9`8&~PP7Ut0qyl^gON|A6(9Ywza?=GYgvEjU_a-YFHWbId~s>i z{^D9zbflWPco6LI1`36O?DIxGS(P!r7m4%4w7&)wQ?qA#)E$k5iFkHm67V7(X73S| zh%+boLG||d9J*i!kc8@g{*rGP_X3sZf>F10=aM>!y@D)$WWj|QpLP&}(1vTvgSCuT z4*+&3o!M%U1V8Fa)Z>*+QOGuQ+`arBU2xDtbPAx17o|Z&E|P=o#j7#XrSNj9rP2)r z4nN_NH={a~!N=YYTA#~PC;muIiO~`=3k*2#UV5Z}{K4a%SJiOI4l={yyLiyGa0$al z1_1(8a^tc*QKdO+lL?Vo578XmhZ4&Gg(8>LF> zY~;AjVH?u67ivH*%W7<6If8&UbUZ3mk_>KR) zopQvF4%$=&C~Q(;k<%dzcTRn7y_KTvPp-ynnIrd&pS(}Md;!A#c#;PpXo(H=_gbnz z!3l@!&?+P{JU@$_X2RgHcu(yqMx3Z&TOZ_ z<&>!*VUtfw_%ET0MHy0DCMD5)p8#>f9YRs@ldGP}%Nw64E;GM3`Z89F0UuRjX2+6; z=DQXBm->|B78FLFgc^1JeJj4_`RPJvUkK%l&MYUjV%OG_(oA$lO`~jj4EJqxEbRo_ z5SrQEkm_5Yv3X@m;w-x+aWt0e9`!FU9{Eby`^|~2*c)~)c7j5;Dku0 zLeh%}j*@Q7&>tEe;j%j^zM0EqC0vS9@;cwo@@6kIcp>pVYOAA zdx9)-6jkeZ(qYfA<7YTj{XyZ`Y}I1Q^9di2LNGbRYvCi~gnXoXfuZMXMa5vJ7`&e1 zyMf!<A~Xm*FZa{>v`t7dzNPB{XG(Ao}MQTYIXt5B58O+LTD$;H9dAIrX#5zGs|U zO9S?S!Te7)j_zB5{`KEL#wlU)Dh#xQB7Sp!uVw(7!VD2qud zGn~HY2B2HtH3FHb`j!EyS2!INFoHM@4>CR7Rw=lhdztetrCXa8=m_);ZYKqRC2Cm0HLIUs6KTQ43`mvzgAJROO^&18 zo9fl!0+DnqT3SYmNTK)9RT0}vd2(^xomKWV!#x0~#Q8pe zeldO&qnDuvul_^@;B7j-(RQ%f3Au+e5wbpKaoE;_ibq}1>MsS=S4A>@mDyA|P=^YU z^!iV>%R6$FA;~DXqg}GPd|XUt)VNJ&Db_BlrL|snR!e^ex0*`mi8>|OrKwn&7{KOM zi|Z=x^aT99f&p5ZEDo9p>$U%=PP2A*>v(1>jm+`%{qYppF@-;fM2!>xs)OK{&S^mh ziwC=-y^);36!yO{7B^H*{&KkfkxBlNkC@qc)fZ|sbibe0N9KmvQts1Su6P8#A1RqyYZ+8c%7d- ztbDYblSwZ&<*mY??^n~;#mMMlJu~p07$RqS1~VR2HQ1%RgVT9A#vPmrS7PeI2=vCS z)aCy^w;vUL*|P|&+T!546!Hh(M*Y&OBQ!Lm1n=DuXn8hC2?M|wzehN_bdeqBzroiK zs2{J}{{2Vm4{e~DcRv@zi4TADlGAuXxJn@dJ{$14tygp+xyb+&1}#rXehFN?x=q2< z)AC9_BG}PceroIMS5r~OLMwxSf3f#)f30E2@gH*uq4MO^9j_lR5X&C7qSoyyx(@xQ zVGIDc->?ctQ2fkjqt~@63ZqkDw3YWP2}~g^sL8`LyyedVcIu{%@q;fdILMCyf%n;Lm2rX z=oObjd{MyrS=Y%xZ=Q|j5YSM&TUbJvAkKK{{P^o(_|V@TkV^wrl5ik$R!ANW8fp@RL*48^%6%NO**T z`Q26*$Kir?I4M+d23w<5a>!dsiYm}wyWQCpksgY7uF~T8a0dhfy1r5^O^CUl8 zlj9OIZsVdzmV}!vIUf|b(Y^0M3!r%!WD_)mru2<{vKdx=j+V=z?LcxQ%eQ{p45*Ma z2&C-f*-G0PKyIV?x@vacD-RgTBQZ+CP1%9Lp}y8r7`8U!Kj&$*%|>?L5QrU3eSb#c z!BkZpBz%Ol*}?PGqDmf#W~7EqpJG&U4Z}v0%K8|0wx5mEXoGvHClfds`jL4Hz|>?y zIi0JOTkO&)F76g*`unE%`~R7vUO>GOg_aQGLpDZJEBf`6!7h6 zmSY*Bdh=|)CECZd59;~X7>S+W-yZk|-x=3%YM}vD0iQV-aiE)vtPsm-V#+^DSR?d@ zRNje-CnGcse_Z(Py~L)u4u}XVZJJuS!?v-{kc)m!{Q_Qyzqta@cskOBy2IULp_TQ) z)G7h*=?$Q+5h|`7_wI#0pRKm%qDXpF94yr&1F_?MxNiKwlUvc}TeGHE?SI2C3&Ap0 z5GW}$Q-fw}!tR%p=MCTsQeI}S_jCd#YiB~GWcMoLYK!j|!p_>`t$cf;tIdn7H>Ge+ z#@yTM5N6iT*_4~w$@Fsa7dbgO#bxmw--zE0#%1ywTHnWt`4bt3$M+FK2LE%a;YEx+ zKZ}N%@uljZ>p~U*q1St}Q{&V&27-8Q>cw`y{*t^tG&{kP3|YGr^g(gidg0)C)eYE_ zM)i1W#}A4h?!5Xa(!|$?%EPO6RC;kPw}Xuzi5%Sq^2g}e{lxT>1i;zRJC@xp{SFiwmiy z=m!F^{QqMl8_0nV4C&DpQ35mpo4a?{E&L=>9jHyv=jxNQCnwyaoxRJhKwb!^M*|j#pbe4y^g^?pvN0 zg%B|ccyoK=fM(xfju;W0hxY}S?)G#|;m%6gcXxLW$slXH$0&V%G`sH+M0_F~)qVq1 z{grF6_fWx2RGu2#E-6cgq^BW;t{h$DIipsc@D>X^S3AES`w?-vxapRy{MOZoW-Tk) zmIXriI>-S##uF#uTy@1aR7j5ot8KkDd01Lb5mzdE2i->PVrl!Zea@_5GL9sAKW$+2 zZUx#dDu>InK}hgEGB9%9tO`cjbE#B0a1o43_FlrMzh3%BDw+Y>4(F%PWZho|F))Yi zrQh({z7o{4YkG0i{i3_(-?_wqkcfhg((d(E$B6t2kkx$x;)CMWv8wy_Qrlib-BZ^- zi=L*s=UA__z>rtuZGu2ZzuIq-&|MiN&Y0srS;{rtm0oyGp`Fv9;x?Z7dhy=oBY;X* z^P649zv}eSdm)gy)%%AjoN+(;wfSx_FzFLTeq=~)aP+|E4S`7WAN0uDW_*$rY1dR7 z^an8561@+gQ#n^%%01(tM}icHu~X>zJSWm-1a&qj%2$psJ$1F~G{_Ea0F%SEn;kdX ziv|csm!6sPTDX=MRp(+;09=>|A96gpeDsya!i_O4sodYzWH5o4c2xjQ7Mv<@5A2qr zg}l)gsa`zn?+-3cVx>~ae8Fa@?y44XCqZ)Mi~a56d7sPd03p zb&eejCPSXzqK47;i2cEo>ow6ks!J@?_J8O4Ll44wfFo_c({>2bdxPV#PH45a0m?~E zSqG8R2YMGb!DCZv2I!x0661p$E!m{iy7U2rl44iswT#3u4rCJ-Y?M?zVhXB#LX%hV zK*%I!`abu|PwtesEiRa&j})$15fy9}MrtBG(a{6H--4H005|E6d@N$n{bF~~~ zm0cczF7=!7@wBP<#4EH*hpeLG1~r*RMBvgJQ#sQ(@OjWvbU`6}?Qzy@nJ~6I76H=# zsk?7TuuQ0R;)PzlpHYK3-VZXnyn4IvTaK9@-PYf@ZQ$eVNvNulAXR?vaLP5mn(;as zgU{WFLzwZH@a4*>2H;{}ILXVX%lqS-RxG|Drl$tsY~JR2K45x6bbce)JGSX(@I*v! zor-0@N=J%Dzk=6dt$#Gar{|nnc`=L=DGy&W3UHRkR78tEu@lD0CU7m<d*7-Xi882{sv=avlZ`$&)C4J#2N)H`pMA(R;Q=lAwBxoI3dQ#$-#y^N}?kM<#l zu<7g~4s*Q{ z(z9o-KQ~c?EuY^KqXZ^N3XPFyffk)jhDVG926emQ$g20%cebg7A{Cxwa)o&?gW~UP z;-IpODPDs-_e)N2OMP1|5eB{Kww2{jMp#u~r~(T}|2_mXs<+72n9hoPaH$Ujf~_w~ zD~}Ujk=u-URqibZCLDsHj3{36#ViyiCorTE&XleR0OU72gM=pRunwPT+hGazjCXl} znz@Ld734nl7jLQC($yuBDPZlrH>JVoEn!g*|C zuX!_j2EwUnaBx|18j!1*h&bAkwMVhKm{Ssgy_ib-TwP$^V)?0eoG#(p<0ot-((AX| z?t_n?A>VB&kbUwUCA*t>h-)8G$N~2MFB9bq|K}qjD_)N%&&X?&3o3QXLZ(#NF2mn7 zf2IU7RcvU3uLnf57N-M6gS>Wo#DO|wD%sbG@d*5_pHN#KEUx+3@iH{iVfj|1t#|wK zdbrVYxLF0F)(x*G>#V(B^$^YXG;!e^pUjEG{8p$DgzykyfH1^c++*Bbh=jn(zw<>C zQ7E!Hrhq56D1vaL33FS^f67gs86rMY+K++V-T-NrSzrhsU|HtbzA&skD27#5h!NA2|NyYg5mj7OXBHox8hN(j8`KQnT#Zk>oE^VsC z+RY^RwUNZ03s4a2R&JzYztu3XDQ-oF9-sQ;81rXo;#bMnF^t$0X$)<81p7{!Xo0}uj zLV62VHoo{Xrd6GDuP8Q-qgSw=YjETDP#w>Emm)22BF<#s=DQ(uIX@lItF4WOk#NF| zGB}-n&`-g4kl4&aW2Js21Ozr9scNO5z+|bFAK!cy6Nv4H?d>l~ z+5AZmZPkv9)6o6qyn=#))A+b{RhqG*RmD@|YM)ORF{2!JwxqDP4HxU%MrG`*@_du? zJhxGUn@9xib~!>4P(fND5MiGT!YF60?tVLs=Uo8CwlYG&4;sSvA<3rR~gnZ z)(K264R2?&y2>M*=XR+^%7$_#HyMyW3Lnan+Y#}``=IMWQw6)^IXqNbF0a6Fks|@p z$Sf)#BSa-8z7a#nB0u&yBk(#}3_#m~Ok6~8a(HR_6SJ`dCukKJVW3}RTM7(5mYaNEb@4FsCUPRptJ=B!`EtSAwoGvw;xqZxeJ|k) zdMKka9%nDwj3}3v7j{@^9=a2FMb9tcioSjW3OLE*L_P_;e1s$ZjV2214}O(J1}8iYFy&A)bH#t=@+Sp@bJ0l*)!L_KIua!* zB#xR!=_aB7264JA`P~G}bW4~OGKk&-#;+CHG)Sd?V=oE(XJlvb4y3OJiFb4Ygc-QA zTO!wRQroYAXkKcYaTI>$=Nb!)zYHk?FZLw}W4E{J*WtD6d}ZW>IwBK<&!!UK8dM+EPu#qF1867q5W4~7)GjPQE#Kbns?GmMI zOR9_<{yt@TB57TW8aCr6qPD0O!Dq2{=JP5NzY?XsoHd1^+LFQz004tM7+(4j87o|e z4sSqFZhGCz>uIh_E(f1RZVGJX?c{C_6<%E2sjuIWtj*~Y?9@bJfkf^3QhDHLjdR0u zy9|}K20{;1oMv%SswQF%WAOEp&a*OTyR;w341s@)8{h)yJAm{o19d%8CfBNdHULGrid{%bLF^CKZ(gyA3akJ zJ|T}Uf;&IG**JZbipU9cB)t6DG|UAclDH|~f|JENQIzH3c?{=wB~QB|hA{=*PFrj> z9dXG7Zl~dcYQ(SSr}v8=XBX!!>vY{ONA@90`&-Jw9%MHEfG32EMp#p zBJf%n;XaJ3Wk7hAvGl`j#)()b9U&#GXh86X4+yN(n$!llf0ADo1j4EE1yRhBqe}gV zPAKl)GwmzoReke?(er~j|ejH?ptD=4nsLBVwlPG7qV-+jkaFETf~q}Qns@5`4~ zLOwJo;v=>F6TXH9wcedz=jFA(cj&7q4dv#b+)%b;(D_kikNT{Wdn8OM0y`L$pXKGSJ6}${krf1zn&!JiIpe0RR|kg&Wr&YlG=T8E~Q_{&StklGKdjyC)EDr)n`e=kEQd~%9t+0{ls(%A%;K4wdBcv7rh4kfAHm*zU$hAj+EM)!fsNn)y98ng{vJ^4Ow%hwwVP*P%QtG zEZE%WQ6T)w0lkoC1@*^B;a$D8FVO|XFXC03LN-qT&U6u(gt4q|YEI50OVB^QgjpPq z@P`}3z5Q2h4=XVVUemi9vuQ%E`Y1g0%7Qu7gRs@iL+=|Kg5*)^fx*+00r_HIBrJwQ zt`6qg+Pv1%%Vd4b)a7p=uZ@iZQQa!D^j!B^R!=d#%r-Jw zQn;G{m=7C>@0OwLdg*hq0Mv)$LK{S;veo==V%FxqPX?pnseEr~_n7NB^X6ZVQ@OEo zkE6jW%awxU#S-~DCIWruEaze=1Iy^G{s1dXd>GQOsBH$iaDK8^!;#NHGZz^THh#6*`f z49g|_*)n%N-YR;7+EZJSC&*cS{>)<&uTwp~`qgJ>WOU$apb5jzO**Q=O4<|^fbZ@r z_S7!S$S$lFZs$B&Dm6NMX8eg58}^zFN5u?!J^;E)6tZp)KXlqC`-udv#HGe$d&M#Y z3cXr_zF(n1p#y9uy7i1Aydkg2Q1V3uLF5q(&%`GGt(Q)!XMSLYkJYyw|N9i3K^8bM zA`Ri*&24|{jn0LA)|TM(!Pd3xM>J%l4ouJo@<_Pe^d~0&~ zr`ww%e0>bPuW-Dme#f!=wmV_Vm!)=}rKC2bU42juQL@M%B)x34gaTpf=dh~RU(fUb zfK=iT)@f~+UV@sQUP2-8Ts7WT@klEGH-?~`DvgkBzqcE+Lr!rIKXJ0T*~Dq#Muzt_ zrherOxd-97s@lbV`He+QH@I(m-?x;I8J|LvA_?tVTUB<%vLh>f89lvdTsO-j%rba^ zotlo7JrDhZZAdO?wP)NRk^e?xJ4Qriet3BN5Ym-5{5pU}3;r6@q%}=fu)01gdxLTd z1J0~N=f-=d-~f35wzxRWr$Fm&qm392p+P~Na9BCtemC1{>vS1xSpokAmhiS( z*P8A;vc1kbZWq|?Kiv*}ci`a!jlpD7g)?Uht`Ee-Hj~ydE8JE~cV)q3KAfd3dG?Ac z0pW~Rf4)$nQTE=w^l9>|eP=#y*OVegU*J{ic^d)K>NK9)-imtdrkMxEwNNe?Cv?hf zw{6QG7!LoH^{l=JVjA)rTwiCy?43TeE$pm}E8sop_aOto3oFo-PNscR?FI`rl%w?9 z&z51uiwmmswMtCh&0K!~B9MulAR=a+zX?=jr8RlrXRDnL!Y)#2llV7P8WcxXa)?kH1W#8{!t?cZ(H?CT}1zp>a188V6~3v<1xm6l-7w{L$h@3%w$gKS?uiCJWQ zh&rQf&HIKU-$ouIRC4~oR0O^M6qF-mWzTO2=o{0`x7ejpO(G_QpU0l*i5-j0%W33mWrXSwnryGAK(BZ#A8}}z;_fF2! zgQ5m+y7G5@*ETYb!q%}Z%2GPCFE-wdc%?Md6@CBxbSJG(n$60hxj+~^2eL8G^!``L zz_h_0QB+i9vePyD&%^uh=%#(;`$ho7cQn^Cazc z{3d?W8dqgkvZLV}Xh_pA>Ug=G$6$OD`zRHI?v??_U!ELp>2run*_f9UsC1!a{)5hW zMhZR7P`VL}@cJlYQ*ru9*|Z49V2psDe4W`S|cL{XcEC{1OtkC&s?9jZ>9$OP#itjR`8+`B!l3OlLq|o`^+zkzU^1=r<^qo#F>uK^1JVy9xn-lPdOAQMl*}%rZR!!ZfSNKEy?%b^{{$1s1 zE^M|8bN^VMqiHfM#FmL-4H=bR-am`IqVevq>wmmb@UQ8MC9AziChugDUo!PCX>oG6 ziFP^I`>aKJIGL>d+1Yq@V0Nt|GvqF!E~9#lKba6G%gP1fEjgG!Y)XG_vk<^2ClGrq z2;rx{G7I1uGaE!&t>s%}X60@%xtxBeQ7>$L6IU>N+~U|Zi8eBSwV6wL{%smxJJ~^{ zOWitLlq{grxz_YUhU$~-a#9Dv(v?f`FtOA;NsCojV+cZgSaGg1k~L} zdO?=;iLRd4e<_!8?;0tJ5F_4@9Ciwo5!OjG=jh$3DR?hp^46NK1U-J53aF0>4`4*i z{-~~=?;nN)?COaThTYze-hRvi@ns4gkDZyB8Gwh)lHyC0Q|q{f?qG%FS^!q^?R~mi zm#q8P>NmG`Ha}w4LLiR2!~=kcA1sI7^u@@o+(c2Vw*9YId0!3nU; zdC(X_<#T}HU!cs{46cqpg$Y4RHYf)R5g!cbbqM01M$oT~$TUtCzb+#Z4dYu_w`FS; z;m;pi?g!f0#y#&>5B$QfIf!T#)8LpVW};R5ZkW;_RP4lAAyEOJCL9^D*xhS$;HvgJ z4!E+7+%5^+EcEz$PG5P{t<+4}b27H8Zo*CdhqG>iqIHKeXl31{88WzM7_TXcjR!+M zJ=pdb)C!F%hU$}3Wa2&8?1uNT{1oNP__C>o!C{Z9D}fRQQ3R5lD?=KP4V)T(5yWyT zA!w8^uSB50aZRr6j-w9#vm%oRv$kKp-isw$iQVU92cZhy#d+YV^qKAy`T&7hZwmOu zpz0rXjKi%2u+;_*ireFbvo^IkH;&Fkaz)Mx*0y1oVjESUodnf>l&D_6W*b?W9Zj5@ z_4)YsOe&q!IzfseSC+6i=61DFKT`DsbrM;z!TgAMa7$FbY` zXjl8hDd{`om(iAunsj%B+b{?2`kDakym6cuJhS7Z*q{#1!vw#H<~x9%c!i+egLqMB z6I9E?3hMKq7}^cn;>YkKy}9X@xzBm;a25|D`Kf|QQRr5Rx-jcsh&p!Yk38DdATHAL zI$o1Kii%^`20~4CBMC}H66>~gV3AW`Qy=0>d`>~oE`nx+;wTxXwI`xZbKbaz%B^yr|20Je{2Oupo_o=^uEn*5 zi4aXVrc;>DQk-d#v@P}1$dJK@(K@mR7ry^3;skfO@(y1S6w6i-Vffw5Zr)%fxo zY3$>x;$Oo**>Wa-=-&zMA-`%&bo+U~YubQ~d*q=3=)#oEDjvqWzF%|5lW_~HgM8$B zbMwf`R!@A}3v&H$Ygi=8rZLPN+JOHUm|rH(BMNmvZcSL_?JqY{BSOz(^JA&mQ`mP9 zWnWg?miEc9T;y;=rx9)tZ(pXb2rQc9BD}9Yh%E$jr6+aC=H89TGmgajer>Z!SFT<) z!aim%0C-hno3+@5BlTGKByL^I&bFxy<0=2G@iK1FVbfpp!xT)Im$Y+I%!MSQ@85&) zkBlK!+{}j0-NM}D6Ro39?J8jzkK-sX}SK6@&t%vIB$to(-^t>1AoDV~h|Kdh-a20NmG_n}%gRvepM> zG3M228Lh$3C*F?%l$xHHDmNAC_tS9)$|+d0`C~*k77cu@NGTI(&#(ZNx>e(Pcuylm>Kln}PgKaU{hJF96u63aX39xnri z{{Ow-6M1oR`0Tb$P;>$3l`p?Ys8e_N+as!(J;=xJfwcp#hS1z5hKk4PT6EZlnKqrB z>2X7k0W@i)u2>^KlnNPS02tqqK@BiEjUV`;nTMqC-|cwd2H7iO%=3&pdnby$DJq!c zP1t0QSKhFaQMOUmjYS&SeEVKV-0mYk>I$?;&u-4`tuU$?+qf4Smp)ZI+yYji&EQ2` zBCD!5)XEmu+{weC*`)00fHGq-glfk?0n4PZB|(ojvDYcLfeQtXXA#VQNU%y*FM(LP zH3PNj)l6Rd1-(3GeA66y#s7BC%F}6Fp726qA{|z`)jqOHIUlbPnDA-cuPqJv@0hY>EZ_onV#VLEKzf$R-OBHt zkQxW({?@j1t_NWF$Xoyh4O&262@{Cuy}(bNQ0<70EM?k3#D=Lv%excvFq*c53w@$Q zBjvk39TvZ|c4@j&Nnfw-G^aV+=Apq%EIOZ2-HW}P^@{Zl3Dh{qV&8r7pW>4!?fRjr zNTK`LTRIuxN8g7(fY|`6MFviNU>W3(iyPoB;~%|r9G2L12nQp<1YAsY^rS{{GqI~H z+PRDTjXiirVI>_%I0^eeX_2pJ61+|(IH-KC$#J8x=BT7s7IK>>UmP{dbZ46bz`g$9 zQ@OM`*GlXNqFO)Vn%q6d_*TaO7#@2yqd0O8(zD8pN8CJ|*$xAA|K_Cn$&5%jt<>H9 z=7_Ct^?SQAQo0B==l;K62j$2D$x-`vhLV}eV&3g(h-+#~OE0Zn#hYHTTK*_-YV9B$ zPv+xpAOSheXluVmFo2jmrzclF9Mb#yK{7pxBg@VJR@Q~pN zni`0yPr1m0A2ZvSw=plHl*!*$p{YWC+toxU!L)D@`7>?n17t2 z++=f!rS%dC!9o%{c(HKGtIY8=t4#S>l{$HF0~t^jX+m-eV+b!0Y#bIWpS_P#5gcp@ zx_V$MRJ$~MtGkjC3K{tL8KO4c&=2%ZaD7`LFG)J|B%{joWf+alEWbpyKHRR&ekK(*D* z<4UFy$EBmUtzvt8JFZDAfh7)TC=)_zfWm}TZ>4fd=d)0k+yRY6fbuW)2)5j1ex-~f z+h7X6H!jr{t+!ZAzb}R?`%4Cp0TOx|7Wd1ilh>rBfo>==xL(P(C4QO$>+jXcKdNtI z5jith{uJbo5XFD$Kd-(?fcx2U0T`hNP0`9RMg9jlN=~B{nt|uB^g9S9>pHh)2OLg& zJVgGRG_tE@R{YREVWMOU1NEfA0c50Hv4)8?_V3(1g5FA&wP?SFs6CAx_-eMpht_Br zt8(OLA_gAWUS^B>dcLG@+nxiNCOP$;byKdGGF@Q5m4hZWLw5p{VjlhH-Km@h$)9v| zj1k{}fRDFZbkZr%-Km^g@`;6M!`?igDpe12KJh-@gCfiTZ zrU_%#d#aoL`-|b{GLxU|`Fw@gT{!kQ;H(^Z@`?}K68mFB-V-}ddv9mDt)`oe&!IGbQ0p zBhQsa$xQIBHNv{@p`ggf+KgYSUyKXTde|<}jzviKHVekiMdja-CB-x>A zXFR~FR4AnzUAI#GOhGWgc4JUlOcf7AQ};oEss&pcdtEiC4kPP*-~Z54e35ZVPr^f?O91i7-27~ z>t7ZRdEI)(Ik<<{K^@YI%9RJO=9DAiZ7mIFB!Fbm#h<+9(fJQfj{}$L{d$_hB#=@V z^sOc(na6oM2S;D)+sQMPg4tl!FPQQseJd3ouyqyU%9O3I(w-H^kry}ZUm7^D%(Ou2^4Fx5Je}u^ zXFj7vMx6anhtU?^OILHPKiLvVbLTvKP#sENP zpRAx@*Tj2N;!m{2FPY83Ka~GUlyPCyfdQxO62{A7qibM@e)qni{+k6$6?fcgxd-gp zpIxJg@0`h;&^xx+A8Yku&-J?eSQ3dM76ZRXArP54^%4hTHh^x@P|$t+tc}5CFiYSy zt4V46x0g#f%Cq%}LqmSn51_oPTL~}NOx0>;-!Ff>@LJoS-(YB+6IJDPKT1!AgL!xQvYQrUPr&;HS?^XG8)DI_?%(Fps38p`?mYWt`N9n) zt+$b?QeeS>cf=I04ui8kiqBoBSUeMB>V~t+wL9SwR>ck8c+~iQalXJ!AnY3P*@VK( z;rKB35@bRO{5W_~WmT+y^Gt=p7D57d!pTG)B-_h;EgJ3vfsG zgI$PIhLh>yG__fs=m?j7VU`~XV_2L(<4(K6N41QN_%_%W!i~A(Ha3r|yw!BKFX6>2 zu~M202{ztkj%J;LDHg~YcFGUXSgqJZTMkg4DZg|6TKs26eo%QO^~ck2k|SG=U#Lnt z?+iy&Jl}tDb-Ow`jH1{0vT9S2bvT#-B@}*yNP=snr4* zQAAs%iZM~K;gf2ZLr~h@>5>5ebkm42rZAcTwDX52{JwsW-U&5vk0e-C3rQeLB+ZVH z%B^{5kVf$3DtVD^``o`lYNV`$Zxi_^k^VI5IhjysKWScrI>ok9P|Q!cfV6c@#R*&5 zH`_)lrG2F)UAiNfaX*pWX^>qDze*(WDi!?+%ET`$n8)=ek;p^}Nf+nIrwpT2<^dQQ zrY`jCz^`VKv9g6i`rXE_C;K!o-1BG6(Si-t&z@I_oAb#cOF6WA`uoox#KiY-ZuHhA z)%^Qcp5>+tt1<^lX=?8lF_Od4Ruu(KP2hPn+FgZWy^*c!K)4T`Q7-FVRsddtBa6Pj z!ra5Nx2hM?poY)8(x~MtBDuGWq%VH`@CejOAYnh_&}imiW0mA^_}L%4^iJpF$d{MKIqxvn=Bx$L#Z));Wd|1iHL9{T(D8!lJ1$RD?X* zy`OEN&96Kkb>{}!=RqNUjVluIA2a8^^wITy$vIx{;Z3qwd+1|B_Py|0=$Un4fvhH1 zcR2Kyj0Lc#db_iD43@h*O`Mo>5Ax2cyj~e9@tB#wYM{q}VSUu0+SXoONXt2;b5Bga zMz0g{ceN%my5EZ&LyeL;egW{%`wXswV4NKL-aDF^*35mK%r}+}0`kxLh;?C{=F0^E95DL1>YbT#2Qtg^6O`5 z109ysgR}DY_R^i-*h)fk*XLwLrk~+yEw6?=q}F!So82H)V1fV0w_^wXeS!LW03T$| zAdzDr`5(uz_ii=FF`&4|rX%RF-v;fQ-n>BE+US)*ir|-0{J0q@zB`)^54B5!m{@m+ z>ln_(20$2z=-rg8{gW(VK}e-aWK5@-i4xw+lK@sC6{JQiVKjY?QcNZ`?@=ti0qc53 z1VraySn#B|`FD+(p*%Wlp_6?AO=!Op&(}9aZjS_J+#*+j?-M+9=gnsN1Kf5y7Rcmp zJ4UeUgey!x6QO+-qW@Fx%L5pDm{8uA=e#4c9tVHeo)0A7?EV5%QsV8nMhv=w#Q)?H z^G#fQk~JWuZX@lx8|wabG)1%rJuC z2;sgrBlkY8zJ)W-8_`IuB(GuFAyQPq3l8QX9$8o0PyZ5PHGVHA?VR@ftyI08>E&?f z&&8hDPj@)yUSoZf^tv>xqiY^Lk5YSGZ%?g_Mn^ds^>etFT5@WI<%KSc3Gb5mwwYb} zeeLKB``x_xtrUi{)YSsJ#Q3GBNuX@LcWl+ZcP#AC&&odNh&!{tSM|~fh2Hx@&nn?P zx~|+4H+2mNY=Cta5LK`91yyC)zJO`fxWKv!AoxPtBR=mKTy0~aP ztJT-7<#|E|(0pt+&~wXfH`U`Br(~NrzAZS;cYF5{<<;QBiLgA;rs%>Sa$EITh&#o3 z41CGGH=LhBNP23S!J3*~lxP21OxB`r!j(Nj7_UO_qt1-qhdG0iWjTea>|XM7OvGlYPfO{{E?4Bu9g_Vd-4{`V~iYaRY4K1J)K#5=^Ot9CP~$Jb|m)6tQ&o| zF1L=%#y&GR|-Nu+4Jz!josj#qsXp=$U%Al#>No6tj^el3ooUNLz#$|2QYWJP5Z7Q2M0SVp; z(beuR$*Eg3Q;+rq-F9nVx_pzCQ7n61`zP^K%6MXJZ%{O~bn52G&5M^&$@vtl`}^QW z&JPPOf*XhKtnY?1jx$!bf*md@TpnX-f)dq7!TTye$DwBLiKwUF3X(*1#We5-w?nF& z7S(L>rIDXy$Yaz3cgwlSka0rvG|SqP^>M7hLu*)O0nIf(54nfhC5lZ^z`9(i#oz}p z+yuLA>@s5Kk<599fh1~ANIv)+kFbkAoIR$)DIXXz?r21r^iF|k{E@d_@Fb93JWf*8 z+Lr4XO}@>2;X|`}rW?uk!u!;ReWHg@U(DTnF)7V^$n0FN?)vgWNWVIUR|M)5v79K6 zxAQr5c9}gB^kmF&%bV6tD6xCKRH*+LFy@taFiWMNUvf_-5|_gmQ#2FL*_RPl(D_Xj zqn+Yemuk?IzQ=P{>X4cWOY>!sj(6s(Z|3{MI}O`422LvDL1DT*8;rlPrauI3M(&?t zZP-^dEnLl%4%FEH6d}hG1i$*QmoF~OI1Oovxo_W@LY10*p#(F9z%nXcy4usnb7ccJ zoGsqp*JJUWZcIxg^OZwb%{5=Wn^z|L@P$~4&yY~>z?KKTUR)}C{_nqj*PcPNJhtFd zOUC5Xz*34vPBv@e8yF0O|FNt@S#s=fw~HHl#NGa?iLY1iHsp1qDl{Oe1J7n9f#oq` z4q^-s>=$8^vKTaIeFGQ6T=MnZl_&m+{Hk}Y<=<_OSAIqZ$z;SRDi>bm$JZ@+hIwn7 z{^zTBby6(Vaec0Ab(~*p%O&YF{Cx|0I<|FjYblbszvpy7X5{nzt~ImNwAiL5 z+Ngr>>{pkl0iDn{#mU|qb8oeKiVf%6RbS??%(C&xm-9o!V;Ihp@AnNVLtVGZc-H4# zJOSYIiZB(bmN~_YZcoYIlDq!7OgqE=#%Qr{=$3y!{59KVm;++rzgH9;u@rCvWlavAXGuA&)6&Z0WJ zMxwHfRAM5qT=&4Quta@^SV*`G3-Z64T?Uld&X>snQYuzRwM=9dogFiBO@S$s{ZVGN zl7Sv8yqg^OeHi=o2S+yG5qli(m>V*e36d@z5P4LEIt1~i5>P6=L$uu<{Rf3=Ax#w$nl6;OJ;LQm37 zW3IKB_t{W}XS^1*=Exs1KM>MLPHb{YMUndHyW|c_{q1J0}NK&j$%(sLV7!j*gmBu9aB8`*v-iXSM!W!+W>RI z0T}c|2H0i}UY54NI}jP z-|pPUCiKJ#IO3{?yfa}*y<6-LRmIlgS}5l#^K|A#%WSa_iD7tqnMdNELosXuAGT2L z|F-39ooeqwu@_n(a~<~*bcls?oWz*{mqI)VTih_d-(H3Pvp5I@{hgEQnIQD+{{kD^ z*Y4RHlnp3;a*Byx4L>bM=Ax3n3_tD2JRuI}`GeI<7lLsRF-E;*LxHLQi(^lPb+Zqz z0yi`-Aw(EADW!-@Bf0+VUaKNOWx>O-Tgnt{r4 zG7=Ij0?F91@3vmmCLFL{?zcx}*OHH+f5A~c;WBRVvtpCW3fDy}5+K>dequ>aFW3QK zA*?5w^8OlV2Quzo4TJ!!JpNmAPV>7mWv>+9$}uTS)f)e^W^W0E(eQiLs(;Jq9N5Dm zGT)Jvs+3YmH$ZAakDuu;pkMSXx{o^~5b>KKK8y@o)8+7su+1Y;UBR6)Cf7=vAeKV@kg8=moeyk5!#Xwv^k-to$w)1W- z864dxg$9&&mHvYES<+odE6UxUVDZ(gX6QWK4P^(rCf72hny1gA`4cCL&ZyrYMIe=$ zL4i#zFzF^y&(=k!)EAourWT7u96&a^ZuX;^+zx(I8mpFxS#Yd|ejT_MJ^dx8;^FR4 ziwT8iUIoisXXiUxx;!QxAnFKvL@#b~@hU5CTgt7lA%v^E;P_xBV0r9C_PfQ!Z3w-L z_xJ9NefZZ}LO*>LHPTRyJsV9G3{;r8Gax~H8*U}19v1^8s0gGu_97XKH~F(ES2eVa z^^EI#A`>L0dkSWC|L~=QQPz#j=aIF;#7%cSzw8e2BsP_f+3=#fQWh)PE><-2aM05Q zy{XDy6Ze~Zgp{0Z2*#kosQkKL?{F+iN^u+)rHnNn_gvYIr>Vp!e5A43xwW$c(^I;u zy&aa`)6xd7!j;JW@(A+%8!~!r>l{34{KeARWnNruMsQwgd<|UcyZ`Btw<*9Xwx>o* z9=yL#`26aC-!R9rdsCoDbU|V3>4+c*I@()owPxzH0T#OQ{afkK#VtKfgQQDIN-lmp zKa@e#6+GCc@@2kuH>x(!YZ4RPqQigZ2{rYFT&TL5@3#K5q;n8Ty3P^Sbed3`BZj>8 zPMk29Rv0()m#QJPtKZJu!_yT|NAcE+ST%ipEAN_%!(8t0IcOVyfrY=70ZQNeWs>q( z`qwM0&^7@T;B<;+5C61_U0+vY1wROCqHu!tg-?cG{W{B%a3VP;lTed-x`?m{b0TeD zW(+PGGUQZA4NdbLbqc5Reql zn30f@q?#ZKzYqnR%>`;#LA4**u2_yNvIUZ}E6OvpG6T)(**xj&{Lt@Y@?ZuC=X5btdQ!nZnlN^gkKU@N;jKa-wV{TlpW2+j2Pj1U_v=C0s<`JTprXlRu zG;4yULeqX_D*#y`<-J6=V#LBbQpfe{n0C<5Lk4Jl<1%r~VY`vR6$61%KWP~s&Vq&v zes<`IBu532B4MA2~E7s6C_hK&6wi#Bq8LcmDb~f|J`&V$?&)3 z9%X0hJ^nXMMgx(`in*v!PEdsOuFvt7?hB}`9(-l0_MX$|*wmkDOlZE>lN^dC<*rks zeL8n?RC!&4FUEKhX3@_1wU3PC@!nv*t)GCeFe_JfGi46a2! zirYx`uE;4A+dlepB?GrLa`29q{#j&V4`1#^O#XbKM>u`Dae^6@2^m{hhWl&m-((tx>jWXaeru2)X=O09}3ys_wA6;3l1{lJGYHj54>g=**dLLba1nDxL zO(#VZ9%%OzH9QMIMmt|r45JFN)^zLjn+MSJjM$3bvIHLFoF&-un{Z}n>pr$9ld}W1 zDHE9+EA&9G(5Y&?vg(zPMg|oi0hn?f@p|Vd7Jmt&aS`g-2z;#pgUw6^A(bf!Ckmf- zpQ9cGh5m4xy#=SjKuE5bI4*G%g^C}-OwPwKU)~fzulCPRQWOtGHw=c*DccIc*POp> zIIydHxMzTAJ52V1p)W(P!P4vpLwN9NB5_L=4GH> z5K>@t(tZicW~{lylYY)RUG=XK`}EINL+E{F&x=L%XEnLc4eUp}b~I7GOq$ZjVNpTY zrdLAQm5YnaIp5kgUxo|-4sYubS7Ts#y>MhX`0aZhm!aO5<6FVoGSWx~Z^Gpc?n`sp zJqvgJ!*sPBGk=J7V}8ud++t(tIG1C4&LG*<1~Ixl-sC)-8iRU}m#10$1jFxX{l>J$ z!^%YMWV!$VFkvDIIEfMiQIufaC^|&{O`gfM?g~f;<%+Y2tHYJ@n*Qn49YSO?>^QP^ zcu{+xheNkId~AMN5%1o0rD2iU`F_K7me7!Q{!er78vHl*x$5OE|B!lu-B*5;d=`Jj zj>F-O))bkyZjU{91#Ox9Pz4&-lV47BwG%zT(qN3)Ww#ej2NWY>prO~Z&{C7ndzdct zftbwkhRJ^7rtVH29K2DR@d@%cSmu5U!I_^g(-U&)t4Tf#Zhk*|t9#2*4PJ!f`0zK# zZTDJKA9DvAo(J)i5Abs%n5k|ugCXw#4nMs0p&*G*RXAn{1-nUQYhZ}i-xqOmXF@h_ zY`hy9k@i+5#8G3vp2~QAq zW;mD(%*`)^>~-_&YvhIbzfm{kdV$Z9`6F5L>QQf{K#l(evtG5y(}z+Ez6y*Sy`9Nk zhXOj+CL^uQg};v>J`vn zO4SZj$Q5!8mD%poWcXei4zoz%kjx0S7I^i}e}pVh{IY4;s$GARHFfhd;jIXc&`GiF zN1Uvi{rtuk+Ejh{mOyG-m7_xd%oDa@yD3iNA>PO102AdKS_)a3kC(6U?V|qm>&}0V z^eM`{6uYAWLZ1l=G)WVSQS+~-imJa%bue`*WAO3vgq-FCss9ui{zF_$7>Bg5n>>*f z&S>3F{^ELCzUXD6(i1G+vWp&_*wNUYJN@Tx+>>m7&b4?GQ?RR(vUy4MBbmLsa5#Dlrk8A3oL7u&K{ef^IKDgD z>MKphHEZ3}x8VMPvb4<$oxqUP9A~$Ssb?liOo3Wx>6&&PxIOshN6;RV;*!_ZxCd$A z!~|>1O=A_Z?3_z-^-&~f=Dj@GEUu*E-T>7{hYh@E~Kd7@Zb|x1{9`4wvyLF&Ai`kr?wy{YG^ce?sPz~ zTKN}rvL_(@_}Na*r=Gsnu8&B)v~`HM&P+cYq);u=a-(LhLbnAxf;q!x$vN3XxIs{R zwwtbN{XgIO<~&RNjWt)%qi{rkK9P>p35tG zen*EmRDQS4S(N4@Ypn*4NWjY|NrrGn;A6yuIk_xqU1a~WuiYDp{eO}jS2iJe7r(cD z|Nb34UE8@zSjq{aZtV#O-wdbZ zP0Z!TCTj~e=ATE89)G{nl&R|97mvPLnW2k)^H2EhFG_92kEgXq49jU5!u&5iUtpE2 z{tNb-Jrq_OoUpdBjF0o?kv-^&goL-bwiTM~+T6p{MWbFrN@1%^A`pw_Sy-=S= z$%ev3BZaSoppN$|lmJ6lq!h(4l+oZ=0>4-YO~EN8dPkRQ!abV8C)YB>_Hq3Cb^b{) zj%t#Bi3MkJX+DZ`IX|CJZ}8jOSt-GaG|qXgB_*jP75*5rVH_UD9Dh=l&Eg_!U#0|g z{xjge)lPMy13f73+FO=VAh)V~NU^gR{O0BChDWGYB4_WHu553C(tW~##FsP*_b-$4 zFP0`Z+}5nrZ2_$&LDBC8H!13$#fX|P2R5XzYRz~p&7jM@gev5ib-KEHc@2`uJjr+H zaBtUm6m7e#gcx<@#spD&DWe`UDB)h zY1*pmpQYJ_<>|O@42^l1GwzSNf&@Od(bcrYvZcj{*}KGrur~d!CQu=1|FZihD0l4S z@6Wq;vt!pk^kREvO2X^;IhPDg_(wiF#WWxVqbd=4ypy-m+s;%5P9Hm>s(j96gD@i;QV+=S$h>4_XnPZVD&KoyDiURlJbK`@!&fUFC0|A=)&OC`yl?glL$_{ zQ#5AsDuvq5O4~?zK)&!7jc~ElA4%194I7Jno8@auZ8#P3gzBnpMhwycGn=~N7<{zh z2cZ4HnWo==i~HyIh_Miw4l%3}@{^2aEu`noqrUx$if}=eJTJ<-DLu`7aw0vY!#H`lN6p%!bfGNFh@Z?xzQ&W}nHa zvJ%%qxX2)wUQ9+oFW$UbL1>LjPLLI09egE}Obe+nX=!sCzp*zH;c%Z4I>O*%MN-JI zii+HaVrbV$d>4={rFu?=@UKK%v#GC*lpxGzCHRJ4g~>%MPXy<2u|eoKhOd;sJW`Z9 zFS}_&{)rXijFn=rp<5sh`NK8aM*9GVQ*gtp!X`RA1scmC5@EyfDY;Kp-Zqb?C;pb0 z#r7DiUJb-ZPw{a}HAHl_g?U--&hMF{SWH~k!x-1Pa_`a_m*+vR{2(96YFnk8qlFr0 zR;3Qf52a{v%ex4-X4SOjRK}x;(5Iv}(`YL$psBY5%Yzo`*2(ukB<9Rtabors!>)+c z!OqUXuff6Zo%cQ@``L$>ETT7A0tHnRJz8)#Gtv!@{`^QzoWC-fs(x5nr063p_8 zwx!v+&cBXe<7s==z88y&^f$V%bN+(kpCb_${^3GT_rCMW53V(Z(K*KW?hV==?S+de zxA@3L9BkfS_o=B+GTq+(c=jOa+6q*B|%sap%+b^ksXQaB8saolV0rV zE!NRi-Fzh3F7iK5(Vn#~T=F1Ay6=36<_V{X_7VPup8wK+arl!cHMuD9Cm9krt^Hgb zMDoGje(&wmwmVz9b!&2DK4N|D=f7ucE!I7Y`R?o8Ti?5>y+~x8*SkNTNw&aGYg08f zn~hLmKtSmhWThGAZ^LUH{Msaxp~jfuhz=lD+Fo=*@i!zhng&gW4D#WGffnbYyY=E- zE7Iu~udarR-_vcon&(Vsx*7QDE2<-IHmHn70pah}r*;H2^v{2lAs?hoSSRp?CO>Ct z<~4KDC`y~O0w6b%HMieEP=-an%F;dBQ(?=lf2;{t2iN}v&3UQc+G!pCQF8g-B>qMUw!o+{NYg4qpuf%SSfgeIxNG^L1ehnSZ|RO#c$Nc*5lR zaM>Gf8x(NfaK7b=e*5?nb|p8xmOH4`4^lI8592F;6dx-pwKH{^YmmXm3`Qn?1Xq;@ z)ZISU0v}rnq5xXcrjAQQYi;}f4r^R%&dZ0}Jp{hMPk!`16Le7v90b&dc5-Z)f`#vr zkZ5RyQniM4@$1)G(R}Tp@TF{}3{CyqwVJ>m3*M#QryI<{kTjo;?A+_90{LpSd&9wc z&)JHBfmU~R@`9oRzdWOhp*d$@L5gcb$?!H-?dQhckbK^ZDiX*bn%|69i$CKSypJ#-cGL7YvmzOvGRc}0eri6V;Kp+K7bnGnqI!d8KY)Y`@s&dVVw*+)Ju z4DTng{y^z3&5uiEz%H{|Nh$raH0{@8!D7H^DunnWlvN*F2KRQwzQ&E*^d2$n0T`!0 zccqFPr6KheIt2%&&CEO;Ga+iI6M9WB!#+!FS1S{K;<3=<3e_sk0^&qT%`h=6x5S&I zD$@vSd8`KzxJLhb7zTJu1FH5qfFlIGis%^#!1vXV<#_ZaVs>aN}pHk3{;P{GIriTDF;r^~k(e$b+fzXpomk zfZrvDSp5>UW&d&HMrirU=iYbzx2=D->Lf%a@V8)!W^_OBomTv~%KluA=yY+7{(g(7 z{b@V1ywl!i<`xr%Q;o`e0)Pqiq$$iT9BMGdN>=6n25ZMlVB_ExX6H4&kT4IvL9cio z$&$mkEXzKYkMxJ6orf^dZEJD>-Ayr5Q;nJTm^ZOup2QQ1tujN*>Z)j#75l;X7xQ74 zo}@rIH9rVn++TboZOEqZOOM;jhCFJWK~9AwAk(TPc2VE$SmCz;xNoR)Yzy{j#@8V| z)dz{3vQp5HyphW7g6xy!Ua!XE-#TRU_*mU64Gu1Hip3ed0E2@D>{u82P&e!}?YM^9 z^NN(OEqC`}boIM>rUwKs+D6jAZA;K`ktw-V&@$TU;DnJRP=gFZY668s9=JwVIdNV| z$Cm!ap59Lr^dz5sjo2so^gD=ic&~8oDlk}xm3m{llK2+E@3XI0Kc!%j-%)gkd%YO6@ZA%@$Yn_!I zRD4|bX>xP6h~oIHH!bv|!(#v`(s%r<$3`x`=1pUpq1S|30&^6YG(MxD zJ8=#voa3Sw^aWf;Un7#5rJ>h$bXmV;(B*g>f&SMIzD1x5OLsYMUb}YfTbHgu?NeAo+D-3tm1Q z&acDK2WPx<`I(a|vDg7->;m<|;Ca-BT%#rP^NoY&8_(}bvA0tvJK6XZ(BB%pRt~ym zf_||Gy62NRI&*FfdJ-w``>3qTV$l5pbK&SttgY`i*aF}jvT$_y;0DW}`!RNOk=`nr z%sCr5m4P(lZghhJ$)s|2qH5_-$9Ye9<8W|1{v|&H@^c5%>mCnT{d9g}rT*0Hisxkl^Yg*jF3>G;?%8dMwKkq_ z+{NR=rY1c6jmGKX=PPqN+0tR`9+CulNDSR2&^2q&70eIgoJvSd_V=yY%mXp(q@?h3 zy8s;__w-oH36bVr+tFppaY>=t=m)UMcqlHX(zyqunqgY$ZPZWT4;cR}|;~kqgTd4Ow41Y{G zo+=fnn2BNZ9y<8|OK0_+i*18nDb2Y%PyVenHkjQyKiIg1!28PirlkiZ~-Dfp$7ogh#@N)UcJ^52~s-xHK)5)qWW>%ca zxs=;nHUHHj?taW1ortF8GYZDhi}_Xzos(e##)D`en=Wp`(Hlg1T*=nWg3%cBCj)`6 z+b|#~wXf3^4ARh@MWk~((xT7Smj21BM6zu5Z(sfHGaDq;l%ja>o!hs6QXFo&j6#tB z{o03iw}`z{_=$D*O8X06h?q_h=CYU)Pc9n0@47!QfdTvY*@0i*%}(Uu?%HDbV~WGi zCJsE>PGXT5@7!FQSVTT5%1}IvB_@^z5pV~t2cTY@V18;P);in>U>yE!W-s z*5%E1ch3nsS2;<Z8Wj(Gb5@?5hmE-`eMI5)aFJ66-F-t!3&ZdW)LfXk3`V^vrH zKOZkgHv?})f#eFYJ>%lyaG%uCnRK~ERgkx?=~eS(PB-WQ^d{uEjN2DJyNdF!q5y^P^!U>cpW$xK(cOCL?*Ki(ptA(%?d_w+|M(&FRYP))(4%N{-En_> zix$u(=p=Gvysl+B*TN$nZD#F$Vg`)o^B@paC@3Wkq{Cy^*J5q0gBZ~R=0t4TQ#M71 zqaPd$HnpBS$)I~nKcepe9LMsPhd2q@ONcY)r7+y)rWbDtJK((DyRzb;l)aq-KJ%_@ z(r+yDuAt$TeTR-fcc`2zDegdA;T-b2;`5Q7Z70d^H0S4Hp-n~r2McL8W({b7#ZiWk8j$ej;{(%fSRNTRF7o}wG-MgpVyEkl> zKCbk1_kEh4ha6t*8SPS~G&xUr;J~pMp7m-o`TRr2q()>va}<4ey;uM9ek)>oRkUV7@gy6~w-kj|iYcG5G7UU1oIzKj+enGJzV7&Uc8P6(4Inj9%plgWHnfpsm{iCU|vGImRGjkbq zfV|b%IOy4GV$g9>$8&T#9tO{yXb_x&R(lgTE-Z>3({Fgn>hkUKRoQNj__*F(uwjGsU$2LI``DknmQKDuigIE z^-CWHT)dA>7Ut*!?a+m2`vZbpMP1uE7yUJOcX+p=6J(bjs`oxIFydW58(w_(U@N=1 znFWfo11Kny4<{1qm`e}FP%|#Jwp9=KKzmzkTob`Hz`uCD@y}ke_!8&@G=Ywoe3wyR z=Fh;HcgT?d`oYTP%73Fd0$rut<;WCpsPf}N#up6wd@jzHQ4F10at|4A`+Jj-BcWhaN|&JVUA zJ(-xB_s<`VrAx7QW$Z32T`54JaxRYBm-nU%nV?@XGGj)seOTvmDwghc|KQ4~;pou| z1Ug6GD?lHK!pI@ddUPDU>kk8%2hU=Qi*IuP>*(Py03XD3y)d%|jc2_`pJ^Khmg9qq z%&IYThs1K(Ok?9rXMG8DGHpg3w@ngYD3lN?&Z!!Bqym($d|oyBFGD3_E4^#?45V+sfy&oq#kGYdvZ3osG3lGs-;cHbM84 zK!?cad8A%g>gu|Ihf^pug8~w!qgM)KUnKEC$t zzM=WaFXGQc<=kAD-0$g+q|?MXE+56%OVe}iTAYZ*E%4um<-pythLm;1iO(E6N0<2K zMz{7W0!6cAR>P)kcOnH@E=;cB@uA#`eaCD^29F3zZi|I-&SX_q`iIgQ7 zI-2U}0NoY;*=+({<=g;0>Vu`-yf7*`x&k^hhO)1rkS8|~&Uv+o2`@>bi80I`BX*uH zqbHv9Var_-v*XtLYwY!qL6L(FyTRaT3EzipP1fOHIukU2k$Ug!jH|Zhkm- zFgLd&3-2xU7LHzS=%JO@AiR72)!xeELZL9Z^J+B$txo!?W3C&dKi*zlUS3Ww&ja?$ z*aS?V3r8=4KCC9cT4&2J=rtT2?KnERm_Q427l^Bl&XU&I444DK1|m6`WT0i%>4Hj} zW9htd6FLvszw8gfDgGXs9hkVQq4We853S`h8UTchgPYb9ZCeB+wZXQ6wiuh3*^TXbNXU#^ zG9Co$3_3m$j?SJ=Ca=A0T0-D`s>{$rFNNQ*`TXTuRdY4;y!+xt+0Z@n3NF69oeiDA z_gV_2v|K@bY;GQ?KOa*-C(L&@3KHXKnCw$Zcrtkztj*!l8hOA7lkHRdVoM@nT~!G zmR_V>VO|QrDn~isB`D{CF#fse1x@*1$dqyVwpqvIum>n?N^g zKeyX8cq8W}*#F$z^wOJc>e$r^h|1xc0%>^a&|M7HjXNgg=p}V$QihqM%5!RE=wf=x1D++~Z8ho^-i&Jxeq8ov*r`4g>O{pL5T$Lw=_lr6AoR5=T7;o}HUA-ZWqA?&wxRanO8L z82S$8eT3zBrOeTL8jW zK$3^=Hr|jbIA{6&=*k?ObySXSksT*Oa`Y0I*G!MI)pm4}T2{n8avSWdgp z&Wirk4oBAo=#X6J-90gI-v`Laz1^8?!C?&7a}$ZRkz58N^-OMSeQ_eeC$!AWF6Lvg z@f@*fEUkG96RpUDtqD4wyt`OAUux>YV4p$vz{8g|-!|iG0AqLBPP1a6sf$x9*_D-S zr3#*q!z6vKw|8jLsxii#kgH^^kRW7RCQ0&lNh(N z>In3*sJrlUOGj5i?y8?3KSyWK70Ed=emAg>2XZ-u_}o{+0}3(T$N>3l1PTm@ZrbP07ogZ|73*?#b>x^6jL zJ~u!=2hauS%sI>bhk*eN3We}8%0Pz-1(%k~NAn3%I2>-v<52Vr{m^?#%f;FTV~bgD z2*1Fy@{dGQ)2>$Ec?jtBaC3NnWAoNYFN5BRy)Ce7l#8PBuF|jfn$jTeRQDSP9wixX z-Jz7mfw;F=Ov|`aj#u5iDA)daqwuP?Z~oPM9}Z-h*I)ws{po!+^y-cdQQdZ$pjW7_ z&c2LulR4(Dp2H;1!sL<@3fscZkE^4bwIgJf#W`x6_$sHfqtT-fMcjkB-#jGMh|Teg z9Iaf~IgI`^fldo*6`D*u>1eD_XgV6*(n7hB2N_~rbF5>Yp8s_yGqSjrfZ7HZr#H8VK0*@Ej<>e9 zwYCnft?jPGNPmXE*909?fD8=X;|1GTK4;MJuDP+Rd5)$A0&xX)&C_vl3Si4}u(S~M zp?8tq=NX6oc-~lB1M;p(l!@}d$onFF=?%Zj_bL*3<- zNedzD{x0u#ef%JonNFM?A5VnWQ{Ir4$rToHT*Er$p8{7#{UFf_!cC38^N=h- z2jLpvUc!|8VH53S^CB0Xme)MmpRuN!X?a#vTq|+IG9*t37>_I3HPO%7vqslG`1t>1`Vx@RA+9;qmf*ye?3C30>XK zqAsXO8&r?o@|pl0FLY3ec|knyP?as6-?7RbO|Sz&bQZ$47WjEm3bzZeF-s@VSps#$ z2rO>BDyE=HUy_k*+Yddy9qZ63(l8+M&^4p$%>)JD{v&EfIT(16XoT}_Lq`U7U z1vVALBND)g^)g3yJi!qy5AtW%P%I9xM+d_1TxSjv<3t(SW@3CX-5TO%K?`!57Gf$jz9Z=0Imn%^FzPI}b_ zT`lVz3a$2b7gCOm?!MkhM}a{1`GR|q-hQwibS(_^q+h#($>m67AsBQyA!XnW#cj!Z z$x_ZSv(AQY57afZoUjYdYYyDS0=jj~{a8lb1?ad}f-XtQkzvwh(1AC^?p+IXz}?_S z3LTe|0pb$n)cmTWlimw+QBca#(a{v0I5a&Dwl2$bbCUfJNSL|oWZ8pUE0SbxOV_#A zDKieqVU{(kq@<7*S+546w>wy+>Gjv9(Q7sdyvpxksU);6Bg z;OMa}6GW!FrlU7uoKDr7iS&oWb;;2m!qNGkBEnVFH37PUxzO9*-W|6-+1)o((0$&a z$f}0_WPOivWyW;3Q`_(B+4f&9MEbXbc6M~Mijt5{g4=0lOP3rSmR{HLxt0bC&?RxK zRo%8GTl(>Jbkn>^V(6|^aYl-}G-Q{v+8J?(Zv?uPfVZgKUV4cMIfJh1{g>F3b&*xyq=a4v%HDM#q; zKr|eEqJiuXenc_32aB`VAqK7}v3*g?CE9Ws>^et|%#23GATywl3FS^S{y>19TWV@- znxP*Q*4GQhFA|``(S?~S(oE3b2+S4ELqPkv%a4u#E~mTRQrGJ3~LdjxKFpQ6);p*mS=cVOLUN zNMcfsE*Yd=5#tTam2f*>^b(Fbg`84gKuBG{+(xa6^{Pw#myH* zjpFAl*53|0aLF_TuIMPClPttxe1hi`YA5o497fXA}Pz!?@lNB0ZiN zJdY6E6StR`$1#fk0O;Pu*o=28F<9?iicNq&5i#?~0@Dn7aewJj7q||2NZ{E7-D^6! zCFPm`J!Dd@aGnai?Tuu8?)l!{$9gE0(j7sEo}$?b-IDIO+i7>CBk5G!z1+Vl99@?| z58z-7swRV1nRo|b;KT_#|E(yTD@k$6&|R_g@@2YvU!~j|7{2Qcw=bppbmYJ4d~&L8)P~dDV;Hu(gL6HR1=Y-j zm~k0&T-Hf(;zV@S&ub^e9nYw{n0#L;Rvx`131dgTU{P*)t?)cc>! zgzFQRgFQTH=GoQTcYc2R`bTx;1IjyBf%&!DT*aLVJeP!&xCN(4y7T5JcA<|#QB@zG zAgYl>9o=~3Lhg7Y zivd5>(V1^_HovVwDh|h3^IHb}sAuUs`%;51f^&I=O?7nj$i?4X|L}94e|r5_CrXaq zf?9IdE4$-{Y`wbf*%96D4I7KUW-G-L_W4&yTO8TOckVXUv&y3dvO(IuWcox?YWQb1!ND z>+At_n;Q4R(G|U~qoW;K4c6Mxg^HoLA6&iu`=?j0->EYXTZ96?yM5&&Wq-Qyc)$Hp z5p;C-E{tBdtQ^h;1xIk?k{^hRbXYeqnn;|zdoPEiS1n}LhkY=Zfbtm9Pp^*@a$AdS zvDVhs*tn-iqXK&l>6}V z4)xBt-K)rlySRr`-_&w%Kk7qmPz=(M?vE_M+x?Tskn3M|J0JH)3QkD30Np&Ar4n?@ zTp38M701nL+nOb*uYh@MzIMZ-wehv=$gtovre6s+j-MX+ti-F%TR6)T`cqW zosC2aj?hpaNXJy&y^Z^PU+;VfIRuWOzBGm=PXA;-Nbi}{ha>3?yKX&2R|0X$xLw}2 zrdD~DNVhI6gX>s2Iv^LAtBzh-NP=@`$|xs^0X)C02Ks3h?`A;?Y?441mM)kF=rZ3z zpIZ4ibupb8uerwd)6YKo@cQpQXD85EEd=_vKB}{mWKR;HgL4D)bC>lhp!-ImX9@I_ zuOtU)EgS|-nh&#a_r0wLTRAV1QS`+`D>k%*+gf8W^HyANFT45zBl5{$d4F*esp z*uQM-qKF)&)B)RFs`89ys_aOh(_IE#E0aeHVPsMNmba#M1d9S_5adkWBmc?$CNhe})SVxYNj zA*pbfy2VL2`Y{Eaah4oi{j?+w$(a_<2;+ecA(2ZrTL$8i0Zx_iQ>HufErsPc>VDLg z{#mm@))n5(gm*>NE5h1N(Eo6O@@;G?=t%=~Gyq}oMxE4cB2F1bUg1o z+;i$?x6qw2)bCe2t^kF(B? zvzn&Ez^Rcc99^z9<2O*39GyW|NH;-uSAy;fjko2E!1l~Qo|e~0dCbu7hD`b}nkUSG zv5Gpm_D<7zoSQU=GnkAIv_Cg>Vf0Qv!E!pzY?$+URN1!iY7n`1Z-P6B^RGE0BzXba7 zcK@=?z7a{sgC#@%sj{+W|Ky+pZ&hO0X0dIp7DobsR}!!?pHNRyCdk1N7>N@ z>3-@M^ec*^>$(-_c#uhVuP4F}aGLrotb7qs`h%I>@wMH95OGczf?b0_pk7Q*RUEx_ zaByn}j!u+gE)LT94gE^&XqgL*bpZ5^u4dqGIyy@R-Sc*;gNEyK2pDKn9f+&Dyl@_r7Ye;? z?s%w~Um0{`Ip(IiQB~%7mX_22I@Zz2!!wi5QQ`UG&<;2^+kPFvCbd*( zs6UcTVgJg8&*_dEQFou7%p&H#)1S7*ANTh=>~Zq*Do3qVWXh04$lcn_)p8Ru*^;Fb z&&Pa@t{k4mey*ytdqp}8wYX^RZzywgX>wkX5=ZOkew{d%D@Jki+my-qktEHiJJ?iD z7nNper@TX=z`G3kwM!ob>{U2d96joGv7<|J5a>zlQYVF&7^BzWOY-o1aOqZ+&=(09AKGTto|g2d0Xm8lMZjsnw~EZ-ZfsJ+Tk-YbNJ znfc=HM#Hqd7a?_D8R&3ym*MCD-RH&@^!Yukq!&_dS1R2X8HRu3c;uetz4WpT5B)x? zlsLP}xlOTiWeMFf_f@d2&e4x6=nZBQP7YZXc;sAmh>k$zD-h7>V|MgP&`(K1ZdF9+of>c&&* z0YDG<%oA4y=&6bPV#twCET#aKkVdV9y0@mSYjcZ>J9HjX76s?12z-b}@3|7_H@doZ zJa0Qt`c)&|n>tqLS7z}~mkiwjM}PA$gWm!!((9qu^hDD1!mIfWhgJr9?{jzXb!7il zZ|^q#>sHdeeaMd4*IigZr=h!UciC3@`j)qQ<}pX_Nyn{H;$#Us|73=4y=_fcI$r-* zj$VAS77xjd2kvntlV$);1s$kcDCqwhp)M@FhNCm?pMPIV=drtb$+L7`O_v?LHt2$M zIQqB#dbz5w7nQfN*~!tP7tW2{l68-?uzn6hcb5ses4w*-oS1e!h+)I3Dq-z4Kqn8! z&DJ`1kGQO0# zg(U^b9U_eAe%(9ob8Sc7poGkx?sElSrGfWyhH~89C=N9Y;Qe$CIz`=W+sn(gYM={A z6Lfm0YUtdmVdiVHqkj!?e4eeF zfjI6zmD5$#vHRh3bk)+SFv7EMU1`5EYHVVm4mVTS#qy*yM|Zx6MlW257In_(&q86k zW-&)k$)r25WS+l!=Ff@zJx2&ll?K(MoqBrCHbE;~WYUAB^&F5lKu6!vgf@q9J20Pm z9P$GAZj_Geekn!`Wa=SY;%o+y80W(~ciFMXZ?U}QM!x&X(0sv(oh**M`Sj~CRD6x= z_E*cVTuy&59$HQBEcb8Q?fc974b?Mc)EPQnR-jw@xnk+k#hf2sM`ssT_n%suxxm|O zcn@X zbqZyI&V{5}0=Cj0yr|=+q{P)6xvOH*|6hjg(x_ctIv1%irHppT)ae5m^wZ^};{!T# zBS@&5OKRv5e=oOm$F3a+=mkJdBH+X4&R#Ga{oJ{UpXo3^!MXu@>TcpJ*3H9-b-aJPWaR+^x9YX!rtOkWp*?mmgY3?{+Gp_+T-{r&4Gm*$U7h-U$ z7?>36=^(q^9oO-QI^c1a(+%`(JL&#m=U9KwW7E+6vYi`!AYON(|SxAeoJ>l<(}>%>5Isc3Fl`O*i;>2=unn&+v&jsofggKjD#Dsk*Vfp_&LhM z7@!xBI@7T!IIjVFCFuHU-)k2Q;D=qo6n%Z+<%FMO?Cy-mo&LDXhOMb6I0=>qiA4Em8fx^;G(Xt)k>^g5;VGm~Ck6ZQw~?N95R9Dt{) zXQZN~$4gaVb^ikFeAw`C7oni1=-EXQw5MDTKK}UKu*>%eK(};u1{s>^YP#XjJaF_b z3dU(B==8&_Da_AT@C19O&pWWrHs!5FI5!D$mW~~&1-CZTyZuGC?gZdLv^a&}bp4A! z+)sZDNg)2%-?G}X>XegZ*vt7sYna4g4a;LK0RcL$T8=L7L$8>jqp2sT9`{^#h>OSs zB+waW^wy59N*%hx#<{pN6>}GkP9JjEu!gcJGg0(Fv~o&CZ6f$+KTz#qo^fZl}}tPhaml1U#9 zfn8+Ar3|XLJ2d|85I0$<`bvA(7A>SIo-!1y?c86 zZKU_D5C2*EnT^%Sq~GPYJ2v*W$1q#>JCpv6bl-A3j^O*G-C1J+yD7FhC#V12OQlTx|cgT9#+8l)E1YO&~j5`=x=UhEzmZbtWd~Z z((Ar?&Al*$?Wq}z*B5&F`g}+ea%SuR}t- zV+`hhqSmH0wX%$w;x2}Ca#GxJ2ECy|GU)PL9IdvI6VZGKi=O0UI3afUU_*IUJI~6I zIPA_J$>^4ynt%%!>M?1zrUYI$DG{h-|O$$Zt;)d?U>!E*Xhy*bi07P zWxr?j3pi?RQs&&MfRt+JD(SMN|6hULK%krPb@T#pX^LLlyT(m`&NoVg#~WFJARXW0 z=QVV>-b9ejW%D=4pwmzB6(C`5g8r#y$BAW_L#w}c=R%~ve-)2=Hm3~(_tj+FwcG<^ z-v~IJ<(%UZYMAWlSzYZZK0dtv9AyC(bndu6^|Bv#NSA8Xd98 zjviQxMqk{uAtD#_+aHCaXJ-Sh>ZH4TDT8%frlisHVPe+Ngrhr(pl2k|F(=P3)j2Gt zcf2!lW2$4Wnz@jRI?CA3%tuht-M5be8@$OvB+@_Z@=xLHnmb<~W^ydH3(IvQ5N2jj?p$gJZV+Si? zD~o_O#1rz0I&pp;&j)4D&1pI@&Xfz%Kh5kodGJo1F!O2~Rt4xr-|)LpbN5;u7Az@Q zF2FlnZ1=BRd4ziz>h|>u;r1sk7oNbn|55mCc)?{sddheA(X%HmpO7`-A+7KrSCvjP zI`_~3-61ZM)o|*FAm9j|)A0 z{ljjY*3iH0ob18EIp*r~SV~9SecOI-A3d!edqBw(yKA<&&j) zdP;Nc#u{61oYAbt@eDd;tQ(GwPubBSfZpuEzalh7N$>Bx4oRV5&4*IdymQSLN`;to z?p-MzS$l{kqRb4;9hdHceB0^BwmszE?ukUOMK!(P3haaZj-swX%YPdUGknab=R8Ilt6p=^Dzhg8%jouaCB*oZpz%=_H~wt zx=r^dgGnU_XVz}#j(BdM)6Zo`H$hL9K(|Sb&X^0(@zK#w($$($+@vfkzNyRU3g{;o z^zu^VRY4yf9TJ#uq_ha^qb#0G9SU# zwf>ci9|QIM0y->~ZMgIk@b=|}79<+DR{MKahg6e@eXVto z6oIh2*mi|QRH9&gg@`4Jg7{j+xBtxN%wf2_+zz6mCbG{=XQl&R6ZQMyx%b?2&k>+| zAfG$B=x;8<@ATHg{4J1&dOCX79koXtCBvP07Q53NOvr_!3oD1&_nyIj4!Q==J;E7z zkNRt9+TvsAjJSWb8I|-7pocFyy51S|m+yb@Wv8uysQK{0L%^a4z+ms%ms-D~5^*NU z$JIkFjPr@e>F1{-1NDCl`WxFp54C}Q?=Y1)=r|;Jgfel^_Tf_3dB}h#7Z&ndL)fXB z&BjqQYPV!YgOFw1L09Ax6X~w28c```jXL5LMhwK`=cV?> zH%ea?%{LAw8#+_2X1}VlbUfS;M`xftN5|K;)~Hb1uQ2GKob*m+9p5d-{p(bpHwzvZdpCCC~$HgHP|Gg8t&YANX>ZL1$Z^ zd+@N!0@kXhFAhIgc}x0e_{w})t+o$vd>y`4zk)TN-G!kh{8G5R&0qkEu} zbFZfQ4`A3A9lfi0R{0gI7xv$Y?khqV(^qQoi7b3OizC&a&OdUa@B_ObcwFTLK+$OQ2t$qqC(;p3b^b zMOYQ}fC@S{Q+Cr~>BH^&zx-YXU9kT00}s9&Fxr}@KYaLbB%!6LKGm-}ddi)cANKce%MBAlG%ZKPSD5+gVH4#bQBJ01 zcF~`q5aR2>F6+@|0fg5M8`WxK6D4087`Ia^c&lr7d?LO#r6+BxSgJIRkB=**qLof+ zq2nCr#(3S<0!tf{8%3=|<9UQI^_Yyg=jWcIE0MU)Yv`S_WhBONh5qjyogr5popgQ| zz%%CpbS9lZr(cayl?mC;Az|taeAx1$>imXZc8fIEbI=`IxVSt=XVS0F)x{Ut)CKC& z$nYZQ9_|`}{(eN=p$nlnT&cp0E@}3CA9%>acAb-9A3WF!Mpeq8t}i{98yG;@ILo|1 zI^!*#aklcP4wwI-f0By;eN367d!VO=w(m(^juMcc-_2we9^IqH$8V}M7qZz*a|#WX zkn9H1@9fctz4@ps2wC?U_%;PQcmHhJl^U#iw`9?O(eqej5tsw(JUtQ37)c#HulO9cz|A!^e zMYv6JbWFq@>&((2Q3O)Rc8GHUx}s~(j)Q#mrC;vos;?91AN=_(&3%;Tna-Er|K1Cr z8(;ds2X^Rg**YQc1Zl9Yv#Z7A%=+B@_Y5Sy6bzY?m~KymIfmV%kJi_ZJ`eZ4CeX?1 zt*$%G3^t~x-bjIeqT6~?GXvE3-$|X_jS%%!%Wi%?9SR~S#~WHEnr`|&q#5tr#oQdM zSMI#!U09;IYn5EeDfs3bqvA9Ce7s8kTNwrx)K6}4k7NmSez+xq0^x{I&OB^~a%8Z+_EI;J)ZZbhe+&8UFDyG~sifok{8nQ485nXUlP;&NaLW`+ z0(<@?G~WUAR2%5`+Caa@px-;ssJAK$n44$&vxR*VZ@@ZmE@8$V-+dG9Ha`aRwJNO- z+;jUOb5dL>8Z_xkD4I0tdo)a+%x!E!8+ZoiCna68^;m^sa~l<1Gm;qmn0u4C&0x$q z$DAb|mHnymaI0;Cu9kAYXz4fH(N)ne(G>lR^B@WLvxuB+MeV!HxgcAr@yafa19vNp zy8^gR!i*TYUt8E+9P${>Z^No8Uq`?F^Y?x57f(kzY3sco_!eU>u3H~?@TC_ZT?gwk zXqNit0BveR290XzX*6hQjS@EkbbKKLHA6D#Zn1X{Qv7KCQ=j`L^tn&XUlH_sZJ@J4 zk1#o}E$?pQ*08!Pw0?AF=Bvc_5+>mYweQ|BgI0(epTp#%70l8p;y@?u)YG~uyBqOw z2wcO&PZ~wtaqBy4v;!4|q{z`7O%Ly3VZH9CoQsgVPesVvYO6Im{hPDgbVu(HL-%m&vdc?d)qX<$H&njc61+0@3}@x=O+QV%(@hMm()(VXqt|lEYuZU z6$;9{(4;OThK`c2!yow7h!b#TpP+oqduENKR^H05)sg;fkpMGNXz0N{8Xn%UtSBd) zAw7leq0!*ZR{j)*E1@nY$)p5JBh%WzyzagOiQ;zN#I@cy9 zv86NUOuC4{L19O8db?#uhcv63;r4ZOAPvw}L+9t$Uqj9elP(8k1nE+LRR#6ie6)+L z29=~f`@v6Q+ZySUqz`z)?v*dcbXl-*l!ScvaBJ&J4<5W{_~A?V8y3qS0BQTI6m*mg z+%lafiQygg@==SkXrRz^B0rGGKN>Yds@T!zKljbYU;GC2MUXxp5}-pK=s{o5nRN5f zP^KBAzbE&OGnsv}ZAjtH)Z^LO=Yzmp0=+9fQPauK-L*r(`dJ|kNm?Cq^yBkPxL7rZ zeQHQNaySepNda9H-{&X)Qtt8$%_~oEBS4oNJ?sxU6c(IYq-Y!I9_V~MZd7@jLJ=oXL9=U*&-Yb%j>K2xp(axrSRRuX8m6iFOK z&5(~~;pgA@{uiL{f8+5tKP4Rf9)q6h1bXP+ssMcy)_pjXS;9SJb&VD>O*7QKdu11E zYK70^FAq*3Oh9V!f?=H-I>R0eS^{%HyP?%LLAs43S5_6^H#A37nb5A0P3nK@VCe0%kGjg!-AK%b9Uus@#rc@?H<{bKrX|jXho!gANhs zj66P&uysNDbu?~)I(AGK#}ynsKp%4S0DdKQ>vkAY*4Wvu+8rII_N_0G-9P+tIbbJk z{dW%@?C2{p<)4C9V>vC#CH7uS(?A*~( z0gj&8*HMZrD1+XN$4!e7KVL(9>aDHu-W4Ik-o!lJ2(zQ#D&`Ql+Y5x5Z?#>6|DHAf zp}U$habq3*E*#z8(9c1)fcnUY6?I96(rKjQ&U;ErTvQ zy65M({F|J=Ryg{l?NxE}DLANt&Jy7xSvp_aLATv~TE@uy%h?f^gCv&N=0>cSb6961 z1#NR>3soTp1_ly*(7ClTZ&O`4tIJ+bB+!2`GWdgget9&X`sNqE@r56M_!FP_@Q=Un zjW5E|{}Jf@3h4JngS(m3QV4?ypkBys^PX7-da0I;Q`YNoNJ;>(0QzoxVpL4f#gjB} zPXm44uV!K&>Pb;^WnwiQ=ZP5*-5ONRRY#Xc?%v^V={M8SLAqskAe?Vviq5%P7S79c zwkg#C=-^wHX8FoF2b~|-(>*H}WB2QsrTceG@36*p`>JJOH4k`h$;MZmF{h6#L+JT~ z`)i9&-ZT7N8Z;RrJ6lBgF#01uSzB8>+R7(fC;&dobubXr6ipzhoX}Qv0+Ab_#_c?}Pw)Wg&Ao>UCY|3Vtnv zuOSKmwua(|LDLn5*zeS+_j#+3jW_7uh>M_`0GtmAa#b2+93yqc)Uitq%;y@$-Z;LB z|1mEgwu@yopUo_4+?4>nF`x@ahtg3yisf)tw@=WW9i3#+oedpQ%EywU6VrOQL({}= zU%jcVs)G0cDWzIWB-Y;EFmr3V`r&w)!DrL#x;>-H%-0=NjoL{n% zQL8-no?(pG=hhZSb;rT!gchIW6ZaE?k>N@jk@!+RGB6nFi%eS*0@90 z0)q~4TdOvBH-nE3_Fa;vrki&VT{qvkdQcr51?a)nz7^z9zS)J^vxT8dHHe3$%uoT@ zt|n`=Sj)vX--%ih1@vRFlA~>w9;KnXz+2$1Lv}~d13-Jc9xwv%_Qq=i;XPQqbDNXm zd^x}ETMM#tp*Xp$46Zi=^ma#2+qj1t8=(UFXxE zbiIt_^aM!Xnm1PF?hhyKua!;9O7g1ODEJAMp3wmQ{;3@-=g~1SF*tZMZxVX4n06QR zFSw5WOLlaM#A96!)`>VT!3=Psv z5;9+dd)aIgcRN}^0mv#|kSvHmM=(y89o^r~IfKyMz*Y^PAPW{nJYr#3eDy-WU74Tz zMB=Wn4ISsrb95-o+s~Y+YUuLVJ;3)tLP)F}0U;oZeN#FU-@0oNZAV)IVEL|^tW#?PZF<^hPb|1X2m8tkS zupxHNmfGx>u=PmfsFlYcerIQO#sqq**xFyx9G%AQKfxUR@i$+>&_Ob1*>sKa=)8+b zr<-NLz^yJ5YzO&o1y9G_)VE9=8`lmgxXv%T_Twn;YMJ;NZO{%ik8{X~nuT~xi(Z(c zd-L+NE+ooD#oQtkRB9#=&~+!~pdzG__^LR%l;h4YjzI}*~L6vCjA;H7wK_6<>Rhi z=iPgj9sMad|03z}Pl@!H7mxGzbRDN{IM469VfOrtMtnqVBFi$|*oX-ZZFu>>K%{S= zZwHH8_W%kdJx` zdGtr-(RN`do2j*I6GMOQHL+BfC`?7W-Wg0=L8{(fD01`$LQZdSa^fVwpbJN5%q6vL zKEJIwOSwtZrMTs20bn0PFXiHG&2L?e4pnN$2HC{L^G$N}b9EEwt}O9R$cwvsq^tQd z;M_;OI4*T`=2#M~wMV02JSl{W^|fm;bss6TdxjzHlTs{Q^iyWei84a$iy*yikNTza zbb5F`e19PEq-Z2vLhSO3B;ZOfmTk6m%Up{L4h|+-cobQ&s%Y#^`7*SSP8l5oYx`sm8&h|S=3R}7;2Ams2=U27-8R}Ka{Yb# zr8@NutgGG9Z)@QG77}Obw|tAZ3kl-o;vw|Kf7E`Hf^qG#!er@=4QORUmq6zvSdjqh z4c)~KbmimVV;^D@~@gySm5%^ITYxhzUd;1kbI#N&9HoS^&0=5C`}Hvza_ zjgME1s90F@55qwM_CRqHL-&|=JAh|wcPaV@+{HBNvT75GsS9tvS<7o-&(gIhNOu_b zZh<*lyO^bu7<8tb9R1AD#r{+|8O}q#r-P$2$QMdXcum)$J?F0_Z-4#u#NC;8f5}~x zrC!Rup?Bfw8VsG*uqh7q1z7rs_!pX={0HzW2*|BzMiP#0uF#s={VgOYZQV~mYcFYi z-LzLCg9E2>I+_pZNP>m1t7SHX_6q-!FgqBrn@EI>XE0gMVo&N2058-EHE4rAIRxHa z*6}HvRJ-1JJRV;H=$>Sl|Tv+8kZ)zOA2(Z(Ls6SzY*)=3Ihr+0t*A zqyLrQm3kvWcyR`}1iBP>#|e2AX6^!XCR{RfY4IH4cXf@;o0!u_2E8ZxI`=1bw+cEq z2kF;r>GJ7YLXIUFiCTujbM&WQ|LGUM@P&_k0-`_|#_qCxbtavFSVS1^i9rC4Kykm) z%5#{kueI(c*2>6hf3cDu9>|vv{K9^=)5u`o3hez-(1RfTZ=d=c^r^qidV5i>tRRU) z?C&fc7qSpYFVwaVTgOP0xr4cQD5P2T)SKSfRs4J-J_&~(k52&j+-7b(*QEt@B}2xy z!5T}yrPK1-W=T(0;q9exj4==3P#s;W^ToriwkWD3r3=%|iMs#J<(HFJ>EjiGJg1u_ zh!2aBap#!CrCeG&|_M=q5Gk3jBE`j=H`tgm^aG9iIuu-@8lEEPSl=lMG`OEf({|`2!&*d zNH)DF8FZXo)Q@=-50_Kzu`R!%OYlwwn^&OB=3%8;w1P%)AzN6oLXUO}g(c(jwZcgy zWT%g*Ps37cZ!^cRH%c}^PMmi%^q8iJo-AB<%UzU=?P8gYwzKAryJNSn)RaYWbiuir z4n>&XBuD>Ux~-lgfzIFjBs`tP8&tWv+<5`_c1QQ7=*WV_&50_DX#T>`oj}LXT@^Vx zbS;$Q@HO;XimA( zX@agNF~LoaB!*i_!_qCRYbU|EIg{A3P0J1%fPEFDzmPyzdB3`X762x;+g{OPy4|Xw zx=g%T8QmQ!)X~@(Eu<>(>~?S}4$)K{_D%PSU5roUikePyj@>Ls1??M^m>qR&U6J^k zM;9o@j6+T!tee%6>0pl#KB@%gN{%}ps-3^)21nf)bd~7KEj@Zc_CHkA;c>m~ zpK4}c=7*m*ZSzq!dl;p}oqM&3!sm@8jO9;`X_THlJ_hYyGm`spT ze6y;lBX^=(Hgv|E74sr+sGNO1BVq|3mmV1Ls~P%@bMzmGHs)6*-LYh{(=1J+i}$Dr z(KpnRP(bfdP`^ljZ|?=nnRA>Z2L0Mt_g7~_mw4|8I>hVf-#A}K_Z(fEYVaLSj-BVv z9FQ?GH$3;`D3bUV$E)sG z-O}zx4vmq!v{)xUSJ&29SX&F|ftW+z!vUXlHJo&jyEUje?&^*u`hK$<{ddk)G4DOQ zgY)!x{}wT&?>Jx=l85dj*}q}nJz-4?yRW^E>g5c(&^2Msr|jtp{K79fIl3Bq=fV5P z*MBN9?K=AnU@;2j5vCa^%6FIA+x zmqYESD-+ewY*lBQS0n5!+0a$cuiw+vE~}9LPBzuk-CXy_oa@d3Wi~F)&FR7iiI<_C z-yhgATtiPhn^`&iN-~+$XPyx9iLD(pST^QYXUY{cXh>Q~{n=pOk!D&+FA@Gq?v4+$ zF9}O_*oA(uFti;MDxm?>$sKRMqWLJEJ+w^|8_-)+-Wgf1ruDVe-Q!A8Z({v?G9E8A z&}z*#>f;le&RN90idIZ00cV^K;M+XB-Pi-F7U!yNtHBKwUw z2J3!~PN;{miIsjwPLXAnIn|ZFs-opJ%`x+Phs&BxWtVz@YsZFki=RbnIlA zwcJ9n_EFz63$ycbB(e@4PttZV+p?|%deBF9*Fk^tohbh}D_)_r%x=)@@QS}Dn41^P z{djiok=tUeR+_3|-Q2AWiS6f@uJ55Pdu~c^U<6NC;}g)jFrFVTX`o&Sw(EQsQaP6; zVeVhX-QUu0qNBh20@uX7hUi*Y;k=hQr_?!?7__%%xt??1b79H67izb28FW}ae~mKI zs(SEN5UzrL{j2Hf5*r_WrQTInXH(ax`14b?^8lkRTugTKr#j(2GB=Qzr$x3WiPNx4 zzVSStcrxo~wuL0QEfbMB4B>bB`iS{?G-W9p#9)$|AK6zx-NYdpnYU>)crdga{$VCN zfBx!gL-7Md`(gGZ)3zH!*;S_&&)nHW!HMI&()>m?H-WB+n6sxFtwtPw2Oh;IH`AcK z%Vw-?Sot2J??pWFa9y@^m2h=qs=uR?t@EhOwxOt8)flly7(*(5R`8adFrF z*a)q%#m)#9D$mS}xMSNF_Xp-Ix+jJg@g1DQ+E)zLOg^zgciUKuL}q9MD;gxPSXR_B ztQW0TT`35!QjTOBo4O>CQ$Ele^eauDSlr19oE%5jjpAkW5^E`c(G;{mPE6TKX$T zS3s8pLr-h2ty}mI-z?nmtXW#`TR9_!VP`KFMJ2pjkL2h$WOEnfipO3X?{SL?s^xe) z1iJZ9!tU1u{8eJ_{-~?V={&!b<&?MH^WZscSd~}WM%IQ=`gIX(Xx~CnI09r5xvph{ zU|EYi(G3WfG#1xpUEhi(*cJGbyWj}BXP@ApmGLSCFkUQFR@qD%}5#F+*^GwIfzTip3AjR`n%3pFey!^Zell>H*o&XXozEed~E*IKQ>E zHHYd^9I=``N+d>L?YJ5olE>#LvV(_}zG*V{sA2C%BB}MG^`pMX*bbmyHvl$C3bWf? zt>(g5iq%|RSm;kpAM|GsVed~ZY;SKL7s2u#lkPa!1 zet!4P>xG;TVd(ya-XYmn#Pj9VrEY!s;qz{N?!lt8u|*3k&b?-s_edepEa(lM6DMwo*!nD~0C zfBm#oPp7A*50)zjse{owpdHX>`ln69qROpU$}_#nZ^SnZ-7S(qf^{1@ZCt@`nuo}C zbxtM$I$j!Y%1w+HkFjEouX+i2>R;21r|nrlVKis}?@T(x)px!DQ8E#Ex*z6hGevcF zeqPt_m7{k8{T&yB>(u)ga_n==Ir3)F3o$YGb#zE^bOM|YhMjruku2T&Dv>U%oKEp9 z^qT*wrAwRBJBHp>&M$)g{K4F}emD1^j%OCT8~U?6>7GZ$*8y~DNDMqDh><|FlNc@< zMmY~(C(%n|eQ=+$WCxE%0G?sFp!d0~A6e;jI66R2x(5qoo{m%fI8K-B>C}uB#g9!p zy_8BVzbm=Fu<*6%gY8r%gTE4XFzKeTp9SvSPD-0Yg<@2Io;IU;4UJhU$$`8IRkKPwSJNt!SVn8;6phQ}Y>@-!#Jcak^<;b!&M&Ira(*!B8@yLvnV&22Qf1VX zI4R`rZ|UbfdTu!AkYj0@ea^YJWM;u_>G5B1bZIxaC;{j0O;cy=AYBXr7;hPIdVl?M zbe0IbOLv)e$=Jr2#uYC)q)L6`U7p zT8j!tg^VTS$n2T6UBMdLv8J~+P$y2Wj$>0gx}6j1j^R{KHa0f)mNv#WotS;@=)AS> z$3OY#$3OGyZ+sD!?g8#z{2Hd>rWSE$&aW14hi+gn4xpYqgbt4r20Q z8Vk#H{1#MEMeokZdZ`5+pi=bst$w3a`@#21_)X1y|y?;_$C*ZU$Z9j?Ai zQFj{VulHd9e^d;$S5i-!Wy1V)CT*Fs`;KUcp-3tjp*3Q z60@J`hHk>i&pIVnbOQYh^${5Q7rsh+ct85SZ+{UxemIaJK=)ZbR|D=iFAuo$$<-il zF6g~dKD(6kN^E)EGrHp;xQOoQr4ZJ-t!sT|r@)yXsK%t(QQTBW@0bZ=(<{R~G-i=h)F+!MHL{7nZJy zeOH{#g_mhY>%_;$6h*+c?dNylc{ zQA1B|OiVWSHZggh*re6KW6a4ly*>_OC%WC`wWP)p!MNXi0wV!?Z{h$3dQf5Y^ z&F7M2%++>^S5I-jv5qdH?i##S>-Hm5b##D^$E!=qhI2HK-jjAzKj+mmmvrV_^>lAI z{~sEX zXs`tAq&l)==E;(~#!~xJ%c(;Pk-OE@ObJu*1Jm42AtkPqq{zHcX;rJosY2bbs+pnf zqTc4{N&P842Ik?|iMqS=rdoD0N-_Q*cBhr}t}K#gP)Z7u^X4uxW)3yIG9KSFVwR@G z0KOTkBk?t$#{hWu$v8l7+p2b+FvI<-ZVS*A&Q;Wi(*{~bri`46++A*HQ@ zSY`8%y{lc+1?4!y&l3Yj_r*yFJgQe2vmluXo3qcEWddJ>hF5AmP$5zle0qMc^|G1>3 zlRw*sg+dmrPmDJ-yfm~{c6`H1BMOK3+awmzv(&qxL11qrH6vi?dIL-8_l_qgHw~TQ z55>9oE;X1j;AX)dXGdnj9V;? zi-b7AxQMr5t-gbmljzQ*D}#6b@QLHiGU)%+iMlxCAe?_XthcvX2Ayx3dL#Ygm9hcg{LG7S6l&xX1ZeSfW(4pJmBBJkZZFX3?9xyPZZ`=+r=Yn+XYH_P_Fy$Tnoat61 z09UHYD3;Ew-eAz5Yov6VGTeJz-aU|MQi1I!`AF#_7Uh(1q3p2EX)FcibSBok)BO zNE?gKR#tWlinr$@b={Q~L@Hc9qzlmL<~{{Z`^S(yJ5}m({a1I~H1(u?&|fmKV7`D! z`I$1Q4CTrgwM@x?KA#zC#0&$Z%hHp&wRG&(kU`-6uu?BkLv&4lQ=_)=2w5>mebX_- z&rO_!9W7ZbHBM@iIV8kxY)p=yR7;gj+z@LBw1JM3wyFM}iflm&~ugjqx)i zQE3#%`VaNc!Pr>prH!x2`P0<)B7W;IRViiqE9U5Yv0f@wE0tEOvAU4$KR77t-Zh}& zQ1+l?8&;HO>cZD`yg=3VS}l%cb9{2y@rhiuJ`%h2*02+M>)o*7pye@!_N~1WxA{U2 zpUukMD!DqH$efFHn7JrLAQeOR>7*$6xhFM>iqlPyxY3UOmtK~W?xvgVw1ciVy5PE_ zqYKcPaM{sa;63!!x{tD>^Mhm46HDEZ{vUSq%d_4kH}R`N;a)l_;3JI{&{q-ziJh?P ze_`Zbz>2ar=6P@uJ)&j`gG=Kx_?KPULz7C->p$=$>dAByAhTolAgRmzy1HxTP!cc*QJMkTjE<%uZqF&2 zj$r#jt5`iqEj0TVDw)*2?iw~rpDGM(*Q-l5C*6dy$a&qyy0{^OP8Ww7wYc?SrYgs~ zn;Y2Bf^&Rg0vU0Wbmgq}$&Jmu<7z$GMSac<5M8e{szora#oFiScoxKq9*XL)-9O&0 zkgjC6^S2v6M}PZy;;J3RA*HYS58ZKa)Lnc(zYB!v5O9kJwsy(V)g^KPx_}#CzZq;_ zv!nZ$dhb|RP%Riqrwq4$I;?+*DKS&W6+pa%k0(lD4KIxF$Vxuaw}VugdZI7CQgn2K zqH*&Pl$0q`nyJ!s93%JCs!{DfwOeB?M>|X%fX}44yO3I_uBI~cNuB8>-$p6-g;|ei zTQ6m|_fbJcSDK(2wh`FO*3z0Db-8yOtyq|x$ewK0Ha1}Ewaxf=4upSgDXE1it377G z)!UQ$#q2~l`K5*q@l|4G?KEnSyEH?(lIiJ<#Q)g8t2t2*jOJqjn?DZJLmP3G#}2K!h{ zk5LMZ787twlq&=GPW@Oq+c_ll$%5#r&d|k|o9^fUU96_H1>9B8!xuphN2N>{MqFAf z7jLD$YAWn>Fw&66KE$941827R>K}E#-X5zyMdXywyNtTh9;}0WKtDUmd5qQv2UZLR zr~vG6fIL#y0z@?E*Z1)Yu=Ht=u6YIxJxvpV&T{{HP}iHO({laAVc)_Fa0r_(*LCA- zsZ^bU^Jg4umCVqd)>_EawlkUSI&$aZg<>TeuQ2Gkkhzyd$Q`KfSxG8G4{2H@j=Y(G zTU??9&Zk5JWPB(WZiGP>>k1ty_o?;jCwbg1*|%Cz=9QyMpv#O)wci+YdCMs}C`IXu z7k0xTFP_Ubu8h_haH(LF;M~i86^?!ZZ{}SE{kjA1R}Q^P!FQ>nuaBqWP)XPLLNvTU z*gc;gq?x(yXof>*6Fhj(1EB2X5i;UZ>tlm`(@1V3?w(3R_w|0NG*w(rrI7Ew@Pha= zstCxH^@G&5!^?0ay=!6UziOZ>36{klHs9S#wd_*4jXqAQ!0r#~q!oNczicPBhcK)^$d+P+?r!$vu5C8z=y<$< zQG}#M4#9W~=--6BE69rHp}X-kwwI$+IIXxbF@Egr#*4++#H*9zRVOAmSAh2)x8GLM zkkp8~vNN@VpTBZ+23^@Vt|qezN5|>?fQRlpAD2y>U0xcpGe-Kq>*y?XhR&xNk(2Y@Wz_u#@EynR-$GC1BUt%&1K76i%xq!&j>?bg z&vwc$U%n`}mXF}&s0O*d6VT8G!kHe!2A976b$OhPZRcfz5tH{S>b#~7_Ce@&|H9sC zrhip4&I?U`Ei+W>(rXJv9icvi)vK9pgMJpMWe;^pBZc^`;RAe;Hov8(7f; zw`o>BLqRtZUE$xj!^B&9jED+!gZrWbUoXenIA_rDMuWF!4w?s<%t6r(HWyMuL#dL^ zN;*ccHncFWzsxk>Y8Z_&8TUfQWjMSY-)s-K$Mhn0tZ;sDX<0afU9E+Qkk3jE|$M zTTr!bddr}TIeOcb3IA^d}e^yB_rdSA=b-5Tu8Y@rQwa`gDbo;Owp==!;HJGxe-Jeh`Ocpc?D&YcSO zuZgVJPIcP@S01;kBX|G#Ilp@4=sut;h(ls|&Hc&i1lbgivvhoytG4rpAy$ybD|w2p zR*ngap}Vxa#$0;_?;Sp%U-vrOKb7`Qx7!^Z{YuMhD~Z7;JNbdWMKB(VopoP*v2%K~ z4$Rk&)>g^}roS+BhoB3?%%^wIiK!@0(hvHl0X7hwPMMgd8+WHuU&F62Eo7>Wh_4N0 z8o-`*vpBUvQL~JZOvXF|Ju3y{jAUVAlgnCRww_B0>o#K5I5Of8igRdX;9LOQInIq2 z1?Lw*_px-<&DCMM59i7Z{gtD41ij}{(806|x`#Dhpsz!3j}mhiNPEL|d5@YWk8N?d z!MHFU!Y<-8beAO<=jZ=8=*+o)@AO;p=-sEQvQS6PCEE3^$N-Kf?JI7KJ2Ks8s7s-w z62|e7!G&jzE)X;H7*gF|JO#0E^wnZxVK-b&H4VdQ4jov!ZE%Rbni^^%{GQAVnL0o> zo#dd)p1iQ?*5iesK0Xbzq5BPzpWEd7>C4TAQTonf+J)Wk!6Cqu*AZPsGC+o z2<1wVSgSIUZm0YFXs7u_sPa zF5LUO-^`gaX9^v?g;R*03D8@NwJqb=&e_+nGtznjool;F!OT06UACq`XE(n`N3T9k zXW*;WjMS@Fq2Z5jzWK@TC-t2L>n8jSpqoT)s&35LZwR}8)8Hq*!Hdt2k8ffCYJNCB z#8!%rr$7^BY_^z18`KTFr+ z$@nt;V0x59hU9zk>=tDw7AB!JY8MmFBz5`~+7g&Ui>%;Sx3FpRF zU$4I&nZ|SH$8+qc*!zAmnsvyeBV@vj5Ae}(ZxsXTn_fRnm(Tem7<{_1t-HTeyop_` zhPA09lgUpKA5IRwLl3iVSa*(6&z_f9tn2;@E5zcw80{*U*Oa60yi`3WB{f9fgDyem z`Edl}B_ZEHto!tc$|ne%ehn01%x z_9#!E2{SuvWPFP*Z-AW|KE+)+IuTACYV$46Yb9uo-aG#3Ia=_N)sX!*zB$Hjidc+# zh)NT^8J<7Ji@{(NEe72g#r2Mwiz}Kax4W6=^XT>mW{Gv)$t~8Tb|2lo>{I3rb=)VX z$dNL!#F%Ne`$g z9UW=zAU$ri<2cAzrcbb1Q51Dz3x7cE(7qc+P{4U_+jj@(VsxHgDu8}(YuRw(aF>>1 zQ+G=HF3SZx%v7dw0g%V3Z2Bp$9^nJL%x0XeIGTX zrpr+>T%J$SY0p;AG!BmN{qNr&H25b%`Tg}3mYZLWCzJ8=^bt5;&9)bxHWn9;PO0L= zX0tk_S`&VvC)J8cFY5R>=G4>xU2U35*S(`VLC1v*J!f1B&!c95H-)2Hc}>t9M?Sf| z0KL@-a52Vi%FssJk1gjdJOJrA<4k)A=rRtMaei-(USXV$L&x6TdSXY43eoF$!p!cB?_6Mf*0?bdFyY7#l>o>%j{D=HdALhkz_Fm;nuio#jE zT@qgIuyaZN9`sVsAxSxNOD#iLuhP-!z#CBuiV0`|x{0$}hs-_-Ec)4{m7x!g1>z1z zXEEi~Gvn?pM{Z}1*KO(cpi327hv8BGAAOU7r%lrJI1`;|#j271Od= zugcK@E}mL`rT%;n$ChxCAK+gV z=`##E2jjMYS5bG1c;@1{p>xmup2VDM$?8&^iCrFkcZcQdO=^e?SgcYz|8BcZ3*87l zoXJi!UXq_2o%c{EevO#>;exvD-}>mdo5pL6gsN1WFWzV*_yq+-Y<Gxyk<*|3&qnBYLsk(gS06qU_N0U$9dh_tDpBiXN z$Si@0a$uhi80~d;jaBM{d3Sn}#Q5%}+ug+E&!3NPmg}3v4@VvZ^KX20+e_20@Dusl zh8qg=t(NCd-6@@7fTN_)L=oRAac19jF94lO#;~39*xO1lA9SMp0n3DVN#IIRFb>}` z=GhGE%*(?(7)Ni45jxJ7H0UImCS%S0-W^^2xl&T3L)?GYsF$0%V*Tv*pMLx9K$<3( zp7IhkWEVmVZZ^am+u(8b$zd;@pBNLKtOmV4DIKhDt{yKot9jZ_r(=^O@+WCY_qI(3 z=bn2xDzmKf!_om0Eb^J)TUfc~s>beSrXzXI$}-JduawkZ705-HhgZ)1+CHkJjyh&qpAA zw>~>QXYfob?BGIhLnflD&2#mAG>UdSSD^FqRBl?*?-6uHy^5#vA5OJ_y2|~kJgrV8 zTjIpqt&EQL$@gFXa)1|Q3O9${^}I)Uaos^bT&!f|^s<@Q!8kUzUV ze(&&lXnAw!+oZYfLcJa?*U71`O?BXjq>xrV%6 zR?=CQlTntA>)z3wpmY2!D zo%UPg++U*>m+lB_+-U(scq87Et*a`HkZA=j08y51Ma>@Er_bdJ# zFL!g*f44VCZs(#@p$}hR>%-sPJ3hXCUcSR(B|=T+2;ob zlX==(;RDEcALjd(v%tBz{orDqyA$}j>om53++MIpxZ2V zpL1ohBcC*DI|J%8pDMK8#q(nE8g-+*5hl{cA-JvuVdnOUhfP}2iTA>?KQ;Qx zcu*Ui(w@kx*xNF4d(naR@>w-|REokWJI~!)=Em(ouLe2@_YqtZS*_d_kAIIxD)klL+4YoxCea~^cD}# zm7_E4M6$%0?xERQjmjU(51i$ogK}2bv2S9TDmCzm9o_Ox*M|rF?suM{E)}~vuF}$- z^T|X;IqbQp30{d>Ey;T5oppPY8^qd&=%ME|4(t z%Al*q40zsDBXMRtf4X<{DxlNw+}PYW<>Yt(Bhe}-9iXUdhQ#wwt^Cu*zS`(<~{Tts>Z)aS7y|Iv_apzqUzOE z!%+cqrd%xiEjRQc8uy^9={3I-66RVhHgCExLeKmIY+L7rHenf_RnlF-pc69BtdiGz15-@U-mI~-d!d&qva**w0Da- zQK_F84y>y*I2Cu}mg86KUc0Af&)ws5xdjWuj-$}f_n^znxF9R}DBK)piCWKYwM{L% z`vE&Ov|ZVC2Xw~z9m9z*3Uu3O8x!V!4r$b%dO-0C7gsGgDRy*@zE@BQ=-jI2=w7() zBTHuE=zP~ux4oa9rkD7_lYb~9UaqYP1-I^8+gZ~2DVL$U7~g{~LI1TG5$3KgI&ZDQ zcWx2I^wlu3c;h6}MFBVky4%s)9kF^%?y9ID$N8|14k@P9A4e5`&$Bz#I(1hC`06SH zJ!_zndv|aZct!CvJPRhvDXK}0VCUvv>#^(qGvS=`b|Z-j6~ z1sy{RS>?7Xl^tDLdP{&V*Q9E__^?BC4!)N!W|J8o?zmj3>s9N+m$FGw`DqW??WGVW zzVAd^GF~AA&MoyGZ(=7t`!_z zfu7-wn^P}7Xe_RLg?iD`nmnW!cXtZSKeYBq7v-J3zbzdNESKe3&nf6WJ0B&BSoo|~ zP-YLh9HYDTEl1{ikvc_-*7QCM+*Y8w9sL=AZW~RWQgcE6V6iw2J9|0~MjW^Dl`ds< zVzr*YHV%cO98%YQIi1gH{}6HYu+-5tGJ%c>Abt&;g!Ok63AvYRzC+%+RCF0Wl9 zhHj|jdT-0OYPrp_%3sj`p|F>8$lY!BdOCVN=)01kb-|z0HNEegiAxR%?RPmkKwr`| z|1PtFPI#-Ew*k8S^>atvPHVP(*AjKsaeik*o`>5zG4_&hxxSg+?Nrd0h$g#Q(IA0K)6tKCE;C!_czCVXmFQ~Xj&FG*{w}26;pmf2s!6%*2+zr#ODwJl*+_1=t=Tc@ zR>w`*vd`C9I(V(KW86>ms!BJHMBnm-GX1NiYMP^fr?EUg9laj(4;Pkf^UAdd$9HSG z8qQ$dPP1UuzjAc1_gA@RzR0f~7Xuw0u1CvT0o`D?MA0M1xo@Sy*2wzu%U2m-V zN4u-9Q)dYuHtw8rEDkS|!yIS8{dDw$praceu|B0WJ3hO)&@Fc;=GA?41$qNLm;37m z`qXPZV7Fs4u{a)PbBt4qlS1#bel>Pe4FZainu$9y)=Iol3e?W`P!2D<3GrtVz`!$2I7 zbBR!=sGLu=wAVZQ{gO{Zj$R&1P^dc!!j)E}T5hSV6Z8=Dah@gCwvGYk;8W{$qGPB3zuFJE1%X9;LHr36 zcbLEdc|M+weh=txF>aoge(m39ep%|?lF$=J^WCbkF_h$0j?SP9436thk9(brC^#(BgAU~cDIdh_O4W5S+BG<_?-SdMb@x`65kcYWpbZ`Ii`62K=r zcP$w^A<=@NoawugyA$+X(Czt%V`9)z(u5ImX})7zP0>E$=&KH%V-<_i>Y}U7 zXu%1mxWxT*%g~zLgBo&>jb!Mr^j5Mh6_L!ZJ_Th zMisE_)WYxWwcH)-G(N<%b5}iV#slEk(fO88$2;USU9X5Zz74)+dGU(DPFL!QJ3(jA zH@4qzF!u|AzN~$}E{3zj+lSsD6~C4!J*KjHN6**g?q2VnQ{Sh>ygiraGU$lASrPQN z)pGakcW>!Z+C`jP9T<2%AR>w?|Bj{KVy#66I*z<~#!)Z@#%^$HvaDJs=nT4OUdQRW zBmvqDrp-wz`o%z>*6UXLV*tH>zg>gi=(h579;9YW*sXo;H&dSpboO&iz+Nlcin)iD z?hT!g4=2$OrzUb|$cqaOf0vlcHd*XSIcn3Orq1)+P{ z&bRWE#O0l%+d&N7X1gWq(|#E=x}H7WFjvmG>8?qHvv#}ah{#EaHbepO(O?^ro`D&N z9i6M$|LgU_C3f`kkLNyzL!2>Zxb@&`Pe(rvx}vVU>zO(FX29E(^xfwT^kv@8(R`QN z*wN+gtYwz6Rt*y#&R`Z!oWG&8)tTCv=pKp((#s)u5a1(KB_L}t#xSVkSDPmeUBr*jp=5uDhSR2cp2vkyo|Z9oirqRsWx} z|Na4vgKlg?aP)2UibU0~IkTAi`ZAxV<)`SIJcdoZTyv|JZvT{KMLhqI>w1;)7SbEn zsmS>oqZ%PNMZLjkaq%v(Z0RxW9VJJ29G@#ec-+w?qwB%_=xN{!`2!1N>%rUnoo0T5 zzT0$n3Hw0F!yWpR2I+$<_sg`Tlrq#_`AW}@F7>$kvi&5zl7Fg@%iDCp$p*b_>oy(!@KB$YH7;~r$$lg7$^p$o%qPfJAi z{NnuF5C_<>_N)hKxvKy>ZHQC>yC#le6#in;+BkPtKs1KapvN$DK<+)=7xX}4P%OdO>uIcvzVl_i zmqZ&qVw|vYcFcYcy`fh7(*Ye#^_!z^90#z$y}Y^E4tYZ-G3RT#z0U~8@s|5X109ae z_EqY!Gbkv6EmV|1u5Q~i-d)vG(@Yp1;FkOE8hQ&6xTgj>j2uUq?&j=Vx%Z%>LGLEU zxlVco=YzL%h>?GYP#lHjKs)-Qm12L?N_~U2dVjAEx@i{6P07!q_6Mp+2)RY`4ca2C zk&kiE(b>@R449WfwVrBseG<+YbZyX^s8wI-=ghfx@&|3X8{$6x?)Ou`0V-QNNHW}T z7KyR0Zw0XM`{Kw;>rumc78p9(?@jU5%m+CZb%(3#nwDu@M~lC^4I|0Xa~ytDppiWB zb$Q9u-6i%slR){{cP*!TI+9`x6EI zBcgEmbsbXbr+H03mZq&Kjk*+7mC2`OyLWR%U5C%BIcLlnc#X28Nk6gW- zB(o6ax+Fd2H80bUua+=>XPhg=dt__rQDg4RJ4f9?JuX#@xZw;_X_P?)|IfRI z)?WQw6d3vwXt^VPyk*|UjecSw;5=@^d%EHA$3;I1fa`&W0|U`+kNXcjKOLQP81&C8 z40>7>`8B^4RDCYXFu#5+fJ+Y4wSmu|%jNG^<}Oz<|}F zgf=S4V*DCp zWxm#4uVWvX>OwnMl#}z7m6xgquPJY1d=rP_TEowP)hQ6VlCHCy6BvVztqXMcIw=X! z>L*|dV~))iX{q;giOkSIzsYM8W;Sw7^}_gyyqJ-qB2EeQ)QM4_S-+en@#2Dy@H(m* z1-k`2`S@A+)N}6H)=GxbLto=tawLchjV>3a9ozBX_B9E{P4=y3#NELsfLq8{M^I?w zK2PfB5}Bca{zb2TDQ#35@pw~4Q87xL=9MUu$7wOOnsdROH=7r`MP!SS-jc>=zRz(k zR*p;GhBFlRas)}a>fSon1*C|EoZj>%je38Tyeh~W<)ny4Cl5=1iD3xrO)&Jneyl$L z{ZI_uWyWbvXD=7i-K2SixvAt>H*Q*8gZ`YAEiSWsj017v=WqEP>oaIb_(0~?6yqwX zKX)*6L_Hir7PD(~`qu$3oDUrIJq8@awd0C(=jY2K*0P-1G)B;6Z+V&Pdfye7nw0$9 z)hIhpblZKXhRzZjQ={(?M+xVokSQh0Q>z-=D4v=I!-0_?@DH+ycnSdB`1-FKtOR5-DBd&M? z)&q_IHuxJQ^%$~Q{}z1hMR?b(0e!_#qpu6-X=*a!d_p4Bm)Sgx$0P*ZHQOz{sm^&} z+*(*ew~C(|chOr=VR9|uw9)(4+O@s_Oz_swgF6aRVc>&IBg~jX{P_|Te(V77{~i6+ zgpRHggU*IdoI6byN7+1X2A!MjV+FdEaL-s8=|3$ z#MQwGo7@m9!$OY^4$MgoT)m6dBiUn|JfESX2ajLbGSCB#{tg_yWYDpX&dqjLp=x6bUG*D&2!wSix=QgP%ih$K3yG@0P>#bEhfxR}emrLNf3jz>sgkgxcUh zCtcs&eV$BKle(ySwK_b!IR0?*>GbU3@&5i7>HhoU!`bP{$K#8`gB87@a9YsaS2sUy z_gqnHa{Kjmf3o-O=l(=q99g^UnlhU?SyGlo=iT&X<0So6@o&WkSUS-4Z}bMe9l9Mb jz}~~a5-(}ze8cq{VE5fN!p zqX>i+P!U5fK}quy={c5oB!zpvci*^gjQ8t~dk=#j3E||dz1N<5t~u9UpIT?Z6i%IV zIti1Ikbqr=eqifjn9cEX_D3(E&spyXyk;D7!#{9Gl$pu?9eW}pg2MuLIGu9Xamv-+ zVMo-#1AC2jghyVv9(?Qe4%;1O`}fPP&%loVcd!4e>%W`)U)|upe_ro|DQ=Q9m-;Rx zp$L;yl#o)CSnq`)U@(ad|MT9k|M!xRl#<@CQD)O-Svlwf%?dC{2`MQ_X{ikxq@|%x z$3lOHNh@wp+G%!lqw?7+GP}Z64#eLp*`#}{zDxDoFi+3?>dl1BvTEwvwrlLxH`udx zpM~W?t3!vakDstV>EP&e>ih+_i|&^^JpHcu2LuKMheX_pjEctJzLS`goRXTBo^k)d zKM(Wr9~BgqmOU$fUh$&xWkX|Aa|`A5o7V20URq!OyZ0YPM#sh{Ca0zuY!3J9zu)G* z&o2N!e*XH+{{#MA`!6mDnAHCovi}*_|1U0R5F{j}rKO~0{)|RNI*JfSwb3E0nH-}}_^eouBf&W7LeH zyf4i3HgnGRQURu$o}te8^A|ET9 z7UhN!hy1c@RX3$)!|*bFcJDgmA*L1w~|EJ zni~029!-Cq`jBw=R=ekilh&oypJ&W6_NKuEt8BDTHIb9rWtuvx0N&2 ztzI$;O>W7UHEVYotk3&4da?Q8vwNpr!GzF!B-{}9wvw0M2u=hSZ5lV-$-{MFBrXjA zaDFyOj*hZemSeAWo5BYynQ1r^%fogJnO>8|%U?$cN20?IY+79j%ZeeNcxWf{OBqU( zWI#QeiepYKiZ`d168oSB6}m<99?($ANGs5tEt4eFE%u>r$ctwgK1H!pyC3~t3^J~* zu{mJUCRK{HUjDEUaHz>;-%q!U07GS0=^G0dQH_6uN))lIx?qd=W9u!FQY1WH2M#)v z5JyqUHmX;w(o8#ciia>71WXvwm|kZ2u&FhBGODv!!Cx?b~J{&;s`RD~^ND$KIo z%Awio?OXMS$PdOV$RXA4I0?~NQkSa`4u#jBNy9XJs}k5_Nt7{P#3;#+9N(b>YEc`R zJ2AlkpJ||#tV0|#x;7Y zvM zDsEO~5*JxMPE>QK@wp^46FnnX^7xFH=aZJ2GuR_kIQa69eM^y;X+LISFiM-aYbkvh z{w;!Qx-<#U`L{5V75#|hj(s2|0x;;giKln+i6-?v25u8=`zd~>9-M!CqKMrbtD|2u z!d%L0{X{PJaSr2LYcjk0OM)Tpfv~`x)xZ2$U_@}sJG!AFzQdNVZ_;FCOVuW?LpGtM z{gs=yXH*?qGh&p#-Mcf|yTmWC5GHyEX0mY%iZ3)6;{+E`7Eysf1lc?Rn^-3}i_fJo z^2n^!%noy2ej>q`7wt`L04ZGb_UBJ47A6W!3d-ngW77F?oTX}pq06zAA7g_@oTGf= zASieWd_c_Pisj_}0asltQ3f(wrrB-zN+22X6}li(gtkFECV z=sLHbjEQVZ2{JDI!oM3N^&^+l--(-E?IWv*RSCw>^F-sB+|^|>4oPz$yuvrQ;|O>* z6apvjeA>BL!{CdvIt`=0YmUXt9I4vAii>=Ty^zq49xX|6ezTHvu5n1_(^_sPDf14l zZyFWww@v%SXF^b7zaJY5>U>)1x%d{Xu3y97WFcpK-096&Q>yOS_9D*exlxmqn(e#p zja-!Ng%N*)&;#@!rj+?P(BniBdwk^pWM(Daiho2@BHkv%it5EXrk&LJ2HB`!lwycR zXkBhwRTZwOd?KD#W;gcRMHl_gUNa}D-eJQsV_2-bz)RdmZZs7t)A>2zFa&?f;3yzq z!eo1oNSRn<$Jrze5VVQ z&a?V*0kT+DPGEBhU*e}D1L8Pdhf+U9&e z&`F>rJSmDNZW2#<_V=LF@r9=Gg_PU?;CkUo0SP$n{U(MEr?Ht^yXLUxeL}2pYFu#Uk+x~CFJ>9iK;EhlK5}yFxr~h zL%_;=yvYUE>ADWXrd!(XB*Z>DzV%HV)(j*%R#zK0+e!{5Ha9H6c-dZJ*=Q)ul>n{t z9wDd}(z0>5f4J6-!7En)hhBa++r-(+_EpP>7y5NR9h+->oqeNrchI-LFcpbE)MTM) zIJ9BD5@mMm1b=dx>b=CX2nw?&_B_NdsHqfpGdhPFk7_*mW`BM^Le(ZDHzGWu1mDAX z?Au@Rea+ddLn9{L|F#Af6J&?l*qC+g!NQrFXqfWnuab^)9k+`&=$SY{jN3q_Y)hr| zmb*RDicfOR?2$kZJni+r0k4fnIP}T!@Q19s_(R1~hxGjJWw8>or*z|D5?}x6Gq*!D zG8^lbCk4vlPvnL(_3DYJz`FE1xMe{s55@CDC zN0;IqIM3^WOqS2!T!wmg3sC`xPWyF4Uf1UpaHOfSa$(IL9Ig5HKG!9wKjMnb+U`$} ziq5_%hJRpYQWa{5dxU2JW^bnH20V&Q%g?qs{}Q_v9oL~f-x!3H;XO?$F**f6RYTX*pE*{fToStb4@fHdkA-3#kiYOF@dyC{hQUbRV*0s^bqBALGw- z6L(Jt51KsaX7s07971t?UM%XU3*1JgLS)QmEI+)v@%WX*ujr+J1?QXvDB%uKZHMOk z?3Pfk^P-nT+3(biHNwklO$X28b?tQwMhQ-w%qgR?rZ*&oj`oWARG0WIe6J#QfO^mc zn8Yv2k_Z}~BckUhe?$UFUbwlKupe}P*~8Ly<`|@t0+^XhWu#IM$5cybnBQ-T#d2qu zDabuz0ZsK|DhsaFBkxtL#~5Li*3}V|hp;TEpT}uBdjWV(n==Qc0lvN14#;G*U8!Aw z(Bk0R%FmkZ4&DGf%~CGg-DCW>W@$++F4Fkq%C{d2?7J3?1A18!*IhSuJOp(6@3{An zm4$kvZD&ap1T&y6gJ8==tAZNCMYfu~zQ(+#*~=-3bs=YCQ}$2ZacTehYWjmW&RtKL z28$stVFVY!M*epr>gD)v zM@!3CsJ>Ip$`bc}jLWf+5A3*vz3*Ua815gT3g$Rq)wvE+A`T)NQkSM#xG4dPl-q3R z~=K|4P(QeHS+;~`PS;w_YSIdqV~t@ zJxSKg`F+9df&_0a4I9*70*K@ z>EES(P}t;jq%BS0gE#LLYXA-zzu}Cl=-k54a&neY+wqtec)ABz5yJ(4{fwI3+0>rr zf)=2ABRnk^YnDZ|+zRIB(N7Ovn)4dpf8{eb@vxx`JCZN@icu z_KZVC5JGu@eDz5xD~+V}OjOW}NFzyEpb!K{Ah!=K8?p}5Y6~o0)Y&O2HzE0JAMZgb zM86)c5x-xDZ4&2m+<}jNr+LiOU2m){|7}1=y~7b+!i0AqK7u41;oU+t$@D@@mGd;q zx~>~DCBvd;l0vlIonQO(^dhZ!SxWQNjRZ66gGwt9cK54$29$&(C(dF$?X%7r-G5gc zcZ1pp?{;Q_d)$||1An;r%eKIg3%uMO+(vv1joOsE>;xF1n-VWgE;YqGy>h*nH<`G! zGJXBGc69HZfB3%JdF}0E^+&g;oi(bxXRo_QW?ns;#Jb0PWwZ|4NK_)u`XWTNNL``1 zD5=9BNb&ncYljRQQ$G%wadjQW-3buHTW~2w+j*k)7V;nX8?UF8>Qgm zwcy$n6P0)8-y(8EsDF(YW5?*=zi%H$xy4V!Ap@X+`{y7;=XhrkGtWRm-`X+p1XU-G z#)-r;b^^MXpM_S(&W_g(4kSFQ+_Kg2P%%7x0CwZd0rMR2(*fIt8?o=Sl;6qxItuMf z=R_IQ5T-ogDDY>QX3%MxVaWT*4qSRHaKzJDv+ana$U{{2HG8T7`<)-@t9TAR|9Pg}I1&fTxjm`~i=t4QwpKvnwjv^n_Y(Vk z)j?fAlSN;y0ges$g0X_j1V`|JWufmpVry-AVW=+qPglEAzpD(I$IK+`MdruWEHXYC zoNa6BXcvEKEt+tSvfD(|&U}jlcT$!S@HFx8#zcrQY>)-jvApr;c^*5FCiNh?#v4Dl zBosRDMa#G3%XiC z?^;E2n|X2Z^;eYPCV4q@L-XRJE_w*;@1a+|oBMs=&_Bp&ZLN^4cKRj#%82W`DniR4Cc~;#+j1P&=;OxJWl6F{haAHZkfNQ8h3;wEervh zd2^h|y1rE@p*>KSBsQL2C?uyDi+9X>iazTYg+_a10G3Z(n$t}T+VcLmJ7+#0Gx@z& z4)gfAkHc2iA!(FKGc!Yc5P&$RBJ#jOhsA^t3mjf*bBpBA(;^nF3MaF7j%~9m+LeEC zoRdwk4-J>zDCbeF6RPi?pc!i7;P}i{VlAQqfBw8EaUI4g^0fv6vx(zuCIcmw-Tf?> zyj84&H(A@=0-zGsVSdalST#s`OXs#<9XYBxn;8M*_739%d#o*MXsxl79)`J-T+@wL zI}mrhvr=R=4_$szjI+0$d}X_-Z;w<7FF!+gl!Z~n7dLk zs=N&(B1+#AJtM+tT%UcY90a7QVwv$HOd%?}S$dMmcIZKD4r%4$TRC`LHJ5T1jx%{_ zN|>(a7?#j=Q0LNjshjQ5Pe7d;yV>s{6AH%c1S<)Q+~1;$sZjx&V4LZ+{KAgO><37lU;Jt1U*u=j)xu9 z)jec$P@-#xs}zA|c-O%`Y#kQmR45%FbGPO|ligj3%g6Y|GIP>7>K2}ijvbniOI zG{cT5Dhj8>PS@_9(0+GTFWy-iv)V#4fGJ5n8Iq|3%UJoe+>L}iU@(x_+o3RVi?x`D zXGlqR@%_b%V_T>GeC5P#p#TRRDtu#f$i6`mXxyk!u;tg>TAWT9u>9RRN+|gG*;S?k z{KJ0$rUS|>s9c)v@$%Vfnu=*$UhS<~mDJQe;A|YQdKN6_JOm8VY6nH`D!JG7+w@GBc&Bs(haA||Kx;W|tR z>7C+DqXMgGZ*P;#GV!}R7@AAeM{j!#dbLlRuiWs;zmzSQ$eNCVnuhx@!<)sI6XW0* zsO;Q=PvrCe+wh6+3NhWJ9F(M3XNvd|rXnkkj;e zZr^#o=W+;!75Ib|`ngtKO8rX@Uq;ZRul{I_POR*=m9ehpjM*7a%>J0mFqvOWfj)-d zhK)7kS_zdK#hY^lDDnFuS>9KQRiVRjmEb1f#Mtkdr|M_kHsbrrO2InLi6WiRukFqr zWqBt+EZa8r*#&(c{$uSpls1V~(Q1w%1dS05K{}re?iB9OxpsaJY#D-{mnhl)00@#4f7QhBzm1V_a5%y7HGo^1=Z&)eEO7-U zLakAYN&=ScLh78|FhW4FzZ1N|^_#04o`dM4ubU?#$r>|0L4~^B3lZ<1R~v^PItc?v z$q*YkgPF?c=|wdnmi&Q5A8>_Av0C=vEkgRGg2n1KA@VKnHE7Q>^aSrc8>k`Z^U&!+ zl!YRg>yz25h6{|~3{)M72!~)s`u>X)RSSu~1p+Mso=0Dj5?q@Wxci>*Icmig+)#BoF;rkk83-T@YHozb~iE;;8S zwAS-`w3WNXmelc;-*g?q;YVIyzy_ay;4n*!#C)qt>yN`3jM!r?mS6Y7;GZ$Mj3}z$ zv~U-+Kw_~do>OiXb8Cl&P*RjFGz5OKsOe%O;Tectp;Fh#JtcJqAFxeT@pCkyDiGUS zxHt{YbssG3coTH%snX92)Lh?YFS&EHXGv0jKaEmZPu5{g{k@s;rZlRgMK+wD#v^4x z0ux+EJV=7Ko#!pOnGbST5qR|z^?*FTedB|g-AzSWJfk)3AB#~$Ab)UOzdZFM6sSBwywjjGvVUl4j7P90l5lj7Qc#@gOhbAZprtLF8$(M zz1Z5&Qu2a%Vt8PKsCu~qGxd(5X7FDAE<9drKFEvWa#FKQI$*(61q*6oW5~f=R;Vc_ zyD#xJpo|{*RB#p7pSD7eQZPaZ+w%rkhT4Vlwi9kqhlMXONWHR zO*Qf_0Ep$=K-n1{Gn;yC9Ts1~%FiIGfr(ssXg;iZvMUx_QyO1jjmsFd1uqu-Xq5cC z)+e4hx#2ISJ{pN>ns2id;O`LNp(0#zp}}$qua1T8it!3HWXmkgLDHElPd%aOt$BDj z_DV69&V_5u4~D8vuDQ|45NXGIauzr|`7=@13cqgiHMqPL&Tn&(Vn;7+fVRPGP#HMI z^oEoaXjgz3Zt$v3u#R@odqSW;BnNZVdrh|ntzEFbURaXwiS9a1Fk}0iTZcJQ#%ilC zE4Xa<;JoAhfnr1$;i#wzTEBUSru-!?+jOdje^XS_VY?3Nvg}Y#+XnXc=-7aJpFxCs z64Ic*c%M*q2kC$To18&#;kLWg)es(gkb5`MPt<$a=b)gQftzvqPx?m}q3XxOe-Sym|{-LRF(_a@!WM9*AQH*&Kao{IKeKk=}UW zlut0_Mc$vuyqiJMh87nWE=heyGotyXP&zjf$2;}_lpInh9DUj+}dr%l3c*_w0b z`bnD!HlQi1UB;2pij%`fL%u-|LI$7d;lW9m-Da_TVD$tq3D~@%w1QUBeXzw2@oAp5 zd_|Ub_M-eZD~0Rm!|6@U=T7HF@*}L}-Wz}YvIo`zGY$4Uw^;ZoyR7kK)yoh@)MJCf zW4%Km?s;pmLs7c0*t~CHcWxY=w?hn(gevMZRAUWUsGd)opf>#;T&`Oh<{A3(tdg3a z?raKm_VHSlVPm>ucdd5M%H_HYm$Z-TH3JtLvB9MwO1pEXHa@8IZ7I8v_rVS(f|gW* zuXu*31kFx~c&n%iX-1sH$P#t%Ojhoau6{l+$G_=?*fj5}p2?oeB*>NF7kgVBlCH0W z76q*?TR?^KyoMQGHEYO`-?**#!^=35EI|_a*XYJuzx$`}J|n-bzESqgsB@2$^tU*a z2bGmdmTMwspm}HhYE=4LGA-7fWiVvWd*_FKjhXp)0{RuOd26r zVYBTCu4{hYPG3D{6oAi;jxrz4fyc$0u5g<$@F0W=l~sz6#MSD)&I0E+GO5U|!VA2Q z?$wV{)}Uk)D-X9>RonaPWz-m%`F?Sc4mo}K)y_^^1O6+pQXnI|4ozw*@dIY7&>ird z>qEV=>SV@|CB>3pE(gf2cwCFW&#_(vh&cd!R zu9JSs;hPfukT2hiS(@Z2vruWI%Vbc0HG|;8sY{wg0g`wwKTlNFhPhmb$`V_KNcRV` z$tfE08H`KlV&OjCVBTZ@i0~^B(&rcNB(Ic2y;uJ1T@Yt;o@FXj74UzE3Q_*c5qx{_ z2;h8%SJAy3;uC_6b+WO%63bncnE{KO(`_8e@iMf@25+ZgkB}dqy+04=tz9gsoFDah zck8L=RMwE>e*i*aRbn7yVim5#R0%llsuVt^8{!EzmVo>5rXjHbDAyNlUDHlUf8abG z>4R|`dw^-UU3-hJ5IlKjqI7tHI(;X@XSe7F#={$zIRN`fpMh);bs(R=1VrYQbLHV( z7&);fKIYu`Z#Z0)Rr z;YVn88mPuqp{fC&D|*znIh03X2lgF&%q1Z7|g zT%x1+eWX6{4~5JUGpTrb|bN0pq#w^dhq{UAagONejDHkw!Qkp%UhhZ#z*z ziv9OQ*72SVh0`RgVpW@|$-ErYnUxM>eFikTwy<)z}-RMO|Y9tCTCl~mGLnch>2&Oylu0WH+hB^^=9K+HgyaBPS+EaCg&N)ut-~P zYfMOVzPh>3OzSh*VuYXgGelRRFOO$72X?J)!+WuFmk>M(%e{}9t7FKU%Z4^+Ob$6L zkw(-9FliPp7e{;56+zjBRkY9f9<{YTYT$;IsuyxIUiqW<7ag4qlgFOFNs^kQPfdzt zWy*=4)?sqQUE)E64p9m`&o3AKLo@>Odir(tfKD7J(TEBY1WT^S%B6XP4bXe2+k&^Z z4@J&AMWJ6cjD0;~zZ7+_dgR2S?tYmXe3@7_oVUofN+dY5OhE(G5}A#gqKlPl4!EfI zqclw0{AvikqSDA3C7{S_S-D4y0**V&#NS$kM%lmpE$?^Jo%$w>{s?E==VF-f1#dE2 z7|4cihxiu3jFYPX#`LuwtQ5U12dhp^;C2G?IZY29zi19>c0ce1S#08Tuhu{CwpJtM zqvMWko;f9?D&~SlQB7p@%LUJx&y43al_6F)eZGWvJ{(KI&O9!dJK(aL_OU?TqMUce zQFm9|PAm7--Ocard!;Y!s#i;KM_xnPGar9FZcQl&4Ary`SlnWi2i#qGf!P&@s;Gln z0m4N0saOQVfwfA5jF4t}8c`l{ENjT){QBj!B4u~HR(8d`oHyRL z`eF|RNiCwLEBMNy8tU~u&4RjVIx96ptQYL77Q~(;I0ac`n|8PQmKbhm9Idr+;2aTI z+2@~Y&2zjsgHBk9Daaa1Iz71kGW;Tl0giL1LgffAQH@xMU=J$Sf(~2?jg+7<*XgB6 zSM&V>;(8IXO-xOn@021vz=T|(={x{|ou7U4+mWSz6HYW2$)EJt@%sLrnHw^5v*Zd) zNMQ>3*_6j%O!Nz1n*Wu+=SEexBlQ|e8S}i8gZmy=?!q+8SSS_`*c5ra+jUGrFb&bf z|0HOf@SrGx>W2UgB7f&StRQs@f!}o65j5inE$aLiA8m)eq4^>eU^g|1(Z}KYpd4Fx z#~0LnHFGq*4y{~uv5i~0cQGKEe}V@>=-%lg zcP>p1T1ZgWmXdJ*Z?(qynXH6X^cfo9L}VwKuPZuf1*nAJ;p0&caIA&=M<^=6d`L?%opI-X- zJ9oFvcg0)t`-+YLb-gG>G=%hES+9;IVEm=Ly?q^)W^=r17HhpUp%Gcf4|uBjsVp}9 za7?TfdhlU$QuRF#Po}DnsFqYDUJ1w&BmmdN{q97a(2?qFTUOCM&U`dxl z-;UpCnuby!xU(?w;NweRbnUe8#4uq)ILH6>>xo;p2itk|ZvdHeKp*^+4JClugBP3UMDlEWLcU`5n}i$loPO7w;r8P|9I8N1pjIKpD0} z+>%2U(pne_T) zC{btoLw2hvdD0p@G{a#60Us-Bf^=he9I7!Yk=0HTZyw}T@XJKcR&6j`ZLdD60@7l^ zx=yHD|J}+tFoGpAw;%o$Y|?Chupu}_ZeVZquQasst2HrtfvwZ+ZEVeNX$Cx%7iu*yc~0xGJz`mtS4} zepK$~v-G<_y5hZpW3BHFojiAwouB0@u@0MDhb=-47{7nBh9v%qZ6bxXs5Rk`R77*8 z;=Sc1JE1k{0S!j-q4J#T(n<0}Sz;gN6)Ksy8_#tB-AcrxR1FfVRAXa2)ba77rirvc zcxk*t={s4LsEx04eN({6P#N6~*e+*$8Qj$q8gW~5JL|_!I~@t$Prf^N7OIVHfQl?s z$SNa{Yt@NkAo%gTmdt8~m`2`!+=g!yD-G26adevfBK@Bet^q=`CSx{fYtX^ylpoAb z>0trmTgzSqO}bl+Oz(;#Kq{Fu0idV1MF?&P31B^^>TlaoZtQXX8Bue`&IPN+ZC>R0P6goQ zgKDPq(ixD^uHJjw#O6|@={8U@7XMjzd0e=GlYQ)VvnCn-1w2-q<@j4hwwAISb7Q?xxzEo<-Rf%Ht;0rRr%l72 zWpL?5K^n&Om6zdpQ!CMa!my={bJsUU`D;GO@01DJa>K0o222I^^o=e3$`r6TAOBet z1Y02{uB`#|U&R|6QefJ&SI3egwjb&1KkR!(H}GgN;i` zFouT-kR&DJS>965A#KN}x>LQX;G035bFg;gktz03{ZfINdfqWRna|{wIiX@cKO559 zHwQ7BlEgH&H8SYL2%1>Dh6V8i>2L%yFDRvaxJj#ocj z9z9nyG0Qo0abfQsDHeJ(znAPN>T|Cv)EWP#)9|PCP4mduKBj)$e?rhD(>%{v0Ip<<+|LWSC zrmQ2|wWPi?L4$+2ygL3hQ7$S3;-n}VCfQdK9N;je!MEMaM(WZq3zwmGV9{4<3~7je zJ`@58B>FYQHm_)qFyF1sM!j0Fy2x+rH}v>>)2PyT$RG9tk%E#F%0X3=9=HwzQn_=z z2sq)0CsB@^|<1py9Fh27yjm9#M+-je2JyPX`IMb7CG%+VRHR zZwBul^gGg=J)PnB6*dIE7pM~7iBY104t?<$iHBZ?Wii#;B1?&ME3pa!K~xmR@>&v~ z5l--!Np1cwdY*$3oc!cwj128c*`lx7+X;Pa?D^BwhCEK+amzYJqW!Bl;>lwBo#*#z zI_ZTI8Z_})G1DdK-H*yr_SX*AUeLpTcD`wcI@Ls-f`J(EW;(1=+tv8bXB zO!RIZUT*&4cYcEO>Q+hF_p%FZZ#!B>EFOJ*eP7$)Ov;b7@AGRRPT|HMrsnC`E1@>k z{0a8-cZrs@=jDOHw*!MK7cLeOOCf`$$zle(_%3v{V^BRGHl`S$7CoSL&8+?3r(?YI zZ6}6T?!Ml5J@KPkj=hJ{{&%WH?pb{=Bh-8-rJs#bHWu#Aa6BKguX0b2)O5#g@i0aQ zX&$Vj$ooS>Ws((aBG{`E_!u;glGtX;K`UWGG}Kdss4_;s-?MoQlc#&{_oG|tn^O#L z-Jj^5U2{BpoT+YfNvbdI-M&j@^^#6wucMc?y@K?xx8jX|pZKB7RA97Y>M}P@b~iZu z+x=1-P1bWg9=7v9g5->|p?aG6HLbgLN{*+~58xCUk6P?CQs}Pz^zP#9Z9r0c7}@va z*gC93BE#foYaFULa~%d6x-X664rU5G@L0Y-5J?fLkX=CCVuJiAA-o#jO}`*2ve|yS z6g)NAN#eb=F=M@fG9l?<#|~1rp_Etf1f?~pFyQgcAKu4h&$xBUuq=Q7o~?+DsbtcP z$N6QqBVYJ_IkRd^8ysJvM?kI9x?A2ySd6sgZ{1&Kex!JFldH@mhKDxInjke!E_Vrx zg;&O(#aoz439&q%Z}t2X!;o9@$M`-{-q^fza@`h#B-i^v)bt&&oN&9XP(~&Zwswe~ zuERoc3_rHz*PxxeIa;oAh>k`fHkiRY&r8jyrDq8h>lnJyO>bYlx47t(U)4^gO^hVb;4{Z~z?$=#!+>=9pA5Ls~9rK-{cW!5LZ_5yLT&Dstz5~zzW=n(^N&Ixo~G2b&McU z4U*V>>7~94cVZ_VzVN0b+ekeFW6s`RJk$2(ZQeMPuQh=ej)w1QJM}=)xqb(NJOKorR z;nS1jwRf8)a=#t(b~uzK^~;=}{8sdY8d3)u1unxm7=|wA8=1%D2l4`YsfyVo&oL7n zV=t)FkQBJQaiFsm(}R+-3IC8tE3zFXXcZz;x1UTDhbTXl~1otZ9G~}UX zBM*H3einbEd*o<2{x~}@E|ky}$U-ISz?)puJ7?8eGEUyg@Ig6H(n2z?SK{9yD2z#NtY6kuI1EON2G($W>hFbs6@(}Dl zeR@3WFGoZDBMoSXq?V#>uP1(m-<)7Xu23LTJ2?NIgMaNH>ABZ}#I+P-)HzC~>kggp z=cKOdzeq_W1*A4!o$n)hWOJBD=5OMi$>q{z@uPf9+hlEIEQ$LmA3&XaYzM}35E)|I zfCsMaz(er^!{9a3f=Ckf*^~0YufwIyFyaPIcHPSO;WFBpRKZSn@N$`ed}ChnRwokSiHL+@?>l+I=< zV3AfpXaXJ%NdUaeE^9Z@AdMW~hN=XgFRUzHurvK#RTblvnW>9|jhVnud4Xx&gPRU3 zJW$a03b?A+l487tuB@c~?MFeG-v^~~7uX${ZqrRu!ebzXNAA>k#$h@GlQekhtlZej z=1up4JN5!8*(D+JddYX~)u&d!DV_FG?DLt@PEvr69B}G&b+1 zZZs8|^GrLft$Q;=%kWl^;P|Uu*JboNr`ss1{FN{maF@@6&SPEcJFH_a6fQnF!kcs zCAsZ_z5_EZFlmX``${)P=w3E6-Er9c;pavq+?#T!Q&$C6tONcL(*0&$&(rN<;wxbhChLxN&F7>ZvW@w1XAUTa{h}j$&FE}8@%1M(O9F=;UHu(;vd4L( z8Y*s%pL>(@p|87p1pDRW8b9FDa`mqn!^rzFi&d4fCbkySPs-`B(Ni> zj~X@CWs;l)vUlc;;T-OWioyeR5z56TfDbgLcpJ&D#)v8w8 z3$rQe?suiOH;p#U+%{>BJ+`L3ukQmxztFjA@qGtP^&jl6lm3NH&nej@ZhyH?3%sxQ zz=)lsF0^_#rU5d1snav!&5bO1gQG5J7B0DooQ-IrX8MgTd-Xz9(DM#c@Fvr9{MPxZ zekpHsIl-2vUr`*?H{v|Cykjg`qOS-HlY|+d_)l!+&1x*SAf6_+$g@{Y5{Gw{cm17q7`btv-1wW#>(acU$Z} zna6*IZQT;W*8j?tkbe31Ytwlz+mENyBz~U0e+(eEYy5_7{`P#N7dfKmzvc49X(jWw z`(u(dJuW=@XpAaIkgP0*hYXVf5xGP?$aW($tn48FAk>S6OKSEInsN}CYIio4g|lMi zdbi|ZMNeNYqOKjh7^)d^kr?@9*!f66fs#j_ldcjhR!(qAOyfaE<;{3F{$uEvILZl} z<4JL)!@;OIS}Q8G86mq!+KM!t_Y6TpIzthX*HPBqWb8?Q_nLg7{dI=|m@{JVw$X8z zD=oz~yDwpbaQO~iKCD?9o!z;ERn~Ct-JwG;xx~->{_T(eC^q=(cX#7~1i!70**-d_ z{Unr4KK-572vs^rbKx&ep*mf{ZYzl5EQFsX544 zE`6i%XL&=p5PCZa#lJneQvEG-q`$%-7vo;9?H&+aBF>qy{)|yF!&!For@z= z|NqA)Wh9n+xlJzJS-FmzOTYK`_Yc^dopau=*Yov!K3<6?8fTe1n=+<$zj)%Tp77BcegT~8nE!uw^^t2x z;2vajV`_^#S5yd>HT6eDD1kOqRkq;5tC=&82hsbS?!*_+pVJTtS3FU7tb^Sdp4I{! z=LLuPr^p4C3r4T`=U_xsuR(9X>>0^l7qd>kZh!e#Xa9z+?jHM7P*Z2~qitX$cT6K$ zpzxdbrA7E3NN10i@qniy>=p3~8IJKoNr2&T{HB#98O}2%v5YIuu1c%3Ab}RUD`z&= zq)OdibMFZ5?(SJ-P1TmB0n4&v<0kW){Q5*)n8;1w$HuDT|M)}(_Yn7Sol^x)n_??d zch<7Lg`VeXR_`Rg@aQUbn;X2|{ABd8NyB5D{|Nn7)?Cn~BnPe3&NEN#;q$Hnc`Oqb zy^2#40DUz&ir_aedJak{t^skDdU7tsf;*McI4|!LqLLk@fIyq~KyKBV~;jdul_41Jw&+L$8t8}$2*ogRpH zD7XMj>Z7rO!$V8;PDq|z|4Xjfl6cviO=@-uDG-%FKObI|cC%9LIy&EX`Vj60LGNYV zU$SF;)|{SwVSSDSOyCdVy;f&!8wcx_j*4bKhGl6Oe4JmOF^+ZopwqQc-YH1;65*#_ zh5`)cs9XJDM---v4C!E!L|T)i_Cx>sh!su}Hql2%7)xSa6I>Oh2#$o6NM8Di-b*+J zIk8Z(f+GWUY8F{p443qy9N|#jVMmSqQ*@isNip>X`UB}u82onx*KZo=t73Rh;83O} zWQ_y|Q0q_KXH_LI5C$(sY^zv75vK+$LESCcU0c3ju6`!h&c1QTB;essTdY<6vxvZj z&N1#M7`*SYjiH_FxaFL$OWJ`lofh-52U1%y92Kuh$5?&3H>Y0xU-8zAZLr8Ad`(fG zIdbU;P>9}<;2^+iVF&3mWv)l`t|)-Wkn+>u(txS$19!{kUH_b$GcEO*J#*lhn{i-o ztdpi$3nL<|LBCQ=XhgUSdRsI>6S~C0MPWF$3&`=&vrtgEe_#Q@S@?pW!aqycJw&~M zO8ff0+w=X}@sTx@Bm)xfXnsPu!o}OY`!o2>;coAZ%SqpI;4slmvS?n2_~)0`Jqo*8Y^>HQ3xjn}{U!zdY)-jv(qyxg$(o-F z+48to+%gI|hU%a~`+F99`EK4sy3bxAndDmtnQ_mZ1QQc0;8kUS^1y`xtLFA3qAc~8 z@AnhbrYH3MI_&9;gv%|XA=Ql+Mz1zzagF&~jd^b$ebPd9`NdPOq3CKJ;G%UZJp?(9UT!|03#TU9Lk}fk(A}bm2)n>Q zD-PTNzL?0japtJ7T$e1-+ZL38=A(dLI0fPU7nDnBFu#`@z?O0LA|Z|x%(-NTp1H26 zozZ*2ZK=M*SUSkoO4jk&)uEip_fu{X)J1px;XradNQ_OBtZ;AJr`UoCI5}arvYG-k((d=yN4HehaIlNfb+jCRzxXj!M^vcEDD~&B?KT{fm^-~7a4xwOZaLPtMA%4IDz=toA9)Aup<2>u3 zXhDxSqFLr28a6R+`zXP)z#tVk-QX;2Y=xymPyxR7Wb}x1p#G_^D_VOtPk8ns^1ga? z`RN9!!3c7FP^1xb?;lLx8jfz79c*x>9!6#t`HPiXi<3U4FwH``?*um-!~4Lh>UN7G2SDuusiJ2L*}T=Vj4J7On~X=NyA*^2RAkG*HUI30}Jq6@4{`6P~GAKe1Dd|LQU zfyk}I1J^_8cK2J4airOQIu<6rS=drWE6<$+9`hb=&d7rp=fM|E%~#0ot@W|X%2o}O zfVcO~C*;65nDG@}xA65p*p(jcK_Hu(gV(SlA|*&)pKu~id059A;}t#i-qXWmxTxZ4 z@V22KEwtmeiWHTs-!{!^2@ExBG?A;40J3oi(4p+(65o~r+0#d7$RLEY75@xNZ5|F0 zMqFrp{JD83B=s!Mp>n12$2Cda9M@mdY1?$9w%8|$@yy;~=#M*)(xfX@6M;N`u4}AG zMfh*t(|(cMB!uu9STW{%o3;CxL07pFeVy^{+`UuwjWupwKOaqwdC2UXS`KY`nsq+y zfLj#d(D~jUZ?EkHzvR9d#{}n(3O8#1R=&JEX&T4pi;K(uy|Q5@FY6<#dp=-s)lqo# zYd!wPz@Asi$yyEu&m8jtJvDrOJ0<>jJNRPg%Eax%rnM~v35w#Yqr5(0Ekyx(JiP*L z{9c{`XEw!Bimxj$Yr|>ci#@o_#{#)VykzJ-Pr1RBq=43@1w;_^S92wH-`p>-Vh)Dt z)#pBwel=*Re0eJZD~uvai@p$d38P8VMGPg3|4Q_p()4(eWQ8xJ$DIK$j>hJj6-->J zYHY<+HgG+~O>G^}$vd=`E*}^^Jd#_&_f`L=?WPJ} ztm5c8JV(yTk|T!!{D5Ih+{Qp3q*;T9Y=u8Xg)1YX@ESrxRjnosA;0s@tqr1yn8k)r z?UiGLy7f(Y^gS>B6Yn75FcY5L|6q-no;{+y3pvX19Dch_^j>Etxrh#Y) zj>}6nNJPx{(3T2*ny*n}So*f>Gkxw{#DgCrb#2ZfxuBctZhC3+eA}!q0lLw9o_ZYL z@PfTzPO25htbmQQbjF@HViDx_oH#{%pw{vw) zQgY@soI>~$Dlw;8lfN|cPu`+?y18rpt`S>*$;PH)1i+uZX$N?{JirE^wS`TV zj?)tnIA~Yn;vM4&Zv+T+)H33)ypdv(QZx^8WD3%Y1s1^h6Z@%pUbKl5+$(%rGBx4m zGiCWFxFc&Cxm(o7s1e3T*}epS_5k>^gA!zQ)d=!=z*7{^!k1 z96jUAq9%BJZ8kM!jyiEb5qYV8P+-E8QHF=|4U(CWNasOgp~1Ip`vd*2>O^+(-&u;q zO&{5=&IwmnpoPbDA7MwU8N-Guf^VbrJpT1Ul^|6Gtg%{u*8mCc@UZ{;eW9% z^Br7{A4=>Z%af+bpvoqvkgk!F!W-jS9azp=0@h{Q0$5bZ#&Od;b0b^S6y! zOzT7KUfjn~wx`6lzuU;RWN+UPg?WX9bl{6ETokWa_!u)?ND?RcVO1gVc5)q}3n5ET z<`7ilQ`fBQ>>R6NLw$A)tuc^e0qd0HqU-dAC(T`-K{|+@Swz-v%rMZ$=}#G&MMWM2 zw*gNPpckeyupB4woVj+Q=p^@kff=SNT@pZX2+$>t1ugH;00BEK^gPYQBxpBa6iC%~ zv$Gur>phblR)PdNx+5N^8t4S`sr=XOp1r}oyjzHP()0uD)^dRg_Z+X9Lyet=JWB*& zP(5E4PirT^TJ^bwix)YF^j2SuDRMArT7x)ywxwq9RZM5|$zx9DH($PXcb~WZ>uThH zn|T!lTE^&XNSZB*=O4n0cT(%B+W$xJInBL+m9LoPFxr6^@=wRqM$Vrl6}JAKIvK}P z!>t!s51`1`3k@!LVw^vZ{S=OrElO6;&kN z&p5N^v0I+{HVs)?y@KXGe<{Z8f8udh%lzYTcMi;y50EZm=oluW$5W_D zHt}CLUIcWTjY$ATY0)+y=wRb4_pw$~7JLqWm565|6IEB@T2a$DB`3ez zI~$r+5>2Y)0&`4rFBM?v?ycte*HxNtgx)&TSl5)T?Eu4+V4%Gy8jkMAVh9uz+TrZ@ zuJrMnq@;f^8KRNkB=E1G9x!BxhoB9k9aK04K1W6vGo%QIAtO*ShTolGcY9=}7Tpih zZ-{;MtRM>Cl|6c)5fmEav|lTz{T zNaM{GZI?;gBcIoJ=~Y&+y{&1*UZ)ZRN{%iE5ol@$VS>wx<`)qsXVbp@gO$91!S}x~ zxN}}1`e@4Y>h@jt-yNFBlCVuX@JmZ9?(F-I$80|pq?lh(h|+V}eMPFSS8kuN-jID# zx_Sl-Zu$EkY`*8vY+CRbxk?dMeE1&>9iP``;TyA(1e3({H8NKP?R~}SFyt&SX#G9C zLH57()pZfpnwXmkLs|9RU?WA8ur+~2zgS-t(df(JHFothal3YuEa_Aerq?QJZRKy2 zI=hPG%eX&wn?bz@!r$t%9ChA&1ZnLj494yIy}GjS7eRXo!xTf2VJj0&;61Tva3*Ef zf9Pdlmt6ebxwq{wztcUvIy~)I~X; zi+L{)A*9X1Zg^MTTF{99=hu|lB>zg|@tLIM zLnhJ|j!$1UrR`9HkzBEjRRU#Ck<<>hw=0d>fL#=F279Oq)ae&Fm?)g34u|H=-4!G7 zoVGy3DlC}$lobU`?dMK5=YH~0;VRco4-Ho#X@FwH?p%}^?lkpCeSSj*M%t!YC=}I_ zNTt^s?_5{SDgh%j?9{i|oQUK$+nK=LIrpc|eYZ&wq!P>je%ma4PpSGLrU5%ICJGvG z(9&;$?|6SlM%)Nq5tq^3 zcb`$4{f?m;5H8c5%{Tm+XT+(xZ|SNS>etl>NI@=D%4JuXPJXBJ!sOXDoZk9BJd@Vj zwj;m3J-sj*;0pWuM7+=HL$`&(uavZcEgL4k&szSrM&d-h)J>$zeFC9$5J`83bFe zL`73MHbMG78o(QdD{KroV{Cyc{EE4Tl=`V$+xc0&LoJ~mDd#=e{KdB+Yj<&I;&D?b z`NxhZ%Td-W&2vNXjzP{$*QFh=ewvede%Aeb zP{9f1WpP&GSC`;Dni|}bF1MOEqGznDUn|zAvh%8MmBVwJEh=BtII5lygI+cq_jB=n zf!=!f%#KY*&i1Okf3`XE;Kz&)>hFGRUV5=<0pc$7&o;u+!vBbKVUwJySiuoCNdXFN z&r=vQSX)uTH+l}VqL^)P$NL8L`ktNU%Z1A1S930>|8l!W!=LVr4|`(M5?TG!tL8&BL#9AVU7_!CzlxKIU{cGaU@&Z%VS zyUCw5>Jt!M$&Uk=E1R$P6|4EF^f)Ft`2^ZHcxOQ#AF0>2+Fygs5;uUe>@oB?Jc+4OeC8&N#xm`M6Dr-f(HMtST3MHUi=|G{KuINx{$GrZpyAXn0B zI^h(DhW(GXV@QQ|V2*C{lO47eTM$Ie_bZj-dZPC&7)rIN~ON^RQ9?IzoJ_ z9rz?a7VwShczl~f))jYI1$x%>qjm6s?Nlkb%XQU&xN{9mD`Jh4KkpprIset8o9=`? zJFMk65os;{ePtZvjill4oS~iROM(1dT`T~2is{DG`6bmY!L6a-@GLB+NTdNNuWLx3 z1@yECwWntDe;9Itb0u>mPO8(@Lp5~`Z@w<+4KBq@dU-`jGpdv3Dca!s0@7Z)M;%fg zMMtYa$KC zzj{V?jU<{2{AFq%3c&_rKeC#pkuFo z(g6%|zd(`e+WyVRXV{=#>^27qT+e$73NERJL|jq3*04z+e9$*nyV;ZzhfmL$+sHit zAHV=h$aEE1Q6w+woZ2FMOwkbRfUZ!tTJ8dKGZkgv1Wc_9L?I!*Bd9~Wl6kG%o7vM= zyE>H({qvMZP4?{^x-?7(0OC=HjU~_3eQ;^N$!!h6MdZU9$=c@jU*xas(}_p>yNk(A z{5c+aDRJ^j%CcH=@qQR4bPWORP8TRJ*ra%otiVTDjqzoqAeAxK88OJs1R2hNg|igF ztC~{<(senbOkS386?}oAXuJc)b!z)uSbngl+@>(>-ujz>ib~u3pj$!j;XlAlcjzYt zBP>*7XVr3onUqv!3nS43krLv3pMkh*#O#ZYp@rWpDP}lK>+f<*i#!@=rLj6;@yxUL zgZ;Df5evO7W!sqhKrH(unR5mK?J;|Y1qN2B8>DMhV<3W(%yWmNSja?=;vQ0p`PLzQ zvO845y~dhap6U>nO}-rhQ2Is--;k!C zGZ2)$0)0r@EXsr`NjNj4(5X#hsl#1WZu%)`kdu6h*L~%R=MS)Vt+LL z2$WlRMFSK)fdn_bg9PFfVrUrQu#k)r94o;a@xma?w1($#Xb0=CrSB^&KCrut9jQ6% zJrqJx{lE%>pKBYK$#AkAh%_;f8GazO*U1PLu>vrQe)t8^W;1Z*o(bbx!?@&jhHNY1 z1>U#Q91&%?gHy~Li8aUt%9JLYGec%Y2J=;EN4X|5`8xD@RZ;%!au6^6R84Q|{MvJM zQ}eq=Fhni!AQ1@jHPj^|{s9pkS*s#YsG8nX+})SPs5R?BMmOhG@ofS!=Kg>~484`3_}_Qxh(pzOyoOVLRh*RO!D8g4}Uo2=3Nz>rTtnW))LQCSMv2TZ4&x9J(7ycCO z=gzdS^>;B59p9AnZ@kpp17;u&MJ?4FQcT6-tH$_q!#nfbUuMuP90N3^l&uj{>EmIT zD+r)5E@HuNu?=eD8wwl;seuSOn+c~OV=*#sSDqLkP)Dg_V5jeUrF`p z-N#2(_M`U&7~%7~r>v+0prY%bY>Vr0J2*DhwRggwqTGM?gTb*QML&7w#N3+|mapXd zU#5uj%)mhbgav4K&pXV-AkUP8>On+fYwQ3%y^K}dzDKk%fGl4>;Uw-nM3EO*aYq_i zq$I5J&U%#t^WY)s=}cD||6GwO+tU}uZvOf`Yv@b1(;@gT$H4M=pUvEzB_Y#fmFrs@ z(n}e1f6;DEo2ZsKI~1hToV6_^aF70q)^L#inuUMRvQAyrtlY8hx{v;Tt$m1SKm(k| zInp#x_atvTw&(aQ1#v@GH~o5$x1AeT8^Begx$c3fX*GQ%YK`G@dKVW6 zLt5hRNCD~eKvvaKFjo%DNd29~sVR1GtEA17AEupQcxUF9^pZKcW%HA{kWzn5#GrA7 z_R9Go-C>+W3Gv2=WBmbJW!?G)IT)rJS)aaS!G$wJ<0w+lESu5gtSr#wM7HG?$D=hC z@=(Jlf=;n}K+RFQcREj@S4fTwD6|O*!WA7UITE4qCo$g)?r31|MO!d8gZMl^#g`S=zQCEeZBcy!f^FiMVcSEyIk~ zrHd7BAGKZJoZ-gN9b8=K=fzAl#9_5YS&-z`eNsE+rIejY z#a4iMMW5T!Nfi^o`9{DHK_UGQAA6(}s_GEIt7``8$|MC&<=*vt7CY$9i>-|L0JV|33HORHbpAStG^8Bt9ef^CaI=Oc7~Bv zT7GdaE+lZbgNBR!6_<%$8Ru&rZpqr|aHZ)lM@-r#9ah`Fg{N}?*YJ$*+c<}tI%b(P_jTd z$Tcl7&97kiluMZ+T&Q#?x+#N9af6?{()>Pj8s*BBf$f(1#aBrglR#F0vaV zZ+XPGFkoFNtAe!zhL|30)nD&6Y1f0x{1{dui&))KdjQhckh1n`A+RS`m7i028ztehJ)%!B)&Td=KCn&;DYcYbj{ zf4nB>QM(zOKQi2db!&UIfr;%rNpEI4$lwl$k;WZv#YmF>4~Ay&C`f^zDUXs>Z< zPC1Crndz7bno_F z1u2{uU2Vg<4757zioSc>m&2ZRjnS@a3VqW;k{Ju2N2Yc*s#g5?JmJM)m@Iw!G`JP< zdH2(Aa^tEyuYlRKgr>6pFx?c6cpECl?~FS&a7=B>>5cy_2!s2a`}%)b$@+^)P96Rt zdDfO?6fLE`Wjal2qxvn$1#8U`P6WHE8#~VK!gBHmvQR#o8jCjaG^;`Da8wd-0*i9z zix2Z1uj+jvTh-yj9DNU^;b&qElyWbB(*ZdsIerM?H1zZZ=$G`Vmr3zRMM4BvA3?Yv ztRU_OK9yRE{sPf()I=uKb%?Nsc^^r`#F&n%G>$&*!xc>)LD#+>3+PX{tT1=)&!Z^S z+qc9Zr=Wc{h99J|#%%UWj6bN@_~>L7YugpGy`00(zifFw<-fTy_JV_Mlr!n|b@SZb zZ}ZKW%Y~8J`_FpKyXu^2X;S{!9b#stBLDerv`c8NzDz-$9vF08(@j6c`3-&i-4W`uxAp5{kdMV~ z{pGa)O-Ij&kd_SA@qlgOOs8FJEU4lT_}`rjMS>X@)lJz3CA0Op8BRCbjt?q8fB& zC8T;d&6!@zYaSIuaL(DOI@1u6_;c;ZEdsTQ%H*p|z>T{{LafFPkeHX|!%sa5ET$FC zqj$YEQ%P!Y;cdE+GlkJp*=vpV6=u%}pXrNC@s6j3mHFH+FF9&9o}Y1D&2U;vpexro zdFR08AUj)5%+xn~-2r2i-1U17#S7*?@+NvE4Q|fXWdIce_n$aL*hsp8Q5Dg#+i!~W z@G(aQTE(X;#O()LWjGN`m3UnwCz+rJoni9!F#F(h`*6k$3+4~ARofPiYdD|zU3?*W zIKLr5{JTQb;l9Hc+JAoj^>Qv{!x1aASJWL~cr0Xa9R63iZXrbS-RKaXogIG@uvv_F z{$O+~&3g`&F@L^D`w{X@oO!k4fe(@+cymg5>idGoa@114xG?uq^%3GO{+8GWKXhS2 z^?$IKM*=(kVLX`ykg&xp5MbXNB$5+&apM>N6hjN_UH5(a>quJ`=myVNnS;fU=L=}P z@0ihC!<13Y;Eo}hBI>p~5Rz?|Of`_onH*Y&$KC#ge=zY7navA!j{3+h6$Qp;#nj5m{Lp{^lS%U{WuZmRdKC?YH~v~9 zGT@hGr*B7}GDI8(T*~k(>v=^{2(L7WJ9mg}7HQA#&L5nIQ5b#mMgnNj>b00fID2TV zhjL1mlzsCZ%>BTp(odx-!CTueJRLYRN*s`S^j!AngKqaUX4K{k?}_f^xaGd@_iO`h zzD&T`?PXv8ux~@<_Asx80c-IjSY6A1Mu`1d)_>E^7#g0amMl_;wT3;t88oHreBe;> z49s*I9`ZdRjY7S@3?FO)arStM5&2I#kzHxkqzG$eN5LEH+4JU+@Y#9T#ec9o#_ZfE zx(>FLNKRZ8Vm_$U{v4D7F%_`?fn z&PcK`%y(h*Ef~eLu61JyVFz}2PgbA(F)ZD(NgtW5b@tADO(f4IsO7_6rCW8dO;`6G zdpkiD`|m^(sg>e&9}MRz?`%{!YJb4eE(dlSF36C;h&^BJ8;mxzfuYd3|C_q?v4L5K zT!q6{{rwEvhMx4(s3rTzBAOQZPWHoHrx9X!5E@QxMRC~7mAHCmWr*Fqz7!uP&|~Im zN;D0JwE3ND<|`FGF^l$;w6ydp=xEsG``-TIH4vZU4Az6hcOx|f2uo#aE_3`2UEq zOcwlup*XztaM&)($cXBd=OLQ^U_QD2u=WozYyC{7+f%SSzjWeA{&$%7X&gv zfIrk8syk{#yF2(nskGhr1u_e%iq4p0)n=WL=5DHJc(TeGhmTYg= zcn7CrtK*ldxV^k%+_=IZ&OoGE7exZf+fP7oG1J+d`g)e$oLW{xX2{)#NAjWKjz34Q zngu%>dd#gaXUi{`U-?Zlf4aoX_!ebVCXyEC(5AE7sT1TLj4WCluW$lzm4Z~>k)=Wq zN_f=q;rd+HWBowEZh7Z{Nofte^rEtD=6bkGVlfsU$FQv#xYp8auQzB=9A}^3?CBqc zIRGu*I=sq$^j_=kI$Jx+e4hrY+ru&(m1>-flykF_XuM&0m{n$F|) zG$Q}u?(g!sY?p>wPYrX{Z-V|v^2 z(~yUV?|wZD1i^;0&`q@9_>}A8s?ZuxqX)-1VJ&i7-)rce6@a>#XzT?{h0wI&tFBu@xDRLHmluYIEm`5)xilqpn0w|yC+msA%N5>#xX zt76_1urXWp#vPk2DrrS~Vae=SuwRDfP%V62iSbVr2kf_}s{lU$PYM}@p1uaaX2qo{ z%KkYIq?m#y=Q#6gPui1V}2Zbh}* z`yGnBXso1Q5|3kk`?F-h>F%S^&qK-QYoJR-8vB?6y9%D;`D;Hb+L%dZ{a+AGR^%xnijyGB(a$3*u_2;p&m>W)ZJWJ>q zci%f7&EZxvZbQ#>RT_HF*d*gRRkrouz4hU|_%i%|eW8KH^RDK$8KcCYc`px-57#-~ zcH#OdclKXI_NOnhr!nIQkz@eh_J3Q)s(85oCVx(Fd`RG3Ot6CB(^WC_7<}by2T~%W zcR@y_o2)_9@uquy3m|Gx1_&kwubUCE6)3c4>?Ot8u-b|y_(ujn!YroqVUQ=A+EWEK z>;Mx)^aBXNil^aHk@D!Bz+UXm$>!NMyyY@_oRw|3`+Px)`yDEIvee_;?-~PSYFo7em$b|MV>QLP>D$LM#&f*Zchp5jbC4T=ZP+-cScbn+n znW1}@sGMpVzEXo3yQj=+j_}C!bp=*2l=IfU&@fb{+Wsiy8wTfeWH9h#9^~sn`v3%r z^QQ{Ax`~3E2 z7+gcDXY~23R*GCS3_TqJp4iPgh7H4$udDY*iS0hW$@;g1F7r-Hqq5fX=L!!uU9~-$ z@|yemzi&!xt%sBO8;Xm!Ye>y+`!1HQeirvz$5vkFc>C9FiWilhh(ByObJW&VLh3fBs_*(mF_q2jPqw`L3lqQLx6#P)+X&;=~_^=$0J+S}3DW%sNY_1)>r?8zw`6^l}n3TG- zuW0m%*^URXF&o4-UW-z@pRy(D;}KXU%OZnejIIvcB}mHW#9-bYx|W3{P%`y z4frb#mm`9|8{g@6&f@-e@5XM2J$0{-ezNxdWDhSC7(f%;{OJ%zB90_T>Mk=2OZ1D z4!X6Dx*4C1Ym94pjVrHYxLKnG3)qFl?JsCX=f>@@^IB%kX=cdq01c7EBy4wWxDQ#_%dPw&U5Az3l!}B=foi0 z7wrbwea>UBSPZxbX2H|x&3fif*?N6&Ax}M-xr?@GsD8}gat_NPkY|6nE_*76a- zjKRUV!8}TV!t>Wx9dv(_YYeDUUc_?il^H*V?`rdE-zV#aYD>phDqD&W<0y>DFM zoyV>?qGaII&Y@OI^hUgRSGa5Jjk|?VqDuT~>Pg26Z%Cbr&YQ|0ET47kn-xG9&GVUxCdSDUEKp;^-I8QNcu z$&@2D?h6duXv(!sC^!J$wvjxE?8Eqkudl85Z(w1UiDY3o%qgRBTMs*Bb-`M?{)u3= zaXV&qa3o44?Rm%&2d^)&jNG7rv+Gz=#a)?P@|Zd^_> z5ALz#_ORFc#R@Sqz_1 zMOaw=OaBk*$f*e;<`>s`!~2YV#zAeFZV7@h#lvSFKFfdfUF`kllh?l({VFpRgRP=s z8B!uWu*jl27F6?yXiH~6{>|If9HAlIF{Su*B_zo4Y?EhfR3FG6o|U#T)b1)CT63sa zx_mzD+MdlVyWxvls5R6@KVSf4e|>yj*b?##1E|lGY$yKBc>Y#_9@hm-wgeV=%Lnz+ z1;*)JE3)4l{d`|{F30_DpF#}Xyvb^>1#-2?)uAYOGO7I_$&VTfbfVn?qVN?34PB@I z;awJ{5mgpQb(l_ETtkibGiMDS??6-#rTNddca3 z2!LV0#Cv;>&T?H3?t`Uu6k<>sbqQnAn zGerr&XGNLZV)mEh0*K$~ny1|V5(;2eDNo5DJ2aIdG*I42Y@Z<1xq-*025%Z%Oeyv3 z{MGrO3C1S*A|GID8C`uzX3ikW6g+^I5XgXJehU?epZOWr3f<>QI}Y5G(}jXLJtrC) zo9kM*xP+nu0VBcK-i^u)jgPO(@2jZY0~6YUlf9eZJBUURdodixrE_3fH7D96`sB)x z618jzN*QF-@mS4Z6JaZb>e zboFBAm~-6o+4rXQw$$f*^jtF3x22stx4&A=91I}<-XL0xa1=Z#e(B?(Y;3o4k||Y8 zpQsn$gN9!6ZJ^|ytkbmV^eCkG>LG%*Xj99h7pe7Dt}I;YH7;Uo6}b9#H06XD#AMbb z+qHyQo0C;Y?Glzs+`S9D6VQFWylB89y9+l?0=A)6S2ijUxP#C;AkwM+yDKqhY3S}q zPY;+Y-q1%n=S|IgP;hCi=2TYh57|*3@GBrYUGYHu2iRXb(dG%hjHsJh1LkzTR8?XR z0Nh!+a!J<0S40hn)*YU}J-O3jd=)7h7_OPfTtUu#KtF!@p)dfwXQ+^ieooi;L8Y_I z)obbd->bopDYuhtGfXG}@}FlL9UG%Jy=`)|&C_einELiqIU->xrD~(+ki*`B8)Mhb ziv3v;C?u)-it57AgL!>il{m~*ec5yP_2Lffcat_tTWy!C!s)PzAiWx!z2Suu_OjC2OIu)jX604vE8aaZ z&`U_I`HNFWS$={&YUsK3!1yV_GwRfoBz!f1zlr#s!EwaJVx_2lRSC%dVSknI^Ap|J zhEHgnA<-ttllD8%JTwNaAu#$%urrIUVz5>cO`Ps`?OJcE|0c=|xfOpR{#c$g z)xa;JxDNJLO(gBihY5_qJsnOQF~!QDx8lG2?j*?)cR*^kPzh7t$SuP{7}!ofRa?5C ztems6s>FJu$LNHLLmvCckV;tAP>Z+Gr%4c0!ixK3&2c;0dQ8m!&UihAAGa4>xN{OTSNS*hIM>OaDSONP3@H8h* zMeYkceYt<0qRR1I2+jyv{QT+QhMB6uKZZXzLAxckS}ahK<{g4u z`I`w|kT=)^f$UbTN+TNL&oSX~CJee!=j~HCf&oKG;Hk|xmJjXi-e+%Ou@AKcy5neT z8-eP;{;ZepWf>Us;!XZJGYzrf%(E$nLxF7?J#;O8tIdGa?J*kW9Z$=a?xy zD!5=I(dfuO7+Bc3i4)3dr1ixz)LO+~;`idVKst_OmfqN){u;2)L1m}iDsZibUwS20 z0S~9=b-E~H{X$$p@?8{j+*+9$STz>IQx#@Xr~fk_O%%wKdlCDQb%Vsg9Mn660wJ<) z2xZKOU&p1jqKeFth>d1;U(t+3uB!Z%RJ_=gr_k?3Q5ce_LLiPz|gXD~H z>A-1j?3_eY8hn|==NaLOk%16lRLlp#N={sjKR?Sn!MM{F!$zJ-DokGuUo_ zZN$=s6?jiIl21?2$Ax8?!$c=7fjno|Ejqx(B-h*D2&%^@K{0MzjWx=%;>ygTLQ&&u zrxLH8Kwo*AAl&svh0Bs!afg_x=WWZZ)d%G-2om$3zWTc?ZxW_^y4c`RY1{_e_RYV< zDO-r0z}v5DXSG5LpzND~TEGlM5v&qZ{^bDY&Bmh@AowZj=pgOa zq@A`weTl=3#~sQn`gZ*vNw%<6Nl=&o$x~gl1AlLU6~9=*ZQ+H0Z4ny+!ZFDvRAs^y zj$f*QK*^WW!RTn-G@58oF4}`P>oHX&CAXp?OWfyvP2h<;u7>b>-)=N~FAaZz=t)%Q z4O7t4KGCaAx#K0(K%Tm6PdT#vqEne%_Jb=K{$dZN@>1z>pxYSMOfGH?a=-2m{JwIGksr%_3cgL?K{j*b( z$t}+o7XC9H3$l-BDZ3*J!<++^GPHAck<5X4Lw>vu;QM6?c0oAqtMQ}rWJ#g~aJ+&| z!g9gWd{(-C_hZmTsNhd+BOLTA4S4+eO>6S0wzk+^7g=_v{l%G>ah2(*MD7IdgD_XL zsgWM?4+a+;6TT2^nr?GpNzT?ajanjJB7T26}%cO*vTz8=Re%F+pwQp`zF&Vfg- zmAxgLV+o62${B2F!7lBKsZWo+mA;x2PrY$XjhjEtMe80XTrI1wXcF*JiD&M|S zSqKCac)Gsvc3{=BCz;}ZC!#QlH&<@1EU%zestl@dA=1y17x=xsiItT&+|AjDvX+Ur zvthbTt#99E1wKoD`*ap05fTd?6*N69fR%e`*Tp)P?zxv6P*utYw*qlx_4@Br%m_i>YJ}2^kqPQueI~MT|l!Th_5p)~Q!g z){K}jBeKtwqftHSuHV{$5h)5;2>=K7+B5u?~0Y#Dzlk=3b$7spYONSfeT#}SwGCN#ZMA< zStt-Xod}>wnlRlYu^D|-QkVuofId~+G3 zBjPBR8o(*+cgJ&hos(kvNntJ^%6IMn;u(M;F0Xx8oIx*V1OFCTH>jU;6KmXusHbJ% zRk8UD<1S(>LZn`n9auUa-xkQ;4X9)pm!WRv;RZ1`_ztl&X;=P_p)fmxFIx^xt@Vn9 zZimhKVBAZTYvF3?p{deD`-H;6_Pf4UB!W z-TIk(js$ewDO+{!4AnSC*JB(lgDGor_xUi6ah-YXsDNsfwv+?ocnhvoK&V8Iq_bbg z!M1PpOTOoH()P)i+!m8aU6IV5ctrITWQy}WPt3oa4I4yb?5(hd0^0$cBZN*Dlpk{9 z*oj-e#FdWP z+=lWw0W?w)PQI8w^gq}Riv!o7-MY5+BsQwAS{4fE_Ane@CnlH#GW)1;1!ZYCyLTlW zI)#1vW5wk3W^ErRT|O5hk6<~XzPI+#6uF1#1W2p8EBGN{mxVC?i-#I7twR*&$Oe^4 ze0a2G=QOMdB>6?^8S>mdcnA)O{TN`tl`Vr-D&q@+y#mx{6|Q8_+y`efbOir?-|Yw! zNLR!5%U^i;Q*R!Oc~h5n{g&QUN&fe9lyeNeA5Up1sE#AhAEQ`K=J%{05N1r+n#+N# za&&~2TH*E@qk)_lw89j6s1s0nHuTuey;9SrNr{rZsxWxp#l+71{#1EOy?bweoP*u^ zL-u8AR?96}%2pt&!df|dF83|kYOfgZ#~o_g5}R^{FV-~K_S=4JjJJqx9|THh1J*qJ=h zEp+@2HuK-%^Pc{{#xG4#uMM1>3Q2}RlQ%O~1ySQY-;>8tgmRVH3@{1V4-nfzc+xNy{c#_J$D>G__ z+KTR0a~Iyg{#G+kX+{}I1c|n2_OYJ3ZPSMU%OOJ+nskq>vYV}Zhi$z^9@{vm{t5OY zsF|FtaJlFLEb!&IivoSG)S@ybG((I{a4F`oEW<~v$ zh@4!f-22%B1N@ia<{KL;1EZsWMw8iZG%!2-aXjL#+43JV()bSB{^#nM-px&g0lC?! zO+~w46#J9r!YF&9z5XvP!7W+7s5(G=a-k3ob|LS-Q--PuASmv!!LA$3z+*G;?vsYC zG+|d@nM&&0b9vv0Y?p^5My}T=oBplh#c`4fu>@F9j1+D&x~3|1ZW^;<@ZS6pTHPle z5nzCbbW|NB&g=}RsFph^C@6?T7!*us$1BC%))xwgJuZ8HQ{bc_jQRN7tb~=7NoXueTxtFZu`5R4 z3rcYO#h(RA8JbWE6I6SbZquiqN-Uav!P%Qo2|F+={Ke_>Z;}8d?`6Lg+H&U7Cv@8m zeagRg1QZ4ZPo()Im1wC;RwQEn)i?_$g#^aEdK%m(ntX~2j}bkeed0iDnULhi19W!Z zu6b3ZFqkAToi;}y{|9p^+Uck#|5R4AJ6kbUWv|>DI{SrEw8?dWX2i5orig0Qv5dPg z&4m$_dRUS4&+7Rr??8xDe*gS8kQ1Wlx;v_pc&#*bI;qSA;vr+OmO2jI{?e znKolqwL9nVH@xZ+4bF8bfn=|^t=`T@vvq&I^V>TfGe>28ZY06@v&Z+quyqCpsmpJ= z-?eI)<;<=wpFa+lKk|Ly1B_LYT=TIxZKPI9P(1bCPXQqVeN$Wb(|O;;voPmOzgfwO zabt&+vm8U)5W_67D+1^zyIT=;RXe1mRxa+Cy}ko>sPjwcxI8^H2DU{P-)nzj`6TT0 zQU^VkI(0J!-O(zeihxdF&7ECD^UH?H)2|CydERIP}8sX zma)gl{#zJy@Si`QBFiynU%Qsv5lQ^)x)`4L&(I=>`8; z1!r|=vW#Q#qP`tFp)D4IDI3!v_zT30Yy`upE`aDD(l>DIEUPRLuesC^xoj=b8%sYYRHIp`08`-ov$?AT&VbTV-4Na-Q&;>-M^!P^eAGI|Tmpl0xDD|eQF+1pc^0vA=b1o}^|W;Eku zeO%O_5)?mt#m)k=tJ_DHdspZ306`eKx-9sIvzZIZ;{*Stzau)d@Ly2U!}b_uoG#mJ zT9&=9#qc-gmuqFs`wWi+%+RX1Oon-ULs9$d){=hZw5L(k7f?Rh@|hRmW54i%{7w|h zo)`;g{5V@qFzz5bcO{VH&b`C3%4S>J^Iin<5Bj6_0*$X66wWX;YGuWZ_f(?0YY1J| zZ?S_tc`g}O>=bO9mnZFa%qdHa9XA#iVEhI|z0>5IX9LFwENj%nQP;hJs;y8pyW{s? zRm7h^rq?>o$q_z z(pRA`z-_VA%(DDopD;IcxM&}&9IQ_*UM>tq=tQ7yn8CMGabD&;x1LDuuN{lvcU(gG z{BpJq1p6=OjbkH9-xWD>uI91>E;fl5Y5iDZaPIba=E9_`R`+VbFwha>4|h%lLu$cX z=}O8;hZ$PRi&;r7-qed(edK<~mpw3N@IAy%#d9tI$AZ*=GuGY1Nk7-P``NaRNx4;{ zNo|pVKRZ@R&Z(ccxactV1EvbTf2}*Pm~RALB%w~h_n<#>XP%-8eKFQlZmfAQS`bith%v!hjuMArF1NIlh8KK3QF zpS}BTvS^olMz7JI^XX%B;=E zzrOSRNx6Lj?qK_<={RYlfP$_QAdS6NZUOjzsfG0X#7~1gAMRBC4(c-&4=;khA_tPw zEhGRcCDpx&XA?5WUTYJoGr1HLmCP7VY73|555{0bt-I44x={ikiNZA@-ofelD)$tl z-EkF!RG3Ch?(!sFik%2k)Vd|0;&2r9cbG3~#6HqbN2lO0fPG73U0f5u_n?$i-{9CRK6wa&`veyl_J8vFwwLal$v@jJdQE^A(S>r|f1AbKp$-5MNMTRh}OBopm z@rLRI6+IStMT7;4S`KK1ned(hm8Bk<5Fr5Z{SY6s3k#aksr0%sa6G#MOYzFp;wYsS zr$k)^UT?88)n9K?Zs`Kp z*gXL!K`*G1d-$czovi7BegDmlgXKinQ(~W#4w?GRB`J)!e7x*^wW;Bd)Yoe!fzrJP z<29ei$<2mYO=e$%8F>F@MxwbjSD)!$@k;Fzd#+^W06%i{<(#&3Qxc|M*b?d ztMTfO1VK}An8w}T-$QIfE?09j_+MP%>1}7LUnk7CSRCY!P=C%9pyCJ9EKV#&Y(L6# zir3XWWMGLyu(rqCQVKMv)M-mr}-0T>;fJ`GfQv7rlKXR{|Jj1_mP z>>vYcF|$i-5lW*%=?GtoO-tYgRuFf!XE2Ss)?yH(4iejKCZQWO5j4*5sAFU#@=Zp}E-6hPguLjq*lE?~ma&a? z>+>zK$`*EnDsx>14Tkohg6-a7W^}~m zdRZCjNu+s4eyC6IW(*u;zdLLS0n-_G+!1J=El8hgpHIS_Vx?RIBx8a&4mD(`IN*5= zN@F0C`NxP|uJ^LargT>-U#F=&^i%#)#?VDO=r1&OqdTIWuM{;FFkyet)WOZ8yqDyg zNTpvyIN1*_a1ATl0}9N9gyeeQ3esS>a=+V|H9Ib1-P|$MP@sZYZtWd5Qbg8bx>PeR z*0uN@Y?9Sa+W-GpBaz0sgq)Dw^yB5>#8|{05TD~u*nuT|9tAmng^1kEXx;bBe8I3;*Z1ZDvArkd{$M4^E^&bFm=p0rrZn0bKoEQu~m;oIoQ3f9p>#h6Nc0=X# z@uaX|8LS%`ihS+c(nxxZ|ALq0<*%+8O%^ygw~SgRwrSF*Fju}cEVb&Jr*u}kXuKG5 zdsbmV`8knP0^^Od2D$r8`McWL0h>E{kJph@Pr6k*+jz5q_sl{bL={KT$?@CX36*{( zYe{Ap*2?mD_kD4oi3w{CT+f9F%u0_*yOD5rk%ZyhpRKrAcZt4pv3a7q7`z}U}o=Vt20!dsul8~1z) z6t6yTS;HR&(KdT|X@uYXdZLB-p9W3_nC~-}#ned?P#JXx+gb!Ck>OOr)Df~DFB$@notoE_D`cR;x z5XK#RgC98NPY}rOlUTsRSktBKL;Z}gBu2HQ9Y97UuPt8vy>y@3i1ZUTNtHcHbH(Dn ztOjWNSru|%C*@Io851c$JOU(U!1o4d`MWOl<0dJKQg%ju@$bJQBL|w_F zogPsgu3u|B4`4`hw=1>7fypQ_q;VXc8$$}fgf&XXOP&-Cqubrub-ZD_B zh&}3G^=?;sx3P^Ay@!!y4dZRr;cr|63ZeN%w)>1pg=F_k7_XB z*z}F9Xb;t}-2i?dSwAQMp=N@fGm+AZd7D5S{NWyEVp@V#8pnLw_|4GN#du(U<T%@HODFjgHfw#Fu9t3$EvHg4#H?8{dPeN2eJE zOZbEHd)PHCS_2hxtgu_>Siuj?RQ+}(Q_cV{rQigaZJ$SN$Gq9yyV^-L z3J;nOzHD#JLik=61|AN4A^rf0bqPZ~wmHd0rFRH(b$B`d@eV}zl$NPc%I4WkG6N}s z+r^gs9#(K1p4?(l$BedLUG49^cV{9X?}zuVPu`g>6djqWcOO5gfLh7EzX`-X0dyLTG{>{vmg4?Mc9SJ$&19}T$rZM_j|REhcA`)gKl7F zd(wuS-3+PeNuX{?{dKU+QS>?l>vACS+%OkaiXXN(IJ91AYS%biQT&a&pH*g;sc3FN zv9~>^+KNv2Lp$((ht`chtFI&Rg9kAM@-ZUwV`S!l8$-4yY7K$28(G6ygWP1O8HjG$ z@A7h@2D};bm(w|zf~(Vt0ai^TdXJ9Dl4_H8>0V|Wx2y$D7yq>xSROE<;1A)CK(?G9 zTkr*G5PzwN|C|drI~@`bs@nrL7Je9{(;;{L;!Ax!G~D2>IHUQXO*JGEnbfeyd!}l2 z&9F4(bn9@#d~@9x@@zq25A5$?7qJ1=jRc$;mNf>_x);aMhO{pM{)+ulA3+E>1MGz+ zdo#pv{?Is+m`J%eg4Y5LAz9^An@wf;aHOEy@q!s-0#)evHsZnhiKN!%(FWr}5O_}f zFFiE~s^d5VxOVH-jJ*0yZ@#8& zcr^l-g%Mc(H12W86O6&_!T;^J%Frc7yEFwa$x7(VNZk|47V93Ky!H!(eV*yi z{2B>Bf;L}p^BOT`N z0cQuB-&rUHMY3wA2@_qIy{=qemQ!~|(sn}P?Y0|i!OpzFWJ4Z)usb<7C(r&t6*hj@ zKGJYXSFJ$xQOkQ6)XFgK%$R81EMZl>MR27xH*I)LM#%k;2GBLL$9sUt5hH^m%#af; z_78U`gXEd=NQPX)sh55x8fl(j!TG(;4e^V71H~>~YO{^1+g~s&;|K%e`B+8YG0nx%&y_pN{3r=gbWa&FHKGQ&}C5?2YIqrjQ#5u{Z~gr$5+u z6FU&xOVEpPvK-DSfHB6pR~n5-4K8uF?a8Ah=h}_+WtX{l_)B`weoXi-uar(so$^tV za#PHK@veZoe8!am|K0#$8Q~9sl~5-HHc&B@VM>QYO>=Mpxd8yFYbn&#;u+gFM(Sg$ z!C*=?*=^$a$~qDdV5cT&3ri(d5oZ<4b^HaGp!oiaj_KF|j)lWypH{JKf;d)nb0>?| zXPRmnXt#TaP7Ly!UYBum`@-xHM0pZ^WxL=ob+lXPaUS&~7 z9c&4Z1c`~+^Z?KSjQ8(7Ao&E}cFL-FNER7_3peDVew9K{)`wpGV zm$z-=`T)Sc(l7FzzwGh*|JsjDYt|1N=KIdRZ=BK+`BMfI9G#jddWD6|Sp5{C8&yMG z0mfFL=bGKE&yUx`;r+#n+a7QI)Gj-FIMe*{rqX7my1H)6S%;@pol1oqBYIWH)zt>EHF5Q^;>f5sE3!_%rw87i+j5vP_&csY^)52&*q3qL zJ1YlaDSV*S13!M|;1<4{T(7d#)kb3_w`Ce2r?AuWPEAeeGj}A_qTXKo=e&^Ry_j7T zuH+y7?nk=@m-X4)2ZbWBD|ZRz*5hL_7PH{Ndx4uVB?lrWXdt+pqlUAC{s-e6#W_Ql z=(-{d?^e8-3#j|{biZDZ1k!BC5d?3W?0MD{yZUorgaoRKGg_*uUKO>+v;b?1i-fYv z(F}MZO~R-(_s{E55Q{hUk6TRWWxEPUT<;`3H0MuTBsWqVi%W^igYRrsKdNFm-KZB#t1yJPRT(IDJ6*Fp{=(8(gQIg5Y7IkU8)>k#;^Yn z2$U0HYS^5-5r0x=7i{)L6_b!3 zD$OwAqt?J;c3kP6a%A7JzGDM0K{GyTmJ<__y7!s+dzj~o*)7${g-&yhPnr1}*a=Ny z@U{E&+3V7O*MZ>!ETnVY+)|`ketW*~_jc4Vf|I9J$JZ4niHXJvPI2MxdTVGlI@a*}hGONh zzHw?=&*Mdtik7qn67}_Puqp}mx7?nW&p!&4(Jd-DPJSMW!+Lx<3Y+u~gf#Ch9m_MC z;A>%*tVg(RWq^DXyRuv|+MtVCeoOFD8vCY*^=_)Gw>1s9ok~ZX!`%lj<#qfRrfxg~ zoQ+q-|Pp(f$8Z*pON~%2hyE;Cwm08z~phi>gM@6@YJyXA;8zGC+hn0?&TvX zoIq%`|Iq=S^4^=tzXJ=@-U99lAyS>ZO({pOl_ zV-lml+dOEMs%^u_G3|FZh;n%{b)zLB{Y=r`tW8ip$KK!%q3Q_TdmNaWNgK2Y4nriy z%(&WjT9J(}XPst}9$IMn<97qUGo8lmgKU|kVw-?yul>A%w$05yrvsvcJmh;nk3Bf_ z=YuHbD zt2gC*VH**oT>NB%rQL|6UZFdYU#$*BTB1rbz5_J2m;>6hP6yB})0FrWvOK|aEiRW< zhej4-S5>}W zNqz#1j{+Hc(Z)0JPdFI-2vA@mirk=3z+sB-Fz4E`%DSWB_EDTxG58+HsdwB%vG#cE z=HhB(aTD>M7d^Ah`qK?PH;O(#6)*s zve=ydQ=@u1W_D3;PZjwUD%R!O=gs%v)!W?_FN^N{cha|UPHv^NBYlpsYZms`>+TN0 zI@o!~OO7)8{A{gZ$tQ}Fjgxk{Jxcj}mUzedaAS-z4cK%;_>Lb>zF27q5O!FjJqp|Y z5ZMDT<>2htDYl#*IlP6CNuRC8zu&-|m z=;t^-rcS!UE`d+{jGsnH5CR7$5ft_k9G(CmD+zW+8eW!rgk@gBCfxMR*^Wpos~>A~ zjF3xpo5JSzw!~#(poGgU=|kA%A&R~Can|{-Ff4K9Ajc4A$9s(OUYpqLrx0YhKslTb zQjnyX35n!ekoVZ_S1jVg+pm+y>4WKf)ro5~g10RbEj{V3*l(wK@!M?kSHt<_iGjAO z5#BI3f|an0<~U_WkHaiZ1O0LDIx18iL56AVHWt@$v#^0G3*t+6&TXjS$~O9OB;(zW7izpcfMB` z5N;D8HBfxRzn5)Zg1k6qagNX^i_~Lb2p}=-wMX}f0L01NQaO_Qp`|n-ZVhJ}eS?8k znso1~gNdFM_3#ga?#WSce{9rq%Og);5#BwuN)7 z_}@?s!5>zTw$Of7wGBq<7h(?z{3;P&i~Zklk&MA~m7}-P_rQT0FPs%r9aV+}Fi?AC zDb;*v?w%uz3kM&Zh6@O$!Z&Alr50M$DN4wzMU*torU+}NTbg4lh7!geVL=XE3Le?w zPkM63WVSW(E>Ei`YcbLT^PNgzxGQ@_|6ZYIWWVSuEBp70fR?EYjII=8?<=@Oi$lCe zulTW?pks!>M#6Dyc#<4LV851vqwHWNmqIy6nm)Hq0^;%}u3>Q7KllwmfO8+yw zo}Yh@KTST~;?pp{!!lBr~$!0s5c7m5#BoN}s*yGuS*JaJs$dN|F z;F7+^!+zApjJMRxTMccst+8jZG?iOc+KH(-e@^SfC~W}|)@(-i8{KbDduVf3@UN=~U84IXA(brbHEpx{F zQlgOd-j_9xt$l1{*Q^i4?A-jcG8poSdo*M<+E)O5=>`88YTPWu5+>C0MnuQ)O2{Yg z13sF9>gSe2UT|Rd_)3r`Tf5J-D&Zw?Hd{+Kv#W&iFA!Z>b7@78!?n={bHcgC9Jj~~ zcd4#_yP_%`(CG}ZM7y6aLMP<|t8PjGO7ayL<7a$8WLaZ#a)4WYq=)EAm8Y-X%|e;X zlZ>cHM=$F*M%eQU*Lx2f))+goOUOC~e&L>+IhGz_HlIGB7Gf!Q!hN!Km+s^H+;^fU zzJv_!#A?Rad@O=@fPqdPI1=aMP`8m(k1cv`>P_~U&0N6$HHNCs)UXOZ-!mDjWFlc$ z#j9gnqf*z;V_V2Kr``nQTwK_4HItNnQ_>&*dt6O$Jr|rZu~+Z|2q9nrhN31D(magV zjisD>Tp>u2#Q|A5h8Pid7fmprX-ViYzu$ibfE*6JfyzxP2caxRKVpUF=0vi|M2MX( zd2;tgQ{nVUOH|?MJAY0~<}A!{Bmzx-7!LjD4wP)lV>lm_s`%#S3L<5r_TBq*L(zjq z(Gv^uhh@O#PtG3wxI6mJ(eCkI=AU|D2I}8qyy`OVTsHcsvhak~yieeO5Ph%cNf>in zvl`E*w8QW{I2|BK1ckPcBd?U}b!f_UkqiuX16XDB>4I;UZnfA23Rgj5?^!VA&f zkkaGuEA2R=T#el=?s5BHM-XhhPoli}yhQO4{ofKN(U(%MZ+@iMYY8?&fipzN&O#k` zlI1f|+HRo-J)j4Ri=B|aJogLA=ykP9&DqS7SLGPZ4&JY!x>qD_IwAK{YwHihz!&~7 zowyPibgLAOHJ8AuBC{KpHn9!ta7v#pdQ%e;6C$A`@CUHtUEs~f?KDp1IabBWT#^YO`xk_~ZN0V_k$aVW_Q=s>>K*vcyA+!?@+gQgiaNXzXurKpyHL|Gv|Px% zG#V~?Ng-(^Yqv5^!+t!D2VzKRF--eB)# zY+d(Nli>8Qd}XnV?=vog^s)q&RVhO>>>PE>G1Irfxnfn3|4`~bN=7?^RQxn|>Pqh;)jURnZK zVs^+3$$EI^>+RBxQJf7seXzH&(PZrr_oQE-_@GyLr0cKI*|QBpfn_aOA1%cYBpWW@zo>mSk!7yYwGwnhCL3GUD}zTgv%;&B&8*>%s*N;8Hqg5#0`8}(D3V7H zO&;=(a21km%)5~HBFkF2MCJGsS!Bt zg*G8U);MKa0e};Tm1>R=7Ucpt(v^d)N9(;IRE4kWR=OrTzKluU=YPMkI+|hU;MmGJ zvk>X0k+EleeAxP+Um!y*2LF#&?S$seXYX<&9VJV%zX`S;=6w%+{J!sbNbJ2AFqq2h zx)Pi6W}T0pFNH;Wa(3bjQ<#Vq3g-YGT)Sa_YGQH`db_?h{GA2`)dy|;e&|&Kp1Xu(LCxC=O9_5eT@1HxZY*IlD z=Bj&00jP%!1j& zJyg5g_?zRD1TKnBfME$dRcjWqdrj20(L$MpjHyQ-8MODi`OUD^u(9C4@MC}1lEIMO zUw+3QKp0>aTyuy~;iZcv?!3a%Tf^#e#7H0WhK(#Yt<3qAd!b#VqZN`brV^Z8wQ%#< zIVYxkcO?H?((*I=!UeUIU2H=?xhD)Mp~OfPH>=GWt49@gQ|q34>cI3T-Wyi2%lZSP zKmLi%y}Non^{<87tb~((_0-3?n0?tPH#3F#fESqFeoZO`(4$pfDluZzNlB|-=MGBQ z7*@%gI4H1Ma`Teo!AJPtHtq#7`_6i5?{^+eIuv8UYzAAuH=*;?Z41OC_i{5Yo#13p1j7LVeymR*t}lg>ko6y z?^6qte5y~u{E+n=#W@2vkgzD#vijU7_1QCl7E%v@p21Pty|xRdGsX>Im|M?KVm649 z*o@8&p|(-4W*{>IEGAhob++EymkYtWY{dqyO<3v8*~}i(_<5vcF6P9AXwY(Gtq;oa zLOuvO7MwfKalH0O{9pWyB$;#W|Cl)(*mW+NDRMwu0k($U&piw!u*)bNN`Qi$)k!6% zbsWI{%ip8f4N$Xzg5I%)H&YiTZ<1os>#t+AeV})k9tn>&p}33%&o;3NlYr}^ENj-U*lmz-c$|AKFDNb2kd-ebV^clrroEI9 zBlr`MLXbjQ=;1Ddg!L2$Qk3ahGDeac!WMBzY`eoNB*>RQC z`IhvoXJD7r0fZOe{k$)ssc;Xnr8gf!6Ko`!mrwSrepDv}yzI)IhJ%c3d4!&6YpM=j zk*m*2$jib~7(wq}jZFIGT)dDMwidMg)KBN}89JgJ%*q_9sT?|b*&IxmG=Ze{0il-I zg-E-L_1UyyQBpeOa8|quGj(`cn*+UC02+4g7+sH2tWIEGv!Qo0?1C6$2+2ibg}9O} zg=k!8kMBl7RaZ0Q*mDF?>FVrCB?4w370>S};9Lo}|%+0TtUYlwXsH zt4%S0!hW3I`R?=~lWTFQnF5RCTExm=G{+P&rO*WUM{s&9+KM{o2-mT&i$fs_nT3}7 zo9G2o>4e1NSC3jtE#42TzvC~_N_n$u+Q0iGtfp7CZxM*b?7*pYIwzBZpVRk_H--z7-K8l$lZ0|_6vyAmp4#Zfh?t}SH4=v z3oNm)E77_o*VNqP2!ozKe^wiWWoG*%TRrHxyDwzxdt$t8qno^1(BGN?7#`ehr>!8@ zf%lY5e~PN#%(>q3tpPJ3gPtJv+@0L{)Q=!R@CjCiV*BRNqq}z42DBC^_S~K775Y^e zCZKoF;H=u+qaw_ZOKnH5C~Y{x);CnmIQP`Ti-f|g72P9Vehey2-nt^dpkCWk=>JLS z@%-ncwhzs)v5}2|2saLLv#+Myzw3*j-_5+n$~XQs-zO<5{n(dw_HHAM@%#VoU&*wM zKj|Iv0w%6OBHg-vyA7IpE;7wP-A1lRuuRD|>jWtb%k_|hFLXvZyU-Y3moJsf{PIFg=@-VoG$p#obNJo%vr23pM2 zbrDr%`w(t|IVVQ`%l*w?R+01OA`a_c>_gIFa*Fd=*#}kketc7#rQ#fvci_=q(bOb0 z>^J*+GNv)GGnpE~S{{c>3S@2s tKif?GROtpOyy1mO%?$KFMl7A}9Hg)6nW7tP? zUg+u_>){_?1s?pkwLuXXUteGN>Qf3smYKIA$}AGNtWZnQnEQN8PRIu?Ix)+;QDUb7S+<>Ip%4xuucquL7yt^U9oP{jb~1WY|f;baV6CNV>RBp|E|( z#>XAo8|G}v%mVEt&sX4R6anclz9^!6a;@+w$hri`(IspY@Yl*^Ak)fbG#;?gjH)#H zfYl(~4s=&RKMA}48#{&T*2|zI!!dB3V9k`Ous{ntdpo3U?#1E2%5cs4LNlG)mLhAp zD8cbK3F>{`RRDfO&5%vK}qhy<5rEF?_h8Wh^ ztFMfSNVL#o3zxrj^HVS?k)kS}Elb~ai@K9}QKXGa07c3@tRjCmk|nlj45e_6;|P#D zOOY>{!8yv^+3>CH3|Npj+Xx>fu&!EU_VIx;5`_T9YI`+^;U!d@%?Q`{#!SjmK$^F& zireJhnR7um$WC|QaUFtq8u`8TvL8pC|IGpcoh}JL$pia6r_E8eY!CT+eivD0#d9S0 zOf4pXn$b-b$Av)8)jNo}sQ3FJQZVCslT7Ghj9v{ijN%6#nuMsx{pHr7CveI2Jhd;Gl%_Yyn( zB_S}t@2zuv_i0wEYIpeUpun~ddk|n$(SYDeYHwy|d_Xl)C|r4JBZ5k= z?0S#KM!?_;)*rwze3wWG7T|wE?Eze&sKJSHIw{pcldH_iNHCLRk(n7Ot9nMJ!Q!Nc z>j7|aV5}j9Gs`*`5X!7XT9axUYXcB6ewPMtIV7hX(ij`#XHp~rs7;+0{s#lgZEN$> z2>AvGVU(WH?13k462(wA5klB%2C?uFiX=fAxx6uxA%z`h{G5C8cER{Q zAs(kaV;7Dyu7n=Zy6@lK&^1uL_Ut;*w(!r&`v;-kb(03?E%l8s4^eW4sCcSzpE+g* zld(>W!ygzW)NoI-RUKGW4P{t_>$qyd%Hh`w(;Cfn8R(WwN6VWzLpEb!QZ~0W^$orP z7I`gcB@XDK%e(lxtg>X3Cwk)8W?U{*PV_T1S45d)G5jE3 zeiTr+wQ&EiDDlE^EPgNon}0Rmi7nH*&vh_yh9jFG73HgPmOm7bR8{%)9|3L)?k?{s z!5<#mv4>?0SdwJXwgA>}_y1t>;mWRrT|b~cs=2B&`z*VLG-_n$z}c}F&Q@-dbyE1$ z^6*moFg^aa-}t{z3J@@a6QT2n#U3bvvDS-B%_H?z9-nvOJ;zJpgrVnbA7?JsZ3 zu;({XaBNhM$V@!3T7wuBu($XPLWOo(GyX=Kk5`knJ}FZt8O9}fO55I11>@Xv`Js`o zF~mFGexga@wV#^gHS{73BqS|A!SvTDCl_c{1f7UM{8QK+Yq|;iBa_(ue$7(W@RQ+E z^yv#k=i$=P^^`!Xf0>R$fgAE+DQOON@r#Fp>wg*fUT@B~&D$LHYD~G5OpVkjd3vqH zp#GGc;8x1!J+PCe1fc5Sz`24iI>{QgVUgld-k9%9`qqju>lP$NN5%%P9%7|uo4N9d zuSx0n0|B{`r=plhC68(Mf$(UCZshCuv6DY{y;beU7pZIjslFrxrXxZhcyBt6@<%{M zZ($jokkl9qfK#x`gCJ+}L}(ZF!zlMGd%nlDkc>(+giGP<8seu-GlaRitSh9=qXCUu zk;I<7(HD(5ei6Y+=`a-7L+LNcHS=6)^5%zei+$iVcj?gK|45k`e4wHQg|Jqc49(6J zgFjuaWyuXZQ!7@F=qv6q$U5+~$s$#zxeAK??XKNt;<%a;9Z{^-61SRV{;vKiJ0^t2 zx=4nO43N8#kjD739|H}ls}FoQbQmBhk66Y%%F;8+<8gr&qoDf5f0k+a*^wWDh_CV{ z=eZy2#p^sg%ks{Sd`E2u!aIeavXJ@HwYxrwbT<-N?rUk=1g=!w703}d_yR)*UM<+r zQEn_uT*tx1Z7=X_q%GjUyH?Y-ZnW)&aq?`O>X?~hr8VIl)~&naK)Ovsz#fESNA6XV zoR>nzwZKjx2@po^@M_5PEYN5Z_TqL7uNOFUsYo>O-goHWzY+EU@`#wOW*1(VeZhnn zU)tCKJzzw;uH=n7wbv1Qj&Dl&>zZ$tZay0tojPWT`0SD%5P4T4Wkmvgr}4L2+vT|^ zm)G%%DT)u3zS=7F8DXnDZEPBd1Ek+``rsrhMS=zdFv!H0J1iz345l7(5$RDRBpfG8 z2ht0|w1X3`ZeMdLFPj_GAvjvS`ZEXg)xPiy85UO_3Ol>{W;?F0^c~uw7Rqr^ICrZlKG=A9 z;jsYpeNAo1#_&t+pkbz6h?eKU9EOX6&77b7_m96n52|Hks$BKKJdJ^w4#pF9cc=kd z^ZoOW@N!t~9=s7-JtH9;(4MA!`HD5hEpfznuJ&u&3T%p7j?wj-bwl+{Y0I_m|IT2| zN+UJWjbZ$UC?7O*c<=$)^MAnBVhfZ&KPy}{WOpK08tb-^EP zk|nt)z0xA<=ny}(d$~ZtY+UAMt=`+3Q4d`mbZd$C5|g+NFCf$FX|G;9U13 z)FYRfY{d_>2H8GdFa=~xAXsaRxZD)H8eA0Hr&>O-w7nHqKY2ILf>`YL#N})yh|2IS zWQZUSOAEoc=KVVissy@TpXw=-zn99?N$xXp(Qm{UDVU{whizfN7z_Aq9Apo7cN}2{ zcM=~Er1@HqPp8BHU<(1~5J?b$u6QbbY$9lN=Vq~wpliyro6*0Wdv!=p`l`(ys~s7% zL{!jNXAl#V%}`zu2tbn{*ROneI75v0*vWeu#`C<)?yD^7u3GGXkvgEHS3j0i)LD@Q zkNqr={C`j>rrZ zUIBAozWk0M=s!7^r3d}?)K=7%D9X-`czsd%qq`MX@+(hEkjjSx%MJwDS3B+IFx{t< zn1qnk!DQe{>p3{K;qktkuWST9eJFwlHGvBa{0|pf zSz8N2-K0@n#R9Z4=WLTKq_G96=~Sv|e@fEF<2UVUqF)s24NP4BB@Y@9o;ZGh#&SeX zEN$L}e8AcdhV_|K zEUG$`Y~2_lJNq3KqV23xN65E^MWl&TJ{KjO;e9Te&X9c_G6@g-hf5`A`?%!(T$Im| z+ui5YzMJzqhki>z;W9<5+f520^v~^Qyc=nl@m|jogN6lBZU=m?xlKEs zedGIjXReQb+>o_x^6^yD=h~{Ea=u3gpLWQG%!ITHU)Z}XU2+TAy7JX@P$V)orWt&| zQwfxT-FD>N06Q9A8ryK*Ey?u&G(~}@DCjX&9ug=cp~rM^$#wy>#p@?e#;N+MOA8H@Gfd^5;hSzsA( zzRmFhZR`uOBu4{NX#<(R#@JnJ-Yqh^s#A@y1yagY3BUIQ;BlHTb2)gFv6Xn|Uzl0b zNWey_u0>%~RK-ixz2p--6S*A0lc4RUuXJ=`HHLY7=lkkjwHX^;a(zV5r;lI5-1wL7 z<@o%(|6Ywd=H8@D#uqAT-ch*dT*=`|QjS=3(*$V|lFeSQyojL;|f`;XjkP2wI zS3z;2)UW3q^K!wwtSnjpBYPfq1k;)!^GnG6(t}F#pHf%CVbQn2b8ZArCI`c}c?R^< z8~;;DuuYg2J%l)XQ`Q5Gb$IFo0)BU>0eH9+CJra}ZfaFQ!NwQ(qxE$g$CA@GbXbVV z09^eakvS)Cl}zQt4xRsgbk8wJkm1yY6$h=>+r+5{KEyhetvDTT|44mUO!bP+%B74GH~9!zm*^On(HTU(ebgx zWdR>@Vr7cay=e%H2a*)xN+%$&on&6(7&kxx^DemFIsJDN(YJMWN*IF`n6y4U+b;o> zSxuTV2zpmEtsA*&=(POPp(gNFZQVf_*6y#>{n=lq`l@8ik6jV7+JB|5_sWBlXXUOO zy4`Z{&mGh~7}$-MnWG`Su*1KxtwEVq9mrh(e~5}zmD{BE6XiH!u;=b0&x2Ni3XyKU zFjr1t_Yw||MJgQQS*LUOnM_fNR~LMSNa^p>N^4hQ&=c9(jDpU8VN$K(uV*oZ+y3fd zT2hM$LH4;q(QoR%a(z*i;Z{)dm- z|0B<2s&aOq+?oCzg%-6rbf>4*e=Q$|w|G1cs`He*tKPcSr>T2yS?Zl|HP?N&uWqKk zq_wJ0L4r1ZU+dB+Mh276x<`JoN@C=Q_eRZBXJ>NugQ-Ob%50d=$4-Ey>1$Y9QC^)%U`k(pRF^Kyjl z8ilFh^d+57f2Z*rBwP_zz{HmKCite$V&@P-nx4DQK%GaQ;dY7Y*pzXNV@iTRBKe_E z@`ERsc6pKEe_>bOp@kt@*I-ETi*uL4_;sRiQZ8vyKP89(@4P?v@Gx;GuqcdioqD+W zD)PS8lT0ltAs!q<_*mQ?6D^EkkAe!bEMfh(u8_fk`Jon*`d5?WrAVVfB}U-O2zQl~ z^(eLF7VrFFcSNDYi4Mdlfh|Vh>;tzj-nn)rer1Yv8?x9OAB{$rV@U?+jK(Yh&R0Y| zA+2%Mum>R*jjA3fnT6agF#ixrZ9Mz{xpT*pP&yDneLA5|P<|Kh)l_dUdh#L{gSZ03 z6nuhgF-YSTm~{8E%rM=XkAWEsN4s=?XP4e#t=Op`HiG4eJ~bX~Y*Os?3JRakO}TcZ zx$T>gN=Ol@ba5wLhVcP!+d87iBy>dFUUJS7~oGWdEq zCMF8Q_d;jITb*jie$}t147&XZd!CSu>l*I$OWZl4bB@cD?^&396z%Vs-AEkBKPD|M z(kG0GJM1Bt1>hD`(2$fugOEj2SO@#$%uv{3`dX$T_Bt#_g#GmSC`D6RZLg;8T>pXVUuKn0sCcPIYJcIL( zKEGRt^s+`cj<@Q|M5_)}uWH_vSxBL;N42^26lx^N4wB*azBji-r`;TW+dA0tEhMjM z?Y?KwIQ%S@I9r$baZHDlf|CWlcw82_VZ5Src^X^oQ8SCtvSj>?Huz>_Q{GEx-h6BA z<~a5zpt^KaGtjHeGutcw4DFaHQubn!l*w$|@3WjkXQg@cnaCz$D0{qnT6QA*W`~f)X|w~OH(5v zN6?8WY$Jx9267P2MA`QN)@k-ZNR)ME303+vu%QKa05G7@?zd@zDAdzJiY9iD>SFfy z+pmnLmvXZEkEnmEy#@dHQ%a$1;{C<|owQ!G2keI%>=@jx>vLPKU2Hpfp?OUQcI!T( zY+`YF+e_XX)@NZu%7+0DMkEd&2)02(ir(rq29>tC2V$`#*tV>A2TF5@{w0L_7) zHQj6YNV{{9%={Iri0Mf?q5g5sN&P6l;+N1{j@I2jr=znI!(ZPSH0yaDi!*g&+$239 z)|q6mPI5YS2rGs~LQx&3=M;Bef!qo9B^zN2VIs4vmn_)mwlBk?cg7u0^eDd+J9MPF6d*ATsi45I}2=OD0OaUdEu4q4aq@(uVso(4iJ`5;yu!u3q| zWtn7EV|JGNIA@xN%9wvD%bfd~B~*GPu~K7UhA69S)vBsLhngNAyk8E3{u`#0y?LR{ z+4qzsqk1o`nvEK3J+dvmzwjg!LqqNl&_`mLs$QSCJpSm&+v7d9d>wchodgOwohZRZ z9v8!3fC^~UCO6hu$cMW2%*=<1k5JNraiCz9rv1n)_~EnWIzMnW4`mJPfbRWQJE4F&QZX6i|lqS-mZYOdNkh z%zb;1tLWH&evEBWs5TV%s`kUkj}PBA1~m*e2d~#jN=DwSN}9c0aAo9=rvD>{{kD9W zVSI&19=vlH0K&A7!t&@lv@(AMT&tnT$@Ic?u1$-TWF}yFc2ryRat?v;Yh`b8V^&WK z9kTG}?8o5x+K;p+Q~Hy1``@>*V*Z6aG!nR#Q;(rv8SS48h}QN{CIswhT-Ppnq&?gF zj%U-3BdEe`F@w-x;R^1ix5hAJH?J{pt7sa3^hqX(w%N?^ng<%Ey=30?7|RTs4mUm> zc`vIGP%iyqDvitj7yGtn{G~x#gQ(NVyYCt!V95MK7F%Z7T*oh*;OCJtd2upsF>W+g zOBy_28@Is7$mhE-Lt-L%e`^pEPqm1{im-LQ!Bf~#Ohm7YMFRWi%YtQNZPqo{@tT_x zHuV9h>-iHGN5*~Zw{!C|W~)p1tNvQv6N`;`3ZM;XoPEf!Vu;vt8duxjH?u0(F`Yy# z!=%~OGd>KU<=SOO19f&UE?XrzeGF{1DKYccRUL6=j5^j7grPkNKI%zJvzB}dKOT{I z>pR2t=WHFYRu?^cYCEJ;UDG{Vm8$ZkxlmuMq-m|n?(|>Phty-QEzZNde#cmDoauj0 zYj!gQ%1HN(+j*wQzp(S|E^3Z% z8CM#c|ACU$P@}Wz^wm{yY;WiwI_Qwl0!&!suRO{+hi*AKVO!raOsMf+7}2u+(va_2 z>1QzAc>AU=k^ z58dlgsCr$*@!&x**l@_B;45hoH$3lZ+V^tbu*IRABfTj8M_qDN4^Ipi#0AJ47)pw0y3ir71wJ3D4Fc2L z#o5CM%nh%>)c1Y(l}AAcc{iP_{#spw$^0v?a&hG4z56{|;kmY5NAI1xxT-qy_AQL_ zf-MV#|AAn8xEhV!EhKuVxy3ueSDnCR;{;SmwRy0EOEdCe+L`OQadEhNnJA~)kyj(& z#hI*qPUxFc9ie7>myVZQk^cKfSM|@3QqT}z2a@4LpFvRj`?F@~n^_E5_H|}XDZ`wS zmGuh*LbxR{ptKv!P(+Cnq h)VOz~RX!jHo*X^+e(^e4nxwbR7_Bf1- zj$oJ@(x>`6Z~&E5MwCSQR${_z55s4`)2;1TgvIkr#eEck7a3pJ_>jU{apyohhg>!bnLKpwie{!KTQ!%9MD3q zCG;X@Y^W|<@kV=0GAM2Wkj2H^Uc4kId&d`82ugDog3G=VRRLA@Ij;YC)FW|bPG0L` zvyZNZ!0EA&>I!HZPk>Uw>H6T@JtWRlb!k~IAsr{q zF@`3tK^e?4keJYZ45HaTqwM#xLG|_de=KAIpFLc&*jwQF*2!G%;o~Cv zBm20%0r|~Chy;RNOpJG5L~GEhZO%eSrg<+`fr-yztCXJi@N8EdZ`HUJIxC`KcP)Na z*Dyom>!_;Bo|?@3E!5`w!#;VbFs3f=@e1VeBd;iC>o<5`wAO2@-uKV$B`I;43iBWc zOTjS%Mp(u2|ENTkt7JS5*Uwmf2s`g^*ZfHF_U)(9zaA906Yr1wDycp$x$mclpM74A zs)JFlbdhk&O>nr*@^xWlnRCtac+uZ;Ld*wSwh*d!Iyrg6&qio@LP9E=q9Y&t=c|Uc zL)7(?ZJiCzQ9f$?KhlD8VbThUAUA^TQeo0aLx}is)W%BV1wF^uH>!QHN5=ApzXf@D zPUqByMPSBnswnsMe|17xL?y+GvChTEp8R&ftgkl0+apJWd9DNPEl3NFFtfU}PgCaD ztL&dr>e=^%V2-V$VF6~wr^#1GH#9D5oF4WQ+9z*e+tw!I0W7ng@~6K&U^-0phblxI zYDQs{UZfd_CYXxdTM_JU8$JH!vGTy(r^N>=?Au2fTA#{SMUB(rhIAfR{xRj6u?Hd0 ziKUR%6$o95S?-%g$8HS%9BssNEf#8w`?F;$mP}oeMs?+dN!J3jufJCdwbPz7cyli9 zj>;Jr?L~2(o1%dBnd{S>Yuh%gI>nPtyH7O^UKI@W5IU~v|6CV_19EMicZ|kd~LufbwL9+1oZZ8p=D9xuO zUBjpTh{1@Wt7xPr(7+)(amnHLnHH)LmoRBTSGg~ddL+;pfgf zq4)5Sx$Kv6C-bA7kGlO|Ii)^3mu||p0i1EX+rW9V3`i^{U>?KuA9wR0G|Rf>3!H$2 zX+to%Jj_IO>NQOhyGe183UQ1bCz5cqIK$?lqSZ_0&G#WxA*{7ifBx%?5c%g@H=BIz z&C6!`O%eV7+|g}nQm;_>_^9RB8q9t^R!v>c+DPGc#L>I5W||j|r{v|tBG3N|yJZ6= z6utPFnLY!Pr}MxGq5{-a6I8c&BS9L9EXBUWKtG7~k|(4FR(+Wz$Fl_@Lgx-Mdzxg9 z_KeuSewHMJx^ZH%pGO^ZBk?SEP6zG)(Nk+W2|m@*;d&I*z%QcuKnH7`o@8Lxl_kE& z(Pv_dayUl|%o%fWv;awq2K!b7Q}+71w^79u!}lV3p3l-gD#rG~kX6JK@cz2V(BZAl z|K@9yoMLA2<$@l(FgV^dw9Ve`CbAC}dFN-g0d-#%lu+ipJa+~nF-g6_i^ItHRy_(3 zbokjNUw)Txvl%4*97g*GP+LD9X;ZKvWAI`n0<)U=IK-@;T0z8YU>ai zN$4;48(F3oElr0#ovrho4l8U7(f!$kH%_;mJd^kU^=XKw8bGYA-ZGAox@FNRb4H+J zRk7rLWxSQgiv~Y1g+bW26?)4otzBniG?k$#t7g9 z5E!wPF`YJVr)_2qD?wLEXZ>6cGtLteTE(ea>7f4mxIyQ}5WAZrXD}nS%%`NRfY2;~ z*~Oat8}X2+V^czpWSUXEBRDx<0>i-qxX~=6aJ1OKc^>}|`?-*hF_FP`pYI<<`jG*H z9IFKOZFnpC%`^trd(ek0v7Vvj@GY!>zR$N7RI`!tt&e>FTD<-~OP=-}F9&4Dkxqk? zu-ziaoMnj}!b(D8;Dr7RZCA7JL1e16+U9Krx`)IceT<<{3Ffq5;-2%AM##j8x7y9S z@HjLwF73?z%FF@%?u7?*+Ayz7=soY(n5iRIe;k1S^t(3PSbYj*)tDylIj~WA5LZU3 znDO-yt}eC;Qm@1X=rj#|iFlp(&?}QOYh6EEF+7CV8nA`Wf+WJ+l?kwFK0%4MsRXg; znB6y3Vepg_*v~{cT{jUnoY~w#khG={lEQQHC0+{SX#r||6Lh~KsI13TxY5bQ>)cSo z6lbQ=_41KMk=f%9VB4Y`fmKE`Vr?mqrQ^dXCY0MEX{15gxAuv%onfkE%oO2tWHRekF)sGi(M_D-`)(@7OnE#cph zNy(+X#idWXj#hs+K+<@4aC6FMu~D_JO&Gzn#)gqwpT54&-d@Q<*DOZC@WtEp**{{gkX_&oyCl&CdVL6bQYSH(!RgD zdgcpruUnGsVhFSmZ78v**`HKA{%2*qm-xi@g3|jJyjNPY5BZp>!Z`0t4uX%f^91(H ze>Xs^8g>>b3TE;z6b?W>42w-UsN6`K;oiv+{lHR5$9aD_^>@UQvD=(%ateAM5V}UI z4yN_x@77dP?P(w0k7#J_n|35$%E?tTMI_)Ec6qqlviRxOr2l$^G1K0W=q{TBAOqDa z#mK1`a~U<&D*rOtX8iDnc91_S+%|p0zrL;jW$ocVjXJXdR-RH^e8wBDZU2%-0x-G# zaYJ&?0rt1;qmeC7d+|%-HVBzZ#XaL%{#)@cEJpFMjkAZ6oq1;QaZ96`aab&W&uv}> zC#l5uWPbBFxy*Q#?9(4H1L=28XAD~k%<>f5^7QXXRiDz|GZS*aVxKd6_=LWOdg;QW zTA$KKv7@l!TEve?T(Zgh<4%}y6f0+3qOI8)N4 ztUqu19?W~ZP`qL1-v4D_eaV~aU)X{etgv`{$9;oN28dyKpy|WmLNX18xJ6+49?roI znc^4dt&_Z~^7fhc4DJqbQ#!fg%*uq7Y}C<;N2KfxQ=PTwQWUWZ0z5{xFkCr1f*plZ zXPWMfFo}hK-dOlswD&`~w||FqbH^%NwmP#sxnH=g zEeY5Ce26z^ZNM+YCtNJZN~3zLaWv?13nd%nb2ZfJnqsK`x5S{QM}xWKy?NGWu7Y<} z%Ng0pTHWGfiycQHc!|Ag@-j98K666tBQ?4<$vFyKxn0+w(`eq#Vz)Wev2V)5fo0k% zh5K_wt<+RKKf<=vHuHDinCO8-6wy-9747b6$|t2B8h*tfCHt=jh@NNAlUt?dXJji| z+9W&YM84_x89Es&cJGMgW@i|Ts*fmI>0}w|{1xF^qka7Rr}M>GPk(78|G=Xu(pZ=h z#S|tS>lv%rUAcU7T+;DIdvn#yV)#h)mp^#{mA(=mJiaq~`s=Y2CcPBHaWwj%e13fMK{>d(S&m1n~8r~Ph(KJQ(y{MzS(;Ep{| zF>@qhV!v`ODQ2x@yN;h?{r6`wUiDzpqUYSaWfxN8JZ!&FfA-Z!^@EC*!$!xpxROLp ziyg}o znfkPmCqp>bc|NyG2apQHGH8g~Yk737UygN~KpDaZ&ZT+LHQL^^~c`Cckd` zW}oMNHm5Xcid5W~jj{)oRyC%$$Fcg?d#aX*?@mq|Vc9XQwRxwhvC5THzLcNw1Mnbg zH0gn!c|V@M_{vE9^lF|*Ggfqhu)KrYc(-1}L&O2#_4W{0yUYEE1@=Y}&l5};=D^D!v8#|8%ulZk-3+&&r#N99N(1js^nF6`M zeh51TTj%@(!{N&=6Bht9oQs8Gse*kL)O2Vd)^FZ)9YuNqKw5oL8cj=huo}Ym0x>H` z35*tFm<#ExVGO{5`-&GLq!O@lbDaT_NtE0xv6dRj?Zarz`ONz#KPTKN=ll7bC`9xI zWT?bw2d!>Wx`GT-Bd1iDD4BwwY-Oy=-OMmYbHC8u~FLF`VAGJ zZ*?g2*aKJuF;!RSKJDh%22pZ$iO*$BS9aINX_Re>7PT+9FGyYhipaHY`jJ(?c) zE61tny^l)v&%D=KhxB0^xHMt|U$0i?2Y4qE&b}~K6{Z27=j$3oK5*Md8Cf3`l3yoq zJ1k4?{E!J?=iGl2)>2!s)SL^)#V9UaRe1wHgRlc}n{Dba4_k$BomuU=^Hf(u*zZ_a%Nmoasw>~xHJ34*( z-Jwvx)aD3~S-#OSkmySgOab(6UIYb<;5v4|()>r8?kAI%4rYA`9XqHRY^@qW*N#OfZ zc?sEtfKo|*_mL>gh4Z+SZO{Y_QkG}Uwp>1Bk}a)$Kh0bjZ9Se}{IzS^@HErpMG3MX z_vi?>D{rI}VeAZu3xt(~?p8#{ z_%Q|ndJTVoeSulJnmU@_kFxyIn3Pi|s@X^ty1FuL|2ya4g|qp~*IMH4mpqA3^6=(a zbejfJ#u%0d=L4_Fxdrr_gd!9g$H4k zYn1n#8Fh&ZsAaV%TmwnGEN?Tf_zu@^1miqze?Jus9d}#txQ(I82Ix^63RPMj=UqJ( zz}BOXM11S43{Auj*X3G89Gt{5EM=@&cHVqsH-I{{y}gB%6*Ef` zL7lp-c1<0KWL%?GX6sQV5{PYgnTj1ZgC;SG{WvL;Lm>a-eyvlQ2{p$ z9vChdUrWkt7v%K2)xp^ZMIE5Wlxh{QNi|ILf)J2|4j>BBqTc&!52=-OzvX~gcRjxEg)A+6HIezyEfH&22c0bD zOt{G#DaJ4rwz~5Za}!rLcKxUe54A!Al$OIcspC;o9_N!j9b5V1enX8n{H?`D*ajVh zPy(&Fjg2Ve);(PVYQR007`OzZ^W{E{0ds8>@&OQDDfY?x!#|r6AS;6B`r8RFR7!v{ zNMu>s4vm>U) zGq4~EU0S7Zxi*{%$Vn5FvMf5f`ZckAq~vXDecP}`QN7cZ5XCQHCyIo@{H7CZF{)rP zFkk4lry*-vkpjxuWB7v{`K9U1r)Av~cxs``SV^78#nmb%(c5PJty^a)^iG5-WvOI}rDsRB?|MLQBB3M9^m=qO|+`}VwzVDuTL z{O%$n>%kpHMq-Pp%D2$jt9uJvnHrlVPs}TP>t?MKo(A|S=lX~dg4AOb+^FcujvCkzC6)@vc>^} z^-D&7r?GZVI6C}j+s1mWOH4C1pKDW+Eh)O`vYUYw2UFEnE*f1Pa99y61MHMuD&u@B zU@bjxG|8g?zH4&T>~BTtM11J1$sBddc7hED4`?4CO?$I8WM7=G)>=1 zjB}1a#9k9arV%NGswE8-xG+6i?kz2=Pfw&YqN?4lWt;x4<3?-#0+94_F}1qb9-e)& ztmUYvN-$YCT;I;pZ>@Vs#Z=K9zy4BNR%vOvX|JhN*OH`GpqD3K?UmsX+pla5yJ}y2 zgP-d-|6wPkz5I%HFraHm;qTs^&vTA@YzhywKRGacqKI2uav%4RxWuy!@|Cet^#^m@ zP4}Jp{PF@yB(+fBQKdNFneEGSPURq00q22ZoFh%HZF=ky!4t_QH-xWL2Qg2BF3uL% z?xmt#(FZZKtBluWBA8+jAe)n~(cg^%!Rao6Rd<=;-ASWv*P6!~{NB?>0!tnj8k?iW z4mY%=@Gu8C`{IJ;(Nj+-$PP55%-2Ooo#yMEahsayC5m!1I73L$Rt49QZdY|^fhn^& zK8P9$2spP^RSSuB<_vjnoif@IopK{PJGOoP;!juCQUG`x&0*cYyu*WRnL9WBorY z&LH~AZu!43f07U~z15HkpFvuTVP&G#6a7~R?Pr#7{A@vVF@)~ucgPm4I8Tu!N&9ku zX)$@z$=o}XBphL9WiV(I0*lxJV8P8arW*jmB7kXteU^zuz#)kd>Wkcdtu%w=WD{A; zW%6Z4e-|k!7jRP5+0qnwc$b3o+ZX%CT@Nr63eXBKB${KC>j}OGX1`Q?*X=9Hm6%89 zBBru9{DUE~1t4cWcU83PPxkq^4gBP*zo7M~@CSJ}z4taZHmxf^oD` zu!S!Ra`qXs#BiM_q7Rm1P#xq3LMLLXc++z?H!Mrg>Gx^iKvu0FbmOxmaN>W$+Egww z>=N6hr@bBbY_!pnSJyoXK@R6a>Uu`-r@nDYywtBmk8~aye5zoR4?+;4(PPnjXg1JM z@=6#1+hNd``(_G*+$;rf!lf;0Sw`q&P7moN<_S11M0IUa?vs8%56?Vznf7s&V2mv! z{!zXlDf6cM7#BMQB%i{u{L!kr1e`ZI*{$B~A-<2u&6Z_qJAE^ef#L|nCJ7rMGxw)D zP~+QhzthtaWy&v$CYME|WpC7vyllN6QD}7+_V-^{ELl3*j%5t;dEJ;T9oelWCm;tJ z8jhJbr)2Zv;*N4zKv0xd@Q{^QpU}yjo|a}{lLWA6rUrLF*`ddE`GhewKE z#QFRL)92&YVX%E~^tVs!%joul7u@qNzoE|&zw*r#hA76rMJ}N zzy~by62v?3RN@}2gj1(agbr4b8J?xpkfFa`lCS2V7{3_m-98o)WR$w+YhkwZ_fL+d zT-$vYh83^gc=dPat3=ugTYSe!euZYSbdRY zJ$__S-<-LV6GrBhQ2$lt<-8Eh0zz|YSip@iV;iU8KkIA$Fh;=olYDrZy4}I84Et-F}a_PI-q%zRVl!=Dle)$-@kxPV~aXqy2!|0-x5d2X3&E z(dJC$N-F+;C3ai*HAeR|=ggBc?z#X)caN1x!= zpGkPW#VZ%D%^63Qxtl8FzK-^SDofbvOuY^kTy|OFUzh+WwCQlnT93)TOb^bd(N+fv zO7@F(_fBi)qtDqsP2pQt&hEe3ALx*6ms*;1+7$jCciRn84S$Hcjo%9qx(xJIK#inL zY;iHDoYeduh+gqY1f(8 zBMVE*d<-@dfO)qJlWaGt*|7Px@XUjdf@?;bKjOJ4ZjHICVa)o?QHT)6Os2gi6wg7A z7Oz0z=e`+fQ@$X(zy|4+SRDxXci;g)sBQy{cjyb*OZgnid}dVI>SdqJxuwX$m*uHi zcVXL)OOH)D-QHhlqZTnRe>m`n0ZaI(pP6idML z3}qho4sTcc_KLYT!@E$0(bq!BxJ;#4=R}3*S0s`q3TbVq4TVx%be5q@3St8)308U zJe%vf?=nujKNXAFN&|7z^b`dZkUatNRs7nunc=BnDr2o@sWu`@dIs%;swK2iaNA~a zRFqP^i*ruW(KRK{BTDxF@K3r&4hV8|Ds<0*27ktN}ou0vJOh>?uWK9 zT@BZHhsFG`7bX79l8%Y1-z+sgeSUOaVIhy-wu5JxFulMLps`%A-HKg$$cEgm0y^(v zyFrsS2Ghy#why*Av-`FA<8@w1e1qoY(UK?rXX5fGMfaizKs zPpuSu7EmqxjhE}y-FxuKK{#}bZzYZ8#`#E=K+YFTm>e0x-GWbzbnOtkszWwGR8m?) zj9tHk<3Fxkx!1>x0uX(>A>9H&Z+uPkp@cU}C!xLo)TIZG<}qGgw*~Y}Lcelt?_u3u z#J>f#>v|Solq2uT0^%0?22(Gdlv2ad#ts)AFSrK!GD#e@WykM3p$3<;7<5Vl+LkY9$-FuB zCN@$Jz$KBAU@G3#Pg&@6OU53DUeIFBf4jxlN@$Q>4dfiwKFB+9Vd}yCoCh!A;Zfq% z4M^?QRT*ttz7cX`E?YDiP=Zrh_2%8cI*vC8BN4OXiKIuK;b~#>}H!hg+-Sz%^G1!zpL-uZx zR-=PGx9+_W!Js1oFPdH%-U;V&y!7h9i6jr+e$kJ^;$P{{Y`I_m6>Hxc0qbMFbw2R& z=~N~=hAJ#DU@I193J3kKb-s~_bUD!8zR$ zUWDl{tnavIxo$7-2AT31#eim+K5{oOav@eoexHWHn$wW~#tSnSQH!?m^+BNDOd4oZ z4SjKn%Hbz$oM{;I@Z0v+P;hFfyLl~?$Y7reqm&LEOuL*~PFM%wk;ObyOY z^x)&CzveON;L@{?=tWO)I}i?f0o*`&(SelzV$m6%#_fZx!)H7DjT7)M(Er>{=k6 z!VpW9#OOD@b1H}ZznF^BJn8qaZpvv=I{UV+b zOU<@9yZNF2no3R<}fW!oz;xJL_&B3|R&ouCzD`40tS;PA&6>CbP_ zUMfEB+&#vnuc+VeN_K4vnp_)74=dVZY%v7nVIx^TmreI#*fE@9@n$0q?fPZW-b#5HQSV37B&nMgVEQ9 zV@20DIp7C=bKyufYm0!VjcRm*)bVU_xRcMqhux-jotct{a%YQ|WRz1reu_Nz!7FLy zo4fH?$Zz4jusySi96@n#MBVtdLKz2&bbnlCozElZ-G>_H*EyXNI#cuqsC%2&zP!QQ z`8Dm-m~{Tp%z)a(!*`zy9(=09wRoDyZ~=&gn9ewmqR~Qu_si~$lU^sKgEqxP7hZC5 zcABdWsr#gB+PR+jSAR{({P3;2>fyG08Q39!96~08W{6uiTnuvRaeJG!@}EpKr^Ps9 zE{{o^CnV_N66cN#jmw`!Yq|o@UobQJXhr9n>B`9gjotb*s&p@>xvSH2j&dFDk?9&VrLICBrr~27!7mNN#dXoB~tf0sFRx(GG zB|7+iX{ASBiK3yUFoZsXG;K1}+yevv)@$u3LDGTc$fQc6OHS&;d*3&h)OWGFw6MuLAn!0u%bAI{e^FPD};Vtv!iDY~o~Kktnfmp8qJ@0EN7rr-F&qBjChCsz;UyTu{=MX$Nc3sokhD@I}e2+h5Yh?Tn zzcXK|F(Oe>erW#1LPXo&(5Nryp3QG^0v#pSKyti_J{ zE>w4`fo6q`r>7l5g^>coTdGrj*L$W z-vy#?{In8GWOl2Fe(02Z!S)0iTx=14fY4}-ppZHz##i%&K+CRs2a)JG8fGlJnxXJA zEK=c*{;k%>QanteDU?Vf3V6F}GRhJ`0h%QMxl{K(MZbdRT^u>4aP>4p#Px+2dq|w));qdVu$=-+@aOoj@4L;cb-KMZ z`4D|FoM}RLVidE@a=WGzmOj8^jdaVs@4KF02G&#M9nrOplUngc$N#!Ok6fLr|IuUW zCe~}an#u9})`Bz_R&LLeCW|+hayyX%tr|WQTmteyyD(c5N@mO%Yb+F460f`@CPcdh zc?22HGNaP6#4$Fu<7jI9&uXf{Si_>6=VHDsyieKrsZwIbuv5g{J969Fod%a5^&NCk zo}`}3^SXNasmuExJj*s|_$JaS7jQ|cbdXXDk1(xklBsS}Eud-Fot2E+E8CSH>78$p z(BPI7`M{ti*CgY$#QPD-m}Xd6(}eC|8m3gH`b+CQ=f%(EMKFKt1lTyWrn@&O<11lV z(^^yz72l?Jk$(q~5SjuQj=~Lo6@QMti@7<|xnpioG(aFbf4A1?tHo|j&kP#XX>^9` za4|n*KM*&x?3*pAkW%lKTKDvn>Q^dkeFVP==`zVQo0Vm%A6+Jz7kWvw+RQD<wqUs^rtH^kE)jiM1{AyRKXASJbLjT^-{}2HP>rST^K&u zuuyn1Gx=@N-W9X>Q)%^smQ4Y&MkOPMmTGxUzl{!7@xu2`N-P7&6aXhb8IHNp;eD9LeK@fgXa-*u0V={I?>S;TYb zNyA_5j|+S}Bkf3|JF%LRzWGgA`oJHMVQ29kbF?91z1oa!leg-5^@&q!(-X5*^l@#a zT8U#zmiNvciB%Q1wd$ifUQq8=|Ds9PoqS~F6R!8H=jYaGs+B$%k{^P9(1!&VWqhYS z;@__CnARThj1`SHxhl**AV*U=_&oG0+_f#FvuF>ksJ{leGfV8xhXM6EpD{*dVsV3} zv{kfcoi;KtiqrfawE6`W1F@jcY?hAmM)5!h;u=tkiWkdHNtG{jX5$j%z4zRzCzZZNAR^VXGKd+d^lcJJAM9$qU9j|l->0+UcB5hF0)p?Iq;C6 zqv=|ZO_yXAxilY(bqRqww zG)%e!-e73G`Sgd4f+*v`!}X1=69jrqz^c)hg~T_@M@i=b3jVppQ;nFz)$44fZ7~ z_t~sVwM#vr17DB96c44lvlozg?BB58`__4bx#s5+K4rc;qnJP{yT|+ZTuGkk<4i6u zt^t@hESE2~>bh{8)s@a(*qIkS*;X3aR!UEvWLuF!$b^^6rb1_K19(ev8DACw0<|>k zj1-J(B+Iq^JI-i``&4Dce(?nEbSZA>aMk0HcO}j%g?@i?@7^)4oBHwqs|HIl+M)?UvRuGS zO{8&el8{#Czc5*jykBG*NrWUYV%M|dQB!|7BjK28nT`4zOWBgkr@CU14#fFtw`TOJ zPYFQ0HT>5;z{?w12QbShH_BGIIVT|-s#~hejpyV8CJK;e%x$b9B(dE2VLUyjYDZ)1 zLF@T4?+W3XwZ^1^qrD~k0o`4uNLWW#gs*UiAb-cUsdUyo(-W}$?!5bNH?2KCu!sCM zncVP-yGnMCd**?(gvSq1m@ZBJ8;RWMN9KX=LhL0%_0vv{}P8{X+6X$O6IGv_-ov0M(K&bs{o$_n1&7|2TfiF($m@fwqJOk+@z zB^OyY)%!Ch$8r>uW|Z(aLc0pv8i)x;EDcCy-ha>1WZd;^SEMf^M{(!n2PV_F-J3(6RnOkr^9ga_{*UjuYp`bzuV1s-)^#!SMK^NGu;`l3d&01Ba9etGXcZ8qCWU7<@fA(M4AY z=n6CU#<$QZ-H2A%2%ic+NyqfUz2kqh0!%+N7sB`36#IoY4d}iu6WZ;X>bv_)8-|U= zj^i0uNsvnaYHl?qi;C+dPBw3*jhdWT!ip!GGk!GT|Bs_;=>-Lbc%cD3Ov-Cx;n?S;k7o855E;}onwnOg`s%) z@??S|U5dPNKF3?}#j7v8zosI#Vwh_$`&(^huTWUi$JBTjuid%PwGh)V^HH7k=jfX8 zkf@`58&^^_Hyxa82>g|Uy=_H9yKMAZWolhjG#8AEd~QEK8V!R_f*w91-LjsOr-Skh zkYooz|6s)+InbD{>@bXzqeu41Qeqkue0vD>2|bwqwe0-q)!~PnI`MmlwTpeN!}U-S zOWAWr$oW6&D|at>gr0kH#}USMx@_j0st_%{|D6Eh7=>*XEZVo;{uFj^H7bJjB9cj4 z0ojKXHqcz;|AU>Lh?a=WSjopC+h+-`W3bY~@4<#l8vFm1R>NSpSz$oKJSzZDc(=lh zQishewwG4R)5xVwG&2E1wR8^pqNieeEGi8q@7aDY%A)l55aQZgH1EdB%*q4uOw}o% z1fw*e@6rMn3;*5=oC^Qt1yc#k+HJ+#82AyUSS;dMEuee=sKxLU^5_nO1KWi{zWC1C z$>=A`5A%GBiZTjZ+1}nh6fHvnTr?`-uO2M+k$`Ko*n2gYNJQ2>*q8gqC!)p0FY)lv z%@o4t5%ZSG zyj(JX>xStZP%{+HTpHEmVVQ#S^8Zdsj#v%uo>S%Qpb6wrw%2<|hYk{P;)7j09VhDF z>1jM)$+PI8ffb4xh3>gF{Yy3=A$aU-Rki-}HCr!#r|VDZp{gq)hog~4IL#!tmW|2h zv@eD9G60K6(9jRJO(oEbfM}#T+h}5o<7UPPco=fZU#X_`edFjxHTld$QKa{+vEc0d z>lWwU$AsU^69i?uRfNKSFkrL$e?8n*0lWz1UWs73j3ZTB@Z|gxP!SdNFc+1eZ|dTQ z6vx_9dhn+huUgzi>%6!kweQDWDHrM8K3`ncO{~i*8hzzX4pu!J7Fa{~1%S2eVgO*m zfPTiBbpbf2Kp%oN4V4*Ktg#l zAz2vrC2BqoRh!U_m-Tk23CNN)NXr{b=zz)D-@MrvnK_&zN$YFP#J^31NIUQQr=OkW+VN)Hi%JPhF=VPn41gU_$%~{?abU+of*oW3jqD69$#&Y|!`W8Y9;2-SN zLambod4%r@{|RSxKa+YuX`L^s&xWP~4k*bff;M5a2t@409~`dM)|(AFP2S22vuLR& zuI7mQG~XDQ^UituEk^Oknny!srrl5MAN2zokKp4&kam8~O(OjYDp5AA6>)paB!vDL zoX}!vf-j?LtJ86wbPXz4Mv~SYZyM;Y{Vc-jI~@4Sx?VDs@!3qkn}OyE2QeN4uC$I4 z146VFKt*UB)0dI`gg!v6?g~i6$RlOY<-GtphGeLTO~kYG_GbZy9Ht>_kgWZVe1l;U z8YcT>?ya=_`7*!OHdt^Ih3@zdrfd70)}FxvX;wQjYuh2z_pJt~wM_5VP(JOd)0Cda zFp7#AX_p-In3HRSiVjzRvY{-$tC3dU?|m$_{As8JW1;5>YzZf(CIggrf)!zdQ$_M| z8b}lFB1zt5pPeVkV3j)^pZo+_hRkwkg8`@Y6%}H|zWDwX-C%k?3>H0Xk%+^TdxhGj zA)jVWJ33E#=$P;V887EOSBh-{NwIEmM-hUcku4G=#>JBkw1!a;f)-_SrW`sDbZ^^S z_#!9Qc45ApV%lsrkrJjl5sB4mAI2?e&Le^eoe_GYU<~2chTexR-+s%G=^l!4WM-0S?Oa|Q4yxVxnd4i+R)%z_fHuO$Y1hxRSyeVTt!yaz z?)St0E-k4|Da?Df#*!pxi_!EPw(0w|%w=m&V)G zu|-$Lv~_M9*~M&C>Ruda)h&gB(785BHj;<+KL+sqE_2bIbBIQ2OhvnKXo?a^7i*9A zef#_2^~ig*XJ?Lt{_Tp5hM@r14ymFNxG3(&d_7Jk;SObblX;0lY*)q|V@Ul;Aq9{U z5aO*N3mSR8eZ<;74=T5Wb!ij@pHRoM%MmUum}##aeM(&Y8rI$gSr(*O^Q%;y2OFiq&gNP<&@r1R1>r^ccvdeeeM+AH#C=L+G?wDYp8 zGy0{Ym*dgdB|GU&{U;SCftTjWF(m|!&>02*LjZgr_7S|-448OENMFKhtQ->B$8sN#e`U=U0cdLW1A5eqH^U=6a;c-P_@F!D|6($Khug z#`j;I__TLg&)nQw+}+kznALmL?35A+1-cAd8a+j8CzOu9-2fGjr$j7}4(5+|v*W%U zN%Odpom6b$puF$Rv)jnKw~F|qc~;o;6R0ou4B4H7Fa_N{Be8FE6bBOr)A`HAaB#!B z+{iQ4W325ZX#RHuOyJbYQvdMw&}b8(1jgGm9Ptl!jk38t2$TP_rC=I91YT7|Gx+A{ z-pSMXvzYBPqmhz*FwugRNkYAbY68g3Oao58;XPo%7e`Y#J8HnHU=&=8D z?CAMr7$=dFMVJSSMx()6EbBOh3yzHN^>zuY1+~5@wM;Yq^J>NsO*3K-u?XoJpW~O z$~}pVz86}X*||F!UQPOu564`@N4@4 zN9|QfZQS0G#;Ffk;YUAThT+eb*&4@+JUVggw0wFSisvyj%u5FmE=efzmV7d7Hz5b~ zC2y~o@hI);ndf<&4}$^lCEjT9K{D(_fk^Aky8e9{Mk-mYMSrX$?{--|n@k)#SKaKH zE^ab)Wj{~M_1U*sRl6_xS7h;w2Bvj{T)oeu%ooF>A`VmONu?ikr+b{ElmL9iDMWn80d$?8mp-m(>*-RF?ysel?WG8usN4=r=gxVVUwdJ)oNeq=aFWY*T^ov|k*1Zr= z<##L2;)J!)`R7C>7`_oqslnTkrf%FlAjo3Bj#BFcaOXT{2GeuW7VGciJqC^E^nmW; zQ1TZlPq1uuPN?$R_C~)!(UJOnRoj^`fxQ1w9ykM*m;oV$6C2nAa=P+?rK%FY+I%Yn{5Wze>VrT z38N_(eHwMh#U#oSlW>6*OXV`%YrD~6-EW|?*T(d0j8?yO-D!W@()6{`J?e~LyxUDU z5j7u2e5;mvd|_j=@mmbflzh2hsFWda-UZVmi|(Q%WRv@j zuY}bzdWfmX~6FSe6{Lsl@F4dWC{5gZBe>Orq&_gKr zAo)4@_}$o(2X;R%a#-eSwA4HJE!56Pm(v206c2??n4Es2pE97avd`gcEd_nzoqz@S zzG#RaOr^#QHDFvn3PpBBxR>^AJS6c22T0(@xX8GlxhuT{8Kk-}Mgmy zvRUot^8veYVzE~5NFzAH!uP1-4Zh0AC#4u{>5Z`}eeq1l_8;ur+tEt5vCbFKQ9^)0 z;R-%s>G+^(Gm=dLSup<0t$M;0=F(k~a~e8`Fm3D{l{zJB_LnP2As1{?Y&uO$zCG z)aNb%mM#)5t&RDjcv=r$d1N~PZtt5G+4y^E{TFW+YQBs_zXYG!WJ+=}QEqv5R7#RS z4=D}MX&}Ku1Q!XFl|t#h8$uLpf2Rl8h9MJ!(|L6URa>ctPEb6Oag~lI2L1%Ea-Pu+ z?OJp5yZ`o^p>}r)gc=q7ck?vA{Lva)zFiekV~@;~j%^q0K0eiK@z?EY+*A5H+w(ZZ z;zMa)%(|+C-ISFO;{nM4KTJT>{9*5tVD5(&sTu>OXeOCq*NI6*)GY#0UgV;a5l)uw znS=|k27&j;qmF%Uk3TTZO@!vkJj9v4{#$;yWRNuk#C=-bJv^IDY^hRa1R;e2vPm)^ z_3H>kyXZ1<7ucZhz^8(!#5kNTAgXehi@nu}L~)!!$?OsyHCptwo^?R9q(9z2Ksy92 z;o`I2IIMZXDWR_2U6Kbn44UqTamI{44^6^Q4^8A4yMRO?_R_>px?PW!(V`t4l{C^v zb@!9&H44ctESYot)$| zWh>KS)l{ezLk00h`}Bz7Ue_>7O^ti8v1fu9iVGv6PUvu2^QdR;oO!bS0_(!(&H)&% zkfn(4BGlWhbyh>M-9(s)K?xK?;5|)kA<4NeBahOV(KxL_?DNYxHj3)+xYPN@&)am( zPliRLm+C(}Ki7R>WL-L3+V{mL9?nVb98wj$@?rR?TmzuZBFTcVMdq|!3|OLwcwsVL z9_K@^df+DZghraySO|SEj}Z3Cb~S=DDh7*ujl^mc%PPHvsgZ;3Menh+k*bPSzGIRt zH{mo7^K&_OyaV^0LY}kp+4tT2U1HA%DcFBUu4qeLN8W?Y!z?&+gdMgl%$VzpCvg0u zUc(fcfksS?h+*v9j^6(zQNHfx(VWcY3NM&k@y$Jb#g8OET-TM?w(5n6EvMCdblAPW z??!7!X>^}SfZ*9(@>L_n$Be@Z1YX=CVLsq-mw6~nXSdh5XsVeI5Oku-)W{?^;#siM z(oNgJMo~)WiiR`PN=eYxHph#16*CEEUU|oTzd0TTXkH^b8>Cy2FNuC`ZmL|TpSzrwVd?;UK$=w)*6&XSUro+jrU1^dn*i!ek zU<7}ZEjp^nr)a{EiQTBIsT!hDczt9}?u;Q76YC9{uZw@U^-QHT@*aHXA56#1Q~9uj zvPU1Lb;iYl=L;RWYb_rjFP8kM$DSu*n8!>o%yvlWJI!Hu z18=5XM9FMBvBZAP$L`NX#uU(bzSv1GL*>YQlZ$i`7Ia-W?aOypc5Zr{`&m54t?dNQ z*Y0_Yyffr!$EYO5uw4vdaj4h5MJsdKDcwdh5a#Uot7kWDbR8IbCi9f7Nt@4ZXN_=w z&tZ$*9o)`zh>JOmx$!$Zd9eTBljOHZ>H~oqWu?qPSO`h1Ro|0FNX*|wzf|<5D#4A0 zkcD(R$vZO_-l}H2!e;b1;xk|SteHHt*yEj@Q#;rCp2hcyFesj6uoFZw23MmgAoMr(v|N(Zg;am^455kaivaAGbwpPImYGM zYMNNJe6yS0dz&#Hc5j6r@?4s&)lHduz^=HPYE1BNbu);`i%9n-O0jjlS3jP49+Y*2TQ@DupmMTha=awoy7OZ%;S zRW^Ha)X#Q|927A#*`Q`5*a6Z;qFo+)^_XW)FUZsur4o*uB;^Stw%QeYZJHkt@YYfI z{Pb}0<1YXlUPutlLOqO$^hD$%+yCR)MiN=>xcguK!Q3jZF_lL22cit-(_0uZG}@)M zQhQD#*`C~LIW3k^?6j1dUL>_K^X* z@t9fzz}S0Kx-qYF^4m3_A!=_%Po)e{t)Y{V2U2>uN>E@S`)a-uW#8#ytfyU5%U9KQ zgF3wlgSz9s9(xBH)(<)e3j9#0sxEVI^l=QS)#S+}P||Vss`+q&fZB z-C+kz)RP}A(j@=*rTMrl^%8ppYJo^w;f04cZtU%=JCR6paMXNwXn*DzO&=Gx)Z^*L zE_7v<-87RwwXxBm(01(7J;QY)_+>NX3n0E1SfdQKuI#ozA5bOJf&T}$6xtB3HyPEdQ}9*d0xCVXVWUgE-p07= zYNyzM zm`z9xhQiOsp7R5dP7N_yNBp&1`Qr6AVMyMUoh{l7nVq>+u)B`mRA>w4@1OU`!{SRI ztNv~n?>pWPo$_wVCNY&C6>fC6e_TFrda2_>0Ux_AlMSddxZ+PB+1&sX+NdOK z5&+4^(>T<=$2W%8?nwX1LBMfVuYHi3ql?7gh!f5(&%?4U5{CtTP3&aqY zPIdM>=NC;rAl|dUtM4+feRBDOv=vXPPX3b0@&A%7pG+8MOMegd z@g)AU{3 zx+!A8PR1@qRJYIs>$W2YGxikoGM?SB5WI3GKK<13_I=BT(@%w3?fW!(p0V-!`=;`Z zeZ`5XtkX(jODp7R`R_#ZkB#)!%ezF}nyA}rsheBt>O6rhf?~p=Mf>?A`p%z)zq--Z zd7uD3Qw<%&topJHxnxwmPA@?cKarbOLR{Sj(ttx__!HPgo3W4;lJKvchGt6f?f@Cz zo#4iuV&uNB_3uKo&brpsw2rn@LT-t+e3#SSy7i-_A?qez6QkI%7J|)S+(I9l15^iFkL)B}Zv=z)$^9r+)Q)uGAI_=Bdi&PuN zU1|r0b*$gH)_F-LXCgb{FfnCfTGMub;ytew$6Ff
Dv z&L`4PKSiU9cHioR0!Zt;ZTgAI>KB2UsO)RPN>A5j1RYK(!uzJql_{ON!@sQtF1r#s z5cI+t{||#2;J6@B?zT*IfV_fnVhzxTSw`*ow^IvLddPC@Jeo86fpebCzhBfFT5$2N z7Vmi>S2R=17LGmOTz^|Ac(E#qa{)7mU|4UP-MWSJ?Ks+s3N9=8FjROr!opiTQRgb({5tmtbDk?Pz!Og4 z+(R4hl(uOH0XUenQc5Sq%nGxu8J<1*led^Zu`rCdPYWr58_0x#jPBup`itL#gQ6;& z7dqsd39WHaSB5NOrcVNE$i2?K4ZmcsUpsnd6P~#fl=97b-UO#_kUPv4(*&_;uCrMLC0bVLOsYINXxMtZOjOE7r`Tex0}2Z3ubvmW>p`8@%Cn) z)cBQ!P;lfAY%W#1>*X;|>eyZZB(xuQnwnGUqih_jTbuX0PEOeF5&B8sN@~+py6lti z4VBd;TmhzORf?GySB!1R@MMKBZYFL1^~75;b@5&Hfq_N&%INEk+C)l^ z-xkUbw?6xzcH^))Y+yxw`#;E{qh>VNcxgxC`mlc?VK~H9k~MEVs>7MOnNks;8e(eX z>^o)C`XKRmLvtg}kfLe%t0mmKj{gTw#+NLH_G6N|-IPE*y~T{Y&c;^x-gM+TG6^y8 z4@`WYO;XsXwEej8hJKBLkN+5WrfypR&h-*k)lPHZgletZ*wDq8p1X1o+M(S3KyKx6U zxVo*b{|7k`633!$-WsN)fGtv6Q!T`i;0Y5MmVofVal}luLg%;KLcT&^S&uEuKB|D@CFeL z$UghYB+A8?96PoRi#%Lk#yf-79W}WUNj=ou(!}h;(-C6c4Y}ojHdEH7N$m$27m)Wk zuj5AQ8SVjlOWJItA?_r5im5;ZW^M#UwCxaR>a?j3UeXD`?O?8e0*lXm7;!VUP3FyX zzwB*xq(y`43t1ZxKeHiC6r3T}qmY?+v3=}w=WFBxgl@^j1yn%lt^h1&HFJ>0eh>UML$w(z% zWB=k@Muo8+g9}^wlq>vdVRAerk6Iez&`E7xdo^5VBfKRQvpTnNeyFeHv99YSV;iTH zBbrlldJqT86Onz4mZy@~$H~j1v_q*=G{W)}XD?chLF$v=J(ozo21)?;V)c?ZO4ikD zrnq4YZG6A_Lb(6X&MU3K1czF!bXc!C7Vl<^(@bU(7rl1P=9T+I=oNds_nJ&CDPG)u zxD@hcaoemYtYk4a>T~k3uW`k9b=!YsbtRW&Poky;l^~G z54310(C&&OWsuh|^OW&fj&Ip1irg7e$ zmy|qMDh|HWO1S?on65SW*<{wco#YTwJVB&hQGRxI!A-o4`6h`e(X9Y?RQ)Eg%Xj@vDe^yaK_YZ0`A`TtVU9kB;I1d>=!(i9meJf@i1RK=dEDGO>&moac{ z=bmH`yJjLb`i!KyBk>7inZPuzp(L`xCs#3^TXl#{?#aC66Y!?KEut@v;%XG?^`^P0 zIpHHB#kh5<;L`dc$6@5yXw~Yfe^u$IO8DjRKT?_R8^;@uQ|eLzb*Z@D`Oh)}axHJ1 zdIl-UEN!ll6+Uh!iBf)M`!Xr zv%02Ek^idz5U!1bHg|3 zliy|c`MZxo3x4=Oyl4WY8pyB!z9GP7yK)S`&#jLZFSezmJvp4ym9~$IWM|L|v`+f1 zBD|3)5hH-3Fn@y?S0w6t3;<#svPc)F1tUc4+E3i+$cu6IwQZ0j^q0;8v-2K6iU375 z4_}L3WnoUS%PYBeLGnlB8yA`4QADkYN0=|rjvrnO<#brX2kjHKWJ6m-hZhDsf_%DU zr1!%<`3bpD@HTgZ=#m!WL!MBb4L1kFhljFfcju>4JOEA*{-RYAp z{PS^~^l)V`$264Qg^|m#x4pDxT|J_{GUYV#$LCQ+M{F_H%cs%j3B4Zu&mGt}FeHI> zX{2ClZ0A?zIp?{L*)~ims%2b^gQKkmSGdX)g2<%HlDMT< zan<7CEZUxjYFPorqO+m7N0D05h+Xe#PJYysUC*;;XO8}GDZjD{UD%(1eHaS31+^E;ni~4%G z8VR~8ej1mZ8YiEc3M@*OLFC*O>3VS8mdCd$Tb zybkO!QSa!;)V16*EavAwV-0J8po7^q#szJ(ldDT& zpCEN3C)WC~2YIInMAGEyJ{u~&haiLcc5Bo0JPv3 zy;&ojdHw1?eErx?lx{%0qA1#eBgN}g6a;i<81RTLUGAqwrXYGEMr#%(|0L)JB-h+# z&uv_1N~Ydv(920HnVfK&o^WxFs~h#F)dve$z2;x{rGzaL%D91S4`A-&P{)pE41WZ` z%oJmmOcN}k;twxA=o)~R3E*IwQU&+j!` zl=x(usSfJc`q+#5)cEXQ=qb>?QaifRK?-?2OLKA{YAbmQD=EGqORD8i6uIIytVirX zSy|*k;8|BMmE3_qy*!qBp}mYG?<(n-Fop4XG?WM4^Ef`96qbNnfh zdMJtyRIUG@od(`<;89@NrN};x?85$d(TOntx+y}MWNy`p>1(tT;|kMbzGvk$vSSd? zsA?G{ThE>|_BF<%g(ZtZ0%&S?ud>D>Py0DvYL1)q%!)>+{s(40!1~;I3S z#N_OsL8~wTe^>^E?L6Q#d!fu}>i497$-&STwSp0iqtl(mq+`9W+~74Aj^2K&{NY(Y z^jGh6-n;PLnfB3{iXY8VU8`e@k-pvNqG*EtBa@rQ|AWw(mjZGfLp^gNhlNbIu2T@nZK+B_?Jji9+*R9G~g;f=ka4m*!@^fAifHUXbyLk68EKUG6%~#dihfH$Wt=C7#xN4uN!=N!j#&f0W$s<-zJLo& zAI95EEAvnt1Ht157L6ho)_$qy%)k0lD2;8<=N^R7=2y~jVvL4%YCs$~iGp0M7@9^L zvbOOf2{wa5Q0f=VM2Hs?QYDEQKrpVuCLCE08C5-aU@J1dT)_$C^(%@}*@~r$|3T7~8al{l z1bH6pcVnYIPy7KhrM=+epvi@znc10Q;=BF;U$J8K>JXMqh4Ixoua9Pj!RB!fF!|D8 z#P-hWjT0z^iZx+rW=ac25hQ?qD(W;^tkn3t5#~?uoH>BczKE5?X?t>vv1kt%e>J+u45q0*|Qcdhd`_;=ZVr;1^xUsp_m4?#i zRiZTO-S3>D%J?^w!NX9*1J~t0NDA%F9H56KRicqhp}s%6qU_k6hZ-_UoQNO1d50Nq z;)R`EVcnCL125byT^MY&^z-yP(E9n;uXO(28f!Stx__b5B%iLF=AJAk1{De@8wzD~ z&dXz!pR9G#Q-L*R@{6YX+WrsMWwzIp@+BW?dmzI;lGvL}Y*Lv3yZcpRj%s)LO3149UhAW3( zxm7Y(U1_Uc6d7_T`9#Bw8xG1MuEv*ZbKU(*?i{%67zBlHn|%0`v)}4o^OK(ME`nL` zIIqqL)yHR59f_xY{`)iad@%e@x?Vc0_9v*G&v@h8@jxX*OPh%8As?F+5tizT)HWh-ieY^q3XIu#yLE?uC3jZ-2EV$G;#r}$i@@NX9L_6C8yW#>*YXiVNcj2*h46lPNWd;fY;&5JX)0HN1nuJ z2x)&~Qrj(We~;l1`$J2&+isA|@P#(U14RorSY<5_Y8tlxshP<;D-MD5>#7vM)}7K2 z|G~ED$z6+k56a(Cl?c0G%6wEK_;;Hl)p@C|mB9FuaHJ;Z)2Ql$w#L+*D`xVu< zXiypW%r zucmIrOncY+u|F9!8}FE( z+~flD24E@kURF-ALung5hohF)kvc0S#%;IO5}%tl4v!U-P3YaP8F{bk8##4)3903DH<#?! z`WNjWsP9M~7ue)>;Mz$;(XvFPEg2F#75cepHtlYb44iUtT3S&xraM|N7y6HoLVmOE zsWo3wQS$QvS#X63hM&D@`pO&4tNZLMB0M z{+K8udZ|maw^vK$yFy>12knJO*M-}w8D$ebEq0})R(A>2Bg^V{+dj0{W$g)WZfJ=+ zNxk#6{$S<08_IHT~L+pFc&_Rz_1EKT`eS=l8e(-lLLuCGqHoz)3^{kzB_y6hT%33Hx^ zOd%Y6pJUH{!Nk%udgM*j@3lEA%An}I3P&28o3e&Zmo3-(xvhTncfaLv9>3gYHDrBA zCafL`x8qap5E77x7K<{4KynxHJJc^kmgx(Ji~TFGgQ8D?$n28gHc#5j>25PP<6;(& zR$XS`lPWjWEP8lSb)!`j#&}`?+Sp_Rh1qZ&EPUw~Vv`~`N1Kb3@rOStyftynZ0;W4 z2eVV$%fQx^@C3wUO9A3|w&b@xQPH`0HGwtt^=e5@WX=KXzIHNvdw?O(me z)^CZAJz%V!3(4WnA;UFsc5zRK`lzfc)AD)!+JPmFy^|!+s@xpW%_G+oeV!_rEe;HP zbmXeF(^H(2%BkWC=$|#$V-E2N*`S9Y)&3s?^z1;llXe%5$v1Pna^e+2g+2J z=FiL2Q2bTp?br6~tATTNG178plyKL6K9-ffOE#GQAhdwz9?)k!2Z8}z%p~S82*@e$ zM+$QdnAmP)60pf6%UZbWmCXtTBq{73X#=(*<;!nnbB8F!Ms!#s`@VB@=6 z1VUSm=efUe>twW2w^oZw8*)l|ee=Mcjs5S1Ie36F6kfHwREXBv0nFJufQ=`$k`u)w zQDF)7*eT#2l?*^!QXF2W?rU&bvYRwhKM|ud@OCkD67a)UL-@-qd_$;eDt6;}4vn>S zjgmuWS|Vep`TXlrEH!pBgG5nTap#zEgV{vtK2I8~n;Z*jZ5x8A;=;hhs%83YuaIK0 zU7fqL8{!?aPYbuDNlbc;*WRAJl;(=aH9cj%F3-8H z1l(#>657tvZG)}>NM~!sw^!lVxUP(@Z9tC#ye;q6xNgOzYr>{;MdhwdnddV2ei=)K zcqkyhKuXfH?WvZf7_Ep9Q*#q~YFo6msN=MY^X)0;X;)V_m&u9SwC`e;^F)5Ig%r0LG7 zj8@UI&ezvdtwm0_t5&}|lz1ya%jQ7l(ToBkbqLyHp!8F6uh^>0bN9MXmAy`|wsWL9>+k>Mhq`{+`M}IpAVw%!+%^$@2vi|HwHr7 zhAw#A>$IaqYoDsAzJ4sG7P9{P?|$Krs(S>o456@x`S%0+HRyjQi-f=cU83K2_z~odE}pkSuGdxy}AazrB_| zw%w-F;$<@?J`}|v`ySXt7;>ioL5bV0FI}-=o7^^;obx_dnPBoWtaz|3@(R0z zS#=KRHJrGHjJ&S+%f15@p5_RzmNuqj{X0}Hl|HyU*XW#d>~Z=TLF{rX=K!x02O#K} z0L)iJqvs0XGWinm`%N5~12m!7HhG5gWw43$P+2)m^RsZQV>EWd;wPWg%zVKqF!bXp zS3&{~u?bG!{Uojus(s}Fx8|B&ZgOZ%^#2W5cE4qZP-di9i)?Z@LxU2P){T6s2v#op zxW}38{t3Q+V#Kt>r0?$+!{A;51pmnh&&3_W7|2Hn+C?}(Q80^7M4=d16{fwK z{Nf{|&E{;3BhCrgnN}ySn~MSK-NmT>z9{Wsp%I5&Til4W(tzR8=1-f@$oG&L`>t(X zJk3_&dgBNe@y!^vn8Lt(K^%5A}Q(`YQM0)7szY$70x0^bU|u zd&ca6JoE7n)36US1J}&=Xea;q<3G(|4p*MRJf-ki8g8=YU!+=sgi-tPdUl zM6g3CHy8(vn?y(a2Ytk*;iV7{G!k+9*)t5|zKERY4mewqO=d?Thl}c3#sBvIbn_mc z4B55;kq>&M55Ic6^KPuxt^bvun#wo%k$**F%9*!!g%iXhcgPO5iGWxf`@uPeWLA{# z+tO$gA*!#207~DHDXQ_y**9XPA#dL2_SU8ZKw-U{iSGd0>w)*5Ox1=D-4Cxh2!XC9 znHqc0{PQpUyJxR?t`EB3`nea5cmLGken8q2;+vnULMz}M$H(OjJ}o6e=R{6Ayw2Wj zE>}QQf)19$UrapdS5^uF@mnon9}yG2EC6fA+~PU|bp<1c(uNm8cw>KfIQWcw!OfWjHZm)X;iR%R1{u=pe6s>jDG{g&OFhSpcf zHa4wEz8;ygzAs0Qrc!MV-GNmiD6f$K&?fi>Ooh?ch-5*Le;7i^ZPp&8XtVl`dR*V? zRHwq>K^+3@xA^prX;P5xdy?3{<1qedo9Z{59tLD~R5?Dm_K?Skj|~=Y0C<0Y)wSC| zSq4r4K8bhLV%#0R30^y(w<`B!WkYmcXzB+)Yss9kTF*g()rvgtATWpW&%n}=pueaQ zTEVLTaB=VN!zNOPQ>(CyZq}>1IwmZRnq*fKL)*4-#>xSXvqcd4s9wXD5}9oCAJE1c zTh+hWkBUtp3u5n^%3e#m2K??WaE#EpJNVU~iIKi}-aPn)H)+!78_kI*cT|B2v!m)* z!QWITPmqy^-d`(~>HwL^13|c1Baa|a*2 zha+nNS44Rp7x{9or4{F!_pZF8sky1?8#s3b?buM zc&4lqj`pjBqg$WnGT#J~Z5Pjj!*SqBlJXlRz&7gZzT}E*Xi^ZF< z&wIGZ_xJD~Q?t97`l2?Ib`XE`ztV? z+HDd>@!Tqe=-)S1$c8JitC8DcO_1Hrj*&zL-i1y;Z*u_IFl4v+J?Y2SyTr6d4&C$G zwLh_*aSBhZKSvdW#2Pv$_#kcCQSJ%xz>X@!)$yWj1^`h&4&lw zIRBTdX;6kX{nWq8&~LPiXsq68c;-49_DMwR!~;kIMef9@vIFS$QfOo5YHSoWKo%v@ zlM5nqovB~#bmOl8AIm~F(wilt3!7d;s$mgP%Mxde;_t|4$!BCe7t+MFiA}(t{H2>k zW4FtoKML_Jln_ebFnIyThJ;Pq!%gmp{RjR0L#l`5+qbv)E5qkIgqLsy+v)uI_dUnu zJ!z5m`6BB4m=11+V4aWxfgOS9TYKnCbc_PJR3?^ld)fiYNXe#%MpqJtpGhCwUR| zaF0*BJS_g2zq5Jv_h6h_*5~Dqe!p_EMDnYYQ@!|pte!a&C3N@6FwU%bUeR1&(cXe2 zmpaqDF#R*AbE(f<^id?;RP+0jCz)E0jsJt5K_?C8F8MGK*1<9@|}jr4? zBaEju8E?@^llD*RPcyyNiqX!+(v1A1x` z5xxBgW?BP)LL|1+lnck1?sOpT1qfbBQ8k0_yfYJUa;f~?b8iyyJa^w#_ZR0N>8q#p zGrBG)ohp4CeS)6_W6_zqAd2oT9(0j?CoxNJdE+3;Oi==LevefW$GKCncJ3L?_Ci~K zTyJyrFxOtkLR3o2?)`I$wu8{d<6jNC9eQ5hGPr8Jp7ka3j7cR`;?wmHA8TT?`qN=E zKq5h4To{-!?lDtfj;|n@nFz4TsohuAxfZJ^-l2>FMFpowXkv#+3bs4|R>!D$T7O%^{6- zF}mLVDrT4O7F_x+R8hJ-$Q$UZ{r$tuIG1W>&E++&W5DqhnzGaT^)B%h%6|O|13v;ZRBfEfYtV)sK z?cG#|*1wU7`8-Pl@vVe-uAiqzh)K}x#VdO8>W!>{TgP8kzSe8h ztF}EA9gMA0?7~lwc*6Ur0K3}9b>+yR>{)|sD%}?HA@4_cA(EG9)>zN`%-y)MNo)c} z4h_{ZSS9Yh6^*eWdG3HK>f1OtHS@LHd%Ut?;aJ5_y^m5@${Rex9yVd$iKvfXR`aCV zE21m_$`!(Sz}-8@JAt+wuK%%Bfb7IW+_R=x1%LHgF6TaJ8?M9eckxmDaYIr9*(rLv zzO80@rqj3K2E+?DaV=Sx6iVb;QfH)kvdz>6+e~Tlz88nu>`T;uOG(1Rx}A%wW5Yd0 z9`7@X8tTWE>5U^-$5w5U{uSle6djK~gMdAJG2UpRP^vJWcG-ci|L*U+#0@;2UmlEN z_HLVeJC}GwH;6%42AVB_I3)YT+VTRUYl^CMnE>$b;eUKY0 znK3ze>f<4Ep^oc5QtLI0Ov7lFNv|5i15B>a)oiDy)yvC(_3u=Ff_M!iG{674Xs@`` zBimlR-JdR88Pb72yM93Xq&oQw+<~9)avRe;KSwnDhS&=U+S3e4{>BVLllvIY)o5r; z>7;>UwDt6WG(CI3_*9{Ezv{B$Q=&PYM?o8eDR=tqbTz9N9LlK#oe zgjP|%jhL~SkNF;a>-%%HrM`N&qB0;Lz{BdH-}}PMMcR5;#!(wzr#rXc&G`dIFZ}FD zQ3|R~KcfUiSul-Ray!)>BIpxC7(?`(om`$OSDOt?4Q|}F%cB#d*)~_c2YUx>9XUBn zZHxR(!`s?yh!F4U74Lp6RhtzvX?-l@BJ|f`zxCehE&h|8iPf&XE}|>x{(OTSA$P9- zqm>RxnKFVl5vc~>VEkJ|!Dz82L&dbBMN`;AOt|Ss(Rmg58khm`x3QRZ>^NM)oUadt zA>x|pl}4MJL^;Y{$T*Y)$K2T<%F&gpI>HN~(D`8kTrM)q6hz+lp`ZuWE?FTrZejf1 zw_Z2f(EhN?1gGq&bcuV>DG}NMBU+3Eh*!(y9b(V3x)@PThgBEYMgRa@4j50=-Yn2F~j%ZtF!5_&jFJ_;l}g? z&efJ^5d72TD_uW^Il*qcGF}m!IarJG>~77R>H;=3vUMk6acG2C!Ry|D2cqIL&dATb zXm+jx42KY(-uew2F0Bmb=Lhl;#cqN?>Q00xbA&Q)*5<&6dr7nlm-A$FuHez3i_n3C z_oF|X-IGp#&Ho}(FiJCZzUem~i?pl)u7kkND+Hwga8=qYncA$#vCT*|`##>jRz`e6 zU4?(=7FKLAJ4Z@iN>ahG5ARn~gg`Ed2hHGj`0e1!kSf1f0B@3jzB=#$W^dA?6d!3L1Q1*LOwyl+kyRC@*C5;UNzT)B8=y`0>k5Ah1uuaYJ>^LE zKUdA4dpKOGDZcb1`K{D|vCBnf+`VFd=wM(zr1kB~v|vWUu{mt}%j1tdvigq|yi8L` z7g~K8Q{ScIXe8`^By8ibY`IN;xa3M1HU#w`c~@b|zjBZl-x<4b9)?F}5?wsoA?Zz) zDa>E!oMtDK7El54ughw%8QKAy~HCga?NJHni+9Md=Id-YW;jf#LdoyhA`&=~)pTcath~sF^)Le2Fye3p z6s5wESw#K-MgXwolD9(ZxdW^W_K`=KaeJ>`U}sV5Q=Z??);~vzw^sK1WOX%V1A@7N zFYww4ybah+D}sNLR&o*62bOeGsWwhsgKHgFwySs`QRghH@X(DjXBCY<+pY64FoEBM zn(5R0FNpS!!-?qcyLJMJ7Tk>md;sx#w94iCVr`F&w@m0<`w~bf|B+FMY3+|R#v+Pd z0iF1{$wL;g=FNUaga6&i5SjiWn=Z2xu28zYA41#(njQSQ7nb3aj2+BkpZ*WJNqz=A zK7PCN8*rLT4HlB;aoBIb6No3zx4Dq|vX5^|{!AnbGGZwdmv$KJT->v{9W4(#n=did zynyTY_-B47nr{kvY-dHZx5wIdt27k~^ce>$OMj%}A{EOPh<}2jce9oSO>-`@X z#D$a13rUC0H|}0(AXNjSW^aH2WO(oJW0`c*|InWY%`c9ulBh3eB=rtU7HkhnP{k{2 zu;=aJ`8B|FzsvZNO};u5hpc01cgM#ek9(H-D*6Kb;R!(FK!oRyeyp!*Xqaf>)rt?c zAeD2YJ02cLL@jL{b5s2ISfsGICe$G)U(>d*K`bFByE5hd!BNJ@O#S*AmIA*( zD>#!xD34;Ci|tK2_VRmVd|?l%mEYHoF%VNB?c!c7WaZAS5)X#Y3;M&XJe9ve~-`C>~cw-8Aeg3B{V^?XE1Y zmYQLuG7S#ooe9Yp&-JM-1wws`t#tB?f3$ERYzMitM;Em`xEpLrQzqhH>9X{CPOlyi zJtLHrYZ7?r9HYS|zC>j~8n(yAPBH406fLBjmT;w_akxc~baQl9{%o0gQ0xbfhmDcP zGo_x@Fwr+wXC5-biXR7_DR)!Em4A4wwSTFoU|v$X2Uni&I8?Bzdw1>$tS87Rz|EIq zkRw`>%dGoY4}_`1cCfZQ6np^+y@JK+wwFbg@hrpw`WX}}{UQjm4MZli2$s1{run2N zq);5RE9J^W)h^k)>0O5CJ+#^-7<{cI1RFx{Kp2QJOu@A<4SNIll-@n2II?%2Ha$Q`sK58xb1-3po@pO- z%Gv9enax?lSxRw14a@qzYh#T^eGS_(_H=%Q!hgVny)zD+l0&dcCXkTW&Tud}KLRXpk zhL__4T@0|?aY;%4JIogpIQ4xBo0TB_R>l1X`F8AQ8D@eA?Qhyr+kmOMzr84dHy)1@ z+wK!KFyb`eWPuf6uPs;X>ig>C09{E?aI-gUr(-$WtB(%~BI!$9Q=>V}eq>$ycKjcRsIVEhn_IxURET zc;FA`-DO-T+xX71@G6V&)}nZmL`b_jlp!(4`NBzN%6kel(o0NFvJErt=)l@^NEzE} zLj`sKYWrk&S!{Rt&oXXVSO_rI(2^ewV}BdAs!k>c9g&5k&s%keHM_Pe_;dNiv6F^* zZk_*;r~@IY0q+kxolQ&t@@oP(tdg(An-XuoP^#gOAxd68xE!q8;$F-)faO8MyV{s2 zTQNH!m@li}h(V|Y2b%Alb;nz+b^f6aCZcAn&Cx&xtcV!w5P44}ulP*W^75+8e0nW+WxYcdMk#0y)F_YE4s!j;Bah4i;ixr#T|Fk}<^KqnYHRbGs}_+5B|mA+GKCDwbyl&-wec z)V3qqsx0aWo}4((YSZD%T-(}IVIS?QcPr@gp6oHy>@tH5Md78k9#-I8EN9HS5NrTp zs3-^sy!pB4A;S0CDf=y9S{(w8E;pu#J*QhvY91z^^_^8@KQHj^!}Lm>qqxTk~j~D5^8%`T#bF#BptvxS^hDZZ)n{;?wX^ZALLH05ESmFc7xG3-6u~hNym7zur?w zg8|9@9>D{43xH|}cTE98$E}EdVN#!?5S?u^-eO!Lsu|l8s4w!fByFMkljCep>ANjO zN2&>9ET!Z0Xd>*>VWdWO45A)Pb&1Ne&D37X#Wro)jRRTNQMxznFG<{Qp>v?qW=1R0 zV^%It?mj**+x$6AFB37>lr`9SA=OVjOuf&ru#V5G zAxJTa00V41-_h**oS(DH>?h`+SJhpOU7%3l9Pa@Uz%T* zQ2HkM49){(GXA9217X!@do(5dc11kn%YyYI7q?u9z)%nN1VZkM!NwDY3c!*PK}3Pa zPGf4_V;9MX{sb7HT6aH1XBWuHktVU*h3YqxR}!wZ8W_G^ zNg8|42fY&zuOzXy|%`gC|{|!&ELd3tEjxawDxwY^Jy=<3Qws{#3`Hp^ zFK{!T$@Q;F%7BFSd+8&m=S7E^MQ(Ky+tUl7SZTd*_WRqCTiku`n5j373iE_j!g!>F z9$53g;xx-Su5z<(BPVC@i%_+F_%i_?wnpKmV>4b^eP=>h*z?DgNWyo!kmuBLlJvW+ z?!r!S+7{mWCVJHBh|2Z+YbvJ3;d=K-a-5u;FeqScG}I7x_zwBlG8R#FJ7E_Uoz0v_ z_D#jrTC%UUagPO?5QmPTvAzBaf))?Mu-s^429NBf`XU_Nblmo1W>INo-?WtLkLd2) zlQ60`b^Nf1p^LTzO|b(V`cc{j#l9EC!#Ky7!}9s!XZm{g95pAEHSk<$cfz#C?_%S#4ciKW}fvaZHqI>H8qxescH3Oypt5Ue^ zTRd-j$6?mZX0V%AGgqU;)zD3t>uJuhwh@ieb6sZNH8@0XRgTh)-4(Z@e9h-|{|sE2 zTtAiGpS5NDU1?)!Z1+`Z^ow~8&0ia(GicJY%i7T3PyuOp*#LQ!Bn5F|LYZ&!e-rf1 zyO*5f>GZ{mX2 zSij=WjgFIu3$$kHjZuA>1BfntSg{M|cX#T}7QnlqM1q0MJ6yV=_O5^>Gu@CZJyBVX zZd1IvF}}cnbKSkOBJVt0`{KbLgZ1ju1Q_wALSFGNv%TU^sqxQ-!>Dfj631=y{$u~W z2c_Qsneo)h2#T#WVw^Nqi>Sx^FqX-{tE8DU92n>KpOuPMC zauT>3ZaAmH*qRt{Y0zx;ilu6!EPt#Po@g;8VWk zD-$yrO^6t}J4LU{?T)nS%s2LOTQa$m{UP5ZEg){fjK`YiVCW)LMNz>ct}BQHa)654b5=Inq!T{Q$wcXiw|2X|WmG zDYk^44{&n{DV)1ls}SbGi|E*&6Y(7n zjQ_M|C^WHEV{n~~?X8hHjnG)hUVRaAEr@;Tb<&DcRoMCNm(_NT#j`FaSy;v1SSp;7 z5gpi=(+nXpLGGUm^WP<@z#d$>$+CR&)MX*JbJMmR+ftH`kIY%Y{woQ~;&O=>&keX| zldQ738O3w#4>js;sZ_kk*^_F{4?-1J(lGB!QA+P}SDJ98g-zM)coL)OMDjU zel|AMbLuLC3m5;qqUR?$6t}pyg^Cw1o-bP6pQ!dYVQ1$q(py@5*;8Z=-B^%~8 zszYP1JOA{A8u3*=l9DUxVt*=i=GUnksrRhSayU4USH^4aD^qQy%5d z-FnS?cV0TWImlk<&7-gDDfebG4^4mzM~S_R8p3uccM07hcphs3~rv4M504A_hh8l z@4>PfSX{p`L5JM|eRN)B#__+$@}AeF9)Jq3;}TsYEw0*5AhsI`1D@@hnK`2y51W&& zubEMhy&)I$<}X=4k!(&dei&)RehwHYb)h&y=y8YBp-CK{3bP#C|KxDLh`tc~Uebz1 zEEJKVvwYG;2Y_^*v#tm2=D_OAeuq@As`cH-iYPLuZa+fGcIxOWzbfJV*MDdEB(tt} z_Si7lE9>1RrnQ}=i7^mnxnXdi#cVZAJ0QE`&`5Gsuxf4V`5T|b6I!8=Q=mKOB!%s` zihp{`Z>E5nNDJ@u6UM2EZ&aJ%Ua`MSGdwxFm*4&bjClT+7Zz?p4>)|kbQnpZDH2f} zwB!To7}>PPQ!FI+MiqXv1|G4G>_m!Yv(lojQ&RFp7v>$O`o$OCct1_C^HFQvo?EF@ zuM+jU2OroHqWSH07%EXv5gL!eEpB(TotbQrDZc!~S+n#L+8@<6T%I=nD@G^2*0CxU{JSI-=&OOV54{@v~l8JISZuJ zFY@gOiA)~A%qp7CmDLy9&!g;CtY#eymXo>7Lw)N}t8UEzO<%=d`n|4ZpZ zouK~@(}vC^Afmu>xoAu-h{+{0XL}XV{z2zK?2*0gE#Q=n^IhbezB4!fU_rWm+#@H2 zID4G08PWlEXo4iw1bijkMaS-7A1pFU%1d7Q+)vr-{POwp--kXYi;I4@@2nUh5}bsD z`;1HpEINi+i;SFO`(&M=tH8Jc9Yn0rtd97uEP_$@`@VaDVVf%2e9*h)GF_aRxM~M! za#eR1MA7C@e;fwxY9l{hmKhVvu!1 zGiG=~!f@aJ{6vO_(WjHF`{>F@_~)=N9kfENPGnc(ZH&{GhZUUWTr3bmlpdHyB$4`h z*ACw1-PuLP;RSY?c}~XrJqDmc%V)9NbBqTU|I{E|P0`Y_7tdX;%m~GF&}9zWYLurV z7_2U&^6B~|tpxP}&lwms9YEF2+M)-pjcjSssyh}>n1%-nepnNn^lJb3=QFy?v?A(a zO=AR)#r*((cF&i-kT{CYUrZQ%D*WK(` zh7JW2vE(7x>1a5NT;dp;FAJ8{XC4UfEk~hWo%i5NdA>5zCq0eRfeZy{$H0}Jo@hHl zmm?2#f)Co7@+a11E?xTmvOErW?!GMuQCzy&rZn99l?(2AcGo*vEYwVlUZL2h_g@U} zKUEo@qD{(y&myNH7@T~x^9M?FVM_U{;PYg*c98NB?|v)Qa-O_Kq09t_5MeP#{yO}d6&*g&jUU#>;UXs9j-0Lgdv(7NG2qF zSQIf7b!}QmrHEeo-oiveG6f^+$JT$1v~Xba+|n+kD@ShDx^0y1e!X#}ex&5I``3Wd zO^wORJv()|{Yp$bE*nAMN`nq@BrwdHq?Q~;wlGl!Yz|AYu0kFquSxKY6Y6&*yYST$ zseny^%Os@l(o_j|1-l*3j{Try>?^(i2#YS8EbSfTx^F4hg&evG5^M)z;G4j^4NcWk z?^6)#TLkz^JgaZ3`^`yk^6@~Dn42FKVLN?h_F_Eq$7>C+y{fYdMuw8;NWI2)H5=}C z-JX^$#mr0)O~Xg$Z^FpMx4+jSxZHKi@FjBt*y6RctoNm>lauU@zd+aad-+ohaq}CE ze$rV~>YqQG_R&+R*qU>yk~u6u0YElSFIqSXcvTT6Es*GY-{|7lcX303=`5Ii!Um+ z73cTfS?M^C#lxJt-S;xLJDoqJ5FkHCQJNP8(F>NbV!}U&3EkXdXBqWCFz{FI(meA& z$gXCjM1;8+N3v@i{m@o3DqznS)bJ7C5Xmd%xqSFG5n>TJe&XDhAEaNmVjyWGRcdX5HX&G3MA+{cnYTKy+dvY-4dojd48aAOzE&b)p7% zr2s?Fl@>-oLJP)xrNlZ*PFn7zk*|XnXHAH)i$~D0JxMg1O}7)m3N07an~P6Bs%?r_ zI`yC(0V{fb?MI3xB;R-XnDd*Pf)=ZTxV%J46+vkJ_)nY>BI8h3;*uK8JG)uqvj0N6C<~@Y5@DYB_ONC@nZ> zj&Z@^bWXpo@MGwl_DVt-ybcmOaC%P)mxYF){X@f6>;)aEd${v_O#5J5aJ29+spa^A zCuZg!#MyQ%_!cE(Il@rbd}#|oJLqdFdn0GRwkIJ(!!8xo4$RA3XhgdR_}@-`>zr<7 z_9KNum*bk*jc?)SAVc7x3D02k;wx@<-i4de7QsjF999O4Q%FlsbBWosi@`U@V|J96 zb)^*34&BzC#_X$YqsST$pHk0m2E5I7b7KFG`6UklqUiHc5T<6)z(TrYnX{5XU8DJ~ zO@>YFoMaBMJU@qP?2~#HV$d9y6bh3uocf0D*!g{YKTCD;<1y&W=5gw(>OMv$G(i0C zNv>mCePv#YNpntNiw2my;UmtZwh(O=Az@M$owCX@%oHHyvS@gSmdv&a&|fadfwXw| zj{NxHAttu-ZIj>1J9*x_?_>-l==TrGym{uznSSfIZMkg^SX}o*rIC~XJ8id)7H zwjYsqJ|c`LRCp(6ru{50)ZYDw7E^8In=gs&OaaSyb|Jz0h{ z&(e;9>ASQTPXOrjz~wGw$700CHk1_$1i6BB#&=`WQ)6bnE)Un(&wjl!e9o?iFI1wQ zh?M}%vbSJlH&spY-&iE)b~x&S9i7}q5F*3ZEf1t$G%>z;p2!w(q%BNRQx1zfmNU5p zt-pIZWp^5~B!kQx*~?k~OInn61Rsg4H!nixc8|I4 z-%W|gw<%hVlQl{a9 zc449l`H8S!2$~FX{U<+)FSsE1oC_@4C zb5HJhw0y^F_YWLM29GjVfVt;G#H`1r^sJO0?Z0SA=x5u|!zW+?~ zOPo?#yvGHMAN?i!}Fw9Nf1)CfB8LFq$bD48M*b2M&d$s zZufD2Ow4a6Sop1fuvA*mIa<&20-;vhVz&M1Qq`3Oo73^xd`yQvpnAH8C%j&)fp3Ts z?NPJ#R42fUG4I1y-F}IDgpH(Pcx&*K!_*hM007I$5JjIkQ(KX&H=y@b(z>RtBJx*= z?ymvs@ReOR#I?4{y7Yl_q1e-Y`FY7w@gH#J?;hZUDPk2o-4^VVy#4xC9qt|~^$jU* znSLgG-z1@MBV;*tdX{lQv6S&SyRKk;9xV2Qhp&R=$(F=|^+_aM4)qej=j(2uG}-f& zGDK8PP}tGiia~MbQ?k;1O8uT>+qG1UDwYWxTxNV3D_2zZuuMv1+TNlI0z#+#(pJ3g z(1!Q{vpeFrkD}ZF^~qtOPY8mFruau}D5x8Vc1ause3YN$`uLjor(dNGt~OU786+zO4#?iKC^koX@qya5uS*_euGP%fdELQ%;?Wl?gG7Gid4< zb&2BNgJPzujv8{(4jjmv@5u>f-+jxuRwXpYYJNW)d9taXepl;ybArX2W9#X)eAOc^ zY4az)*RttpB^xmjv#b3(zFGo=MGl68&N8f+vTQ|rN0D!DV|De&jZt^EaJhY#_j=;n zSXL)o@BW`nbc#WWmx)*O=5?veP^3GUD*ZG#(q(+ug5jX~x5pl390dauC*?2^(=P1# z#aWMl53XQmPS}Z+_mIAiz^DA73f5NCym)0XU{n&J&a_=l=f|)bGSif);)mhyRqWSL zRe+Jdi-^G}yJ^0jdj&8{0Cz0D8=~CHP#P=Cl?BDiW7j|*hB%p?^ZC(#wF5twZpksx zCjOLoHE$+x_kO$3wtR}OlC?^@Nk5Kx+*tkL)F*l7`-$*Jp3@}KC|zx=O|6d{MATU9 zw0LlEdrZAN+NO2Y~^kfP4OkqXB;mSsq?(ZlAWr``z)1mk`va^ z#`8VcpXA9(LtGh~wfWwHB9?QYZsbWpM;hN>dgYSx9(@c|9X4%()7(m8BCJ;RQh!4GgF$XB(fY5Aoa^JJSl zzU8m%7W9UIjG7$5X$RBMD!vO7KIGOEy%O&cy1&xOvV|)$(i_9wW4fD z!L{2qS*Isk&U=@-GFY~iG-B~A4%72;v$U#jIyI$N_B@K``sDg+V8(>KZa^VtWP_ur zO|}^RQ8{?hyFvEP6ZJKFWs$WuCfY3MIb0&A6ce>2-#y)`vEbY@PTn>fW#^HywzrUv zY`pjN9iIjiG}HxN5hVE%m^^HCI4Y6-*K-e^8xyxU(B)nGi!T@UNdBFzb{)az?|-tt zoNuuCx{&t#uGWpZ1M7<^L)WeE%b8)uSA>)xkWov>UF5D{PW{CeL~0t_=A zxS1$5)&|FXYm+47N-0&N?&(^uEngE6(QR-?d-^GIcj~BFeQGJu!rt148(Wk`I^+yR zg5fCPHJ%IPP*5)%meJ)8&IWtYcn?$=^J#m%9K81AY#h?nBB-m0vn9Jac11w5;5;pa z0<#<4+L1?bca+$>ClG~mtKy~O1+C3NlH5@v&;^P?fX#=Hgsj?H=O9P4V?~1z6!Y%I zx~&bx1Gzhxk$ubFY>qkfE7vR-B`O9;P!pJF55_DQBNfA?!Mc+xXQGwfVRIi(x^i11iN!P90Y^*|yL~W~&3_=CC&@9@3Gr(Eu zz3y|&ID}3m)=hFoo{=+4a5oz?l7im9!398($DWIRAci9IfstR!3CO5g3CI|bZr5@& zOPnUHy#9^=XF?y4>Nl*3TjFRIi*dUBN1CPas6}M#$KP$bHCdwH$9M5uvF;&yT}{<< zzvDj|m;QrtQm#OZz&2%c_615`PeP2uIdWSP=gXyXe)>g9!mAqb(=$xR+TP{42hL9v z7ua4bs#B%9v&~*aCt06-PA=>uKmk)ESm9dMB0eQarHMxOevrj=I1oPMB>8Da=eB3+edwCw+TwebYN3^ZQ)o>ez#ePj3iX^R(8 z8?3Bsc6R;gjf$RI(zUbdtiPFvsJh!5T13<=3+P?GqpWq0Z_6)V$T}4MR*!R8Tf1Wg1jA*$_(AY8V)2^Q z_dz&dk&Z@mBhJ5)!mH>FhN+|WZ@AUbexZ(CnF37akT}kuUmfCPdA|`Smhi{)x>xVz z<&y|w#^*ZtC&mP#ikd0mzX<03M-@9@igl>yHbLVWf2T%M$vCeVXb+J4zkhK6i<3Ff zvc2##o20&>z2@?GADeFRSpMwyFd>`md+DA`|7FB+$-?#thhc`=c8noZwI`h`or)a> z8S02QZ0)rIVs-giy+E;MZ9e@-jm=x*=cb zN#&t{L#pjb3aTfMwkpZ|;iF)`{ zEpz?LL@xLVn#X||`0CE#?<-pNp+HmI>8>R;k2Z2(Op5?pCC(5*7awJ4q1`*O*vQEC z28JXWYfVm#Zm}+1K;ON-)z`#GUuwN_aCK~uO1`6>swHT6#f;M0z><#b(0T|Kc4tqT z!p$w2bg9Kk&j{&choWt%*QBkNX_wC>Uo8kqc@pg`gH#jv3Sso)?m1IgD%$@+GL+we z8KFul%mzC8KKsi#2*>$n&mpf>xUhkNYBp-Q_q>4$l#$_`rMJDv5|*^TCmFV9#O_Ty z;0b_se)tWI#{nG082yswKEaoSd&{CvV26&9y}sRiDJ)ijP*P<~78G~^L#i*HFh7f& z9s)js@MOB?2t#-3u?O*;VwuL!^mey^?(Sy1`-k8=!vY3Z*bJ^L%=e_!OVsA{l&$#f zV~Y5@&pFI>QVJW)%ph_CAlLVlvjtxPpA;92CgZ_=!s(jD3v3jItHYQf3HFg13*DIx z!y$5=d!|;mNIbnF=e1R98znC5{+;`Of==WYXLdt|g#;|bB*Abytgi%RJ z$##$C8-hEW;7NSvjW{xn4e(phE?`Wy`MoQ>Uvw2FT?wmuG6+w*^BI=D7aFP0W&K}L zMR*?KKQSNhAi5O(cofQ&b1F>cmrKC8Zs9vWlRZY02470tVK=FAy3Vcn=@`WNMJPq# zXUu8A+Yg^tey)ZbxcZb{IrpD2F}c0K1RZfwc+vK2&NjIt^vREEC!xk8?-$r+|r)McI+O*s#Q=Q;B?_K2 z3#=2z3Ft~6rORY;F5Cz=)JaWknjiQ)=MgP1>D|l))WTLQpR;S0WBa<5XI!W_3fW5m zqea`JB=n42Wpz$b=AxTYGz<6&yMDU<3@6T$!i)Af~4bQW@IxS??H(@|eMcd*j1x@D}P zMjd%FnsIuqO=P52{7i`7=IFw%&%D0k@axM>w&nD}BRMbvb|#g$*MW^+FdS_OBAZh^ zNAQnC4U9YYpBQS=0+td4A7doR29w?LcgkjOPquJ7m6U$VCv3qVH-&6wVV49O@Qtoz zrB|ujN0Y4x2=zc}2gx8(ZB`O7%;_Bm9d%aGW~$O&Wfjw{dtR3DpHG+M{SJ0*h!KY* z%qxQmhZ+eInNR>0erH&!mLjv>LAf?d^p7#rp{d;*-3iNc)zwTQ7?`8R+fG}SYDACc z9EfEhszzMK$~5>#YK3axCFRtkXV^0dQGXc-6E8YSC-=Khc~le}$9D<&OUwohn| zuzGmtpA8~rUJv0BH9LPApulE$J-Lw0p&xKS@t(9SwjDME1#;?c_vQS4#sPfd;8xGl z7MI}D&D71q)Hbc)3w5=PWMpWY>~AecS@(vP-|B}5$2T?2rT$;UYSQE!6~~eWRXhJc ztl;IJ!*<*QRY}=DsJ@^}nI?=D(hq1ftXM{eVGnH?EJy%;9=FGsUc+n`6Li^!RDBODL z2WdaNO_bEGw=F0v;S8lNR=6c> z$K6xdwaB~B&C*|TYKZz(^+6W@)g_YK=|tEbiaJDrueXAE!_gO@S8{u4d0m-|qB>9( zH{=g2X$co;*A|5Fli=EDZ5(v5(G*J_bd()sGG_Skq+nWBqjY4Hp9j@#QXnmjFaA+T zH|*~ZeO1X*LQjoUp5@`PCyRD~q4D!R{Fy4vrx*E7dnFX!fF-0Mg;50f#@~Z8!B@8v zs3EDn39K~x9?dpa5R%1BIUSS`<)~7_Cf8d3`b-qjy*tq?ipNsyi;b(|f(oZHb)6=K zuP)zs2lAo$%)0#~AleYgfY&tY@rRyVlj2qXeT65V`0w1fHeRA1+FoM*#741jla=L* zA&{xVWF~P2oR;_0dUYWogN4(akX3vCWU+FN@G>mJ_nRzJ(C*qw*f+yQx(-R0qywqO z9qM|`WrY$rSq#8X(g-%RGAWU(9|X4wfHnMpbO=5}jx0LMlNCbC$av_a_SRS#>#q3a zzQ!M%1+9=02no?EF8HtG*4tyrW_zFg{#r-kG@h4A51Ji+Aenyj64T*XundVC>R>(_ z>QD?*4>ZxsjN`LpWh*iw67K?!kw6|6;oZtbAB*%Q=m&VPbY&lVyuX=CN_W$cmm=@I z&^N%!){R_?`q1U6m_N7SU*I4*c{>jF7eA?_KG5pz&@{a9uyS~_w$*X%LRDt{kjs0A z$G2~|P9&@9EP8%4C{aA@eBq}Jed)VqI*E&jOEB&*>M> zsDPbK)*0h`K#8b}cZ!@mKdeBg8}k`pm5iBTvsI~#P{?oBH#=tasri=mt(%--)eRzLQS%R>y#grGb7fV8|q2F-F$WTGDf%Z4mvAN}s zfuthv`+}?F^CyEaArN6%CPyd-qIA#$ohU0wlRE7>+RH~mnR=1evgZHxf|-)@u0C`$ zh~QL2_@kmWM_NQg^Fg`gL`$K4=YNor+E`)#XQr$IZPSi)kZ9+81<^RD)q8K8F#Yk6Fti+j-WD*hn(x;r?VL_%7^Xj11VQW4FDJgkbP25mO7a^qI8XeJkTd%`C@CQqZsR1 zm+n%0;pu~h+HoiIQdv=1=YR(gefqEvAZZq37Hchgkv#N!v^JkLET08`d1OhjYL06q z&A+pag}nDtKoM~fY=j=im7d}&ec;DuI-724b`pOs1SoAuTx?2qI>|DRAX_46(*9uJ zNXWq+1^v0T8@(nDR%%j=*F!HpCBv;np^QY zgTvllV6Tef_ox0i6_%~#7ls@l8fpcssUx&>@UP8b7Qm^#a6hb$c+4?H=B^R9P9b zv|DKzo2-_+)fnIDTxC_0W~d15=>y6qCZB|l)uvV6r@XL36oJVyP{h{zpYU;KQ?r2sr#Mp z3L5Tc^-|HU;!k}=^Iz5OtX9U}8XJd^Jg0L($nk*%u$ZQQ~4X-;Lq zH%1w9Xn-odUO%$qA=lOIc#_O`xfE|~nSHgYrI9ij%N(IuS%7}Rf0D3VgpGX&S_Yur#e61P$y^-66+!;Ni$> znU3=4O1cO*%SxjaF*2A@N7><_uB|8uZt~oQV#7qiaNjG9Qml25z59!rmXVUqIfYp7 zQ_!I@O>t=}BD=ed^&DmUAMTV)*Qi;6z)wo1V>IhsXY9t=c&8yAdL<6($U~` z531TS%x6Q%?CHa%)SgL+vm?8TSFdijPYfw&A%2rr*^CfFm`rfdU)HE?6zoP4#c3(3 zV{r!7%TS}3#?!5dU{)Yb*9jpZA_WC8%9V{ZHIX4KAe#+YIKUE-IBk!9ihg=O$UVsZ z%(HTl3TXc?o>1W~c?Be&X{+FU9J`He1Mb>diBM|4Yu_Sui4sWzi3b z<;VJnUu}a=(UyqSA66)9xg$*6kCbIO2hG_VJArp)=p)c0oogRF_h0hx|@A*vQ;0BeI!xBM5M? zokbiLAs;3rG%5EY1lP_v{zicY;v)lj0)Kfoo|}fhHMyHG=dOlU>%quST%w+MY;8(s z-)tK5d8_m@sqAe73*|dS(2R@;qp%;jdtOv$IlABF`H%a6P5 zXM!+i`uU(dwvEf`YptlOkuwb9pd&K*4>_@SKqU0JmAPUId9FjU?|(mxRxe4WEtsU7 z<-aU82S5}s1~!x8%|Sa@kZxY&W>;54l8WpzX%XBP z6R6wQp9!gp@+`1*lmFP~@^~W2t*$P~Si$_<$^TMw*ikXrVut5}eI`l(v^blXiet#R zH`$~z$9`O0Fr;V}6Jysnlol@MM|O?^s!(GPS~ATh)!D?~yml((q8!st6BwV2)CYyf z-P^S===e9=%I(MmL{{ihwAj7Sx$rzOm0nABR!Qw2u*g1rA1LkDWo0Qqq7L}= zyi}^nY)G=HN5MYtxxbFQ{FaNt;Mc~sdnJ&3i^}DO%k0x^Oh2hKWauum|5FIQ2jhVPSD2Z6}jAe z2+(mGN7cEkuaR{;j!~0Ho-u}^67m5$wY6l!*k?0~xLwiSlX*6F*r+FGF287qRhBC4 zI1{Q2sBppNWPC5So@L{FSui_!`6XS0oDu9udGv-n+qJ+u`a~VYKK)w)(Ts(=xmb+W zG`pL#=_y`VyVNr7?e57_pT?H3DBg~EYWG-;@fFAqE(UDw==>ZB22-44Y=<8=gOvf{ zV*$)!YRfTDFf^JU;Bz_rSc3(^p3>Hw|1l^2R3v&+Lr5m(Vp}sO6RoOb^l4J>iYVah zxXy>iujW?`_G_glU*r5Kshz{ZIkoqLk@fCdP*B)7u)|SF`pLSSc~llj9euBF$HAcH zEBf0m-f5906&1#iu*FlgdI7>-CTuR4lt5P+nj(;Vg5|qtr zDc&V#k?VH*g4s{=yg4S?mw-o)`dy?_v=s(zD<7C0t$c7=V*=_U?t4aP&B_ea`NzKo>kX zGZ}G|%*hVC{zH&i_)&CfQ!z1)^h`~jpM`0b)d*>WC%jp#-0De)$BJUuFvD%kRl_Y7 z1bi3@c(T`ra_Hvd-ppbL#H+w-3B+u{6}yy7k|~)`g$7jyPel+`+2*^`xKzq0VOG#8 zuL4>Ft7b!X^&`-n#LO-7Zuu%lH(ACqq=G#ui{0fo32m}wf|H+O8_!CcAuYOe^&bRp zF~FAB{IOeWnkXgw8rwCvpiRVaC~qjT#tD6Rz(h`IK6r>3nVAHop^K%21%>nADkWjP zoPc5r)}e8ixq|LnCUCKxplR8(O{&L+7>e?nd42?)T>ZnhNp3>7RVdh8y@eC zm^y-kXxo(A)Kpq25ecri#g3r|LK^!w9#P=Y&NZVu(@G%W{rzYu&&-fT-PlfuzFV5J zmGMLV)`|lMr?0_-fGC-;bpBHNpoybe*xN5UDVlLOtCG08X1xdnR;gqt_XKB0z1$bS z#GGPw-5*|&R{WrcB}EE7I-^xhT9!OUUBcbExzx0C>nQKYO&>m}@t`lGIv~!}CNbHZ zmu{JSjg;)Ys2_^y#CB-*BAt;m1tY}HL(GYP(D%J1@Hz&FpS7 zQQvE9vi>iq{YOK2#r$D1`p$oJI%IU9<&S`%a55BDa5Z8zU~LSB==!Gfnem>0#I zrb5o03k0O1IPWP)-*;uL&qyUiMddjf68jU#byQpQxRikBYZ+Rkz5M67OKi^01ZTjA z|E(`i+~T3ZyJuM9^%3Rn5<(Q@OAU8RE?Q8xF8{_vda-NTw@l@1r%x+;)Uc1WZPg9O z6pD7RyLLyWr(WhhGZpVk;s6wYG()W9r!`I^?)YTCEj z!Uc1&M{CyKHAK^6H!e6-y7zy4^CQ67?;iUz-Pen=p}O12Sjl#60yqwz)5YZ)c)jlE zEA#F%4!`CKEoT>Yv@heYwS-ve?+-HpfjeoKb6)$KTE;%&F{JDB-I%aq^spjgw-o?QRdCNwe8ol z3~0;6=!Uy!HVblVO}f)}pnG{dwqo~n{(FP}XwEJCHnlxC2R;Cw9s>CkiwAbW7h{*Q z+sltUme8d{P1|1|&#%iG9Wo^^M2Wmi?{<`Z*$epNxVagNm1a z{imUjZBMYRNP`kGdp`IcDtrP}{Q-qe^E^67KOft}6>LPdq|3!-bB280^cjlwEUAy& zI-`{;W5)PaowjFv=e>EDYvluHh~7)fC3p4VkF}jf`6g7}3|wL&YnttwP0VHe1ei^$4cjmcnZ(7+&(iJBr?&TP>C3j|C*f3$6;m5)0ZjF&^^f zlgdv|*T&wdtK=DAle!_xGJvWr2a!?Kg2N~Uj0lP`$~ZEt&*4Ie6!%+hSNsrs8O&Fi zNq5Hd)tjEwH+``W)&Q`l{8mAM{e zK7J-`VF;+58zZi^w$z_=IPVfTY~rS0q4vP`R0zeHIbNHO$)(GXpQ44M8|R|cI_#M9 ziRFUmGzN!!7e;iX^ik{hZl(>x)}v&7eWP^BT+$L`Dn2*MGZne~>&k%810y8Z{s`68 z)2@pm_(#}q!L$tB#qR9sYO3JjTLx?JljfSHs+N=?V(Lr!X&DO3%Pr3o+pmm*rF6Z9 zo!R#WU&kwdULm4wWm&oxBoE)T+*(SQW?L%Q2rH+7A-I23n_{plrJ!Rd3asGLH4{Y2 zBRt4>yHGi+{$`br3CB#6BEf`%+%`9+-$Gx1I$Q}&=Km}NvG$DfiZY$)y(#rntncTBobA&K zZ*4Ag+<|WVp%@{?7dLzlPJ@8lVmu)UTj{c42^3mDi38mVO5k_|7<^C?2f6jpY9X-s(y1?TQ zwa9KXhKJ5Zu91+VwElYYX)QW$ftK32AQC=yc~ml7{1P$bc2623`aWxo*0cVj)?fb2 z>8I1nUq6HP(@BeTo@(mM;k!1$HKcHx8Zd<+B7>RUe=X?wWORU-Mkx?+b<}dGJE{PxU^LS*K89L$@vROXv(oMMS5Z;|OYyY?Kx+BjH#(+@q z3=U)ACvM_xEk$LG&9~)S#)*UrN0y(ikx^Vyi837tT&OF-Ymi2zv)VBH9p$3v#D?s1Q^_# z=Cs%YR|OMp`x^$=hN^%w80AT^uQcnOWFkOz{BHsck36+h`@ZOtm}iT%udYr_Z2j|n zrXx;_3;!)+nyMA9ov!lwY0)*On}_uM;(lpmS?zhp&a^omaW78z;T|c3K5qUfq&@mQ_E`4C*T5CYKoxnRmcvLc9YRN`Ue}j)Sp4L_?)6DW zySMfqvYZxq%pq;h&ieKW2n(3r_y@6lfV{Gw!%bk?7(&qp!UzG}Ql15HzTTa^GWld{ zEETH(&-VSqkE|b2@m+ZF_oJv)IemvQLwS3(eDAx~6zy-#4x>@ezs(d+D=xPo2tl%l zm}*DK z!kgx>Q5$u;U$oOd9jqmHrLU&w_9N!OY@_&)tyqWjP0gitXRVX=b@>h<->Yo?^4I** zCH60FMSRl$D3aiIQu`2K#RWBP@meb|}w{dkY){Tbbs{w^Ss%@}s8?Q*&C(L2cS7ET<)a8oz zWLR$0^Q-cS#~uhMK-tLl14W=+UtK$X|^qa{ov+v zoGsc^O}=EyD9xY8knxGMsu-v(Q1~#UoJHksgXmeNnRA>zzSM7NWp%M3@$ILG*-BtRIrK^dp4%I0nLYgB=ix zZl6k@>gSn*Tp}fmdl)%+l68inY!yh}x7~~&B0!rqJa4gOg_{|nlgHe1E=1?6yY0QG za{MLm&)@_aFdmhHixfpiQrpX}2Nawv;TzIm=M9(6SyUDa6Ry?~jShO&Aht{(93Ssi zmh^wAxFWm1E8#?3te&M;aM%|~Z*9$ECx$+%SHG?QUXW(j@Fs?W_P%blZPMwk^=>

rFx=i>3Z>pPi_<bMV|F z8C`Y%%jbGgfA%T-i^$Q`l^4c39r_8}1czfqu1qliU#41$&GjgEagDwSyB6C$^N}b2 ztoP)bX1AjxIyAFEx69^0abJE}tJW)qQ;+17o5o-T_rnWj9MQ*3XA6w&Wad$~69Z6_ zCwW(!U~|roIY(f$c?7T|D<`?N0ww)4L!^+Pn<&B8;{ejMy$J!#d8%Ea^oK&K9?oiv zE*FkS_fdWQaY-R)3S0cTZKVfWRNnnO515iTcyfMMHSQABgmr&l#cyjE5+W4Q^uU zJe*t3H-PtUYDXJ$YnnT^@SU=5uw#oXKbos>AG^h_DH`iABtr4-hV z-L7PFb&}?rAF2Eb*}8#IakYLQ)#K|x>2xuyPG0e6WmuS~R1)3aeX3|M3T3O_CQ1`F z68b7VSlyv6zp9UzJFhrtuVd`@&ihKnSbs55!=#K?4lah=wbk&{kda)N!^zWG7zU!1 z-7IItw6}4p=E@f`hsO-`ImQ1$zz@%-(`JvtE1aTs)0p{(ShV&VFRuH0!QsKeu_jf%Ef{=SHtSpeA`4VK#iyq$ zB4_dkG8uPivKwjDy>7fccvMbNfBR(_3>&Q=^qQVXWrFST%ypd1)QrvFo2#BB@&g^b*maIOM3#BO=E!U~(thhQG6jna$tXgO`I{T8YT&T&iUF zW~aPz%?KUbMYzo$!{{tvl-E?7sd&+5AjEZ_ON{pS$!|`im-SYU%&alKp%%#Q`NM;T z5hAxwoJy-T%PY6rZ}{`asmenL%Z#VBP9#WKSDGNlfsL8wgVrrvZ5{fg2XeSh9;y!gm2=L zm8-H8C#jQ0L=^u_w$W46_esyoJ7$iFLIudKJJ4O|npiZlNO;}sd+=BzM7#uwi?`|b zqW2Ou%|RH!$C_Cd5SF-6ELux%aGOAwJJmVltG0VElnRAsT^{kT4$|xP^w%tm+7%EQ zepGtx<^5Y@IFDO2b<~ICI*KpGuG^EK{Q{qFwj8YSy6<{jj*a=G?Z27U=5{74ppw`n zm{oTMzhg{fKIS1(muIp7bJ&iI{i)Y6phf>b53mTo-rCV)6WJ+YxCy_VhU<` zo(tUN1$%yWMuv4j>YufvvcRw+W$4^lBJ)%%bWZr*zSFvVM;Ytc1&uCJGv~mK<0h(> zB0>fK!uzD_O_}t{pYUs=?@DBmivnTCVE=}3fM3| zh7b_s8>~qBVw7x=P4HAl<$<-hjFMfVfOc4XmS8%P`Y;sJ=q6zB{)BU%xlGU6ez6tJ zV9nf(L7o~aY`IJtIy2~}wS4r3)n(FlnL8bS#bHHP*Try&I69vRjLeRU8h;j;px8uH z=;8o{xhE>fV!O!98x?b7w(4;uNb%~gKGfKbbF3;D(!X5LRtUBNDdUZAmXbASc@hfthM z%%^qK#%{RjQd#`q`Z&+URRRc&sgR9(?iw03 zKZY*TheG}CPEq?wjgdhX2w_#Rb2uRjTt2j)BX+haZoR(}lI&8NSI=5#RjrS8jT#X0 z!u*&r-U8z~`S3~$e+RCg<^w&k(2lzo-y%yY8d^g=fnvJ_^sn`JzHGBV;w{@%U(!74 zJY2e|9=`AArKEq~z2%V0*f=?NA7v@hogLJ$u+3jLB*bG=#dZM%UtKMRxvZldh(WIX zH3i)T)DPIJBdC|w2}R*F-q)>6Gc-F!yUh*Q7z4Thl@Mv&Shz@jIZLRo&5mJ9(`pq) zhxXf;R}OY4<0a!WpKS}gbR!OD1u~6yIOQhIMfX~2ht5PV|LjIs3-KP-;O$WC&S(wJ zwz+PttLpGPX*f&zjd^n4y55;gkT_R9)2gis_O>a^G?Jr=* zGQ`x8wP)E;lbfq)t*E3X%y6~K(c?o-T=YE#i0Q+goXT%?WcMlY!fCguljb7{mFo=( zez~_DE_pdXTgkBProCW*x&!>Vnf_gHQY%pjyZ0-0FDH9tiWAJDCGt=Dv;D38iPjwX zGB#D*)g3-dcWKC(zm#&S?$6gX*^?g$w^s>@b~fRP!r#LLlDHrPuz^x|8BO&>({JDA zSPAPRZHJR@>lh#tW>wxczWr;&lI&ad9~7mZP7eE=1Mxp`il3Aeu^AEs+JH@{N7)kZ zuF>dVT6Zu(>oHwJ=Tm&W-9CA1$H`@*YanDY_zEX4G=no_d@A3PqwzDCc0YD!xG?iK z!J0U$c)>9Y6DYyT4wRi1}SB3^h#l>9Z5A|u}Hi+t#I^!GJP ziLH8O(uRcuknZ-DlCYSW4iEXHf~l)Dqf~bPuA`N`eXsmWtSa65m1jOy9sN>-=`+i~ z{Bg9hp#a$ko*4yq0|o^M5e&@TydI4nXgI>cO%m;3?wpPY0k>!8d#u7}G_fe$iH~8C zO&wAy;FQxN44cay^_a%v7#BGA%cBKFa0u(6)UX3&!rwUz1EGDI*ZDEN726tUwt2L2 zSGy=M*5%9V?Cgz_GGY|Q?59y!wx1W-b`O7i`)=qQVzYVka=2X^jmWXY!w0HTwe`C` zb2ZbrOT4^6{TX~;y$x*$(9b?&lk0D3q|i5ciwd#3IsAU}-i9$5!(5pf^0rLYE$9b+ ziun}Bd3skD)NYeEd z#bvz?$RrkaYZp$ZqY|Ll2H4O^i^zOHQ~~s@za|Zw_O;;Pk|nv5o3!EOkmh?tzoE*2 z{J^ne6RwCLzt`mN#fTx%IBW^%l?UQHniS?0aXvdcCiAUI zW@qdUPkYbqA6fTTb1}I&){%h@dNFI_bJG&6zUbmmLzxttLvofJ#0xP4<9A+k(S{48udecwb$ zM3@S0HO=+vB8ry2l0=L$qdBGlc*v=i@a1mUF=v?U5*7Z|SX@fhX4f@D%`OB(9kttQ zo_@QoW$i!6cx_=YPV_)zP(KYMgz3)Eh6)?Kf&|TD8& zYl7~a$)*+G3-(~49x_z0LeXY$;otnda5CFp8`B%ysvMTUx2bGS^4I_OqH4l$qy}03 z4S_mY8@56)vkrHLF9<=4k*0%IU|?O@%sF$VV~oG6l(*y6lP~gPxEflE zBKitCXVg|bm#zKC>%p{hw{**+AY~11MIkyRudLfq z@1qcW0|PiwNt=}LBKrF4EXN@aviE0i_2y8~U~!T|r^qAmU5GU*U&bAbeetcDJko)a z`i9knKj4#^#c?;?r;fC{?`V(@?P+CIN%?|8)-xZ4$%I^00y(*4b6 zak727Bt1&o6kp&LAGe2?vNio5q_;7bF)$!}=Z`E#39wM1gW}8CE0IDrOlyI_-q zj;YoOh^3Z;akvvWpc$5m7Uog~5;%YoYr(}Vf1S=|q?s$%`9oW~KyFmdVK8(`g($t0 z^U)zP&*A%2cc_m6=7?`eJYW|TA6z)n(!UcTx}j=~mRTGHo-ZIwZtEmBLrt=fzOo1M7DE+g;!k=X8sP;p(39eB(vA2Dyr4$xZ}^xST#018iEs6#`;AiLzL+ z1%VWY2!wa|sjN;Svgvp@ykC_dxQTN-$&H{kjaB)ypR9Vg^QZej z^_Gi6lnTM~i68Uru=WdW`M&WEPv0*Q4rGo+nQ~mL69qxPPG?!j8tI7z3SU?N z!euWa^Il|38n}8hNEmvH8#@4fgyiwu_DLhC^KgE;do@%C~H;j;sdUF2)qsy{jM5DXGivGh8J*dc_o4qT@hz zcb+7sMjb|ewudBw$A*59RU4Cjbr&E*w#YR{deOWmTa@(_@e8Upw(t_??oWJO_|b2t z6m+9cbH+(0w+&x(`M4hvA%z#GS+rOHYBigu4kygvQN&iy{~+HgWw@u^NU)b{Ki`Ij z5BfQpjuu>xDtk7B-3c$*4i$Qy&P46&u^;mDTPc*?UKl^Ae@am!E_T~OeP}*CX}$Nv zl@wPCeZ^;Up!sPSwTK9TLZ zpT%fnjN_Tpv67tpGT>eq?K>J~w9)-WP6_L8rQ?q~`>&6A8_y6J*h`2=?>e(6JDs1wA;VBkk-Y&De9C9=N0Fuie?w*WGEW4gyWle9&2 zhTP`3xOCxP%Fr@al|WV=YsVqH!D6}*B55SE1I0ducNDICtKPD4AM#uEcP+Y^;Zavh zxpKAhE9>ftr?I;5dQ=DXO=;_CB9qp?sn-fxsJ1kC7q!;tK`ctAEjKFPL0UChCHNWa z!<(~IIJPKW9z$k-vDKj9S5P&7s1{|io?#9OCGi#TOjIjqJP9OpxF}XqOBLh|8b>JS zaJ`nCM(TqGW7>J|QD;(FNv~#Mkf(oGa463*Y@fJi{KPZb@$5fOX=GYoLv|s$#F}sv zJqzjXo+xdbuf(zfv1^gYAB?3{3vs@CB5}_g)$Jf+81h&HfVc3*kA8FDct!kg;(zA9 BG9~~3 literal 0 HcmV?d00001

#Kx@*p1BP1`tu|4I|Nbn5bjhVp6s7Sp~y3_fe zH^T;10;hXbe(~%Cu|P@6A(9ty(rq5qP07IQ#d*^c`sU;qgfw(FTGops!gkR~3?V%v z?%kakR$&3Srb~>oZsBFcwgxAgQ*+J79)rMhkDeK4`+u-k(eO``v}}9T=HvtJjZN5* zaOznTfDm}SKiOD)pZ+&yT>Ga3f6e>ft(}q&Mil}plBnBf;|`d%q3@&28wSqoBLyyKIh-1znVyCmo6@{c+VfgtW@g9NEnua?aZ zO%wY7KK^{cwq_wpf=cc{PsJ|$>95}F<0|z&M(&qBVvpBH_xVT%_k7{R+co1RPOqJu zFw_58Y0t;s@xG0{Y5@Del?5k>#hz+loqxos-x`vca_}b9HT9q#QsCPdfvVicl$GuU zmtP}fr^bo)G3EM{^Ut;>+dG-w`xxm3BbQXo9n6rL@{6Oq>1{#eYFE=Xb=c5s->J2= z^jRZ-=Ix&^a`< zId4v4OByge>G&qrX`C9wpbhalKpQ>4kM8nn`DAQ=b~~bGVqJT6D_r|P54LVWLzXyiny8cYAUUq0^~BXv#rCT? zSHiR;VEB6c*L*EbELb(mC^dv=lqVQ|8>_nmri{;k&x*k2lRJo_VCT$O@h+ZbV;~fe zG|h?A{hu&S-*tcAIFZG&9jU;fBMZHoL$XV%x_>TZN2H}y6-Md&$HOkd1_4fx4iLfo3>EIw8yMgm+Wcp}tYhc8d{dWrLvzMnKZ~X0s&(nSt zVQ9(AkH&butKEg@$Lkdu(Nzwa5}(bSfw>H>$s(h6}%Cf4(W=@zCOdivVJ zV&W|hz)n38;yy-V8HJzrA%t=NOC(CJk0Mx(nOFv=?VFel*!FzKkbJbPc&MeJb)q7; z+)Km`?km+>9F4FAEwLE3LO`b0 zmqL``YT|CSaSHr!2MW6xBqnNaMgD*ib~*#aKkx{>Qzg4fs=k1gTjo1GUw5PAeQ9v zw`mvr7fhYPMocS$I%8_wnUiumJ3_0zeb0y%PU_{HH+k?Ld1P*-uAl$&paKjZO6b%C zmo06i&@XMol9~t!^(?K>jRlJ}9{NCskxu84RjpNuIi1{t*h7JSPGiKOET@#`#%U(@ zzw(b;w`nuGL>+(oUAt)iGksQ#b@3+RCjJwOeq{qvz0uE>DD6V=5az)~!GrT=tk6Ke z%>pH*T@&K>+)b=R#?fpn7SvK2PzVj5_M;>EYqFw8qnvu0+kRxER)s`5&6Ljwy&6-B z|NSF%Z#3bQ?Ce^8Y1K+Zc3ufE%ZIF2m(AM$3O(U6==pNSb#CajYNn^@R+h^*NAIqU z)ko-izpak;I1Re@1V-PfO`91$dVxpyP>8Ce5qDEu{Uz?$va9RM) z=W}FKCzu>Bg}C5hRqG7cH-)|bh3pLS+mVvEoAin~m8^*o7yo}So2Irg6Wyw%2>+rA zbJO9vboD%~X?04t^YHEx(rO8~QW67u0_(}$Q$j;0+&yX9M(RUNk`|uxH$U?5uG(M= zd%kfhGG4RVn704?HsSf@{H08Hm^g_o(a$_Mri!A22cy#sfSrD3R*xmHO;kcZN+2ut z=WWKaE8@7y10Ju;NnJ}R@c^aEd`G~i4Rdl`*nrCCkK{b~lC!xtcob(V8*B7J-9-<& zCz+sp*R$IAwSgs zF>h(zA^+D!eyJm9zkAxA4)f{*t-Fqp+%q5*^g~zUjJ<@;>n33gk#lD68lY3G1GuZc z@qB)uQMFoeYH8xW+7s48_#ty%?a7$FL=Y@C!%Q1RdDzG;;EPJQAMsjH87(8Waau7| ziOfBLwL48Gr10}f6j8LDP6t^m!vv=(#T&N%TW#024 z7C|l`Ti|SdJB8d1FvqRMUy(%2(u>9XqGZ?(k*Uggp~|Tw`2dSZ9i5D#%8s^1Ea#>wVn*RYO*2S*Ea7cm`)9c2fwY@vQMC$U|^$6KPatbhL9Y7rx$pT$vK zbx|1(-_aT;#_OocNe70RR=B6SDIwBSa`x6~dd;L}o#neNhb@_Wo$pv?+!I(o4C};S zk3VB@uR!=4M*u;j4#E_PizbMnC5ZW*mkgR;&eC1A7Y~0SJkedQr9kd(ol| z-z8j@9FeiwuWQkSO3PaRL0||)Wn3wE5P*^c`l-jhG z!7_F*_hV_{XSmYO7ZzhG{0{XJy9r5|mnH}Yko#sU1aH?8_CSpqDS3y6bn4Ex$4p3P zYaI50u`BtG8S=p#E(QJ>{_lwpADjNU_Oy;lc#WlG?OG`8+VDRZvAb{u$PMpuz)tR& zkxXi^>m|Z*a7eAtdrk$)TzIEn#zwvL!D|5Tx8L@;;oz1w5AK|3K<`dpLXp;L<4J$b zttp+ymr40v!1CwuY)YJWn{pO%Ss+3xV78B zbOa$Fk=p~gFGO9lBy*M6Xy}Q-1Hip_;YCKa6%NXDyB?r(3(Jx3qlB{8T~2(w;n0D+ z4E{0h+)4aN&XZQp1#9>N;vSGrvoZacmR?7%qxN?KF&bATrn}b;A*AOQ+55;5+Zn7Q zH0R(t?IZgq<@=bP>4A{fY5?pjk|>_|Bfak%fVzGYIz0g+MwBMFgZwFGI_IH@&U`?E z0a2jUlXc`5UYtRWm+5N0^x&2!qmGsy7eYfT;4qOhE=C8-mfudTG)FpmlSQ)k`6S0U z*kbS8fN@XV{&B`VDRuvQtDdy)eV5K%J@jS%rCV@dI1ktIQ&n%lZmYC;^}DH3!P5nA zm-l|hlD~)?d)rc6iOjw69*ynIHP5dr#_2ZE^~9%e7a$HNi6EKn zSekTbY?oL2K5tqU1)Vi^yfNvHX8}BG3jxr^we8|%MdQ|SlJ`&&ac?k)FcNDjz*4}C zkk%i_NNV3j)(jc&O)-A`gYKDJ+%|kyyCMq*av4nXb9by~%rt41gkTs*1Qr4iGGs#t zKunUmCC+FUjRXFf6noso=?$#>mM^VRjQMMIkHkPc-~D5!_C9Vfo9F$mGfGxjR1iRN zYNOE3^RQp1dk(mr&;as;0m6EX^IEx7UFwwVxSUcgNqaDz4!}4Sns!+~xd{@mf{g zy$5!CYzke5#~9zFJy;oq1&sw8_T95^-ZT@q9JH-mjmFXbdJ0frLLW_t9g0RY@~O2K zB8o4PVcVF8Bn7-4j(=jK)tqhzI?=dND~=_PMz>9^JGiRjID^2~4(d7XB{SZ3p)c1d zx*rZ`pIeiylRKihzW>{>!1HCJ^rFgTgXZ>@*UOF*Wo0hke2mfy%UlP-BMGlv1AGT& zSS{JPzjxM`uPR2i0I0FFTYqeGNTo@H$1>l_ul4&Q81Ke?*cX^u%~FTT{4mesev^66 z2V2`8JBN4$GW)PfyCW3mo+{k45KsyI1yTV_15-Us1^fxEx-*n2bJ#(*545dLKE$<* ziy~Xw-(mbJOjzn-`nNyd??*9)Dq(!YD@#jD>xN4)XuNKwne^6Xp0&+C*xM%w+$ZMH zAKrrRc)(_5{%=ZAA#|SOf{ksmh%WaJc3q|$`FH=$JZnr+xT>c1Tng4+l&ycQ5j&ABP?!;>x9&3pmQ0zQOk(CVCGBD2dI z#c#rMIH$9EYP0*K9AK*hVksiF7%{AQL2ZcQ_2yiMZ!0qSPKV@Jkg96n%~#)f=T+~& zaY}`6gsMFFdiIn`Ug2KXGoRnQ4*DRls+K^Q&qviYb!^19^Fs;XA0^9q)rz4D_YkD@ zL^b*laFk~NI8eehk%w@B@9DM;{@TF=320~3^Y+bbIKiJIw99;rPMeZ2z?Fq1ujqii#VV#J*IIDTe zu4Uw*vBx*Y^+lDtHfOv>=7u8W#kD@oE;E}OdRw2voWjnDD!HukY|4E&J-zD%^?KdS z143QB(ZbR}UAHXNVk{hCf%kMD_1<`*5ECPPp8vD)S((jqmg_fSVm>VIk5i47TtC~; zAyVoK3lr_2M&Ih@iH#j&^uOj+c$<0}#_yJw3Uz%yo^;;DLd5OE9P;WXAthes3R4cc zQqCO$-uE@I;P8u2Gz>(!b>oeovoy+NBz2>kBr=;va3maH-y5s<Qw7qtGU7>+czKZOCJ~d5OpY5JwU)5D`(Yk2Qr)vLB{1K= zYCtb#{sNnOI##i}tNkz}!?I*6gSG~+6Xr+wS9`wlwVe)aVL4-_|5sBlYQ*_j-#CHX z<(17nk}U1(y{=WE7uFOZpv} z+q%&WKONP4vI>l#RNf~U`;*V*ulk!GH?CSFs~5#9Ro-hl`uRr5T+5y8GX~=PLnaR> z%YmtZ~LXxc|5G2*f4b3`U=s)Fpc{0TadZf=W@VH*bcmW%kQQ<3|Xcf;b22 z3~8FuPpreFsYQcI>1ob05aKgLUOJ}H;@C`g`ICY>`nn)1C(IxhP-C~xs$_%{c9|bC zDmo@WbA4av3OQlOIQ`e>S6)M5_7i$q))8r6X90N4%$oSlm-M-2*4{_rTU$qNm~Z0O zSGz<U0_hUWQo z+W|kgfSHCbn2&G7MB>qzv)du_Wipmi;bT~6qT0$#!2$V@0^Y}c$6v(u0V`9K*M>Sw0dxh`I1?G;8K^)1Fj0IgN5B&IzZ;MqeBB>U|qC z1VEEI_Y*uQJoqt8tp!`svx6v!L%+fyA@`EAEMsV8!LG;nX{B--oz#tqYd_h}_>+O@ z{$udfIqLb1ahzI#vv)T4R{T`Jwq(AQs=djxXf1F>OO!_&g4uO-E0-E>!j?eLN>@4bAOB!}#2A7+ zQhk}o$ChC^a|cJ+ia>Muj(rN1BoMQmhw8rqrFWWG6?&ni2FKHziu>gq|1%0~YGAo; z?|T_t&+`Y}2b^9G5Rj$unL^d`@F`=OMG9BHcVT13mx@ZjA7fu4n@H(&SBqdt*T0&G zRze9EM@;0Xk+T+8#{FGupfhTO=M%wGhNd6=o?5Qg5bKx!FF$G?o;g+{!xXu8xEzSOCr*dGm_1T0re z192^Z296%C1EQq(V8oOFvt^9!W;(Uk@gr3QLcLyUI+=5EpL6l_>_qNcy|2ZA-_G9c z`1FPECL;a5VVFk-KW@s_B)hw7w56w~@I34wui%ZhCk|Ncj^6L46T@;k{xorKl_Q_W zN#=&eMqW{qMdtax&+@*1>-U;8Q+SJeab!wX^_zI|p4vfq_h+S+hxt*Vn;((apRH?$ z!3q#%AemZAOcYe|K42+%Yo$j80&9dG7J9O-lUPe(OJy*Ra9$u4Af3Kxwbe?N$q=nN8B^c*zSujA zs|RJUF3y22bO}4m9!ftqvk+42kl|@PTx37#q8TXXaF@b5U3q6L;f3bxlj(Cp zRVt4^Cm+hZFc7F4{Z{kDL-pS#gFosuug>pnHZ11b$HPhi3)tt81-ecK=_V;=@P7{x z8_cIM_wj>s2N+Vh+@&o}1b4vheeN_?>}O1PZ?lWlUJrNc|8B$iqfz=K_l>ErUQ&Za zBKI($s`3vH<5Wp|xPsteUR$tn>N(+hxCp{fEAvvDI$S~ps8?k(q; zYiRa@{cVdUFjHTe8e4+EEQH`J6`T^M4zCXRTGJ`r@YJ23c_mmoi+5vua|Sh-y;d2?^>HKNFw4WZo2V$hB8fPwSwA*|= zGq0{BT@bU(w7;w6_p`mdIa>qk=a#XiJE-@irYC#&zx43p>@#H|w~zh4a6`)Krl7zt z&m_|PV*>q3H9tyAr*RNbp;5 z(o!9@ZhF_+yjdUI+!S-<;%irFzFqyR@4+khg40g8ixM)Y4Lu;iOiq_WV)p)bzOo$_ zw9Z=5@d?#4(}L6H^v~SGZMDxGABfcSXje){qU*oSoMzUw&BBnUp&xX$g$-aqLiHkc z<4mIClsnO?)54!+np? z%lmbg2=)Bjk`lI5b3J{%Bowbk`L*-OOST~=#;@=}RqN>>_wtWY&zuAgr{5p>Ha(#K zq;mGN65{gztW0$4p&4JMHe^0eOrz9ex&-K`Xp^97C#>Vh4qV*JaqVZ^g7u2%*()_) zao0SH=e_~mo-3uhgmTs+%lgDX7~fmt^KKxgRjwTW5mOIO0HIJKfU7$Q5eS7iAz-|C z+|e$|h<~%>4p3PC25lm;AP1>1>M2Bc^GkWPNqVr_sVkTk$p88BfbPua*F2lJxR**N z9_cC{S8;c{@*V+u+g-*dLKj>#xh${>s^b06(qoxmc8Iye{lw_yFP$WL3`udv@(q$t zxH9~{hye>u_c`q9)Rh%2A|=wt}uK) z&zG43y4Z&aK1*Cl(4qknjU~NgndNjYZN8~9Kz!)cGD<8&h;7W@S6A%aerj~9#m1H= zz|TY`v|8??9PI5Qco$)sNVV%A#I_sK|H0yry!a7YO#S)`lOuj1vCp#1`{g=mxq>bAoxP8krN%v@utjNb^!kYXc-t5E(rQ{(=JvYaY4oS+D;o6sP& zO;gD2#wgorzBuKIY_RclpOZ4W^}heM^3GNm_j%f|ZhNJySGbM(qr(2Kq7B*#&{ z+VI4^vAL$X=|~9Qu+6huxk9(!rQ3eY8`}@C6dATu_l&b*O26jB5xBTijfoYZRf|&2MYBvLBcIw)iW0xZyvhu8f^@8ZPm;j zPBhA$GWD3MliVu9I*FyZYS1U`7#Ka%c}MS8fd2l< z5eKuQIC*f*P%jh+;)nsmA;B;0=;k@j3+)Edj6^?yUN397)rMk*fg}8JFM?PxBfF() zX-DP%s5`*0i`X^Oo*MDG6NZ9Ny8b^7C)u z4;lwQs?Czlur^XJSlm7NHf;2!ys(1Qk4SE&h0uGf0Tkp3LbBc9Lj>Rk5p`jmC>6}k zE$LkGKd&5` zvQ|=6Wp`Yr;CECaifIdF`5WeN;`Dp`)yu=b(pA@((>kClHlACzp` z6tQP&SQ#myiTeGsmjjH&kDeJ*n&1hZO{cMcSe8{4Zk0UwEB;bmkjw#= zOk6z-UaAthnzgtcj)+?OrK)5tO{goZl6FOpN5Qc{ z?R^Rh=#$6QRl-%mcfgDZXJ0mzI4W%Zc{uS-AEv@q4gRS59uT0i4;gg~IkkTjn1|g6 z8#G#IQd~4zQ1L1QGi!AjmZK429rUf)k-Efk#YBB0jZyC7KGohsp#r+LtzW#RCUn6%iF_Fafc6hl1#-id(Y z#_-HbEu#C@@`+?YzDOx&AEZu|Ypzhapx28co7e&vfL^jvWh&)x^H|7`DE8pHvCxC1 z17rD6{K$=`g2I!6f1ivIJ?F_BTUJ0u6p?xodt##1J@lW+yy`?3UwmYuFFY=x(^1a% zE|g-yDyaqyh^u*H>Cq)Wo3Dl`CA%)?=7i@x<`3n~8pvHq6pHvaA^>?(5h*C}c}d2-=I@31VSfAUifhNw`fDUz?S(50|^_qlEq}mAUL~eD%62-)yj=n0(90j~!m& zx6vF~IZ)rvUfn#+Ejd&tkQQ|IKaT8&Jx@@qSG-%0-5iF%0n{rFEv%J5%t8O?g~jL2gUS$lzSNglXe0L>``bM#s_tSXwFo1QPQ+LQ+F& z8MTGK)1G;h9p3sb>C0vl1HrK><_px1>Q}bE*-(>!ZntT$OtVi7za6g9{pMGcC*us` z#%8;e=Fn(A{Q?84N)4OBj+)r%49rf2w%mXE$tSC%$xwpl>h`pt*1H|n0uom8ugX`L zh%+uXM;mx&hirkA@Cc!P{+3e-!A~;H-b(woyN4U^k!{@gsp}$xFrg(hSC&NG=p2e6 zG?*7Km4Wfh8cM_sD+q}&ou~;Jx@#uF^~*zYT&E^x6w5iI)w1%JL(pyqzc za;Kea-YM!+QAzk0tqX5HcL*Llbd7Lei^xpHr_NLrF+y&&RuIfzm*Olaw7Pvprj+$- z^7FkF`1=!EODy^{qmV&JN=}>ds6J|PfxqJuFbweiK0sXg{2#1V?IpAH5@}teTM3bd z^hGKz8;8tqfByDs0ljt8KKZ+rw7q+dSxy%+u!?vaj72 ziCg$mK($JwiOcxOyAI?=n;yo~?v3;k?}ZddHGHbC(J*!>dHJTt{PDIiR{H3toij6s zPngyV@T{;XFhqdlp5r|>6#yH+yavjp1gNtoc{!#Bsf1KMFlJ5R@9rq&s^E=7~Xw9x^jFz(fHBIRiqP72&J(Ol3< zh`}u*5jZs{wuc~dovBHnJ0iFy5omeHhuV~iMWJ7NprA|NGn|?L3^;kCYj|=pr6SRF zz$fzZvEMHbzb%40@!o(6d3xIcUM3j%0oAt%5Gf9IfSYp_#O8;33za#_RQ z))m{g7=64J98-L+&&(047J2m3f=Uz|98VNNp>aMnm(kDJ4YA2SjSssB^$8vMQ);GL z0yJ1{2_+uWuYqHzN!pR0KZc7aeWr<1Vi%*3{2g#=923p>3fRc^_SzPKZkjJT<*<#4 z=Ost7se5EuyhZdmo7%q@VMRjHl1I+M%t2Qt=8)?WN~X=68^}vtU#f?0{|5_1OnK4u z+N;Ids!c3C16Zf~x|-C?p`g20S8adoX=4tSupV71yyUol@0?d!`!OUyS$6@T(JIm# z4QZ7=_!Y9GL%gmRn|qMu=uCRvbd+2D%w9`8M5 z0q-&zaxYp?7yR>SWAMH{taUTor{rkX$91D`J4u3VRT*4WxHX2gq9G-R{C$cS$yfz-My~U3d$a;N7n_w}axoj^EtU8V2_&UB z+`N)9um>L7Cd;~5sLL|tix5Zo2rek4&>2EhLK5S)u7OEEBGouIJ~V$F3gZz|Pb>7x&ayT9F=|$}>geGnhQKQudpUmyyOMT2ipq#)<55Ga>P^?jEOyS?j>zL&ivR2*8lm^ zoH~ix{B~4}eDbQJw>Be!AA3jo$Jdv>m#twO8}})20 z4fs%fI`DDxwHX~~B@B&2Yw7_q-5uiJ$Y7Tz$v@;(A!V@ktkT2=`LR|f&sQF1U-BFi zz85-}5B~aIr5+QE_F`XyKjO=R`5QUDC|4dheF#FzS4eBC+i;_<@Hng<%e}bCKkWiw z6wgUd{0D1d0Ox^q?>PNs2Zt|t9*qQ__W?FG+4kFgAE!PNSX7e%pDvL<$=}L*iU!9H zLI@klBE+|Cx#gNujt`^isIg7lel(4L=bWv1Zb=J`Yqj&E4qRZoY9@XS2({Jz^}R5( zz(d4UT7cz;5jnZ_z$LO}EM=+Wfobz zTZDF;ikrOc%mApnm;(=m|)XexDPGhK-*`EGK|GHtzTk(DLmNJ znBA+lQ&r+P>5O}1$X_={{zUqe9{;1%8dKYc32m$kyg&5*3aFUBb5ljL#25U1qm<6P4@K*OK{|$oVLPMdle}2D7~oje#X0-?ca!7sIu=NGA|)Dk075X6 z@RXN`dL%;}{Su@L=JNcH&AM@Ai%JzeEj@V&X}Dc(9#+%ME>F5XD|qlSUp; z^icI**(mg`?5;5q8r?{sfqKD9zDWOTtPS9F2{03=rb>E@|6nX9+A2eq22X4=;(sCG zwsH|8ZE(om8gedyoZFdud*9RCO-_zb`Cphe5A;RvFN_7YK8kzppYXdKDGrRd0(eh= zq6VKxR9HhO;0E!c0FH%^-^_WD2V^L+(8eX`6|>eJAo6m_Ss6WQh_7qPjH4Ha6H>Gr z5JpSnl_N9Oax-UNUwRaU>jOQw8={?@(Nx9|bLi2`z| z6Szr%;w$ByMaA4f)0DwhFh3MOIX^$^YaG`zk-QF!;)1)$Z?b)x zEoExo59^~Oe_pT)(%7vyHg!4*SIwUyG{|$MMY{?0IYdr4#*$sidxiH^MfI91z6Wrf zT0L>r6+!EI!yWab@9`m`NH0BhrdTk{GPEE z_<^Wh<;w142ZO*)h>2&n{!=P}&6@$zfrLI{Z1}J8qNPG1W@(olPM!Z*DWN^!f@7T( z)3F^Vz_DxZTwr!Nd0oH?k*A| zt?Px(W}l?cH%YBqVp0vVFb!#0nLR2hg1@5+3nEg_ZHcLG3#x8ls4;RNj2P}-|B5aq{mlJ9 z#rPJMR49C26q=cVOsEU7rt%fbI29<8BC1KcR{wQ{fmNr7?o;wFaQ!jpKeRG?nsz-9 zz2mLWZ|rc=6ysW|l?p>!M7mGK}kF}voGEpjxo8ytcxPtcaWt>= z4hnrWZCSi@x#p7(e(#Lb)1=G*fT(4Ef=YoOi4Ws2>mabV08N0yTzM#nRxRhR6xS5T zXm!BTtqOlRIA3}nf=DlBWfUT{#^5(WK zh?Vd2Nt-`jn#CM@=}py}r9-)A`A5V?`)S)ioe{!Bs&XL8uB6&3glv@E+P!Y-cYhn8I1Ax3%%6 zihDn_KGvz+lM1uDA9yoKQWj|<0Xtp%Xh&57`H1=o5Vbz?=G+2$y}h}#9&#Cl3xXWokjw|CQLhOw8PyLOfKF*iCk zZH9fVTVIJzh zXW1O%S9FLG3Rn$UG+EI9FMBOggLa3LQNukT{j(0U!gZ8 z``ES%P0cf%^*|X@5+?=vnoL`K0#cES3s4ccET>X$w*cj@`e>WCDyn( z_(828sEF*S8P>OTW87IgwZ#o-0J|$#a>TaJ{Y^2O>i+7QV~F34g9R-!G2Mc-Py_{` z$h`p4YjZWMDYLDV?x=$4cPG{BiD8_q8}pVgOf>i}jhh4*kMG+X_va2K$4md5Tj$<) z@3M60kE6Dsl_0I||6t-+PZ}Wsr%%$=cT~Xa^cdwT(*~5*88@0our}0+_aTv9XYXq- zw4}Q*43zE5Rs+rF@#3Q$YGmIm4Qp6?%;(=f555lw&FCx>R8hp(i}n{Sc|&y{A}w z6PGt3%`0I>KKW>&8D)ARxlBi0_C){NbGGNxW?hc%O^FRj^~8@TrIx7?gX z4Q+0`!+!JmqwUsXO81^upj`d(E*C^p{0!TWa-H-G4=tE`UWS@=*p;qqZ(=VQ_$yHJ z-W6K0zB)#zw@r{Z3aZ+1#5B|0G`Psj{8a9an{|y7 z+9C7ib6PyofOsAHVb;-o2P}-{?hnBvj%+`?{BhC9NTX09*g_<|u)uuPEF$RFAxE>I z!4t5np+mu`>BhXnpNv*)mMw&;iIJsd(R7t>>q`@+f1*Ug@!SV%bHv}$Zy8IW^4Q?f zjYShSjAsk(2-1JB&sOG#Z|nr{+=TYPSX1Ee5Ci!vLxj#Kw;}}1vr+Rvmj}*;2vgW@I~MpF@Co(sEitRXT*6NWfh}j*ZT7hCRPTtnouj! z&}X5`^h-~zDtX21LiKE#oB#g356?sz?+$X~|j?3FV6BL#T94L*OfuT_M?(Sx7hwYCZ3ZW|LvO<6tl+UCj8;-?Sfo0hg2xTOf|&#_`N^ z6+|+k-PoqR`T@wUi0wmDChXhHUoD8FH#i0E0Z_6{Mqp^#Y>fA~L`7}ky1!oEyum4Cq> z#mn&ZmYZE_n31_Fe_u@DJDmge<|bZkOEIse`_z(J&0hMQb2P6AK31-hy?AV2j4O_rJ;g;`ba0par?o4_kX(UiaNl`|;(I7*qh)Evg%hD2*= z$ocW7b54p$NYHHj8p60fN9ig{Stis@8%!0!zaXIfzg{xk*b&T(j(JGSaOFEo=nPat z)r%r$BxA!k7Fd_zxzR*(%YUv_Ug0FxHj8Ju=8TO)+h!VGzX_sFXsU8j4C{%<(sb=!FZR1=t zt>CoLYW2!LfT+f_c{zLfjyQ3;3;cdo_H)Ag=Id9x=M#wAt^@S>bvOQ)ngp|{OB`tj zrO{9bOfLKR*Vaf{K(ICIauY2>{HmX$XK_h_W`R%G^+L5aQuXb<<2w59uDy$q(rA}m z1q~ro&1yG7o_m$|2zhgeoJ$)@#-ouj9+G40Y4)8&**3(Yxh!xm%jh+Sk-Wy*_F~PT ztD8Wz>Rq(lrW&UgpiSy_3^#DDKk`Y*bYSm0ANz7xH^z@Zma{{BMT<>C>EotnpzQgb{tS74?xonBXR~GMy`$3`^Pto zztOMAfAanQ_4lC%BTDC9eaQEK>kHBtmKTU)t5xd`_UGHi^wyPQ%RGL%#tqEQJ-?4P z+1u0eeT?xfBRDj4@1XmtD%l1N3{i{V4PXG8pwK3d-NtJHrwK6)rL-RDXbbr<=V_NR zQn2pje&m3;nn;DZw`26C%%x04bi?+DxRLA9-BrbQaSRKD2lhPg9af7~&Hfil$63*< zmF@0xPU6N8pgYwCPYZXBMORLq#i=c((FW{F>WSH-ldel1vQ3fdU2Jyn=iFn%WJ|Qf zNYptYcV-6lW;Mqc-=RPZz$3V5QZ-jri;)Xvji-h>5p}pYnlQtwsCCG1{Y+RidSnm`TQPz@C zdxd$;A2rMk{a<14K~#tJMFMfdSXIg_vR4|@Na|h=Qg~0&xPLS9o?SrwwCYjIkUaOZB)-|= zmZnthe$FLu*6tz(6d=zM+=jXPS<94_a7|(`FCiF_$US=h6;yiSt3e7L{?ky$uXK7D zhZwI{bzOBBuU`gOK_AOKy*Dm?6e2%_ocSW47_v=^RgzW>Irp~h1D_PjA0x<`TE1A? zTQnS+}{N?0q_%~$Aaf1vH0t}wDqN~-`OF|?ohZOO1PFYR%`+zUy+!K{>b zFr*IBpD^|L5BoIl1%DrZkevI^CEnyBbRg`fFO2+!Q0=60bZ+zu>VcCD~^g}EExp|JK> zV^mjr`rHK8?wt-3I zH)JGjJOp}@n6`cO(IFY}uJGtJyr{uy^m--UKGW|I)M*)dEGxqEe#F-As^?)ygL*;+ z@IH;;oSoy|gWNdBxkukZ!6nO^zsf1|Q+;%e%|76WCk72(Tkn~=?6lU(6ph_+!R8>Mr@^CZ zkIqZv9f4R2#LIDWB-F39B2WdAPMehzU>mzXq|U{fKLAs8MbHTv*vQG3Hyx8rz0s7& zoq&~XhNOkCjZI1ZaP8QxyjQ%o){s%c<*sSB)2RY~+d&)o!XLnu0Y|{r7L_32nl(l1 z{j^4>%_Cf4C@7-ktu~&L53Y54ieyX&{${n@Z-td}cp4*BwOg$?V*Q_wP(sVcr`F)J z-@$k2+3Rg3OqKS$*9aXn*8x~N{%aVL`s+91Z=VoXn#1Hh<+^1I@#V%~T3oZG|6ohX z{~kTWwI;{`gY6U_?gqm`h->!!|NYAX5OIzpU164Bt3{}<;ya9plw7~Ur@XvYm!UK3 zP{XttutGQFtF3kC05K1OyHi-jxKA5j7nU?~_|fg&={DVi+w&ZCC2hEC?F$+(oEN5# zCh?r+nAoNVrMy_2LqI|QqK0#0q%NeG_Bvb`LUoDYcV(?Pid{)_J2$EiY6@G7wD*p! z%AQ6aIqC&dG5h}g$E%}Pe;S5`ZAin^B9@bWl?0yi8Z7vhiLW@HHgyeV8e}0GMRk5A z79|ARa-5!d1oxCd%G}0YNG~`VO)Bd<--i<@5>&0pK()k}q(5yRG?r-n=Nln0Z@IK0 zym&NS!nn`@roSy?H|)$as>QjpR|^Y7bf+`6{?dA`5BqSs(AEXm4#p-)MLR2g`kZ+v zco4QL#&^q?u&}pXp#j@wXFoJ-O!cg{${+0s2-p7ZSGnzb{ap>iJ##*O*Z#bG_$f2j zq4-yk@Q9yR9D|Vlpo_R z-xL&7o+;O0eKK(tv$kkgShCo0NTx(ocjD^tnM3{cRm%6|J$E}trDQ(0f(gdHyK!j8 z5rto!IoFET|6ZwqsbKqRdSH8sc6{D08!xOuNzE1%`;)rCc?r9*1LiULPj>H9b#z`~ zsc?pKPM`w_k!plSy%atz>`Fgjn$-i-J~Mp(re?NbB27{Ko}Fj=zO6@fr}mswj28Z= z@GNsi680)3^_bh<6giWpPaAV|7k&i)F6%#R8axuE+-}r&$3zu{lv&?Mm&%;N#; z*}6F;r0*)287YFTb*MenfR~%KuPzUHbVu!02~GzDK^%AI@Na7{$k;|Nq6?+cpGC(w zHk~|dA^8dM&dyN6D+qjiriY%zKLx8`|Pyt#!naEJ5!59WygbMRmoDF)|TxCe~G z&rcD7;h*VVj=wKoj+GJ1SH${(Ap%Dk7YK|BR}9E^;Y^^27f|rUx)zp2Yg^3nnScXs z2bV=#(vB#%JXAjU26lY@0fFi4IK7iXU<{SvCg*hDk+2agU7elhfO#}N-b15i6stvX zFl?7}xAL8rz#38u_MF#>S_2u)tRK^DiJf=)bjb4ZVbw2A#-9}g!HFCb1&*h4=Ah); zVX;0SG2KI!2Nu~_1heE-@M{c$jf328ZW?wC2wiOBu`6k+Rz+7dSU$-HF0S@Ej&7rG zT(ocBe7u%ezvNyJW$Mfx3k5ISUqqSXL4}yW62pRKVw`!~QI--{ z7eo_H{GD`)?C3A&&H-RPyriE+bI}jlRsYc0;aFI`eUmkYc3rX@w85P*aQ==E90>4KqQ@1K>I}+A`Daf{&sQ+=HsK6XOMQ=a$RYtdV&%;R1I|AO`Vk<7T zTh9sgfMo5~2NVsx(!kNh^?P+*Zc_o44qwZ$t`pWRKIQ#oz@B&~_R`o|$}-RAT+MqD z#a+!`RD6&4anGj5b0HUJ7!rV>qjks`Y5e!u!UAvM;Lz?3vDnT9qXj=f8G)PqcJIel zb@XSC^e$f*GdP&IbIw704Jz_ngpldo#Qx z>DO3jzLwQY_NkQL+pY`50vz}tUK#*C66J|v{GDXKTo26}CcGPs7z&{2TpqHnQ&WJ# zs3h8|w0p{Or$AdBslL4XUUCMf|LOI~=`>iR8MBWs;>GdgI`W>h25_$WA>eB~Au19y zW@s6ShdrR_?r>Cm15pj**eSf>p*_(mM2+AC(cmDqReL0VZd4D@=gv4w(CFv2 z?bN4@GQoA=uFRDSqiOYt5|x$#7>;-yT)60DZYZ52ZWRi02&L+FM_I!(pOz1#A0zYK zIpmjICn??1(mvH<*U&G_85^;2$-sCki+TZQLaMXa`|L^W!W*wE?((nyCJFaF+WJ7z zYllmxtzgR?+pV#wObbfzaCG1M)xIi$Ay=61DWT|?J1116>JP!fRIg-$b4bn#7O4nu z^5b|VP@o=4*+ll~MR{n&cD5nEYNOZ6RvX0OCt2oM5v4z^r&d?=^ugG@zexQCN43r} zOe{&~5%W!%rKa!}{wS&;qi=p%kM?+>7fm*yZCmJuC(5&PHqW&_bgUh&kJPfNKWgV) z457PRuczn;+9mzSwyw?C2y@m&yM?^F1^Y{iPCr^mj~-I{5RPkIFM?4*RnnjS2UEF0 z2Y8Z)qvy74h8YbaPlm%$o{Sk-qz%UiH^{~#khe495OrE{t(LkUUT`g;KQFO5BvTj7 z+|Uio>+2~{v%A*iWM#&l0Q;_-CZ#@hV6}{K;ggNq%?}QLRsbZjdB7q+UrkmA z1!LxAWGt~3z4F0pdmKj*vY~G)U77ae99k??9ofS~B(;hA8aMRCfp4rjNeO%Jh-C#- zwQN^6Vd{jhw!gD2RW9e=L+dZ47$I8kE6zcXXx5{EI0QJIec3oS=u$L4e^hGA{fDqu?umE(5J|tmT0#&eovQS+al8_e_*(I*reO^19m;m%A6$6&G zIc8zqP2)e=x%1Lf3oX&0j(_Br3q$8c6YR%-uxrEwz(c7{ny<5oZ9tI2r`aMIG|zGD z#kPGc`JN&25IwAOD4qP_<9MJ}f)gt}_r(%M zXI~h@{(|X>r!SjQiGGB1gm8V!yA@L}0Bqb3KCz#@TGR3fCe(Ri{nkMvzPDtiUf^+c| zpT*ZLs#F{kwARYCR;TN5ldWvs_27+FBaFtI*R z6K9tV1-%2J+p(IuiQ6##>_<#^8aF7Iq9-szkb%lvFkW5$p;C@@e(~bhX3j9LEKM{W znK@n4y8<(}0#AF7)nujRGE`YY>_;p_A7RRi#s&Ovu=Ql!7tPgp1f2d7UTPh_eNid2 zp|Uit9xvs`K4yH^chfy1p`LMUTv5fWlpWXI_|;x_I|Y6bJTe%X zBPOBA2VJCn>6wC~$Qwv1C>UezXJ*D_-d#M6MVjHWHsPT z!>roKCjE5Fn0`umk*(f}PJ@YkzQ`|z$CceJO*yl(-k>eJFrgJCa^OvOaA&D*w-;=G zujB31tl_BLrfP>UFksMge8_WRn&@~(H18h1zWxj7h3!13hGhIyYva{An#CXA^-rwt zivMd%=tVk0r*MLi_;1nNqa6qkkbo}t_iG!J8$rFng-7SLh!^gEQs{sQtORu|KfAVW z^w(OTW2oA6SO`r)B!Pm16MaY=dmaj<&4zWFn=x_K{?(Y^9@T6@$TnlmpVxEnrsRRdCa&1}%n7@huRhgHwd<3}sY8-wR16$G8q} z09Wl^Z6S=5p3c{3YW{kSm0lc*lDf4S$r~$s=r~%G_gy2Yvfz;*NGM`*aG#L1xgA3f z5H%T$-JAmhEVoW`SzZNJoQ9I`XZq0H#|leXf}NLfs%?$6vE9|LJA8a}Ur~9X3kCWW zi_#97zsOU%YoCf6pQ=V7_ne8lx%cAZKuehIu52UE4%OxjZj^cEDDt)*IhjMzA*^)W zXVGb)t1Ev?=R~Jbu|7S<(afmdRo}+5^h&YpE)9zp{i3Z7sn;?;eQ#8?x%OVQ<(#1N^)1E6t}X=okUvv$t)P-lg%OcawsmevDKn<3787zM^bO^WYEtK~i`rZuFh^8&t1?@ILw ziG2mbwS$;h#8iu>C&!3ja)$C7 zm&Y^b21=onOWAYxvRtKPYU_l%4#C#+rYiU%`7CGlIbJ!@Q%{jk=I@8jGu^s@CrlYm zNt3`aU*;bJepYFi(Me;PZ+b9ExB{l<+?+EvV3l@FQp-s=*p>9atb z#VY-Sa{3}-kxZv*GBAEdar`j$&fqB9&L0FfsSrby^|G6<$9h#y)}EK)p0A#THYcb< zUHAvaAt@3s>C;Hr*;{RhcRL{uGaE~SeY#0Xc~q7{6y9JC$3TM{wg*Q#jsA;GkAOdG( zP|?aZ#Y;*FiP~5sN868F&=yBa(xwI6>OVD3>duvaIOXSBG3Ty}{b$aE|uJSOoXfHZ4%5^ANyVV4sU#4)gYSG@+(C^KO zQ}$q{*(I{s?0~XWD@XiEvy1tsfbUUv)G0$;DUB#0#(R!mk;P7Nj;8BzBmSyfGXBpcP!rmp0qr?|ErCtp)!TV%JM$HFr|eRr?7dCT z{iCCq!p*#YQ-_Sy4={N^!56a zCbm|>%4%z9ndNi8^L9zwA70Bb#@^j|z}ow-uV8*IWr8bZ;DtWOp64})E$rEeSb11_ zu*~eokZ*2K;jy@*GXTTpF#epD zk|B zJ5)?9i3Y!(nH#`V+*iw2e4u~9+E?)RNr&f%;2WMv!O}ydm|h{5Q**2%rxAMUPj9-T=DX0|ViTKhHU7!GP z>S=EH-knpZ%QQ=`tDK#}9ImYY;COE0P072QFLvJBhv5t1Xc;KF)q-!R)LVh3m4`mX zTJY81jt1)U0)Z+6(_G*D6E*!YyQ8e}Raj8@6_0~&8k?MVY6vkdM_34zb8qlo5q*hd zO}a%Jw88494fXnGNcfTb+LVzyvG-{Rg%i)o4Ykv|NIDJqEoE<3Uq#)dUpOz5D-W~7 zY@ICHVGgthNMHC?EOLS`4XM)|d-a5|2Uzwq$fPzED2x`C%mvePA&y&m?ix6X`h|Zn zmi;@b!$M<-Hyj&|wAMY<$eN78eZz`@6F3@jW(R=3HF|(R!Y~bM2J%X*Gq2SAAY?82 z;mGfiX&R!VI!;aZtEmnm#nkOZz`Fqd&T6`>1)8u_)SwekHp_AnUs%aQ+W7Zy0uQ$Axbg(>{K zU*r6q=A-D3MZj3h7A5LSfa7*!BYDl-08fA5vpfmtX31?ji#2ECPmQ%1`TzLSf!MOP z@2xc_1+Q=m{<^&;C-30$!Q+?AEi8l{|2s$PLB}894C0{{&Ixc|38B~pV+|Psm_HYo z@-bXTZ%>-u4$MP3AXRbPdh+{TEd9b(9#g2oIs|QR%d>u)7cx+CwQq|@-?bHizY}jz zET%4F0j4<4K$o3OVEez zc6R1YI~@fo)uAM@aUc~>Z$C5{o)@!*^j+TvoL*i`t3H`-L-eY?7yH;iSfDuCDZ65 z?9Rq|pL!L%MB=U^3i~Ytw*C0Rdo9A9?IxS_YL90sz`o~B&}O#8(bUi(%W3sS@IMZj z87=pDWuNH9tPPbM0W*B&FI5)O6XuuV*Vvj-@`7f^5#PL)3LkLauk9M%a^o7@1GbtS zDQwoKH5vwrMP5w`svqi}Nt@htbPTfC?sRV^bpaOqY%^%7*O=aTXdjF&W-WZ-Zupca z3MRkn?bS3ObsJQu^MUDI06Rt(vudG=kH-qo35uT^+oYM!DO|@O_1npC35=WD#EqZ(j{XGDay9<5%$ zOA%H(=VuI0Y<0kfhxXZj%K&jGo~|Z^=XaDjXJR?FAshn_^GGSB8jB#dFJI3R-H9rNeZjZ z%V^X0ms^B-`y(w~Yf031${)Mq&;1We=iyJ~|G)8D;*fES>~R!Qq3q33%1CJ1s}PbT zgdFD>A@f8gfQme_3?12*@MgjwA*v9E*R^} zERH_ah&(>K$Y^4QsSG9}dxoZ+7hOgVJbtnoSINXE)*tKb50>E$e=&th;HpBf#XJ}{ znD`^jyI>+DRTbAR_7X+Rg!GFOEsiy3;sZB8^p#$ra(x}P@7ay9p=f1^d)Ar z_MS^U%0-)8LxS3~v|Eq#O9-EMFY0}J$5_;1dV1u2X@wHRen@8^=iKY4&!H*fb*Y;_O6nFjxN@I!TkbpX?AobD z!O7e2a|T82pB~FO%h%6K6@8&>eZyDsGDmek7bl@uaHE44O+Yr_*kWp=7G{_oGo2N} z++f6lP(**Mh^vie!RjRKBv|`#jsq>1hiUMB&nc7@op3*jLEC+&`_53gd*^0pWJYLp z;B|AaY_lpQud=U&f2d$>uyyJ6CkVOrEPa)QPH0(67z1S&?d-H=e`xnOP8ezqT7jfBnYP(Zmc!ve`enawT5>5W6l#3pzb zHb-#}bCKpBV&eGFqAY+`NDWLy9l_0eF%3`6ReXuW%Dy-{F(FrYKeVQ%EHXtE`iUeZ zf}Rd{&!>6geup8-cDD={?C~1qllJQX~8hj;J1z2eTod$bfNgWv;FUa- zh7W0VxZ%YPq$!q?hahuCp+Par#4h|;IUVN7d{Z+N!1J2+>&2^OS>D{x*VD|*OXCyE z@g{r;<_{UEyLouhd*9%dTeS&jM@*O7#PR!zKnRd@i!<6{ z&8mWDgN$vpPhs$tXN5v^e0B%V@?8)eEI$2j`6XzUb4#0Bk`D$c9;|z*rEZD9K3T|J zDKd$a>F-(@y0O^{U|o?LIZsK7F)Gx`KLjI)dn=m5S4cM`7k?UJ9$xH>_oGzaLrcbB zyFtHdkuC*5NS;=_i8S1~KSXfpNqP8EfAdWlkmFvibpI`7V({=#-NDpD6ZXM>kb&Y!FL$D|IQq-7 zu+$ETZkjlwAhTDCS?N4>V{>*+cAMPs+N1jyhY=DNxAMZ7UpO~ZJoi{xM7aDBnLOV| zT!6`w+=K_3%$2YbVD})3(8J6*spuzn=#7 z9KV#0QA8xQO+(rKinreLU-1g5}4p}AvDeY%_g;!c2R&tf@v2ij>^TDqlI`4(>tG~<% z2wmPQUB(awrX;+`UBtJwL{P&k?m}F}VV&~6h$~t0i^;@T@c z#HFQW)W7u2H0$nd{^eMhT5<*iL;vfu92)+EZa3=V1DU%`+B&090oy1{Tk%hpyH*k%$XfHZFibjX%z)JhOQ z`Ml@|^C>-!k+EB`Ku>8T67cCA;*GS+SWrDDDMDX&iqp-6XgHh8oUw6K)SJxD4J>eh za=3Ltz&}YnJk-oZn1o0g9#9c(K5tg*`D*EQ;W2yDFHb~MpV^*$qhht!t5?VKO!wCZ z1%8$YGW2&7gK25`> z%DHvKO^l#4D?FFoU+qMvt-AXh`t;3Xr8!4}8|AxK31I~`pp6^Zct2q5oTR`$ZNOpx zG@*R%eKuc3aS~l|F?6|G(~MhgcCV;E(Id4fO6DA8YtRkK$Z!QF+_i2JrA(1KlpPjX?LOmC zAf@s4l#AqQ^C8nJLtfL@x}tlGezxH*6=P=x~vn zQkrE!4>m}T`o?_W8et$O_@LCg{M0Mj0+x~%eC$hMNyYownsj3`F4_tO%RaareX*Du zfMQEp*Ahg0g3fu z;TOlJuXY%QLoq{Z=YhHzz54(eh!q9Z6ZF?thk*0OZ33-2AAA=JZZ>W+9&M7vVp}f` zT6-XpF@hS8+OO#wJ=EKwV~zP;r?%H!G>;7;KV1Wjq^&uc@&telQKsjBlDtPLhxJ zDg-)D0M;e-Gf1q`X-f6$*?rz|hz!a1f8 z5yr8V>PCid0r1CGFND6Fn2fzi_^?>%%}Fy))j14(UK>(F7?=w4sTmC!4S^dyb{!%M=GqRWK7zmvR}s5p{a^E%GMl74ZA64(rP;HB6h? z(1sY-X+JQcZ5J!;s@!MiJZFP;I3Fnj)rOZ_GA4fd=Fh#fomn_I44c7=V1lc|IQx#X zM9?$H2E>{+>nNLtX-7K_`e7+B93kppJxIP7d!I3||HWo7<^cv3NeW9<6ry)j*XR?m z4FcJFsOKN{n1ps<142QJdXajJB)^v>QjBvH{pr8W9$4=sb;#|OpfymstuzE6g2ohd zb0lavZ>oKk=-$ktl^%*xox)6&fg5|4!(z5+WZVV4;P& zz*M6Pbc4$)M=XZr$qb{*byLvPmhDvrS}YTYlbHu))m}>s|E7CTQ@xVMk7n7=7UUjXUr4j zBxh?I=Tp~?+CvRotm7;tP#^S=f%ptWtqW)~Ly0p~Mr$F%lj-$8NZ=$UnJ5$qa*9?F zV*176%=P(E{;>7R(OQfY_{K`;=cCSkn$L8-=ZUZ=IYL;J z0f@`e7ywCFR{t=?K0S0xULV#XKFkVG?F<s-> zqk`s!=KAa>6MvVP7vG)=%)s#XuSegS4*K)M9bqytvFy>_ z`#C=3mc|9n{I5f~SI)tE>Oa;NSo}czxf>sDsYsw1g74rgOOH9oIl$&+7E*929r^{Q z$J@*}sx~!dfFYJ?*e1 zD?t&!gjTTX!a9qdK*%)uBw+ShNUbOyW|ol`zE{v46ZHpcGNn=ujD%f0J6-_-g5xxt z-SxDVzs{T)(!d*(?jP=XM?VMq4gLx;PHq_qCWdkh5K(9Egnlci2`0kOs$jMYOxa4m z`fQOrxu(oII*-e^Su#@gSG9TUc$)*M%FF)5c+HI$e&6+eAB%#WJeGV8M!B5D81Y;x z7@HMfMYiWG$B7?lwx7COy^e4))P7axQ9f5y=+w}Yts$gqg8FQUc+WsIBiZ7L6QiT* z#Y|{3+^NnEbk8Us%3-3Ugn{7Jeepm6@%gM^OJ~zOzgSgs7wVI(AZ5N zB^r-32!))l<*!&Np2m?qg2H@k#}CxE=6$p4aS5L-e_3Za>d#$Dm&JSn16YzW%6vq# zX*Q;iu?T`UaN;4FiNd83v;xM>QK@bdp+^Ngd3Fth7yGuQo@FtL&NoYGsr1{6TO#r? zJRS1OL@BfmTb=1sHa%Fif)qh3$LM}CaK9naFc`@>Fu7SMmm1}go-bWi^v5)9ULv%} z&Mo0=0=yD5K9#_+Z$~1|t}tHF3_FR!zd<;|6&j0jk!?X+Pa_X@pRE5#JD#?Q+!vGE z&^{@3YSuHR%5GxatEbOc?OCky-+^?;Z)-uFgj?~k zJGbE{5-nlPs5|o50V6ZU%8+9N6!9>}>1uf5rdl4CT5}2Y!=4lr9(4 zskSi;hMD57Eo1DfjxUG#a~>T`9Cgk)4gEnTAsTSq*>N0xu=V3HW(L#KDC9A~N5T@N zF@g9@3puYa$M5=a|AR6kVvrZ<2DR^8r--NJ-@7eWy$f+M)QNA2sV0rOBM+Q820|EL zw&j}Gi|bvqI_R{w*7USWw5V-ex5mw^YWdv$x4+4CH6;m_Go1+M6@0VLS|KG~UoSh3 zQNd)zm6s<6KUi9%6lC}%$;l`seN%U%Hsb3N^JeBHvR|gEH9vzM+~|7n06v0a9>T9J zvE(`5F!jgR_-V(fO^K}2Oc({$hY<-@3|!-3X*VbOGQq(^@amXG^%geO;MqpTjljZB z`8(Om%Y$!-OPP8tFPdJ;dZgsdl|1JL6v5^m49&qF6mZuid~I5{&&-B3eUzF1W#uBh zat%(0civr3izB~(R=9j;m>JS-aw+=#%HMfAKQr0Pc{6X9r@wxkbk^F3MudTWMf z-411Z|2h8C&SR({0NNtXf}oR1x;TRjZbFDuu!R$#ag)EbUWp)^J5ii^iy@}_9p+U+tX}K!gby4sb{tSd zX|T`Ho-kKPm_&{yje!fN752`|4Njy02lj8NG4JN&NaC82zO-JIS}c0{%5^%%&!C#8 zWjS9g`0w0dXg7i$#+IBt#YDy;lE8!Vq8Nw>laYZwiQb2piIn5%#qw=3bZuW|N<7PQ zMN#v8>&Ql%4E1$GYqr6VHPfr~_e@S@@Jps&w@dASxaV}^yBU;wc+MvGUl#^p_k&po z>k5%=^u84cb;?M{PsDo`iunKF6iN8Kk(q0-PDO#*$wsc4n0$(Kqd@DL)uoQft$NjO zZ=xd1{QnhQZaez++7oU{``0cvno7V0611llq4U~PDC6Jd-zRy#NncBlsxW`Ay>nOC zgcNBcO*jKN-qP)Zwy48#jHTg)1^IJaMNd1s)c!3NvHj`9OwIdQQo_|C+sh>o8YWfa%F@-SuqMbG4=cURG78l7tjnU%QzTU-HWBp!CH=Y4sto z&mRt_EZ8o35n$WLoSk76=jEeeMS#N5<+I1a2xa$-&q9GM%dDm@-Q*KdOYTr$;(L={6I~Az+B^b`KC*I+>oSV;``#nnowBp22k=hC>?^F$~Id+AX`t;WB3IE4_!dePxdQdhAFk}|H->7+oP%-TNx{U1By7`HV zY%^vUAnxSkM%=CxoDa+<(Qh^?Jbx!(jSE*9>AR#?XQW}N3aCq4fFE0 zf(rOwoBr^+fH^MJ!YmlSPh-yQm>wn>A#5Ru(oUUFTG_uM5EIcQ>Z4)npCweE-z2u@ zuecG=8Y>fs&6u@Hy6QUk#(?a778u=(bEA%@qmMPF2ECXQYGdR-duQNc(0dPXa`zk% zdk=-~gDkF0upQp6h?@`HN1cwY_H22$xbm$1SM9Aj?E%Tc%j5U@J#b$-J6-?NDu3(r^VmmV1k!j?i&`^2l%Q9=x^etz{*1uV+}&3O(Q?9eoqP}W zto7|b<`~nZ$8}B&nHetkxZw7StgQ37U+Ryx)V&W-JA6WrTc`ngl{+l2_L%U+%)HZ z39xw|5T3Rl%V^krB(r&(8)f~GUhrw{45BtBlN8o_*a^!KjcIlKm`!ROX>{UQIk-GJ z4`g_8uEY~qhwm4D*&f2WJuTxu44XiJi1}+_(|;LXn4(n7ERwO8_Jb(E_Myx*zHLME^#k&%Nq6(X@-B$0QlspZ@rG++RqYnE5*hvG37*(ua~! zVPv?B5ZsooL@UlP)22K%)2AH1_K`MuU(RwuZ`b$t&iD4l-uACMij83c9@j-Ir(0^D zmI=0YmB5?-6RYN;hnszxd-3f3PUkEnzoYf1n7@w82G1)_0Kw${AZW(Dw{nm?)g~&D z;q4w`&>`MTo16RoBE9ju*v<^8;dkxR=Mem{g)!>FLuzFo@;;D3;o8JWc*qlFN6#;( z$_;U|E$TmxqJ+qH`f3)}3djz~^ABA+>|O6+LWpmCoHVw#>j%PBl4E`02ehpJ2Wcmb z`5N1L@ArM4CGz@>%*OVUIQhf`cylC{c^J8NksLOSb==J}9S6+ zZ__b7#|4?_&1G4`)S7Fxqiv>Ps6IE<@x|c=bL8oH)uM%?v3O&M zrKQXnA<|B-*oWK>`!9=a@4wxOM&qAN;%1YXU_bKsVtm+9kH@;Im1*b$Ge1tm#>lO{ zy_s)t?hkJXoZ$O~y5P&~G<7_aM$8ZRb}~EB&bYgddmsY}rxs&yp5dPXV=J6+7Yf;P zsfHo<)RHGLLn}HqPwq}i{x55<+>+pP)-wI_!+U_0LTRSVx39ndvYg*#+XkbLPHsNzQel`p>`2{H9&P4&ErN5B8mR{;-np7=qM>JRw&Al!h= zQ7u)|h3ALWwnNq5UoF|5vDPsehJ}LxtvI_w2iSnRhyrN&n6S(A=$y7_ih^#pwI8uH#qGaB7()e~eBgV#hoISGl3M1m z6mvKu8vg+t3!T^}M_v@oQm|`2lr<9>lvSnbU|TiMdGDNb1UX-YuvtB6reQ?fET>56il0A82RZB%!#2*^p56iuO`}b~g+ID225G`1( zvb1vUkb+GOFpzesa5>(U*CMBB6dREgR{j=g|4Hpqu84R5<=w5T$@#Hx zOqZa<6=UAAIN9v|;aj}+?s{7i^xRJ&-UjEGmPk~Jqh~KT4`8_F@$;oK(GsBwY<*z0cZ=8Ld$14aS-)t~dTvjEnu;kb` zz9WN+XSo5^h*r>hJdIcnBj7V;q*kN+Vt3!ri?+$Dc=&O)J}zdAa=EXl;7;*Ao+2zI zB+eXWLl>B4zSwebhdp%6iJ9LAOg9{UV=6v@D1#@}0>(Fy`3R(xWIijd=Mz9^gzU3fG-G3VZH0NEOI_t1^E{kQo>47}(R6lwiJ1#jFih zqDK+5vp&otUpx3lt2-3XUo0e<4_PXlA!J?jrzhg8I~KVwr%>lI`(%r=y=Mp1qo~FA z-S0e%qc}u3UyTZ0^ENv6%=q%J=iL9U9Vp;)_IiCkx0RfK`Sdwi89pBay?DBeHmAk- zQGS$$rGi9nkM>;e-cifukzSseZsnKWvsMub<98-&`vxR_);WmgONSk7u?;xrn)HFR z$D;f82Io)5z9uK@ex>B(z8~_N5Gg?qvXTjp&`8($tlx9%Y>!iJcw<27DY%YMMbxFM znLo9HriT-Z9{#~89X)z9Pwsl3LCMKz+qY6lDf8OkonChO%P{X`YE%qBs$zszw(s7! z+gOxdvn1QQ{?Pnu{_ypQ8SknqZ#oLXP1639ba>b_7M+)K*mu^ySzSpb0o}UtUoyw4 z6Mci?vc^yMVRY=KY!TBtbik?V&MZYLF^b5=zB7S3@P=vHITM)FfZ+L!=fQ0p8ja#> zL~u_}2W5w&X97ANR$Z?r+DmNx?G3qppX>MhzTJ`zAxzg<^kMWET4uLomIPm=QV|lI zLB)McD^N&xEs*)tZag<8up4X&dKuy#F^sodA>7Rdu79p3MY>rat`_neQ;0Vr?Sq^g zUx~pjnDWny&u=Y=<|J@^rm7Vgn_COpLfFc^QLEOLuxWh#(AqBR0_Ur&V=)<Q@fD&&>|BZGC|%-8RU|Q z8`w|nFplW%ahsNwQN-#qNnBuHRQNCR-ios5uxqOU(C(hy_nhOL?qb^cu&EXD?HEHw z5is9?lnzXnX`2uoLEcW`aI+(eIbs|45%QZ)^)$IGM43a~cOBBxoI2=+na4MS+YBj$ z=JA_gf=;qkF&5*8kWVQF`9sLw+yjpkh(_uMhMv3?is zLXg7U99xTInSSov$pm9y`Fp+hwqV~fSLi@OVn)8kE$AqhTslM_W{E&Zf4#COY8dl? zTGr2&7wn+~KSzk6J1>y{!Eiit24+Vo zxiV+kZd>P6KMojD;Vz>gQyP)-F|D8G`9~xwCd{r-LgC#eUie4#8Z99Dz12Z@?y8L- z6}6rIsmb-0a4bO6;Fy}Y`($LZbv-TYtoR8s9%P8ybOAd6w%hrFeph2%Gc%afDn%FY zNdKdmilYjJxqny6Z7X{-(=sF<5c(AaQONgX&(;}`9}#lwXYh z7nL&z_AldzCZ`O2Cm8xGh6m*b8ZlYTth@Qmu(YxYWd;B^7NOj~& zeDo&rFcah6&3HD6zb+HT zmXi45J-sCBl8wBEyE{j3(#k?Rq`qw$_l$d$yJV5F$C+pt_Rivos_v za3%Iw@xgw+^waGunS3E|X9LN^KGcXyaT+v{r>oP;{XR%taa1jmYtECe8W`VeqpOm%Pi5W#03+q}JDw?O>M zSYfB42(zr_%*?6(K_X^hnsYracHv~rLrWW|;}UJUvp;;=$D>agI}{GXKAAK%0tvFq zca&8%ap-M`q7uGka?9)F&t|8%Hz9cnF6Y}n{LGm1o}c^SJ(oRO^ygQsT*0A)!@|0= zm1fVk*5icZA{Q+ZxTqE%HkLZgJBNmI#-XQ68>=yFo$w^n#pk)5dQeAY@%c%%CE|%n z)HZZ>lRFCpFiu5wQH$UTK)-_f*bM!2x#BJO)+6Y_r+0pNT4JVYV#SZ2cK#W-0YNZr zRpAN$&rXk{`6i6EpHOWj@((n83^Sgvv;23y%qKrV3h9G8|D1{_t}u_V6oQp5ngK@E zVve1$vJ z6W5)%pDz?oR+970XfY++N|CE3zF6-MOOxivCnelCF0riKAFwPa{XKu1++||D&^FPSU{~S-X33$Am#~XIlZ0lOa zqvD;H=Lu)WHd*TJ)~s{w9{`JsHOU%#PB!zeAlU!ebhP>-McvNwK;#s^OWV@|tvaFp-nTDrj7c+07 z`GX74F!0vChcY+2z|%rkXRFaB9!n-x-p`#HeBEBrf4TfdkF*wS5u1fJMTIe+hU*i0 zkXaZhRGe9!R4Yb?eU3hquG|tVkn+6yO~|%E^TCsIw9fi{ag&SRZ*6v6)OmIcOtvR8z`qY%#(2G-##K+h1nNvM@f^7 z0g3`N;aW@hBmD{!`MNad2VC@8(%R>@d1Zk6dUN zWjhBIR>f4O7Jphntp3wP3$3^mNhcFJ;Iv=0Z(N#bTG>5^8vg%o%HV28hpRXi` zd&*Bhz%n+4BTcS$v9!?L^GE-iNI6p&J`GF|DDE=svp$JFb->q5X|2jc@-UlwpLsoA zucE>gV@B2W1HbEe@e0V~tIzhvnMMSdi7DGvyT|$}YQF)$G1rQu<+V2M|6daNG536N zA%+RzRuf4Wu{q1S+824hHhQj7*T%SyxZOM!npG#E;W+p@K`aL$MeFFl%* zfMhZ2j{S-(g-5_~MC}y)%2{RS>3*n8JU- z0Qj3PFtsc61o^Tv%GdAR^fMT(I~-e*W!&wa`BFljoA@6-=(x{XdMRhLm}e_&Mvk(P znzp(aKn+zvd4gz3^d7cm>L~aSI<2SQ={bAL)1qoxHRsyki_@U3BX%Ejl9ilKe2{=X z%nqU-OQ9mO5c}Cz$k#N<7)h{k0@e_e$*&p8=YSD?oHV$pKU6V64fnL%88rDdYFcQN z;94V*f9fy49@j6gRAG~x6Dvix?*83WOYS!=fd%4xN(q$_V?>#6<{wJ-E)8>h7oKP3 z>E=1zG_N}h85zr5dBA)ADI%EFtKz8_nsmorj$iqPos95XOV~IbDDUhjLsJ`!bY?8Q zyW6w|Xmj4o1#?Vjl&Wm?K7qMg?D2-{IOr z7wc36pOVLB_nDbCONTQ7ICN;~FU>?Qjvmva3&g*#?p9f|*4EL|D4$X4#H!vCVcP(( z_jAOPz3z zp;D%V`}L3+b!C#avbOR7@P_?akwfplt^Uz{xUOCN3u^CkTI>VrMj3gk87aE&Ff?S` zPkp@#H|LZoz-x^4pr?e3la&j`upC0OcX$$rXDh|f95FLkGH!K+&P^fS!v_h3iG={# z%h~~&DFq0XvvyokBY_Gt+M!`m?_c9_fEchGRb1{97>xK-EVAdivsn8AW;*wWRpg&{ z&ZA;n|9)HPC6+}$iM@T0*S4(Xk$RVW*ukF%!Wp)H&sO@M#ZcQHd^&W8r+-fq_g=0p z?nk)>UyKQ&r3Fgi$*nNlxx-xL;|BAFinA1hXsXI#7=9R zNF=+J2HsifPcnWYb}8YDCFC5Me2mv(h4c^Ys#)tbvEcFLeSO3AJiXvzeas!cg9!>t zD?^(@N*J>UrZ4x>@#$p7Yvv0;@EYqa7PEl&l0$qHnGeYzI9q1PLX!?lTWcH)B(HjF zg&K!}D>>sA`4KaWkblP=wPRgKF}<}d*1WG3>&q0EXOv%mcD%+sha&W_QMzXw)$*@hn7^X3r|GWWlYeD(BX@#NDJsRjIQd4^ z&*sLS@LnkT9J&vP6+qV}=FU7bz-ZlVs$o0Cmm)LXwV8*o_A+NSJU`HI4BrV~x{66u zm3?$<>5rwrHTw0Y2ZI3$`8U_EH8=A5a(y~YgLC-428cR==aYB0;+;#lb^Z#&=sWw- zW5?;2hcr*xxpphX%-fQKRQv;*^JC85{s3FBah&f-h?wn&7+V>W`Njn~eGSZ^=9v>n zx;f~K!;p$_Z0F`uA~Uh5c%bMO6t@$YIH}mK`!HvUvD6PCy^|ke!)i=vCWy1cT+6Wp zdEc#X5m|w6IKDI<2!z_Q?Rs ztTRv<)JffI)E}+Jv*UR_Pl4`n-0?0w+NiegQZTUB8K*C!b!HI?N4DAWE1fie^ z{F~8=C132960RwQ3adJmPQg)6TRZJfnGcmyZk6nDkpLOfo0 zc+}-;1lsyHJ>sqcU+i*DFB)uvk}#ZUwCV0skT8D^&+ILU@jF4w!H3mcZ$Uo>f&6yu zbnC~8<%UUSPLkOg&(6*t7`b2{{Ik8T`F(1f&_$=%$X4~oAMPjIwOQR@-QaX{c$gPy zfmX~F7M~qGlhVAp@j7Ox2hC6SPN6#^_QohvYTn?k*RJdcH4}Vco5VTOi+hNoUMn-@<*EPlc8`r3 zm5yzO3Y^Iat_c8IHLs&%j-#8^W}aV~n%o1PU;j#pJp}#HL?UL~19$J;;M<9RtH|m2 zxClbWmYvv=#heRj^sKwU1#XDgw(O9N+-8urVHps+>Rb4h(OyOCnAKQrc0^7Iw4>N= zOdX~wD<`fR6S((7hp*mw+5fEvs_nVl3ov%lQ9k%VQ>AxwhR)2FG>2y4?a^*ovOaCt&hs?g)g2AI497<@oXBEOk@*IKS?le z`ndP4zjmY%Y#@GldF0O>ox0M{wliM@JN)V;q0{F|FNvK@foiWAL0kiQw@i4xefQkU zee*Dv4PvZ9x7evOd~ys1%zPyz24mxSVlD$-U5_g{wH-2&Df%V-)OK3TE zW0L#%HUejsL(R&7xD;YGf1A)^m5F1^IwlZIK4*W#=RTMMlLR(=Dv@7?VbL@e*{9bj zqj}pT@DnHnissCew;nf~n`!W55UVAtchP6iyyM}1;*^0D=1Lr~cJ4_wv+?Rkna%!M zjT{Jh6_}xhga%n)%cj*K!Rba@hz9#K6o75)Ux(ed9a<=JsXH#u@SkuHGp-xBmC3;P z=RfgKMsGuW#d+r4%wtYwgxtVY&GakR{aq7a+33z9ep>b7D)`wTYrB+!;QNhM8%K$= z4A{qL1<5%5c-b-MR@yH-U73M#}APqu0ux~Syqv(NDmtJCqRLWj3jawXI&i1mz^fOe_-06~KalvBN z2?sUqVElU-_>8fez zq3D}9RV1Ggn&GLXGr8Mxt+PGP#NfJ8g&K7w{B+o`))iX+Jm zts7tEh?!TWtdVHd|ATTmOzp@$_+hp(1>|*)vQLw7ic`!x3a;}j=p(eh)7OWCZmGbl zJnR)kxS7MkX01`#Ij(iKxA(&vEEz5$Z)$)c`Zj+DcSvJGejeYi}tToY}w(>xl$G+17k9>We1TI1F)97Xr7r<2~;br(sI z>Sx~5)$d{5#kEpvBj6rv!{O(q8&aG;A*KsETdYe-8#9@V7-k^7V!KyLHacqyIdg}4 zp#h+KsIgJ|>Bm#n!Q=%C?*8;JcpgCuc#%^`;cwHpsqbaYFItmP0$T*GL2BO&aC$H` z4ZCpZigL6=nY6KP4wAV>@9s50b^{yr+z+oHMc4@22KZnYfbgSvFm52mzomZcMvYRE z@dY{5!MciBlt%*mGY--7hKg5EvHh4yP*l%Txpp}FXH3SnziP+w&)d1YE<9h7s*50G z8Ih@4%$5pd+yfp(JOf-dlTb_qX{Z~&e`Th!7stoe1%hL&>!g0vchnJbSbDM8YS{!3 z26l-ChXqv62OG@HziQDaa~WB!UT@3HtUi8WYE%UN6d6=J%>q5pMGRA8aV?iN1o(;? zdKi(pn0l%~d&N+6h1Yc&(Ty!jo7)MVQEiirU-JJz=Vspc?H;);?$I0g`b@TtgQ)9k z?#1F+9EU%UIy9wD!gmtWLDWz!2`d8<0`D|Ziuq#xk^w8}THMAxMPph{W*IZn6)0&b zd&hANyB_Xip;6UHtM3!v_@o@@LQ|=iBb)Q2Aog3l@~YJTpcwf3`M_&)B)MR&KC4hA zdH*C{ICM1u>4PLSnmn2W;RuB8D(BNVL#KDM@P6%v_8I1%bJ(YCHWNZyZ!0Fta6`@KE@zM#e?Roeg=kqT)mdNS`mrc8^wm}+P~hA6+;-vh z2co((A(7I|-n+e&gHdUv`W!1MHU|4*4q5Q>^%$hiT%(8X^n3mf?lSSyN$R9{b+uqJkynDcsfEag;W&vAxA{{Pv*@M z;iQ#M=p(hX^=zsDQFvtnEk-|7`Zj0>mcHJgX$=ITpn|2r5$YY|g9(Bh`6zSd zqOyXu2^2v%@eH~U%dt@q1{$zjVyQXoI&{`H=JcJL9{aMUPVp>UJIa@J!`XCU-&g7MBDbEid17#>#X^qNp()TQ=tsdGOx8Y?}WcZa=N zo3-#$8pGq@FCnaa5CIS^&qtd<1uyb;N+6c)Lml+C zQ@FX0$gJv-O>H)_2oY@YX#AaG7|;{``dV(YHuSfdE17pe=RxNl-4DLc_q0W13Hp^N z%R`s|B+VY4?Zb~J6&2I0inH;JE;%hjKY_(jMP;Zs5q181L%Sk)UK3D9 zu2F(qnJ(RvpDBfSiT+`)V8W^Tj@J1u-=oojcW>6cabD2Na>lU4)*Vt1cv(lgu!vpdQbvJLE? zDu?*nE_5UX{nS-?AOoF(cFIVE)e|ghkZ&$S{|gAgYDxBNX%JP^+ZrF0AY6+a!H`5?xSe`v{pV>j2X{dmN+fuY%M%v6#KR^3?PY$S-a z*DxQAvG3&UXW|e$!VjZs29Wo+oYLI;u;h5H*{KF8OI}agnWoNg9|fY<*)wxV zcrkp?hcz?%cunf5-`*6$$=We;&dg84&Pf-0?WfoHN)Y#_Gzfw!MQ+D#%G)|0Q2(qv zXWNI0J(qLs$DRGnDQ8({Z|U9OVi8zJP|~|;m@C924Cq2aFifyX1WcOM&dQ`_H1gSRVID-#ZLZCu_$W1tiB;w^~0cH_~3bSZ!j=JlAR(uNVd`uWIS0^Rfe(B2G`R}TjpE+Laky>hewMV0)ALhHXvFyDkcd~1N#aO@Jzl#F z$+Ib{HrrskuG)}%IW_*|$kp_GB`#+D!Jm$f-=(kJ(BdoG&=$=-aQfS?OG-sw;d;<` zWGt7P<|#-eHit6z-mF3Bi2qb>drxV3o}~XRD@pjF+fcPqR?}j-mQ=dTOFm`6V=6iy z26~knF3fP2qWkJ9{8jzW0T@Uya_S*PYPFk&U?7=9YWA6VO4uxS=e#a6nY_VOQSf@H zWjl#&PM(X>Bf7;Dy>CN)qJswdhr~E!Uqe;NR2)73Xw8&H zHS+&qGzg);qQ}MU>Gd3bu+!bU#$1g1>CJ3q+-1I?FER}1s+(`+oLgvX**hRfAVO%o z^Dl2MdF|$;!KmHq~2X**gbRuOQ-@zUaSJlB1{Sx@kZjt=Lhlq1S>ItwwYz@5m zk2yW>u0N1miLQ;&mx0z7(CDuMq%-8>9kJEFL1=K)0smCtuD+ za>-}d%beYJ_KKH^zv^Y>1nUgV$$bete?5C;J?)jX@~rzI?(G4v0pn-p_{% zJy}e<;2=e5?n6qWM-&gyo6>1kB&SbAu^&OH@O^CLch!+CjLdbzeA$B3fFkv|^Ec0V z6GMVuU6WR-AM-hPZk4N_S=LLWg%$K%G203J1hjSqt1GW<>B4 z4fH5r%S>Fb9U;WZ(ckyG{49Z6B!jas?WCrLO*n>L#@TnCV(k)`H+hC8)rk`x(ul$O zgD8vOZFRY|zX8DlLbPtB%O?x(^FG2T;7OT)&!%WkOa%pC&Efw@`u2Dx`~UyRDa4|3 zW)-DEa?WXUxJ!~Il`~6{B$kk2Hc`%p;*KIDIptK&In7z6a*UdoZIv9x7Us$hpWplb z{{Gz`du-SBKD=Jf*Ym(ReLlNK`0)Z`ALmR!;V$L}@AlyIO%+Y5MnYZ9>cJD5UKZ(n zk*g&ii0XSz!2B7$oPP)(FpY^|-q)?T)kliOG+Bgc*BSthdMz`-+_0VGF zT9KWmg=n>S#$ty4y8)Oe$q`+yzBy=t#+;LSe%lfjn&};5ru;{DI?gmSjStiX-NBcY zE}v)rcG-1*V>zj#_vGMD(dR(b*v<##JdKy-*BjGRq#nyxDaE>a1R7R15+9V92P37yl9$*5eye892ovm%WGs*>QqD(2D$z zdAXplo+;`_yqPVm#k*8M7ugx9^VN!f0&2?bX_kd1fD{>6dHmmCEMU(XlarFDzWl8po~w=Y)t(jsDp+&PU+#`!K?d_76@pI`}XHy9;+Sg$|0O z|KlJ}8V`R_Utcumb(S6=_FC`<7N~xG!2p4wFFMGn6TGR4n5aPk_%gOjpI6L{45cZ} zw5wx8pkwQO+YVDtuRL^{H6YJbzJe0wR8HvbZtFSm__7UJUUbAcM}H}syx9D6YRjcP z)wbmxJ-JXKm6Kf-l;@*W_HEAYaSnF>{uaAwVjk;BH)?dfAGL=d&OHilmixGN_62a| z36yfr^UCn{90^>NJyb`xuwF9jjOg9g=~=rc4S7)E-K6`)^h@K0UKcNZQJe1YObrb^ z+1!<&RIDfSI+iF#X)IqU0+*-_>-0S;k*v`jPo<=ExKOuPUge2pQ(@Ynj?J$<{%&wc z&FP|-CSObhKPScAC~xVM{!lAledqYp75KB5Hcrv>^9Al5vAwXTY;}62Xc5LUY%%j( zh^1H@S}5dM)WS)1W^~B+u#?K-F#YUD8EA7p4Ds)F)TVj(UKl|m7QL`MC~_iZs2DgO zd+$Ge1m?Hy`-+!ymSNk)Fx;$?@k; zc*Dn#2>vLdNpt{!lwCWSyxC5E5w%DK`VeD(Cdm4e!mUkhd$fHtuEVPGn_Pfa-R1~G zX~vEH-2=UMsU8N;K9BzKKN!T3wRo%kY&+H{K202EGJ-0Lu^T%7Wcm9;(I}Y7kU8hJ zg@yhE;KYi=?ot1t8WXt{Tp)7V0@>9qq8GIwga|Ir^vnpE$%u(nyi_QswIF}_>aGka zXRV~~nFjX|mdN0<_c~xPew+h*uy5puw?>i$yTOSQ6xM0h7bR5)@_TGRJlJ(p+rE+e zb0Q6CpQ>@?L#7|sRJwAN@0QLWK+GR+h!BU& z!k@n#H8-j7T*^W3_-<*5)CJ4(bpt+(IkJM{L@oXquKdoQ#5A1G1evOewM@lzb{ZI_ zHenJ;(i)Si@h_Q6rn!kPm^G8r#=8!8{khWdVEFB0z%QD7==w8MP4wUDz}-({yY)^# z7x8|Q;_%_^V9x+-zo`7*iyG<&t0T>C-#zkf=;RN71!YI!PbZ}-7M}|8t(}!Kx3?3U zVgG||O0|MZ^(B}O2K*29^B_^4LmW_JDk*G&^#izW+t)T0lv$I4f2!99MNC67`{9HF zOJ#tQ3uJGYu+^H+o34+*vY)Mk)GvNFMnH*Oz^ez*JaJAv3yTB;D7V-P*GWbhjo2|o z6Ydx4vF6Hgrb-T)z3!(=3_sI|p-S27p_Yg?%pTJGtGQ#WPb7vg3*$Jdg*-_ZA;D2M z(eijnsE%U{^sVo5_ZKR=Zub;?@LP1eslL>4?pZ* zgUFbakFb~Vi^R$;mM#sbM?#Vpei4nBwq9r`P;PLGG8lpk4& z4OJk^&6)b$qRk>_Kyc)KVKtMD%@GvFXEJP1KSQir;8}((DpuvQ)kx^)$lp??iZD;s zpY6LR7etl)<51@<^!^KjzV{zS4(?SjyHVI!OU8(qdsPI?$$T-yO1@_$c6Wto7lTxa z!Q=?sv?4ai_igjrERPv4_xBV+y2E4_xw#}XB08Fv)W(}i4GfPo|MSVu zN93D_Z9ws%vz^%-%UAe#sAbgxsmEkpjfprn2Q5V}` z4>%I+tGpuOl{)f8+Cnm44S%+=2@NU?1mW21@8g1HqvQ}<)r;aqCI#HWkr#~Mva z#t%imASX|(6c=L61^-$D6w`nk22GR$1AXdHA0xEih%X3vz2bVXv~qXMoGC=Xe&@~R zBu3ON`g*rc91mYtZF%(#7o`*=GI9Cs z#LkY(RW}qd$s3;}2J_D4oi8C$KY0Pef%?r>`nrFThqtGPc6-u7bgJ8#=8opZC##O| z(K_8;HjInv#L^cVDkKn9U;j7&xdkXOmpl|yq2?Ag(hae_r4@^C}+ ztc&MGdOJe3rvm4D!K0N}A1qQ{o^Nr`-qlc2wAg^JjNcOBzfa3dy>2PiX3(+%9S<9M}l2#aQpBpQ082Rcu#h2@{2 zwsa(3%wKu`qx|6;8};=bOJxLi_q1*1C2-gl6FwO`DBEMHof`Da^wrgwTL(X@>;YB{ zxvx$ctnZR_{n-4NCf6)Up;T;h7xbg_LaJiNTxp;|obYNa^14PTEa&)=Zpde60 z1@aztsd8^a9xMrJVhUG|CF%&hjiHVtBju*e4e@#CUZMytpSCwQ)P1?xnK@^j{aZFG z<^?&B6?AN9S9XbUPPZzCa|FQHO?6+_*&HCK@O8so#>dbkZ*H+u;?GRfhP$pDY?wr9 zAYW)1Mm8xuDL*~2kw2WSb0K=4c<=>m7WvGxk&YeHj6lvgNT0T_qRo-VBT>?7tfIjJ z9l-l`W5u)M-n>n`=zu-LS{NWE)m6GKsj7Ds8n(1|vAq6u-u{$l6;UA&JT*4k*zVwR7EiB)y|#;H6=?JqMF2X^9Wc0jdPt~z_f4<$AOnY z^z|~kn4IBGxtuz$laM}%rrcRgTklrU%{l`34RMJ2wE-+yt{0_(_ho(_a`;jWc04hW zjOf=<36P8oGqb9?r0NM4o7r-J8|tx7(;W2pi)n4g1lEb z=Q9~~0Y?h|OnkPF{#p~9<@&GSFP&JT0RJdq0wrfI$F!XvNjPrq!2%9o6?xc11J_?E zEI>ii$w6=*o2~%x|Ly8 zA04~0N$rdnC{pB$Kv!^eeR&DVrO@dk{;V{DFl!`br4cL}?zv9pilo_`E6(@g+fLH% z`T%J!Z*y6_*hn~6ridNNY+)F3g1O2d4!}Ohbd9R%V93wv0UICKB{Q6%k_ug|0T`q_AuK4mzK3PuhWw$w)hZ-9i+(N&1gEY za|&RnBV>ALB%W`2idrPYH5=us_yj1<)D{HB&^TzR#nJAre|85|P_tWO7@tJ{hNc+K!sPvk7iA2|4r>s1NGgRz2> zQQ->a5ke6Q@&AKOf3{ToD*n;NN6@6_y?Bc28=*fiCs%>iien^QNlOHt*kew*R-?qO zghpPq^nY@~NDw|337TtE;{AA~+?d_ER#Z^wAQf{?wNKPwBLqe2SUY}w2b%4v3u4%X ztnFV^wHw;k$pRc{8f7MRK7z3TI`J)DCsJY&^#;`gzehG+No0>f2f;XV4Rq`A@VJ z2frXElzs?qz}Zj0Id>Hw2Id=ekT+1WH)#v9OL+h4q%(`2I_4mEqaQ)}_K}*qJ{aISl?@jt@I_`Mt>ED++V56yK!g)m% zZ^(Qia=;3|Lx3lMw+53eQScOoM}bO54O-`@8Z(Mn92&6)oe0#5d@UeHChhIG%;faU z2{%Cv*9=tzdx;VoC`m6jZ_1M5se7q}H+B6JBasIW4(!5e-4I{_wWKK9ACj0yix5UA zO?4~TL|QX17aYQ=7^RsTs=HW^KinkfK^80wwe(jJJhMwSFz=OFx_j@`jAJutG~B6p zmC-f%!tFlN7e=`LPxg6+(22j(x(0C`;WyHflC@HmOC3IZKKJO@JKbQHUrKg;!mvUQ zrB=uH+MWF&$^~nC)B~5>eqoQvbAFtO@}C1TA{!pyk_BM);H6r5FY!7|-@z_TNX@@C zP!&Of9#E@}aIMM$Ob3(2?J5Uq7W34}iAfusrl0jD4p+vMczo;X9SoJ9wM3Xbw7GR$ z@bLUQZ`r5cWonFbCC;cdi|WF4bLP((*(s_X5s-frr+j3Gb;Ok&@0HH)OB&o&X#bA| zr8XoM{tf%BOoNwCn$Jf4=UJk5sDB%shPt_yjF2K;ElSW&CM&(=BFfB3G6a3Jzw#;n zXp@@a*JB!m<^so`*83lEJt0bW^(wjIV)?Sd#q6cm%h9COjh$7CVl?tRNgvGBh>%}Un)HIrDnt8*lgXQv5hJO{M4&+dAUiY zt{cNGUX+g zF*Oni`ox|yNTWs8Z!)%LEtVwT;m0`+UF?6M8RJhwqyY3Wv2nUf%Pk<{gP}ST2+70K z5!T~PcDDwarjx>#Z2}^~!iMiA2~rWzf#!TfKM~R_Uuos}bg8in43+y3E z5|jwL%s4Hr^+EuFunAS zcbRCM#nKisf}7ggl_csExYl2SMVg{tK`&f?PH@dq1HaZD%nJWk@NX8cf_R;jM6AaQ zg8E4kdJ{9h$c>-|S{QQC*mts4YSpX7DPOkx9P5r4WWxp%rmoII@yLFFQ{4?_h`Oy6Smzzhp&7B_oy^@TW$Dr zEPqBSvt?d`UDs12k9kkkiJNkvNP_}cu6IrxC|r#5`}n(=&Z$(2JhXMDV?2x}H8?k0 zcRO3b9;GSG!^Aj_s!V_OafvF?+u8QwmCxayP~YRK8fp++QZI_zq$EB}sryD5osew}~5aP_n1#ef)}5dR3R_t@KU zCi1cH-^=veA)SmaS;&~BIWT9)dr6$XhiGif^N%yPy3LnolKR#X2-;A2o%Ub>J5mb| zmwiN)xQ^YIYtv8i}r<~m)pkEDn{3x$df!Ub_21B$2L z)N=K_oV+B>v1I{k25)eYn@P{3Q*{pwe9Eq9$+(^;f4s3m%_T_@#_Ir&dxPu8q70-W zAdP{dexeja>?KKWpuU{~Bf}9urEPXFvx0(=dN^`SbT>k(*yC|SO7C)Sq1c2u$}cdG z7~)rn_P>!EfNf^P#}X|%h!UE0O77-6ze^X4J@9v|)_b?_L+oiW0oPa+3zGhb3T!CY zmKD}b-wV4R5~ugFB7C*`wP0aY!ZEHgP~Hi$C8)b{_+Ypjr@{Y9-u34xwhP410Mrwk z8Fu*&J^*E1{5k)3NYVO^_`M zxlX-a%4(x@lQ$@021Eri*Zd1|e;-z;Q1LaZW-gChQ z?m!$Z@*6Cx3Lgyo!H{HYRe7L7`~hrdkI*72IG0Rqgs&fEJw z=8{*M<#ML;3JUYt7!y!7!lDq2RE`3_zlx)3OQU>6+kjQ6Mni$2m(w@GN0Kx_r?#{P zm(j;gn%+dfhG{jS{asSbyG20<7Ymig!l#Xp`&uU$420_#L#@?zQYiDT-o3RM=Y~LaNLs2Sa%JqP=$=o$eDN77CO%3h&0K>uL{%-J zlF0C$Ic4Y(O-I<_CC-1~XeC#KwG@EV!wNyaj%+#{4NLvE$gxqeU;5_Vi|&TA^{2FM zxC(AW(@DK+aU@kx3+G|>p{@aS*+LHocp;gVuQYJ{fc;~}TUw^*Z`Ly4Hlv_KPi96_ zAdQmdGQ2mi(gPQTQLrOY3q+P!86bqqTVY~qW+ex%d`0lD)fa!@!drnch)%;O4%t z@S)~N3;h&Wz~8e)Phtsvry@NoTuw??3im4O9(v?^Dze~yNc}^&E2xGASrq=hX|DDp zWWusWKR7pj41XRe$Wntub)TJ8XPuLkd{KSj&CcL~k*~9V6loUe58)b))|Izp{e&K} z+L|rjJ0}+!3EeSo2~(5zorIaiPprS!urLd!?|-%sKcWB=VDoV*+cC&3n{L%fSZtJx z)q>EEh<7OO$W7%xCa~M$cYmsff48r19>_SDk!AMQZSSi@*ua~-Q?sut9VP`hkC?#H z=6^7a9?Vw>5VM|w-1~b@iy4a3q+3k)kC1%_?`|29!}uuXH|T58?jUo-{J-OC>Qji* zQybSNe5M!j_Hu5#PCawuN7LK(v#@RKd^uMi!4L-9W|W0y*fSAB?$iBP5C}Av$#sG?}m{Z_i z>hXKq84v?g^nbTqig8JuWIXXK_^#Fwd}JqTqlXp1VE+9d}`(B)#mILW1cI^4_xK6D34 z82vpSXtYzSY1qS~NXn?(ug0d)Ghn;v*;T#B{fOK-QewHrbJ6mPX;S)yt?DqzuUuVP z1alqt$=ssW_WqJw;mZJ)>i8@#0U`74)V@!j zc(sosDy-xaw9iTe4!3*!+t%bOShX9jzDq$B^Z0_`do|d;gF92T%B!kF29z}4T>q=N&)@2pgH*HG>8DW5E|-N{9888zLimZuvxLJ`0nc1;#ZYl1(JvHpF$hJO%>?rEzne)aprw;aEYw$79}v%4}{wPi-)0)OZ2d^%#4@F_i( zuD9jo^SiD`%vVM?u_!aE#+us|7C6kOk;W`ZBt#WaH_NGno(Y zGEd%&Z6r2v?A~6N{UE?XmQX3llwEw$8&QpU+hBYng$gj^yw_mXlx~FF^wzA0j?&`& zed_|ZeVY8SaCd0fxi$YEX^icmU!^6DEq8m(4!OpntOQ~DJ>7c7v@Iuy26s5_ zpieA+jSA~VWWasa#bLZyunqk_76~8$skq0A7@S`zGwlMQnL+?d5MC>nnW@5G=T%@| zF#n0zHcxs9@VJByQPn!4VZ)gQha2k;oy9y~i*^%t@O6G8;9-~yZ!uOu&f7}Jm~Pxd zl))VjpK$sdIHRMm=wADjMoKhYE}ePlWb* zJs6zv(A|f!`OeH8=iEAUSz_|; zY1gDq3QjBk*sZ^0BG}6PEno{pn5+_a>xE43-}7GnxsXc3$obI1kiygFW30HiB}=)? zp=aztTayK3d3@UDkL$7fi+;7V?cX8U^m+B=6_0mz^lXag^Z7@UvGA{JkhV3sk5$D! z<78grjsQi|iVGq|DAv{#7dJ!oYjjD)OHJ85f3QEra^B#4y?VauZVBPl~8-Ee0W3a!kY=aUs1r|yyu1xC_7=zIQ1Zt zjs0=*dq(5mAzwawb~RQ+LFh=|E%BE-z9M=>uj;+BT)hM{7VZ{h=z!r(4^j*8ltJ=@ z?~tlElV3+e$#M_y?h|iHhW2cu6SWYRro!vwmghj({~#2ElbxNQ+nMv+xl1la?B@RK zZeL5JcR!4Uf5Ac;B+v$tJ@{go;Bgi5stnQhQ4O-4atu;2{_c1!U+qSbG*_(39T(qA z5osby1{kR-LTz)>ZD%&IN;r+aJ&WYoG}-N#&8d67H8-6`s?YA#`bUt40i#GpG{>Gl zO=>D>EMD zwi{huTeK_tVCSov_PCEP1^<%ocjT}dyY(G{(~n!3b-d#aIZM75{_*f(bbHA99!nYh zs;jaGnyXHv=bv;rQr)9>Hu3J|tG0*!m9c}x#=!rxHzMfpUV(>3y8l0zfcd@v!_BRJ z9fjZ9-X-CRm2a3^3B_H_tt@m>xSkbmqf_pbM^)AG(_Oc%<8-wu{H!f4KwVwRYO#j? z2bclEFkTtq0L};(%#|#KdJzvu@bV_FwwQ%Dqzn{+hAzKY$|(_CVNbK5=nIoU^z2)Y z8Y;Vcfd#Z0c+9@gFqX52{}I7-%ABclX7ZD13I=c??tWfT*S-LxT8o^iZ3-YVP4{nS zrrfFyKJaf+T@c>HD4XhVmH%1qn(X32`{|0VE23qWJ^hcJx+EYw)U?wV6;{g+tf z10yr~?C-o{z77ywHMbeXmkrEw9y8Aul-^XVEjMR7EB>nAKD1g%rR;KBWGbbfbu&m; zwfB?$`&?BF{JdS$8<|3QG_R4bKE&L{rWWnN3vl3g+o>*D{E19g(r6P=`8y^sPq-z` zpg_=Lg0b3RTfXdsGxaFb`#oy$U3UDl_RbtpvcS6B(8$guz^9k%9gzX4lFvx5raW%|BcjL0QBzj%K;r%O45emRpt@EkR0 zE(vkZsAoAA4Plg#z$%!S4bvC>O-JnRU${G%+DPlXR12D%D-!(#-!Jw5ep z_8Q9?wR+r3fU$nb-?f(4YtYoZefdigzS8kMXW(SUx#8Ws$sZpxTm?Epx*2JOF;&C3&#Fh3bY}bf#+|1ef^^&Y5Hu3n> z{w4Lt_O+I%=&+GQQ+PKV((qpyWU0W7PZt%>ks=WVqP`7JKVyV0hN;(a4@|Glo6;I)ve<5PJ@jum}Vm z#d}5AMVPeXSVMGn5U-HYRCnJKF(CBk<% zh#i_JU;QHhjd6|>1}Kd!jLd3WydAHAuW%E!yNr3c`Dp~*!3P^Qi|Cu%J+WyLOlqh# zx2g?nX}g|Vn))_q?RZc@`CGlD;L;RSmM4txo-m9N;KqOj{hd|LJ%j;8{XE$J4aO_8 z*b0XLct19dAWG`uRjuML6nmZ@eszPd0oZOQi$N~7g0~(!LN=S(*&>_q$ zKb?7MQgxvpxraI3wcX?JtatEV$v56Qa+-2*BK= zobUm(Z|If(G>4ooSncC{LlBPuHF3`H$L{91q(+Fwa>H3Y1`hhYMeyS)du>H(N}#)+ z|17&VIUk&y^~e#ozp?`ihXvUS%*^h!onSNA4(N~tPXU7fZKL)eT9E{fbDEcml;@8V zj3zf7BIylN8?)H{x#L;d5w@cnRUVoqjmyBY+D>cw&VJ`R)dJg3TbG~MQnCNZ zn!CK=+%h#N5PPbvm!&BMpflT%=b0m#4 zpI_9p*0Yw0*vq%jMAJcF9WXaAw_~$e$`bSI`T4{@edtiqlF8K}6Xs`W`ePBh$;Lui? zqre-E7P05*$|@64D;3jUy&NyAs=4-45m+B3SSYdS{|t=%A#K{)gRcMP4wiW2 zByuNClPa3r5yDbMrxSE1M9NwEgOuHb{ZQ~i&ldUIMd&pKN^}6u++2;w@KRUzIgF>x z?d3p=&W=4E=F|HT=|F2y#e(<9&FOx@E73FeLD&ndwZsk^Dn{`paH5U$CDJJ3xEEnv zEY-|v2?MK&#VE#Lfe<+(_pYyhmu6UYCHel2+&FwBy(ocjpq@qSTT_4lM@;)o2F*-I z8fW9x3H)14xJkn(KBx3t66Gw%qvxz$U;h+hBsfl?C4x&83_fDcmi zz*~~Jc8@O)Y4XJuSd>A^9=s;C6G?`PcjrGzc3Z1l( zqcwgryCbUL?W4_Wd5-WOiD!+rDM<$3@IF=F7jyFF8hj&ZJOX*v(Bd?ia%u8$$Ehmv zOy~FAbKm>?4H=A@>IT5+fMuCBG|$6-o3RsEq?J<~K@eBO*WiByG3N^mIUfLz1@f9U zYmP6|>IM+--)q#|1h|OE5&qQ z*Pi{ju1;)&YY%oKZivnmF#+vN%F#$w`eOYfG3lpbX(DeEemhC6?DL1>D@*pzQ{`Uv37T_>h`t3ty}&Kc(MH*}Ol zUsbxSPLj987-2sycxu6^KE@(kfva0LaQJCtg_{M1+&J^L1(_eJ}M&wdRS zdliVGZYCv!qlCD4DF0<(m+a)GiiTsyDBgG;E`USM{i64sbLngN6ip>be5K6BJgk0{ zw7My+kQgAb%W3DN;Cj=`012}l3R`AB_lar8S)F-*Nh9Le+e^R81r1=~BkLy@)^$(w#~3twNUy#&U}Sr!)*m8D;W3DDKX_MGL~s~Jl{V=dL|yzJIToSMt{Dea6y4S zU7=A&M3?EA0uQA5$0x<5^dxpY2!0OK;`8*k%R`Rvi4jPng`vVE_Ik?!!|`iaa=sgB z6m=a-j_!^S;S8|b!au(a1p$(_abd$X7*KiD$k{i{Mu&q9i$x^hZG3Oa#GEBFw0(0< z244?{ar+_bp$%@1sydH4^iME&kL6UI|~;6MvZXtJ3ggu3}U9gWQWP zlP^1%1>g00!_p%pjPkbc_u^>vorSn7d-mUiZH{~;@x=u}t2zvlx(Pe~+Z{4Ln8Dc` zOI)fYlV2Ga`W*80>}X=0FJ)td0#x@i@4l@IaB=C9V+quE#tp7}v}julMi z>HzuUY-9HAKt=y;OV0IDzvD%?N2m1bWATf1z=e`@jYzIBKUD<^IvwK8;Wbt0!-(}H z-p!w*${+PsF6zN39j^(ia>H(>>i}|3>WVi?5^t4lH&H$E&EWF*0aAPOn-hTB;i+|g zLjWR`E)fnggPX%8<(#I?^=TG*H2(J`#qiD5)Y^cp2f>Yh!Bo(Eg^6c*Qjt&G^LRpxLSZ_N4S4}bbA{1 zb;-%?H1j*Tf5P)!>fQ8^UAcSg0z-D@#F}T=xkp4U_~%5<{>pPJc$w$scj#x1n{sZB zo66Hd+v^c`T;Hez=Qi~?IRyptiih|T6Tj~y4TM>+u$(sVCgA-?aSR5}!VGW2ISZBi z1N~(U$|Gg@y(EdI_Vj3a-5_Nj#_czM-`jC>rBKPas(Br;@5}lM;hAzE5#bQ?&tkJt zC|bDYbEW#VIp8|!35!BA0Fo=5$ODh~209IiWG1e#=$JhqyzQxB0pbG4$IWL4F|`Y_ za9&6$Z)?0}sZ|WMD&{zQ;UX&{dF6%uaB91!oqxqSErWOK<*Dt!2b;7ThwWFDew)$P zKc_g|p>g|Vk7Q=jhJ#1n*3&oxxI*I;qk1*%p~$5te|;shdLpjG`rgt1Q{;&t#8}*p zT+CjyM3hn9V3R>+n1_Q-FH#&zthT33r5# zv{?QW`Eu7`HR-!(h?1@RoXF-kwUVFc11GpLM+6r?*|Ke*zvYCF-UwcuIvw9b5pkoe>9u9}^Z;yHeIe%Unzt$2bcJOT* zjJKCp=#2hm&li*Cxbw%fCIToj#Gth#aQPRkSBz{QwvBA1>S$03k1gawwWF zKFo5SnNI|C8>9&@AAqvpb|S+0BEMyDzSPc^0d$;VZD$1(C06U(^<}|66|Qi@KC)(c z`i;l;CAqA<4yS*~yGI@l%Y-3r^5=@0T#}H6V2#Ry*Y_m6=mA2Muxqi)gjdmJvcy+n zBD#Tpz$~QJa{^~WO@D6|wFR@ODyVf*T)WF=)d;9-9*?C9D95=&b8uo(p1(GFoccBebe z%HZ*oVe)FnK40Iux@0$`5(W7tKrVLGP+~aHud{PJXu;TM%f-3JN9QmM6-@G7gN*B_ zsG*TW%x<9i5M&QQ2g<-1quAcc(zss^39u5Wq`{FLL-S{@G$pU%9jNVx#?0-?&^{px zO{<*?ZV!@9Kk_;$Y$-aw#t`5fwqnM!8*|wgnEr$CgrY{Nl1MsA1d=-or8CPfwNB!X zzPdT5H_6z@k1?P9r1ho%d$w!x=S0)}ls`+o6BpbVwb)^g^S=dn=fR`&L0ud9W;0J2 zKqO!qF^p_h?jUIopv!`P?=8wSUa9*}5(~PvrQ|9obpRnwP@cJ&S9G6n7;hD@rw<{u zQFCYbJxio!AN5V=bQlM7@UsVKt?VURMgIhXl{MaVj8{W2hg#^Ey#vvU1wJW(mDI+j z6t~sjQrDp8_Y^%+lP4otR@hxWWreg*SjtIX+Z^YZq|yvBjGslE&%`u<6SD{;YZ~%i z5&eiMU78c-VpQUNQe#UunoJqmo@y~~i)?_jsXX{4=E9&W_344EXgh#lsEkh7WX&gy ze+f4c`_$cX8peGHo~~!hz7P-6;Z3IY9Ty_5ihiH;No@2Q4s-hcMDzxC7p&$7Ebd9&N#4qR2XHK%2hDQFK{{ zE1i8IpQ^E<369uYUa9#3pYR~y1chq2o;*Jy0%e}fv5$%FWvgnY_SXmIB|ISRuvpDi z{9aa6X>Sg0jZ-CLrRS1$tFcHO@CP-x=U5~6N%JM3TQN5-K3+_#lj7cH?OW@`rgrU` zl>FTE64yr^NvjUN$}%S$nNanpvfF6?N20pAh4p@4DM&RrMt*-s(GmVxCp8bWA0|MR zpO}b0WKz^y2z=LA5;n|^cPUo^CW_RXYUBw|;S*bD$~-(yANqPQ>*N^CKfS?H6|eh@ zzM-Hn^!zi-e0_+t7dG0R1@leA_qZ0Dxi+isFa9$b{A33{_QOPwi%@I~i?ff=m$${R z3!!Budt0zB7xf^%p{Tnq+$5@iB&j{t%f*nr!)g0RRJiyl`7E>&7d@4uFZ#gzK&d-s zN%kt$F9%`jT6|k-g8iA-iZ@~D_e6=VlyHvXoY&@yR)B8HDVn#Z`-2mFFZU+P??wJy z=nbpgae}PSae-!W?qzNiDxt3K5R)IT%-ok^mg>73#+`DlJszyq+|Zo0v`V|1GXBlo zVmqBdw^e8f@WYiPhr0)EvbzCoIcfYRxd`&}m*v!(}%c5wq}I*DCsacA<39in^N zJ}o3}s#I9{Ds;Ynoa5+zKxyc}!ug%cr|gMK@vG>Jhd0a^=LZn znN+8!!gH8P{}G`#m+o}Hv+Vq0jZFRr%XxiPA?j7-qwe%aPk@^Z`^70<5pg~T^Lk;0 zops8GMj~Se35LAJu7is&XeZY)@1N?Njzx+sZVzNBzezl8&u|)Te<$NMPE&Vc<-oZA zGAT(-I-*c^-`w#r=TDnNF)02N3;9&46ss~<5L$sFUTpeO%`jAQ%Q|yBYeX8P7^e2?=Z)#%7f!S+I0YL z&Qf``tO|ol=L6%;*Ue|DLctrJ0a2@|ulj$TbPm)H@wp@X!$th+nP-BlI^<9`NXg<@ zT5}x3 z+xbb7p&MRJWo4t)P6IauAa*If3HR0f6u0TIz_my0A3Hx>GrA;qKurAkvw1~Wq)N2l zbjJT+{E~Y;@8=haVJ39MQ!qhItXLjV{*eKbB6M~0tTui}?U^owo(qjE%*5$G3mbWk zSX~er6_|c@S3?NFT+qla^3Bk18(3L@mP|;)`dqmglzyuqhZ}Z+!Y0_Pxqq);4aq=o5}m2Vp_SyPoS z=59s-tp*nx7!VN<6wS5>?RS6qaB<@OUOVjv)p|!Ce&~-ST;;37Z-)ZCu1);946ud~ zL_N_^R)q%9@*2PaAyFk(stA3t=wq61gim2i=y>d|&}Z}(iFReF6gR8IZY==~p|tC< z7A5mUh7t>?`2wakG5ZK37)7KyUWR4VUQ(N0wTzK8zgUU$aGf-um};L}+;#CnPcg25u7Kkr9a9rZ!u!a84MKe*e{>K&MmmMOa!$&buE78=^$>STqiHO zxRpu@F8Y?pz5@Bs>*7HuAv_VW+uV$XO$O{BDX5Zz;RffQ$){tJ@Q6EE9yh)BK|wEW zWP2|bNrYJtKDcRZR)&4DGBgQtJzf1N(GouR8pbhgB;4sUu#$<${;>O>u~+}h>wjnt z*c)hP`{uLeNF_gOLR%{#j??!^Km96c3SilrTn8MlKW;mf9TK5_FE6% zv^*GVvBpm2b&+n7$iwH~j1h03$OxKkKUV7Nd*NOja{Y44xcTmBN2{L$;b4M~@_%{5 zcx%Oa*1m$WSpPl!^3`^X{*@@$I!H7b1-399zD!N3|H{vrr71o9QGfs!A)w*zF6H{M zMvj;-3(;Sd8i~JHy_8p4eS7xi)`WJW^L8|*O41muNA}RO)hven&_W*L+K+Xq;4-PQ zkLZ-WfC9$yZYhRx`zi36CKDefc#rnf_E4$h+GW#YdIPV#7JrW`om^E^_ozE&Xa7R- zIikFXX_uLak|O#RVA#35H+V0W_pl|au1_a!JF{;j9wS6*T2o{FXD*RMRM2j65Vg-+ zJJ-0|wi2g7U0T$I5{FKkJ-yP@@pklQ*5TCmIhOEu%frJ3)=LYHomZz_)-4!4S_nNa z-PSv9PABXYq;9v)KAt_kD?+!fh5|5Q4s=#f}VIqkN7B2sT<}Tu;;)MOD1?ieEqGvN=i5_yae4PHP;btwW zr$B)Gx69>0<*o)L->kRz_|$rgsvGauyZ$p$S=nd~g5BQ@V2^uBVS~|`rX-SQ#&(P~ zo%_YhMOo3lFyw(zaP5h5CdP^M+4Zl^^>a zg~o@)sqDQA%dlne0A+_HGniPt;EGGe`XOX7Vf5{E$ykvAmOr-K&ouHL2I(VZ0>l0Vp5HM8gv>V|&abXHiT;)HzBwl+&V;%z9&JvQ<4;0}_-Vi8yQ+Kg9dxPt z=o4`f=yznsT#icAuS}Ph553K9)aRH8y=c7O;9W8{LQcHYW2LM1=K3=~&h0ngF=?Nv zdo|Kpv5%!b)GS1k>c9%O$oLJrY9(U*aiT2LK(iIyx2=a0C2n-yzK}@LB>6uqpB2CymXgFry?&-K50`2_F_zOls9XOIxo84&>Z6H!dr5!!a z=}d2VZ}Dn`IL*~nR&-}i9j#s^d{;~kcSbUBkl05IIa!#UT z;`(P~Knn&14&IDI{tUzBLq5(RZ|L25suT5of@1j^dCK`_gnI5Gp!?>U`bQ!UqH2f_CM$(?PJoL zwyf@7S8T%amZfvPWgByXJ=ilOa)43#pEcbNn-OsfKPr2Kneu)tN|!HcT1^Vg!Mr@u z#y(K2VFXb(I>w&IjwCZ3g0{h>R3q)jN@yWMS$>UxL>|NV_-^y4vZ*t_Qof zg?kSlh7jtu8H%gWG(-%|I>H75^KrJm-oHI)AxDy}#`-YJI3 zTS+!9;_Q$6Z_;9c_Ct;>g9!k7*A)X$v4jH{Q|96vRxHY`6@_g2JSb1UKZH*?T*5#~ zdib56C^WNxcV-HiP#uD@g1kO?Ay*>S^mq<`(am`BxgM2g=9S*h?B} zr)!$O#Kx+5N3XpWjJ|!mTrvG!BP79Ap_M%}7k+2=Jc``z3bGFjQjy;Z&(cq)y^V@t zAKA|@v>F~8U~}Tr;VhA!Bl|cx3!uu`buIu1r!VP6IS{AS=od}xnnzqelt|M4)K-vQ zPq=|8>lM$IT(1aqYNCjidG3>bIp=66AAZdmE{FI#lV42O z<_SqK$HJodR1U&5=@*$Rsa2*f02FQW1t)Uhf_FyDR zx@k~rt=28=u0yMHUSxqg05wPbi=yznE$ya?y_2PMvz^^J?A-M@Tw?kTGu(H~Iu=WiVIOiLc z#7P`^plH4U>?G<{UYo*ZDgn{MBBU4O&YGnQG{_;q_$-RzXwy$yobQV{sF*3=Jx2 zU>TF_OJ(;gbd~{pA$EhCF|RV zAFyaIs9S;ma+0oUZV&35FJBg7bY=`k0hyD#SoRN7-<2uK0kt+f97}mdfGg3F58@6@ zPfca09NVijM@DQcInN*}YTe4~c7i)+1mH(zQ%YR&oj&=mtehIn+4>NVRx-Z?kONG3 z$M{SFD*({AMd|s;2DV@LI!dHRZWW#x4$JUp!&XgRUNSV|vgPX2N&T!X$d#wSpW~)4 zdNkGdxBpIuv`4`ZyKkhuVw5mzm5Qc32a2YO zbYykP{fX0ih1h^=w0!E1vU|T9TL$H6!8jCT2X14sEYaZLpf!Vl!jmH?q;g{&!wE!T zpXQOyOQ69jD0qn7<4HS*OF`1keNiAbsSwlFnMV*E=Y{fhP*lDXTchatNnR z^xUf{3AuI4H*v}N{+j&3K!~Nxteo>}QDiwW@pQ=<2b>j-3hzKmMNR3uG{3YvjB%p| zUg*$SObdn#8S@Uf3%1Ug912=|hC^Q=&cyiv`nEXyL zj-2jKxwmE?y*M=)y)6IFFfNkJZWy1v~CQ!b+GQC}V#^NdSisI-|fhbrY;N#95lvU@k2C;Nw_rfoniSJ%jf3 zTB+65?hCu2=WS(le{Pa3eh+>5Z2X9G2PiKS6WRU1G#YQMmhbvu zoCy7zEPDkSAt@Zs&nBUe1HVl(RzDDg!yNb)Pjui#YVTDWju=LaKdbL=b0Fi>9dmBoRKGQaV-?x~}1gUL!9t?U!i8hmFfK#z* zYHsRN=-=ulRG%N|xa}{t^Z0w+`X>=9yaSj_$}>>{alvH7vp3crbYz z<_TDFw1Wc-A28X`!zk4=^Y7B{xU8|%<|e?8d}cv6D(>%=4mw zjDH^r?h1end@SJ)ELA!dS8;2gEyk)7c|T6+^Dt#_PV(oo*9$WfhCA-`jq-fQn&y@P zDaT{cH)yWW6r=%`c#f^gGRAcf!1L#Yk7pb81HKI3-oqAYt#%8&u5sQ*+Dr#JErwCL zGt-4kl?>SpZ1Y^)dEG=9HH=A{^q91H?WXAHF+EY*kj1${s)JgkVIz>2bWsVniHU^` zi3>WsV686B(Sv*ZG+)Wtn4)*hh7vzUq4aVWcU6Cw_Al_rhz41!uJ#Xu@n`x!0$IbH zP5;FC;-HF`{Wb3kAL#E%tiL{|`MfDiNMp6FyyJU6haac*p0&j0PNs#8IqsvwO#s;P zctIC8_mYdDc!zI^z_!HNDu5()&Wh*|KO+h{=8F4GB`FjAKKl)p6k9^0}s`RjZW8X9aL@2ktc`j6$9lX z5IdI7XxgjBQeGmf(5qr}2fIEj*r5+l2jdKwV!J2DnwH`unF{oY*|-$1Ssb7X9CW&& zCPdT92+KG+e|+TQ*UTU7mi57hGRrl-#m5=OjS?!gk_p~{YHd9MyKZqm01t@eOmz?p zFszQUqA1r_M18Y^w?&99qwv(d&V zErKl5JoTk!`D)0O^wDF;3IaElH&XN6tqr>37QmB`aI(w+(=%}TkOI3$0t z+w0Jq?Cx6X>eB6sH;*KG))G2{IoLfjVp#9HvOTRvNM4ja$aC~Cr{5YUB-VVfB^kCT z(3+h-E%rx+cBS`D`Za}~x>oj&0khr{?ymVwSv96V)qi-#hE7V^sz5H62+-{l%*!0U)Fv39`2$SvX*-0L zkq>(C-6&DmkAb4MacIhW&q^&~Fb~Tb$L1;C6>{rOO_&!z~u96ZAn3EjoBA zRXXbOMLiU%n5UkdLRE7oZah%1<$Le$bs|wDx@fVo;pA1sJPlD)=`UUc`(yk#@yNK* znQzs$3WEJT>gg*RKMv0yegZpj(7G+1OG2I7)shHzZxy`~w4ked8W^nm@T2b;Vrg?9 zA5eqG5sVai789{6j8Vxu5_g1koluK-$9%ZMm*%g*7H7#1@aQoKRHIatY<2aYg;HaS z5&c(Y+$VEyRtoZB>&UBpp}QwsnZSlm#PK2%@RhhW9`8>~Gn#_!;{vyld(qBVVbsx^ z-wx+?p@o~T>}rOm72A(T>7S~tt?K_M%|VNgizZU}~t8L)e1w7g_~U~^q8D#>zS zlkiYd`g~PIacEi8$jPRp^I@8Hs~tIdLD_BbaaQ!@%#;fAIq4EhiPe#kr%=4xKsS3} zAmo40F*SjIg9rW-R#ac>0iiy!vx1)~dJ4%)@O3da+tpxc*C$6Q*VfVS zt$|v0w3BpKJKm}u{%Mu|tu&lL#NHe20y_e#7x7)_@4;6#2OeK9SCRVJsV;x@x5TJo z+GnAzbwi$X4%phXN7ek9?;6p86i(e=|8ycKkd&wH1Dk<4jT_Y>{Mk?4MiacY^MwvBEPz z_>Kea1%E<1z6#x0EH#HDdJ4;ccKD*RWZbhdj}8MZ{wP9wOwsk3E3}mKFbNKqWwzAt z2iE|zA7+lWudMRSCN6L>h|18OB6DC@7o)uQ8MH|ijPAWemH{FK6F%#17%apd##XvN zs=!hml1D!HnpSjHBrar^&j2Uy((9q1dD`kDj@@?bmnSFbPvtEUox^g1Uqa{AHgt22 z|1mnQSCs60+;T!XyC$&i)KN<@izt&WgtYHy)>D%JZ1Yjg3hjXke z5M?8{6IcB|#{`MzcCLYZ%)P>bV3xu}9 zGfQD;SrS|p{~cc_p^|ym2D(zTz1pQm4#5YREEa99Ku=Z&vbzK)OZe)Il8%*c>?Z2N%L-K zks~zmB%Wx!U|Iwtml8?yo28tb$YeM?eREr3xhr%9f~^mXB&A1E(qkCw1I&1izv*7G zV@MBez0pl*zH|)|+_W$zUztaz%2bv#LYaOr8%l)ueEQY6=uK&FfzSfzY)J!&(?mK( zdV9_SAuiDk0x4rdb6b}X#=q#3MUkw^uMpgrY>cij-y+GFQ4-8nSH38`7HvFH^vsAfMe_%uFhW`9hxoi+WY@ZEjo+_(AQgEjy^<~ZCtlL!BaQN&zpl$T+_5xq!P@UZ zp|-sByME(u98a!8bMZPwckbS?&>=sQF$+n^jpxoHJ%S#8B~E@jHk-81XYzR7Q6uQ% zrLHp%p|zAwWU46Pz*?Me=NwM>7_Du-Q-?{NZ_W5Ht#!YQeyXHIf3(>u_63h(UBjT5 zeSP((fHCU{1-CWWZ3<_yOS;daIlnPUnMsjEfBlR-$R>+al2=kH>?4;Uvct%eCnBZZB8|p8M$2g_nKN3Nj7R1wS6@iO#Jg%#` zC|<>C%+O`k@kkvb1D-}Mo9AUaMeP6kOQiRjUdkf`nOc-EPcZT&7UpMQ8rVnU`t)pO zCw8X{oprv7R=;4HXw_cw6K{YWPjR{fp+);H9lB)geeqN(t5y9PcW0M9H{#0gual># zx1n{1vs+(ftjX9?xea@J<6oSNQ1*Uh^TW@v#8WH@`UR@YY6h--JrTdZ#RTI+&shES z_CS+=oGA0|-Lh}l=B+2ps|-KLWeQTh>NWe>QUnow4r0rAaWtiVrZT_tDT|Y;m27I1 z(IOEy*_6^@AvHUnj){K3R0I3-13=IL4)J735Cg^e(vL7+MG6j|T^F#rsXL?V`G@n6UcQw7nE78u92 zk!X>t1~jnm?J7~AkF+6E*kTW-(c@}eMy3@H3r`GyRaUv1^@xU6ob1hE*Y%YCKYwGL6jz2zbR! zM)$8R5vZ&r36Hfe3hj#B4VQsI_U*5E<|+ttImV>z#$`*m!pMlRMQp+X4{@Gz$rO@F z8Dx;p_4-;g!H9EjzRGc%knMHI*SaZNWcwD7Cb`d81zs{mU$f97u9Kv}F>bQWGPcNc zu#bTi#cUXt?*C=EK+lP2zLRWxOZd0WzryD^t^2$EenYI<<$2bs7&R(9eGEhTyD&dUeRwnHfSn@!&xt=}t(r9lOT zDj*&#@4{iI>9H{w=26@N#^N-q49nYNqQ4C$sgJ~C$mFu{b}=ya|u6M$Z#9R z1|r6Bl=YRqXHkWAv_NvFgNR`YIxSr7!rL)s?AcEPTD4ySoJ_qVo$lP!1wq)DYS*!} zCr(tVo8S1w6KZnsNo+lsJ>@c-_nwBKB?JB=^sq!!X0fYk;WYf^evA%XZ?>z#%WD>y zI)0=2oyP$b_@8~n%DYaq?GATc(Ef&J5&}!f!J(x$vmI{Hr@M1a>$?3l=Izt|rxvAGKD_Zoy)|LAL2@AH4g*QSr>kThzCnk%zt)sJ&dLP&RJ zurpokKR(^ZPmOzJ;POR2z%m2vw9h~dR*)6JeEibRDEY&d)&b_}k@gn;v<^YPkO@+t z|GSBB^_F|d>+b@8Zwdk@jR zCg?yd5nyQF#-UWA=1$i#Wp-6D5qr;?(c$~%S&H?M=xaee4b`pUQXgP-r#2ax~@02>xMAYszYdFVw}k>KpbZ__R_>WAg( znrt3J>iMeqJQ_`B6}YYK_fm}<;p+}74`|Coqzsypjh2dou`GaGn`Sf% zpMGVmj2%NP zIrjZ%&hV^yQ0B9F4_AAyIv;2YmkM5tGe}?X?X2Ks5{R-Ud!^wTZf$YqpT6{A++Jwi zi<24c{NQi;xOng5GtMc0j{^h2<#yCb(xKkk#s_~=^|PS z?*r(y&Hc6G{^yq5S~?Z$<$$KNdYd?N0A%{q49(mR$EIQVTX>mjj&$@xOmI2t1hbZ^ zH(SBC$QG)n8!<|r9jK`YHoFp45U9MMnepu2WzCPJdAfb7&T`1(SmIR>3MjKzilvky z#la+vV;HQn`Cfox*M{W+hD2NEfm!(dV+9Z#%(!>cf%2l|gdZ z-fekDA+&Bc%fTbGBsQdE^0t|DO#J==`4!^BCsP-Vy>0IBeidE?W+nALyQAxmZ(Kqi z?K0}#gsBo&#JF)S$aOlG}D9?wMZV0hyAofHJ7y9<7aHjZFN~)0mQLxaSLoazU+!L?l#xg zIfce8;&S&d#xKNcV-tXFc~(3KQ~1~?fB&H7^trra$$OK9KBumKZ9MJR%TaoCcJyCn z$u+CT^0#;-k=iEv`LlcYZ38W;g?yA+M%h1GTzjF`y8e}Z^Rp3pAMBkBGqroWB=#qx z4Y>MJ&GXW4C@Je6gRbE3oHp~4n^RcM3?Y)wKsrlr#$+xV>CX4&k}!W4OH?6g0d8)+ zfi1u>Mos2g!e8tCgB5uxydlm#BTLA<;bfGtdH#uk#W^J(O>E>#Qi|Y_&h== zt!;UxRVjTs9+AA_cZC*Zh5d znWQdjw)ec*Cs-nkd5UIhdo;hPcKBsqyAKsFOypZcNhH%w8H|iy%Y|ggyO!*}0-)on zAew5mJKP(0db(@k$^I!B`5;^?9$8hK{H4r~z}=#NQ5%r^Oc+$@Rmoeo>pQ&nsz!@n z=}+~?_VVcjI@+#&Fl<$J*8G@Xo)`t0_V61we_A^9fzyXX`!uP`=`FaQ%NJZ*D7U{w za_x`Izj5i}CAn%nV1>H zi;GWp+*xsXe5~UpY&`V^Zz(LkLGA!ox^@?xbacP!oinDnHSh1(pSA*&H$JN000SHw ziF#2|ooCXpbN>;raDy$f0jCNsuU`d5d279!Y>^^0j=DJ6UbfQl9aFeZ`t^PFv&|Czyo$x7kXC& z#R&!~VTf7$KN2)wqRP2Z-9QPQjKPx^i$aQ$eu9PIJbV9QfDc70eMTC6>r2qL>3Psd zS0YyC>4-VaVvQBUjj{SpJ1d7qnpIoc5&6b_njb%!dww{j@uiXftLT$~W80S>ac)>S z4r+zpJ^eq(W$>eKGqBzsU%xb{rRn~%VEsRzU659A(B-EEZiP>^f|VNbj(hl*={4%^ z(}&qq9UK3iE?{)H9g520Kq$?r-^#f_a06ikycqXL8gl z-#`0S^*i!Nre(?>K1Bh}wxx!1B6uCks=LD%ED>`h#Uz#G?1^6CqAiP>)3}(Q=xPwuDJ3OE&T5|oH2@F3Cog6+EhL>XZ#ToX`JtVvL{b;fcQ)hEV-I<-O zVZNoV-ub%u#r}^A8_D!qm6gw7PzO1e5D)SD*a$4q9tic2;{gm^no|O;hfz=a_%sBl zW&;>%wA%8P!a(Z6^GCK7n5%om`aUiKFUl&T4G^MCfuS{nSSfN1S*D^`|1!ivHP)95 zM{5LK%^CkQ^*ecK{I?|E`)DdwW456Gk}dG{eW)}gO9u^{(D!iB{+hM_GvzAsN=exa zP^wLW%`?XwAvoyY!~mg}f5%takweg!NE#Rzyd48rKlZG6@wrI`xz0vf8k~$IaAXJ- z?Ervj3eNru`L%#Hgb04NMV2uQ%|f@E$3ts+gWKDSdog>rLvvJN4Gf~iT^ho}M8zVB z6L$u>63TioUsi@3QIXjC4=2PCwYi=|M*SlP-O<%Gfp|kna+@YwG9=Dt##`G<`yfu4 zFEJq#Ww+)_bV!(A8<0?!9^1GJA~m2!RWfU!&`CBFrZV|INFlQ{%vg>aZo>Iyp0+-L z!}yvD+_n@*k%kQnn9D|WA%2-|M76@wH<6jqGWKv8J<>qBw={G_d_^19O-6d$%dy&$ z-i911j2A~Xi10VvOEzmf(vY~swh9-0U`9bEigcD{0zsU;^oh(lIj=HH>-jo32%b(L zRSa&z-x!05hjj5I%@10_3WgSiGnR9VR}g1!t3u{kQ3}%Od#0h8Q4H1s#Ce|xmtc@@ zrTdOlY?baCw^B?ZZ&yXny-^sP!yPMm0%;2v%f!pXwG`|8Jbhch4+i)6VkLh4woc;R zf227uu_~rpU_X#pL^!iy_>d?b9_?PVSD!TZi5|7dBp)Y6w|B@rddx~Nfto#_6hOj zfyirH3&y`-tz41%u*kDj?0x`L)IJOLmxJn?j0T>0gNEkES1*6R)@yot4JW7D#bv9I zf5kgKZsfn9RomHMh&_z2CB(B)EUV$)dwJ;V!71Q0IoP6t@uUsiL&*-Ay9dr8_$NGB z@-)J|-1F?->Uug7S6hDa*fp-#bA~(IuUDl*$U#40%*RI_y1hM>GiA`6x$YxE^ukuk zbH|?$9i_VEfq=J6}Eb zu5rsgWXH6?kc60G4-yZM&vOdO!JpGoDgc^efu`}6HxTK;VB=M3wEq*pfn(byY462ezL zo%_UHis(ey~3C;h%q9 zuh<2K>e}#Ex3S&623l0sHQY}>Zgj*n_3S(EjUu9u-`Jo9gpw-!9QN5QV4}w>mh010 z?U$+Vfp@CKZe%X9hI5t%&#an`e_!AKM6K$v-93&5yO*IGT2CAeWl#BAPx7Zr(U_L8!Dz zxijyQQ{VdxEBaJkGBf1rYfP!w%bkFQ>Q3(EAW}~z$DHEL6Z^&mkV((aa z%p4<$bHoc*e%-;$q|TQbbC!gTGEE_*|~e8M%5y*6~+Nh;e}JjmN0_%olp z=<3yIvu;rNX02Rs{rP_R=$9GdoSU``(0C?dnV;vjUd@oS(zgdN9D5JhGPbgB>7KyX zKh@Z~nf9K~QKf&AzC0r3d)y(Yp< zGp+IbIoE(VUfKrW&j+__@hl+W1Ydme?Z<9t4wfv9A66jv2#WyTp4Y+45F$E7Kci zE}d6=bSwKwI_2?lH+(6DC8RQIl)=^p6`kOtA>2BC|4%|R<$n-p+1vW)0$<9kZn~YI zA=`hL&Yh#*%8gzQ35*W7%T}7h1BI8oGBM`*EV4-}FrW-{ zl@UK-8$g)ki%SPGaWPswBZn@Q_@}>xALC8ede|&Q8|gy4TMQQXT+URM;$c|@h-UK> z#v@|x<}RHpfBhnq!*%JSui+gVMnHo2KI9mKOc{PpFKiX{tkIasoMq}5)(~jLZY;I_ zn$=*c#niosZTSfs8aw;L%&3cvX2r)7zAvuW#;WH91jP6nM*ejp@j`KI_Lyf)P#PFx zweGzNpBcr^Zhe@kO<6(T1@}e516J^^Qrw81EAwBAR8&&uv{ay*I|zm`cTjrd90$>F z;C${h<$6TlZ$P3R z9d|9C&s6HI$p~&mY0_WSWch>4LGk8W8&m$lZ@6oFH;+kgZ|EB;3JOR80XdG~KQpH0 zDJ!=3JnjIKF#C$N%_m#@ZaJdTjb#Sxu3-yzPc@{ z4RuVCUGHCpW;@~(qFyyM0{$Uic5kPfCT&3*ItiW=ZPrM}>$d-}q)g9k^! zVaF8k!qu<6kcnB9g3rfe&r&vEs?t}bL%J+>WC;}#MAQ>#*MZG5u&Xf!bklYCR;@9@ zg@Rb@0p@n+2)y(4VEsgHn}=lI78(qh0}g&nTGdwgWeS#V7d&F_d{c~dKkV~?U3-!47xX`PJj6=V&`{d7iUl%#~jKljO z>9GI#{n0MTM(Bk4`_<~M}Y0Ny~tpb?OE)KvmO-T09tLd_)uq>Sc2wrh^ZjhcHjw;sM z+BSjgr*@0xWh#q=ONxsAHf#}<<3&OScOuC^_4@$*>oz19@e^^Zcq#^au5~?WpaBXk zv520BixcNE8z}RWUY|I!XGSU%VVN`rF11vFoO_NpXgf5#qC zX3Y3HS904d?57De9v=8`tpiHHD<0G?tyI($k2k$&n#1KYu2!7~)T*AO-{x_oJ` zK?|To5@>E#wDJl-NvNgIr-UP9NV9HV59iFdVR^@RM87&oz4zA)Tno3UIM`+FBYEqB zMteM>C#URt)_YFLs&^0^9`q}kXFlxV#rqDA_8BjC!+OIwOV^9S=A{27fJ$DqZAy%h0idb)alZB@CtL+fnTE~`b8Y~v zBX?2T!X8i2W5FNY0P41`kct-Q34U!*T$_xmXbUv(4-Z(DSSjxZFzNi&3%eG}NML&y zRba&L;ZSNU2jBxifcA#WtffVsA8Zz;4LB8c^h)u^PccOG-LpR=H5F{6%o@|hfr0pm>Hk6JY0fLemy>{*!5=5#-#8Gb%pL(qu$;>p7&VHN+wX^S zwoL5p5AB^EW>LdK@>X>NiASPUIsXtQR2V;*`MX|G%cq7|w!W{U#Ab2rd&*k;7-5j; z(x;K%j=L|*l}cR*4FXRfmEk{^4j{54*iE{Qq)4UJQSFetQOdw*CQIDeBA|HG|xx#rX!<8?l#bm|icJB=M;fyw#97I{KD< zRe_nJm{XKiBxfnaz8kIq)SagXgKGfTCBWUvx}ovYkcRIC>EW1+SF6K?k$#ICS&Z$N zyZze6=BpOkkgx-Zkq4}Z*Q5mQV*8)Zzg-@@>wuyB9~2kPGQTh+UbS=zRZ0h3QEY`xj-4c(E2FGq?55O~y!MjLaW`cis;AfyS*re#!*epU=S>( zhQGKOd3)c!=HFbkP&Dc^YNs*sr(^VSqmvUF4HrUg|9yPRBwU9gEhrkVRG+8%v3&gZ zpUgf*=P$7~_nh1M)y@A9U83(yy-QFOs^ocORt1Dc+qm)I*?j=lmR_PtQ%-s3$xLgj z9A^e}mhjRAUfwFN)jAl3AbE)Ov%F?yPV$AB#+)7JuD-E-^I;z`G%(apDonGKvyY9y z@X%Md0lI@*Hr&_oHp5>eR`Tz1)DYG~+nzOO<1l#Op0 z-_uuq8YPi>XCBmCv40<;9h=kt34V>%PR+-$_ebrv%^X6BMwMNndEzia?|}aoKNKf(vmoNxi=~*%$M5{N$L(Lenz;g&%a7l>O#1I+K7BKGW_V=$?(fIr zucnGSKP>e9@V`?Y5~%xCq^hUZCEk&-zA$RRZR%?dE)HkF9YHR3_O`{Bm^gT%fGW&* zJ>Bay)GoV!nYd1NaMa7%H(z2EH#UbmA0-!BBoLb0Wyue+l9>xuaEUIRgb`51Es=NG z9bhHmu%3#q11KGOGh3hL+`2xtO!-qR@QRLQ++)5n=*&+vm#oKbAo|M{5}r5E;i=&# zQa|W_L7k94OHZrWSd))xMcpUt?%j^3rCcR(w zw%-k)ADy`ni`CXicB)HuN_`tX$s;Eb6Uh?3K(XrBFB_bQrd`C3R&?bLOj?N+0@ijm z=JXzZYx&9VtEp9;)IAl?evQ|+?jKNYRMe^bj{khXQZgo=CHR*?e{{;Iy;y~UBk|eZ z`>Vdz!prVM^I~G^yG>->KMjYupH}wy)e*+X4T|$o(1*>+{&zcf0Llp)6tWqJv>BSr zzfhWn#?J?oWr9%0`z(&uCf%C>M@u$d4yYaZ=9L4&Zs}Sp6}t{hlV1)Z?4NjP7D~Z= zqu6&=3s>%RxG_QL7j0AM&_AAOC54FJTrf*nn>YVuxSR_cYBr*3C{YjlnC2&kmz#Ht za=~pED5IltvxU#PTa8QklN+kOP{o*k0-WL*m^S0~SFnU_N z&7ebK`k^iV(=Kh;9I%W0gCfTR?Xd%ap@X-YD!H;rW^09s#<1Ii&`?7{{n5z@QK(+b zaP;JY_q9KI+A;!dViM9@rF?nABhLH>C(9?Y_9ypG`(J*PPd3uGdk;1tOh*TGUETwGS-xGrXdVCo0SJYn+_K#}~YvuIR3Dd2>#%Om+{%jeF?q z#~1*!bnniGJA{1x;TfLhAo(08F7-f-krWNDeGToH_eDWh8}(snbTZ?_UK#p^n->2~ zY>gNZDa(?pqJJ1KLWu-REpplf-u7=fxy{jkM9D~Ja?EYYGQaYNXxmzPVcbpNDryfQ~ER5=I1?Z}o&=+FY25@7Ww?_1iLZ=XqYz471!<^ugvk&sPUl-XIj z;ai~3{BuA2P4h@ya%eP_Z&i2N%uGjV=25@Hr?x6dXk+N>yj&4<=8=zz`Kos$*cWS) z(!(t*4b3frpHByjE0v|Lmw}9&(D#eOJ8Lfjg0>?KCCx`Zk=}YtxW4CZC>ZX0U7&g> z$wq#E7azbTVpj>fDqu8MLGbbv0VVE}IYCOlT!E58BJ5+6cSh1sM?8B%XK~|t=jiC< zmqP{>f=mIDlV7-oZ18^at=uK2VQnQs{#`{a--B*1DV#KSx7)w>n_%Wuj4_}wh0Nyk zp_SZ{8IMWd603#S^f~T$fhcL;1$C_KQcc+ZS$4z$lyTXniC z%o@G|u3rAQ0oT2?V%ixE!3kytCX}j$2yh+`Tu;jgI=u_(U6FiRdwW0eBQUy4^NY9o!MUHv_zu3x+E)-44!Hlm<9aM9B}Jp9Sa3TU-)L7wR-IV z1^dRtrJZfJ>dz1CQ~Wibh49~(@qqv(o7d(!D~JZz!~84_=8s)_W-Ox`$=7lKAVfh1 z2up<;N)6_Xift`e%T_M)nzfT>l0#(*$L!zLbK7Vt78oV^dOMwTL`>tD8e%K(-Ai$U zr~`WmSU-oco>|{ilzJ4O{ayxJFpATOMu_jGi+-W}IKF6MfG&!7W+e4bMa4}?nU372SM8iH^sU+TpVGGR zb=K|rzSID*f|VZpV;qDg<*i(jxh+p;KYa-c$?xVqn-I?<$1^}*3jJyImLdIjenQ+H_|C%UR59V^b6-GLItCSO7IbR`TwqbQ^e(=}d%ma|}7;L^wW$o|C zo4FYo)4fxW$?3Ux2f1JOv<61O%K7S^4K-d+@m}53nwZ3_d{^k8^~z7gu6UQOz4%oQ z9_37(-HB8=^Ix_h2aUg)|Fh>58r8-KXO&*<dodAn)3=WH`$e9_l`Wc-FYpVEf%n2{75(Fqt9^3kitdr8whcA?%viq zOL%V#mL(*{=JnUU-+XM>w1>t%lC5YW{8@6^1vi;`Y`bGYQ`BPdnCDfvWbvyC{W)D5zOE>RBLV0`5nU)k}zmsub4O~U<5sA{+m zl<_j&6boc?EBUjqU=`5hFx9XtIG;aiHX%N=_e8WPN^9xd$@CtTu~V-q?yYFPhJ{uJ zKjh&%?smrjx|~9Z1jE8sE$L~Uy#}Dt3C2|&hmY%H6Y&+Jp>`a2$GvBmboL$8Ev!s@kOjQ@nuv0w}z83&$&I*(_sRJzDW&1bV({8{EQ zwDz``8-s`As0Nws6*yzR2&y` zuC+vY1ZkPlg+kJk+C$#%ul@6hE(M}}I=7m6N5=EA-?&W{zJ1nM+3*m!+_&b~e4O+N z`#n%%ZfG$T(eW!_Y^cYGu3^4-14b8bpJAk#fUi59O>sy;3z~x%3$U-nZ8v$~^v9E5 zy;g3lFxc>qh}g+sRoeLb2H&cZfKUc|p9t6Rg5b@YfpLV4@SH8?AT!Bm{b)U~75M@v{s^o?o{W~g5|)`*`)8b(AMgjFc)CSL0C#ft)S@8pkVb~Swmztq zrY4R#Szka%PHW`~@mP?|xHp{OSh|s&9gi9YKX=Pz)IR3Ot~YyPymJqqMDTS;!HW>C zJ52Dp+mqUyBuS}2rs36wIzsy(7=U>07dj&o;MRvxcx}H^WC*`9u=cS|D!~6qwhV+p z>^jA@N{+1b(U#k1Br%797KKcxArOI6xDLVLhZk(&V^YWG&!xG8DB8f7Zv&m^16zi{ z;!$Td)<%vG=5^W5g_aTDs8+V*P-@O=!-?AkH=cRz;WdU{4c#*>ugQ-&JCYYEG`^gt zrtqevU%A2E*wgjn@Uc)|gMQ4tCu!V{uocg0D*(#*e=L1_Jk#&{|KyY$mQ&<3LMo@^ zl+&gIlC;ipT%~eKl4O`oi$zS^PqCyBcKBYT_6CRe?a@x{R=dtpx{12RRRTe=`$PZR%i!vfYe zKT`b8evR+&Hja~d6ZYg=w$#$d>;1h38XEI82D-)Z(u4b1*%jk8KE)Dof-IPU{WsRr zg*!vVW*uwx#YNGou#>r?Tv0vwKg6(_)7kr z^CY78I5Qr|0I5Zn%w-pm!+sQe~6uX?x%gvs7*T+%?GI->* zA(-oDWHSrd7O^rR7akuz1k>hppptFF1!ysP{0-+o5#`&@O_(c&5{wiXZ|e5 zq|Bz6JAFv5pu)R5dM?yM>*_?m`M3D!^nQ(3bqAT}uZS@3@33k^7r5&7&=2XHJ7#|6 zh>fm{BPz}UqcANbBiK$wM@LSKM`DrIgw9j;=-41I64>GsJYG`Rovkup1diNU7rB$5|dOb{~KJ2CA&F_^q+o&HN)DemZ;x7+z1jNzmUU?t?hFiZYt-LQ7U# zEjsK)(8g)c*{K0B{Fzgok=HqBrQE~E!dv!V&iQ`Mi5ahQcC^9a_AeI5`t=@gJJq&p z!?yfoYBoTN&)wwm6AQo#Hx!$WJW zyNvzs?_V+g8D^rRa{4Dm_kv`9{I1VKlAk@_pwgyd%{u6<1xClEZc=`^!}f2D$sMiH z{4m&`gSp`PIechnQ(1TS|GrF5d>(7LtC~NIoePO(=A+D>-nV(c#B$A#9!@wc)pF4H z=BIylqx3=|z@-E)OoZF{ZtQ~icJ07IQb(AA9}W%W+lQC>PT+Pz<7`$}-arjD=AYA( zrLVaw+bV-~#NMSG8_}f(wsmCdpOwfQcyKQCs!P~otkV70;h|(NZ7WiikEr8BW#Pg= zSi%HSWE8mro9Dw(NwasO6DEmH|FU*MuZ~k%x&=RL2^H8BCdg!CXY$oXw2VDY1)7-L zLT?ALZb_>emA+m&nWFQwyQpm-@vP2$tbB|Vkds8)&cFD=$I0YQCEsJzKJpdh-Ng-8 z@Vo-Wmi+UMCZ(J7zw$}pTrYXCVD|*1$VYFgd?!XwB?TeUNm`1IwIDK!5fZiqIym=h zwG%G|#j;mBXwvqw3>#BMI`t1oc}BpnA!J)z-Kxsjh2D6IfuHL>D19705{TD zS|qk4YrM-eiz+^yA5%d#O>39std}m>f2n;p)Ie#!FS5r-!O$;grK&eVCtG({%afz= z=p9?()hK!uNy;MZE~zP?2L&BIB?d)vE`SKpM981(uaWQ3E&)wF^MdC3s;}HRXAx?8 zcC4Y_#`uevn~~~fP0(E0t^Ei4BNaphsb=$679xM*p8fEuNL_!g>-gnq>;RggmQQ@T&=@{%9vt)jdn2)&;wT#R zI4&YYeYfOp!vw2*02843JU`rzvkXA8L6S<3}Vx3XuWi;H!#Wuq5oizd1G3JcCyF>L%X}-1oCSS5qG*FWN{?{?dwCmtkxoLv? z_$?6mMDtBVFG2j;4#6Y5&h)gK!s;l%jkl+^t1fNqiMR6!2CvT?A2v)gcrbkwW{+l= zl4*B=Tjgg`5X&^1YS1wdWcn;@GeEv$=`~u|luiH$l>dxAO#!#FM8Rte_PU#>ayv0R zI+`xumu>_ZG91r@&+uu_!z zc+LZySF|pzJ+0$H*b>!vFd={G#4sQAd;w{}K_|59Ez;yK4PCc4ufW+f)i;iqI;~XK zYa>#$OgUjW4Zj&Hy1nhXoa6OE3AP|7vCX;!dd2g2N9+RD3~H{S_6`-3i%T-pt3?&l z(>@<9zBM!ihFxtb=E}cxi`I+Rz}n^b@nPz-XuOGqUSCTZR7J%Zw=!P!QPC|iozS$V~f996M+TgBj0V@tx38NFkFt~o<4|MP%p z{5;Sj=Pae#M_;p%e4@oNAiala@aA)$L?Ryc9>^|=$S#iuRcaCf9wsya@lK)aO1HKQ z^jEw-^t;P%(@4$EY>$L7UU$oSc1lR}{-%r8c^8d;jOu4w$==MbzIwbrUq*!I1dwSl zBdC%dmPl+|?IEPoX3mdic1kb9Gsv$8~A2Erl3tDR9H69+o zUai-1E#|VH&IQhN@HqFLr|N~wG)H7J=Rer;((5Hs-5_59=-VXeD8-y)Fg*eO{u|w> zCdn1+%5L$o3N@cakWFbyAxh*Z292F|+WF?6?@)?nSytwtMU3Aw{n8J$>ia)kKn$2B zwW~t&OjLF`6veLVn%xBzF$3DiFAc9@Hvh2AZmtZm@=^$HQp>LlD)FA04wluh#S`Ny zXD>NfS;>=)3_hjp73nyd{g{#WDf8K$rMg?WA2QF(S$mrAc$-^W67eo{Ps~>Mj|`#I zGoF-ikgy#T>5nn>IV*>FPS_GB=p=)bf~ucSthPVE>9T9BB&dPvErnbo64btjQ;^WE z*6kne9}o_LoNQ?s^`tGGVaES^+jl$E^K9{~)d@IwyM)I}F&oR?92P?>g{Q#}7u4XT zp#b1fh3SH471odkV!1g`RwR>{JgW`ml@3-PV9Ypkdntc!_P%l|d8}R(;_jvsTDBO0 z4hqD7&^{sZ>s%s0F905q@aTu1#{}W}D)siCi!h@YpjRI_-E+bOc2d2Y30XojFwqTn zgMxmmfzQiyt!^OTjsU;zWz4=m`-Gz^fB6UZNpW&O*0S32aVfp+JICzBw?3SYU$x6^ zy6!ylRP-W>-a=>+EPQUQ#=$slqG{?B%-iwKkg);-Iq6RNQsDGQ*Xcfk{O(^h^~u@_ zrhFAqZXNbYQyX-Wrx5(6Rza~6-%F6%dCBb?svvYzr2-w-_4Uv8KjF=ex;aAW5Z-Y^ zDRp|(vhrR1MAPMq?DrbVizi8S@iMmedtoI{WJ8NINH(ghXq)1o(ARm5a+;{!$dz?y)nE#bbGgpdN|vU|IN?r4Dg^}4W*BJ*XQ z1D{O9QtbM>wR*pM&)SZ$n{fm z!;Z$cdc!q#nC+Dj`LjGZomr{(TJMieX450PJ@y%)GX6V4GVVW@#u9f8e{6=u?-J$G zxYJRdX%lWY-TQ^T z3veSkI0U5nbCo7du}_dnR= z?5c1Jh7IpU%0iM%3<4W=g7dv1#r^;Tz7;^)n}RlOzlF$Ts$?o(g<0fsOlxP;$n3WF zU$*@|{xu+-8TQ8SfO*DU7;j0q#ehQ;N(p!;Aq0A)T@gP`O(Cx)p~U&98h74)VUJKo za1RN5g(1ST;Jh8tfH8mioUimbPI<%c^NYoHqh1S>`VXr|oOC5%!YEwEU|o02Sd?_& z+p@};9d7!W1y^UTmuFqy^1q;xb2|^aoTW;?63N%rYG9z_2}=AZb_jDE{2W>AO+!)= zMl_O^A_T%r=Ec_wR&_trQo1v;g8MsbHMh(~TlzZBa#>`r*I{b|NT-oUQpM*KjV_2T z_18dp8!Npas>@GLP}{JYbOIbZ)+uf32ujTx>|+YmDu-iw28uRnY^;Sgjzd9l-?8@~ z`)Z2yo8Tz!-*?z(_N3>v{Nt^bHNs4*~CFP^-z;Y^>^>n3EQCJxg)v!B2<@fTjF_+ z=JNNW)vq7z7W?i5i!tRZSu+qoz#5Nf>|N1id($Z~)fHr7GC`pMBLSVFo&P#F7HMY~ zx?^_IM7*bJ=y2GTd@gexv$C)MW|~-d;`eG;9N4~Mub?WtIT!?h^gmG6;4o2)C3p@6 z8swrmRG-0Tiv;n1b33sPaGj8=L%M8diYT_pi4m;6WfgMElfUNFX_}9;NAcBn-d7#r ziT{j9o4DVgM6DV&np+5xWdc!Jo>(Kfr+I7=vtgX;eKIU|e-mz(V`^K*F>7**Kn z>F0aiXUf8AL+!cG4oac_`t4D@a^i)jpeoTejDyn{8;OEWE0}y zjp+T_zxFz-ZjI=a%+1WbGp*>Tl|(0I5LEDLULofYa>zY=7bwo3dF0d-BJAUV&b-!o4Sp1jja3aFInG`Zb;SEZ0bT&9|Pl2;@Yb60%{ZRTF#_N z-Wt?A`~{jD0X12eGIVK*lJSOO64@HJ?U!EbyY$^V??g%+ojIgTd7*dZ(pJ%~xp((} z&UsIzl@|UC%K9{N?3bHabF@9T==FCA??IEsiz)T4SZiKZ+Rq+hd~Z0`?5 z+`ZsGNx8Lh={K4^*AyhqExS@aAn_nN_WJ!d37o@G?0a5gi;|PtFd@M{ghf!p9pXsT zvf!zHCLEWXHVjj}B(4Tq(W5q3%|t6Msk|!?7@QkxaI0+1NI7W%Y?f416qNS0ID5oi z7G1+|cL`dMS~x@KENgXAlG_4dD)~17W620gKx`WG~sf~B$K!zSo1{Cr0@f;f^B!^MKT^wEnVC(wXL=gYjT$m*qHAYz2hou@Uc*8khjg- zVjtk8bzUL(r#@Fz5EOL$2UA@)ngY(ecl@6{4d{(r#Q!vNROsc2fLGb)7P_rMMxbC? zYm8UU4n=Td1aOI{k5qFOED_ z$v7y;B?Y#oVT#2J!A^-jpRLN$kTCw0=&0nu^#OmHThKR?XW&*m%;8p5k_laQyzP8+ z=r{lWg`W4??<2G6+tzN7F)aSz|LL23N4}>nBAUWhqN<(YtB=W;w+>lA75OI*smswb zsy0BLJkP!Jt}pS9qk<3CRJb{`w9Yk@Ll$n)Wln2dL?>c4+grWeyLig&KiJmUw&3Hp zlN&lDb#-5e9|6L5XotkTWp%~|I4FAUZf<_uO%%7^p|>ae-5;w0acP$RL0of(r0P@I z-R(ADT@=?IKCq~n@a8siEEe}&`%KRpd#CSGUzy6*2g?zpQTEIYtm6Oufo z6gz=qlREqP_UC4Ma5jQMMR-wX$RkZkG3?lyQ2x9MPdW~FQ@($v>h$u2**(|?*q{Gk=QxCn2CnIA=rP5M zmzPiU8E<&+=i+d4@YSrZLa)t-KWTjis!Aev-U}~5V)u^2Ww7dxxA$ql(EGVPFvVtv ziBJQQhRoW^XdDb^mUh9mmTZac)z;9+hS_I_e~x9wi-~M`2zbyRx1V{T9skEj!i=oo z`9|WVc_Q#u!hc4*9q*-4H<*fiBSBHSW-vmbEiCBiqTe3UaJ8ko%HjH9oK49`XpI}T?G4rX&)cGGtA&sK%NJ`GS|a+mzHS~_7Ac2CFD?daCaCG5lw)aG`vU^Q+|OlTxY{J;gF0SBHjb>`RQ zH>0Nfvn)ZtiT1Ci{$}2q&ND{U=k}NQz-}OF!&b;VHJq;Ce@Fx20IC7Cobm#9Ev1{N zv=)}xqCznhfr#W;QBHSH8DqxUJ0jJyXM#+dJ!r_G`KkYAVc8s~$in{ygQEw$@Tlu z%|p92UO0)WJHt?##xt3AGTD{d=^j_Eo^!plZ}ZQ6oBN;MKYWhY>D%LYX|v=)FNpr! z)4L+V4q*7*;UTfKPDEYriW$!yIycOmBGWdL0x^m0GMuqQh5!kev8^Oh(-8P*B0L?3 z40!jxwRqSq>mX}c#T>D`TIbmvBIAf~#)5%uJJ+)F z6RU#TZ@fRUY;;vEHFGfbRsYV*Fh#x#_}l&+@kj8g-hGw6!hZaID1-6h0`C~o08)N| zJT`RwCFi=oCJBkr;T6GFP1{-? zfZ1@-deSFn!zt+{@74_OBIMtRZbd4M+Aoqm_xY^(0qtMk^%CQ^+!?ABDY&8SudA#3 zOygeo;DHG>*y-{83yZdq_G|WXj*CZg$KqAMGFNDGdKV>Zzj3nTMWhhcjuDweTDZ?F2MF$YdyKSFIGlN9JYPw!v}lLi*tFTOLej|Ga%b&GCny zZ<-ZHYiRACw}6#qUZ>xQkF(K`ySuRfJWc`!!U`S(fHSf#r6=oo1x1^`lc!QkKL?MXvv7gM@7y zEfp8ZxqT8e&6-$}Q%EXVRZ^6}f?Y$)3Bc)&QjLrd<|UC&@-JRtl}UOrHYYG%Ql!%*|C~<%n5K~%J!{<2f-^6qa6 z`)d3q$C_bJe{JA-R{ZJfjv_o4wmJ7QbciPngc*Y?6}%Ye zHV^}Ebl=XX!S`coW#?9`J)qp%|G^e)1G-JKI5WfsEShmVH$us}1)o}YElSIf7?rCly_YG9>ntu@*XS!MrDlI;&tsSX;WYp$BuYm{AdrPtAUB#N1nZRO= zFV#m6l}=7p;V;L|lRU#pqmM}5o${Jqv}OlcS=C{Jt-Kb=?iP~nm#cFCJ=cOL1mZ|!2^U=V?Gs3#wzc+LI_ht{FCX)#}@ zgt=R6+61|lLn9cshuM!_1^a1(jD!^rO1{?C{F&FovEHWQ=dsCAzvdHtnzD~{z^0%_ zyhB*G?slbj;Hb!%Y^4`;!Z)L1FB0tk>gS}{M{rD-zWZVyT?$1rS1Aq{PmWI3U8r!X zYYZtrq8Bi!Q)Q6JqS-oUx!b@d$r_7 zkqEx(s4>z$*iVTg1H6TIupPRqgb3(m4Rm>C@fXgGJsy8;nhf3uFY7vQ1Rs8VM;=F|CC6kOSZ-0SSY98xeQ?d^2W>P2`mfS4`;k5VIoV7)!c5d?E10uvay;R> zDAMs1?hA!K!$y%?Xk=0UoKJKw4k>*wBJExm$KiN_)deDwD0;_{e9%Q zY;|8rpipo0n$D={nk?efS4&NkF@i$SU*F6Q#93cFjVRAYm-4T%XUoE>psORj#hOsq zMYao>CKXwPA-9-z7}g|e?P>XHkANpw3?O4=y&M##$XdFb@}k$(uS>SQWZFcgMOI(` z#pH&RIl@Jt$Go#REYwf`z?bHD^pMzZszM{BHP4M>IQR@kk#}!n!B~!&yaC1RY0h@H z3j;>U8%G~w@};)xW(Uoak924@C9k6=qIz27a!atky9kJO5$xj7|kw!z$8gQ#6zti4~>D3=tVHn)BHibT)e$`bUt7H5*3wd?^5Bq zuMKf+Yx5J>P0-IGrEr%ZGrYxW2143Cjy)SS#l6kRNnUr@^HG3__RcarUr6H_oNm0R z@%ZY<>TCZ$s;e+C7>O%evIWt`z6KO8t+MEnC0nPJ#=2*VF}RI$ElCf47B0j}JS&~` zbk+RYAZSf3YrOxZn^Kjbbl9{|RdfSE78*epL2H%7ex1u+Wy!q8ZN}$Dj+krM#mF!X#zs8e1jy^7 zSv5&o`+>+KEDgVH%wCzbDzO;lg-&lXuJN=?u6`c8?KSgqThp(S!%a7ns#h;Ljr?Ry z>X+U>%t?o_Bz+?8Ng~_U2iqkM%qM0JNtWWaL&Qp)GDKpxGP!rkf(mDw*(4_XE?zsx z*15^ylbTh@3kc==mLp!2T&@3*cPXe~JO*dwF>smVQ{6zxk-0N!2*X!lCAohDbtvHV z(^UF5D$R05_YA0`Lw(f-xvpb;`@imU)r=Y#PKh&~Zt>m?GmMig^_`` zzDx`Lx(epQt0&Bp8mS_GrRX);3!D_8G;rn<*AkTC90)Q!E!+W10feYt2aJ!hEttF6&N2qIH*>TaKbSuuR3X zPFojPs}#Ope~u=+hH2wt?C8!S(jHr5So*zp zjq6j5{m1Mh2>#^@eX&djFT;|TOSZ;uCk_~?`HYp@o%1M^%v_<0tb42s@|=OEMdwkx zNgqIKDwb)?L2>MOCjwS4>%mP5g?AT#6yI+x7;>20g4$Vtwz%6J&p7f1lQO&cOO<@6 z{>b$!%{+Bc{cdt?w5lkJI7y<1WugF3lY`jO4#Oe6!vXU;9d9;_Gbo?NYzM<@=duV9 z;;H8Pgfa7-#Gs0@xo7y{;IW@$V`I!d9rO2;r-(@WEQ0M$gX2FDWrLD! z>gaElplr5?(G7(&`s3^rNb#0g+M~13OQ^Ge%w#~4d-*sMc(#V?70tnnUUW(9ZND9En<1_>X>orsa z_bYmm2hfeWfY-r@x&pE{I!TI|Bi>!sSv3nn8a+Ty*cLduqYk}g0_g8^0ywYcawuF@8VAA=zFJ^*ltE;HTOI~9APgq$iELd~& zIBpg0MmIixI$Rwc60X^Hs?PPEh`^mMKT5hF0OO%`-dRY+N$^03z>-0AV9C>>1|y9H zE%unQ;1)Pwks|QZ!rep9j@$L5*HJ+wKLO2&yYvZ}^(TqKp3N z6%?SZtfrzGqB8A0>BLRJc#V;Dv?N~_O5qgveYJOri|{Y+WYCh4sx2I3V}xugrQmPD zz5K$-C&r0OZywui8N^9S{x_gw7(>o2DD= z3nnPH>wtfpJGR4BFVT^`HtswEQr~M8%*-aRAzim)Ofs&%I=K1TgIvdp_uxNR++0DG z2iE~KG>5^)w6Ppa1S0m$8!H!ihwy_j72vP0hM zQoz6}h#9sOvg?)yc3bsHf)W-040|4c_HAY5%=E%vG`Tf7PW|ZR5Vy4Qrc4EqB^vwS zssawq-dF}>6!CN8sK79ICkvHAkha(5nc?EXL}zC^T2LAGX!an5u+0*_CD67}S-)`P zH0R(CTilQZ=4$?I?86ew!xu0HDPWcA!HinbN&)ZQ(G3(B6xCnR^38=TPEMKLsJj4`9GkwG`jC z4D-=+v{E2)FI!j=ush-BxHo&Q765upH+)1MM+8NcANEa{3k=9tjGu%dULn|?HIhH9 zfmK!wnMRH30_ixyErOD;^zVw)VrtmSgV zhbtuvmA&(rnNwG~fPvKM&WODu4ci|U8ooW0b)5q)@u8811 zA5rO*u=&lXdS^ikvqK*JbJ`{DilI*D?(pqmKb1tFrbQPv=Z^;Pd3Hpm6hDdAVBJHn z*@sW5uJc{Q=pr9~?dp01%Zn#GU0M5<{XbayH4+W;X-G6e=d+JS-hZ(1mG!5hnsJ-; zelKLokKeP~KqbL;qWo^!t{cGSH5BZA5BhaTkJx878{{0AT3UCo1(TJ;@jFDn*!kDt zy?)3i{uZ!i8{H`yO)rZ38KgP^_kf35CGQqv?qLl?9sX*7$&1FuvW+&27>0ky(sThP zEiy6k;}zvGYlHUdHoBuQE$D~M;8Sh3EtmE5gT0&+Z2vr|hJ+>@N69eb)dw|WV>Y;^nPZxJ1c+@CicB%zu_1;fFFMxiM+RGan` zTfOI~WWevV>Fg{i8>pT}TPNSLGwfA4IkixGyeTTPc9#e2j|aD3kO_{GL`F{6AT^PD z9U710xqxYgnq?@VOQ^+3N#$GEE3#;!xTEZudUI?dqnCn~X_&1I4S1=a*%xu8W+5s; zeM0WAS?Qfr$$2-OEUJ#~F&%&SlWB{DO*&$)kBA?>X!b<&?wvj{(PiQ!folz(e_I!o z?wUz!dQ&%Tdqq6X`$~vpPP;l4tx>GZiA(RSoKko2Q*ZS>babpL>yzyhYpcs$L!a($ z?Y$I_AMsftu_N%6jLI4bh}c1^pU=RdqDyC?+ka`@|H0%yi{AWgq`@%YZL92ykV4(r zBQ;@0>DT?S<;PZ?Wh7vzLgR=dUvJ5VoEmBvu=?#je#=dOb{ex-diag!S5y@gQ}&bi zDk$n?mHhvj8%K>zlplWANY0))91e2s&Vm?tF~QsaU_paYv)JEs`0@hBkV#5h6I^AH zH`=w?**VHwVkn5^4GWJf?>C(8>W`4I-uq$cO}zfe)>W9ungn>f#qNW9tUB%V881@Q&zGTi_O(YNO@!q2cTDry>5pz2ChDPD1957vMF@3;DkhrZsEi&o|zt^UNC zoRGX~TrWz!48AZR2zR0%xLGf|b4G%Hq4~OdpWOaodXpeRYuqKeNP~8wI&0D~FhG1I zLc8Ey#LRZd@pes|6T5=1sLLfIb=fIt#`z=Xzx33x3pJzMtlyO0FAa$;rA^Wfb~&F` zrd&+_0EEveLUr&##^6bRhEZ;lVxWPQjvV1KF%z7)rdbx=4uq~*gDDM7~bCys7AY`^v6F3|@Nzy!qB1Py?@NfJX*h&jMYi?W*3E|zm0M~%6<>CzXzA&=m_db7Y0)c5b*Vyt^80yp-|WJ{6fq(p>jsI+Q!|;{T3@{ z#e0b1CIw|nIf2}Q@ZS0SSkIwS9WB}Ahp;u%Nl`9V(1N$*A8Hbmk*?;%YZ*cR!76cP z(BlB6L##(ZR|_G#&C)dM>ecbm5nRyorJG(0f4{|KV5WU&UYN;L6aUAbX}UcJfgf8x zoi@elIq{jK1=gEpI=R7&LGNuIjikxoOgQLHl=6cU=YmiTfVL<-prc{>31rl|h5#4M z#qxO6PUL`gx3N0R}-QWky0#Qt5%VQLvuX1(rT|CKcl)vhWUp@zFKV8fF4xr z=rDP`LJx2=lDM}GBvOd{8tfuL`h&|(JSV6fk%IbXpACE{fr_)p#bw;`J-nk4Q5!|^ zweuP`wlKDA2uZ5_CEPI(U=eaH!R@6%Zs-N)Y-i8Z%`l`tQMeg2YvMM^LHFSkt*oqX z@@{%k(f_29H!P`Btp_wmTy{3xKKuGo()CS7soEZHzK`PjwI#o_0I?dVzcHJUQrHj{ zoUddGc+^QeIUlH`$66>q48Sh*J4! zMvOHX6;G30o9b&mbJqRb-$m|bN)o0Za6&u#0;jGLh8M?A)9OA~^33_sWA?ja z*{4B6yu40dRR|hG8bcU5JcnAGLg3g5xAuLx6_VxcH)DX0Wcg2E!HE~SH}f%6|}(;)InA~CaHMqIT_-TB>VRRlv zqU2ZdL!90;HaOk+>36;Z1-KaWSA?e7EZ#w3C*sXu=b(fie?R+cN~$Fo2Vhh!2??|< zmJae3s|iT4p-A;X7#{&Sy*yCnTXb;Ta|CA>6jcRfyL)|twDKQ} z-i?vrD?{7qS^xY0;?_hP8?+!4r-m6P*#qK!m=;T*=l z37*=Yq7FOvm16Bqa6CE%Cad&bGC|g}N1~`AC>+b>p@*vf zv@1-_uoJ}PZ3*rH-FThxb@I1Al`oBIEbhaw12Z!->)#5R3KV^PKJ);4I+LBP#U7VD z%m;4_`}y>}+4E<-Aj;(IIft(OUtr>nuM)RB-PI4$^;^*C4H+ zZT-%57&xTG^BVO($)br{aMgT!(;8gH_@<${iyLF-IuKevuTJ+Erz(BC9lvc?X6Dm} zU4LG@x}$IRT;$R%++4im$JM^=>hXIzvaN4wM#(lKu#y<2pP3_~7EBGl=do&*2f*@tyaBOm^|lxP!n}D z{h-Rut#Cj`V8yvP2Xg;7FCD|454e`$@X15r=p)$wqVIh*;Cc8cPun0SQj<%s&+T#evg*3Wfw^S6|Wz7Y_2AnGdk3Xi#c5XU}u z5~6Y2pd>nCqK=XE(PE66IV(0im&1(eX`R_QXKFMM6j`xAbxEo5{eDLFrPox8@XsR{ z3UPGBNJK{Ti4{ytadZ#kYcm1tU zaG@f9|H#$*4$_^%ts|nLzZfY310Rm^FuDrr#XLU2LDmr>2`}B>h^o^~yFHPcalw!` zC)8-_jFSZcc6k((4DM`bD7asfwRX6sO~kJ%BCi|tx%qdM_{vk1;OMfH}k%kkn}7|Om*y@4AYhk8NS;8^5LD4{gO>RF_$qQ~2O04sL# zc{ZhHpL1a`HOub2eG|e7RDD1qINYP#bTTGg@fmYSZLEZZGWhPk;)^y?~}{Maw=zrBM;+NXD> zY9s7~pTG#tdZmwEi2~JnooQzd^*6JrgO=3Nn^Qk1v!Hf8g|*8boYk8vviXO}l$aOq zLjq~8>C`2^mhlEnzqo_L);e2#Hm{7HkgNnXLLe#7*p3UT+l0HgE6wbhjBlJAP>&?u zE|wP%g@~n<%5J9E*jiXpq1J=vOJ0JU)Mxo~rV8F6LIc~+Pb2R{;|ks!Ex5|q`QW)D z{9~6obpB`ihqHfbHyd zM&r;e1RN6_<2~iwRvlVCY30!LQr2Itq-G)$6IwG}=PzI%c#hdYcK-y$Yh!>J5R}fL{22#3?np}(Y4n5@e zZfqNEMH}r1hPuls;9b8$U;+%!(&(H`aw|kb_gy7xTeyfg-} ziaoT9F}0QU$FD=Yfah8D=Nv|*VeDhW5{uuo!+)!pj;5NI(#l}JF_VT|PeFAF8%4>P z1b!n~JY(;QWgB({gVvLyf(_x8aVFU-dHByr?b@Mr_W;Lnp%^ak%R10%+--8?U^2bF z+#*N>ZP;d&x&eB#7BR@^^yE9q4h-9c7XPo?$;oUY)=M<-e4*crqFzdv)Pj~4sxGH9 zEDO*&EaCB}E$yb5#Rw>^%`KKi(4vwB&$)$)2c7}Cc9)IpH*2^ANxys9_G{jN@>9ub z@T{F!gHyrWUTm#tj1Y){3=9%cS8Ecf4G)N2|G_{vVgnj-MrfQxMx2LV=uXX>{heN# z_15jSW7YM%*60SBLP(x^k=5tjKi57~ReZV$!_M>NO}SYDaCu%qr2#|U9yrx+zH8&J zTrN)^OFP64Acvm1g!u;=+Pg9_(p~Km!}l$)*MHVryVaY#ay5V8((t->)7Ex171-Yy zJ^)4j70oF66GHt!Z(`dOdW{n9_BiUAAtHmMf!oepP4iP>OAy%`bQN+|lbQ;;+u5)i))p5`2`WgRIjYcM$lnak|?!$n2@W!rFH@{W6gy%cjr0>oA zd~{>Br2k?EB{?2-1^i_794LSW>7!S8JlB>Clrnw@x&NPZjK9D2(M25#bM9aoYWwW; z2eBSmx1+wI=(gAZqyFm`AD!A;7LQVs7DL&{<5%TNv&63HHvC%7SF*_*GzwAJl_p=d zr&N5016_0+WTi(z)f^Brvzf2MQS8oxa>$rXI6&)>!z(S^Ux;xi#JY8ktId3Fj(Rrb zcCqYGc#Qt2Nm)zYiP6`lraO%_6pr{O?G^cnHxN``86#Z)dP_{La}oW|tyojAC$s>? zGHkpJqgAS}W!T2Kno;t@c3WXf%w*{pYQi=yYuHtH=5tf-bl>%lFU{|XaI)!GGO_Bq z8T7xU@w}Cvfg(|I`>&%Wt~xHB+SNxa zUzJ1Sk=2GSoNQuAu-mI3H9~rQzxoSf%SvOz6G4^^wsxug2eA5o(tR$%EGe>7F zPQ$tQ{SM0_ux8Z`eGDpfGCMZsKbS8J9TXc~;p>fA7>~NGmXt-YzO6E5?(69{;pcrc z&x>|`)~n6KsvtjNhmY6vfjw@rlpwupRZoXDVQ9XK)=h=4Df9w3^lb+1GbFMq0 zvYzGdtE#H{gi1RLh=LO}RBkLmlhCP8qCcQ1d`p`f2Zx?ImjsE&#TpeyernPBKaQ?E zkm>*bPvn;Sj+G;oq;iGXbhr{a-A6)5l8|9Gk^5LEN}&tMl{@F$k}|gtv#ltMP0TyC zeSfdd??3$Uvb~Sj>+yU(A4hg2+PY6!eDryS-E^Qzssq}~SNU?A=l(mDJus}%BYyqT z&#<(Y_hB6x_&Xep-=;8_>anCIuKc8HWHhuTu_Bm)u+A)(0AJygu*>=U@39T^e%qi@?ujLfcCnr|I*>Yio%5@>6y9 zf#}<9SZ|#lN_3=8rG#p^bh{s>&zFV03uxwTEOGNwJ~$EAP?YeOj7wu>`) z1^l{TQ3u3m-y^=oi2mX3z|pS(p^j)ehYy2n`G5Iho-@3MG}A~9e}qAiqfx!{0w9Hr zA8kj$a3rb|ZHB`NlVo>ztH^)WRq}?qFQ#4}MOeH>f6&O8cS*H+8B zXdaAhiUWW0EfmM6qWU!~nMZZ6U98TJ=xokh#%Mm>Qm}Mi;-v>x;Vg zLHpQGWApN}w5z&i-&J)E!(d2%kLL&fI?df1_;ohR($^h$94P4YG0G&$k(X>?yug>n zkfP1M?4V$fQYYfu?c7@@Q~SeTf%81n_ypPx7 z6YNAp-zmE>p8Goeh^76)89y^k`8;e-^Ho}PejbW^um$2q1YHiAFKRq|;z~U|aPx!C z*rKXdpXgZ_?q2%O6L+2Ese;sm^t2I+#H(5NPMrB7;|wz@Fljq+^7DS*(xvK7x!-=U zKGTj>Lon)fFUU6$m4ZFv?0-nSD9p(I{AXFkRokwzL(9$Q!$rZ z-#w?-ZY9_sd~ETla)Cz}^VX>UWa?MTk;(a6PKch&N0~MKwS(qKy{g6sQ(P@hoVAaU zRXg-|+qCnB!x#HwW!05ZFC2fua=wr4+0)>Xc8NR2FnEZ&W`C*ObH_5T`w2Jr$<@_; zf;z87Oa-R0H))xmL(3zUP5XUw_N%9VXq(dGGNZrGXm9_cFb@C(_Ok+uX8hU@PHox=eu~A3LKnI( zfROMp874?GIbji^UesD~XhHPL>lQv$xX$CUxX#RH-UW^e>HU0*hCky>mBJIiAz)5- zHL#x-VUW^*&vHoU@%h=V`Qr?}goV7{W&sK$+h*n0p6$PjYFuQ)GSiMWg^b1+1v93o zDUThf7QI_Hu=a4NR&+T4pHd9L(@UB4PhKiVn zKmNM$!}`5zYh!vycyO>@&qf{o$kXJ%R%(y;wRGdeIjWFR=O&4tb(TSj51lUB_1^J~ zXZcWY`~0gS8Sk#RwSxnMnIqvgD0Sl9cI@=syJs~DHtyKGQWAsK8Q)HOqUi z=)m<`fZC%;JTnGNh&gSQEA2+^sAgRh`@ZWe#DZ~%gY9AHq zrX2RZWGV0cIG9cDgt0FIgn|m#Sx@3BkEv9G=ijdvvE zjFIpE$-&*8s2?0;TTRZ|RoB)hyDwkQP(K74v@%Mcs<Azx&(+u%BATtBzIP zB#c$wE#ES>PP%)is@xY{d2H=U=+nEe+pWCN$7yMYCx>c5n8#Yj+ZaFYQ3ooqU;7ja zY6BNWt=r{-G*CY>+|rVzzaFVaTvh5H;l;x(;&Gxv&Aa>2&-@Z7ZbywDkvG4V3>-T>oKI@@%f&X6bk0Uj5 zvQa&1x?DRC$++(U%y^{aFrbcya|Ek*wJ`!Ctykar&8!-MsEoU!auB}XzS*{|#A*2f z@jY6n)45`?Iw&g8?{}j@lxtFNN0vaiS1{;r0ya{o$!5CkkPz`Zj+xQ~?t22pb~xy6 zePN1`!77pNLJb^;i_~cfc|`?MqwV@4D{~s0V+TrIBJ>B|xgV;C3_M;Lc<__=-VPHN za#e2?q%j#s-|cM1_n~GSKn+3>P050=L%3utH-w`E8g#Ww4P$I$(d(NP8a-H5#*gO3 z*FHWb6YseW5yL+^GRUfK}iiCtV9TAQ){R3%6< z(-qS+=i0SlEHRTA2CLLvT`ZAQ-&X}nHxfH95G%Q)Nsn=(4oh|~JHVOA0FU|xH2t2jvN&&k2yVwdH~)hr7it==~}UF0ObF%%zR zp|aoz1hjZOzp}wzCCFuw!GG4^leYx?9n;mZ=nWp3FhB6u`i7uTlB65{WZ;xt$o=@m{%0_g)CFyt)dC69Ah0Cu zy6l}by6qm4;O%OhIf4Ys<2}oZF56qC2wMHWjZmGUaSaXb()7f{Ew*OYy!u#ufYvz& zMnEOuZu8TLv{UB-OAW`4m$b(IwSzHOW;9fVi+uLds!P46~2 z11`Jg$cJ7~T}a$$k6xURu1g^=%aHz#X_QbzflCn5ws69J^ugc~lNDfP^JB)fd%x9# zg&6D&Y>)C`b_Kx`&5{LAjwNKhpmnlU1SFoHC+vAgfF3TMdxh@yV36YKaWp*l@r>NO zpyaC1-F*E~T79~mMaHC?kmG1M?f5?ry%^*&v;r3SN?)h-f3PaykO&{Ay6sZiVgAZ+ z5f|10sD)v%03z7Mi{)+xL@8-r1$}0RM2sgcD4LmmW=(HS)b>-3b+-SanZI0e41x=H zOB7NMqA53QaE3rfP8N^zDU@@ixxhgyt%1~?q$euWsT;9t^E9y^6+b2U{kt>Y?EWju z^I{WfPi>h^k2LjQ5%`{UNKJ#*yPEmZ9J_)7F!w1@d50`gQaaR>k}F-HN?I zB(kE!^$5&m{!jB0{l#jl!mB6^=8QR89udfL`!z9N?FICj53T(({AHc760OUau2Ad|W-u!2Mkhx&Q7 zO@;xE;qT_I+fy9Hnns|0u$ipqeW9NnCg~p7VBNGBWW36lorBo}IH1}?kVhWISVC=7 zq1Ry)+nl_|DzIVZMg+Y7^Vdl2&E(D(?@x{SD$jI9 zl5ZxtS{$>@J)t>!dgJG9*scqeAM)6>U?C?_es;WeWRG~9e5;O9E#Tyy30AL6B@-3r zNj6@^{fv@@IMV@tUv_sS9@gZXst zix;XRuFB2(sR&Cr6=$5ja`@uMoXpgVFZ4h1%@opYaWjImz@i3hy1qead2dzGUgx0H zw|?!2j9=EXZWt6J0ii^d(?}w_fA=%0|2FZ}*4g->&s}`=)qAi(_9vY0G|C4z>p(5( zI3Me{i3gn%dA&h9epf3o{O&n2r>xwp3WIDuTYItQEJ)(h^;A@qx}s(}6aj-y9S>UGs!5_l0F~%t9$mjrF zb76`FI#o=W1ix2p(2-rlACf%hgnv%uMGyKXg5PL?-uR7itpdI5-K*Wo@~LK{Kc43} z-K@4hIbG3_a6q*G{EaViH+Z>nxnZg>FAr1cdnFobaX;u>Tm1mj?V_NN9<0(7)&vtZ zzJ=q(e)AArzcO}}FnaJ2dW@=__P{E{lkTuE@$_G^o~Z0{bsvpVDB*S%1U&hYpM?d1 zl)o7rCx2&|b}lkW1`}2c{Lz^N3Jq8V$U{DFmwz5lq=O!$%tlE)+i5}G8FHL8vb_S`R7w3oqX$x)Zt5K9PJNho;8y? z(UbIQPtij)u8{g1@7cw^&wmyVXK68-Ya%{owcji~CLdmTuI8t&F4r2^HMj67fHQ*; zp|4XK(>7m3U=+#GvS-*Y5GN@smY(VB{arx?Kd+t-DVkSpDrG57qEY>u*5 z--~O)jl>o|V`-B3$2~eP&tKhJiF7&dc_1LyZu0E=w33boV&3 zsJgHxMpfOY>81V}v>C;EkO(Zx$Fe(+d&YFPZxqfrv>#`@Y(D-}CV#P}Y_7a&5AT&B z0iZ&21XP-4ly_B32@Tg!$dbBB#>a0(>6$k~I=hE|%X-vV^z5dVd6@hfTVr|nzyA#v z4b^N)8%e724U%v(9@{oLW_62r+~0Qge44Sdv1_W8+kYXpHUv>pFv*Im>Qo zHw6q$GM@P$E*!gU1P-+KW{f(fy3jjwo9> zt7)uXm%kbPz~lnJ60T!gNA|AXW9Jxp!&upnFGI2i$KS3BiBfVOX{&t%;wFNuA5w?D zHLhdD!&p&P9#))@yli+*w zPX9;4-unppMaIqGdyURtaymaRL)Y$x$G0oxI?z3>ttzi?uE5jURVMZLO>^dCR_qB^ zrLy`Dyb}Fl#{-o9xP^5{b-RB(v6Rv#%(@Is@A_es(h-^>>?)R;FPYR0-@B6cYq27t zNdC%#-^j|8n3?oEqM@mVhu2@fkL%wpnNwe{lQi~qkKBZ~Z^rfX z_IvwIn2Jjdo)Nw?u%|Ecou$e(%SW*xJf5m%8Pm7+n5FO}E$uOXa?eybBpbrMH(Yf+d%yJ3Ad$ujkzn)#*n>lgGdH^#<@ zd~YedeLiWUH9^Ymhtk&}9q#7`=bB72J^5vo@SEd-fK%H-ab zmiBRx*2CqgB6WHBbDNV51CmlHlkyVMcPs~0V|%!V^+eN1a`i!bd+F5cphGQr6~YPj zsL+wGscYG!*@R2KHa{@zlbyp|pC4&S;a(a`Hi zH9P;c*Yu!c?Wwvn!6P@SV>-l-orvwc)`4isw5jP9PG0O(+jky>zZO2nxJ>Q{KHv>!VMUg+NmEd&ymCsy=H644#wHl zOe^8cW-BR1S;YN5n&#J$JC15kXI!2YXQqmY7GsZc#+H@#}b2%oG~dDpi_HZtV*gPev_j@`Uo{_-hzPY&~aJ^nqt z?^MyfW!?}O?!kV>on)ua#DkF;9q^BZ4QfFBwhy)Nlt8!Eex(0d>p1qv3ZdSPc8LxyuwwM0l1Vf6kpAq~P*?9=BN~ql~7RQY= z8mRaaZghDVYXa&257sxofAn4U)u^%Z-+5sfgx7irwcE5T+EV`Bx-4_gTQ|ONZ3B91 zY)zjIz%zfqPXS@KmY9(Yf#n|dBJ|h_v87Sa#QlnIl2Bto17l$33?nfaqeR!>2o!>{ ztG?|;XPzQDx6fULk=yl-t`XO&`UcQtiSIfAyHNs4rPSx=yxeXss+)0payk0F-TM!- z0@tkeS91L%DdU2Il2ElCy#xi;KBxt(Hu&HFM~6^QF?zQ^TeirIu6k{}{Ph!;*?F@N zpHgScr7Y5=+P38j-ic^5=gfj?8N#s6TI|9yemA!D4A>plDb;&%@J z4`z$%7~X^Ho~ABw^uW8$GtmK%rrZ`qnBzgOF^T@og(I(ygn;i3IFPZo8AR5LsU#yY zJ6tmN@FwwZ#~!e`PT%|=>|GeHQxC^bZawoq7*Tbsum}d$23O4%MV|AZ+g#j$p=wD< zytb#|8<&TKB&58#ni4EsKf)}*T(x|mW&qBTEaM0?Nm)(@JOQR;rlkU@&vPi(p;X<= zmWjN2U&{{4-+a~!!}?=xKXZ9^h4k}aYNtQ+2RDlF-1WsyDjyqF z?>yY>n~2dNS4eOy|8g2sgQJwUucWB=oIe@3$@LFr{`K|2q%SY`|J-A4QF-5v@70hW z><8Bmmz!y8Sbjl^ua*lZX0txuv$dpVo5Je})l-S(?f(&|j{sMds)ykUQ;96CV0t!n zeU-r2=Q5a($O907F62lNdqkyIWbN&zoa5W$-pvMh0S(_OP1~n_WOG2(MIMuR z8n#oiz_%Je-{uJ9gJT-Uk^&~_btq8fUz=eaA6rK%V!t%NdBeU1^yC3*Ce4KMGh?uO z2A}*ENJc9VTbzGqjLQ&@JkFAVd=j!A985P5Pe^V(08FXA9@SK6CVVphyo=RD7 z^I922TRb2#zuRQ7rD|oTd}^k>HcU5e$zUh`N25et6eS}dsY3Oyz}XJDeJ$dh*SdAe zW6EO$c!Cj@-!yVYT8n1%Ez)!Y#d%JChDnH-e%_yvZ&z~ZRjrsgFV|{>F4ymdLcaFn z9%uzz@}!3#R?QuA8fLv@8+FLs}xNnZ!;K3I0F33v$%nul1!H968 zHH2F{^Fj{K>Et56f!PdwiQhxwa)BXtw^`3)}6fFmvfqx){~~J2-#=H zwgy|I#w#}6&N*iAjosM_D3jkiQTg0tlw^Ik=IbOwrSvhk@P-b)zc~0u!K>iH!OU+k zo;`fRSf48??>>t3KW}O{UtQX_FsjMeFkkxhO7tcqu!~}OvWq$Dzft@kW=+)|7`kS( zo+Yo=RVc`twKWtFT3e^|qOWWi+ZdP)hyD&WVOD)`b9Qnm`Mi_YO=d%`UdVex zJ@HO;t~qav3E<1E)*wZJAf;B6#!c~1b{POJG}sl2c`cdrD6TtKf7_W4Dldv(^V;0x(W^^40!@fjO| zJgZ#v;!9M|V~MIE$0vp5-rKw8zk6k_I?`ddY%2RBR_f;wMW>?Zaxnel`x@~~9`}r^vYpRrY;rhafHb;Dv_g3k;K|X?gE(^elAA)jZvnaotHbeYyECnODkuyPBMd zNcie_fkC1NP__b6|IAHdAD_kZo9tJu@CtO+dKcIryI2Xmm0LHPDe43>=L97lxq>kJTulV3}yE@ zySGs7M4Uk4baMM4sH_8dqGd`O5MGR;iyB`02Wm|)%lMLSwAQrN9x1_`Z6S*J;ogQjyknMfW+Zi-zGK@N&s?`hVZKS*DEmOG;zdvk1cD_!Q+ILl5) zug*a;lH_^qZ+!mW!rI58FYcXv04!ZXN(Ko;LCgktkJoO}{rC)2NQZFC(7MOqkFnR$q;O%pK zw%`gklI!}V-V`2k`yHTG(Rfc=8TL(MuB82qF4KG5=9$U`-}w8?rep-%_$&syW7v;=*BCjsG3!D7c0 z7%Q1RmAC~E-k&^p{HB{n&&*|GlpP^ht(FE`9=qd}6y{FUWOv@T(@0t76>w=oX#d#1A@-NErJe zxf~>#$GLX9#2-dEkV3I~9Xq5RFk5>5lHA$; zW7$TC2aFy_XwgPxcDnr5dHcCVc`1=6Tx==m+<9mgz`pn3Bj3S}*psAKmqX~X%s1>} ztTvViOsR#)1QBS~4o@j6e8f?7afe>mGYU$5U%B5l*pYp-zN+_r^=9Mb(#M90EyeVgoFmq|9)jO9qtORy3aswx zlTG>h$DD49oe2svdD`3RaK%okWzNa9Uub}aVFzusj#Gw%W2n6&d%mM}< zo9ZNekGldc(FXwMPJy6V+#`=#1DYgVwG$@>XkcP2>{a>Te;pI4f%FFveT-r;@I5~d~O zIH9X#75?u_b3!sMxQO(@rU9jYFg)ONiha7XBh&cst<$itr-^EeJ|o;#MsS}Yarg#I zPD5^p&)qmUBS!e|4mC{y#ziD3zw$lsa)0>Nwi*?7Y0uAaxNF8fqe2YVEb3}n47Y0K z$pRMJkXjk_K;E$blIU+CCg@Vf`ayim|Hf+3f8Y|}V78{~SHo$!U|1oG^H0CEeL10@ ztg)I)?$Way;#vX=tl$0ChSX%IqZ83JpRsp({Bn~njNzVi`h|f9hTo!pM5FYD|qAkyTd=h zH)jKuZ!B>d&d)l-PR8<;`v(+bB*33SvEQ1HIv!1;+C=qBpb32()fTfe@{%juSH|X< z4!lnwkK&bY!9Q-u5rmW9#em+o)*7yRk zMhH`nK>*FnJ5|#=vkoJkbX-yuLJ+I6j1_F32-IF~%s`)0qU7_D#JQnjz2 zUG~9IXhrd@px)@vH{p{pqTlVHqb0Prj20>=fKr)ox^}>#?;(AzeA|kr;7H3P;yQ)DGU)5r{NSPd57px5RUdQz>*PP- zWF_3Qcq!b{0Tni5cj7{KmJfjZHx})wuzi=LjP_oY{H%WRl$jd##=pl^c^QKrGiw95 z<{Evaxc-oYk-98TNVReX%3UD!+9d&X(yrlo=1f&LULK0;oC;8pZ2T6;eW}_7>=|v+ zGp^?jsC%d9f@^1`D+(Rc?$?&_pH!A(*I-2f=xNk`6iC6_xezi)!kfl+DU*2*mLhJ9 zv};2uW9JwZ`Cm&s^Eme7K}&vb3=GU19*Z2Q1cy!k{Ju-k)N9_)QIG zkY14d8tsFy_Iz-xBs!Q<-4^in!~uBzUL{xEhc%3Nd52{6bkmIs2(@X~8kG;)*`Hlo z*1|+h_ig@*w$q3wVvU1oV;O$1epuipjTqEjz&7&Unl%;i7A{BB?M=yFJpC^7ff|FA zu6uv4y>)VqWyiG9N2@H$&MREoiOC!p01@9JrQKQX?3V~Ve=DWix((>$3_LB%*T@m-i2zWH*tB=|FvngDgi69qZACZ|4mfD28 z$h<8gEzmi~2?8*cpa^eg9Q;@h+B*0cxG=(tYE@O@6WUO(pq!0X%QIEA8CaAt1>?y~ zR3nNegxkHwnF2aARR8RYRMt!4S4E4Y(2+jYxTQZE5z^H#D zWC_8=A*xg3!{cF5#Z5-Q9Dff-l@U0RUOuV64TBqz(D)wNo#UD07Zp*iZS?-cLlf%s z`G@b9!yRUhC(KN~$}1>6l>6p~H8Q)z`u+0BGw?M7L5VRkhQ55T%Q=2XSA zHJ;dRRDKyrjya;4@LW3Q0ZDM+XL|LMWJ|S}BitYP3h%b2-Pq*$SqP&4xLpV8q|yt+ z;Vs?Os_(tk?wIl;J=J8?(MRudZR1w!-!79B+j%$>gj#IsI}*CE_a0=|?NK+?Wus^@ zcDFwI%6cD)yZ&kr1Kuqo;~dsDx4DB)23c&`z1t0w;vEVb9)d1BP99$zU|aY^INgFU z8`W#1MM!Tyf;Q-a*Hm>&RD%|WIN|ns5>gh@e9L}L35bjKe*@*P5nk`ht3`6E9e z41a#ArLWafl`hP7G>495`M4>PnfOC57SfN7-#ytI=uh5EQ148?K7JZT$0xSK7~#Ek zijWJrr2Cs7P7<2x%bUX|BD7vP>b*3`ZEmPAc6xgBUOe$*a_1vfmO~uQ*&ZwLTes8w7+!TI7G)DQqLCFNLeQj+&0-WG< z(7>O!c3#k zIcG$@Jh`q{XyOxfsho?&!~{X!|AX;iwa~NVncEIK96XsQww#H!=*=s9)OuzinH9Y7 zHux0UrGFBuz%VJwX)N-?o*EUr-7Mc0Z~OY)k9VIp%+z85tJvCuv9G$=ulaac)k6hYX&9dea+`KFT5VnlhJi}m2wxgWFs510@By54k1qPe}2_ndY$ z((nEo|Bgm6s-(XhW7%nD;%CQ^t??ZuNWidz_o9 zjG&E&^aVeLdEXBvzG9&}4S!OX(JP<1extWfQ;X3{57W|im_LWP3bE{ev@sYH@nCcL zzNy$IksG$~H{+S|97G)Eo{W)NDlmoPO^D=&!@h{6?_4II7O4BT)IyaLRcGH1?epW3 z%WM=5ZaZ_$X2VjTTTOr!68MjL7Kc6qOZy3nBqejWG2TGdPDmsuVjO^t&5URlDP=nk zq%6($IWPreGN}Lro7b*LpxyQJ^=vHcGGtD<+*avvnu`c250O;n=gZm<`1h24PoDp9 z_dQtY547rT#|ebKs1(;M<^&W>^+*HmOp5d>@=yxKhDu5*I6BOd2T;Ny%EuE~MS+Xm z=C&CtAKl>i-O;+fc1peQ@~4IbRPMf?m8M6~vz^NStpS%rbN5H`7LjpsXdSn4AbgDG zwy%#0Ja=JAH`#gW%GlkT80i4(^kU=quMYQ44T^v780OQ^49qZBd-(H*W}sYPjf=Hp zpvGxqZ^QzXZ~?|n22W(Ul~PqV;2ZEI9!NNSCzJpk+J2~a`e3?XPcBjw5a%SeFFkco zi>M`fismhNmj|b1H328s$CcK?rzOhc)x~SyUAfiM3!Rg>p6ED%Ir-Y%-K-MbXgT8c zBp|XO-YeYm-E%JqqGoIg*H$A-JkFI}P$xPg+CNi8lVtquCJ3%U3sd+Z-H0b!c>K3N>xCFWV=q)Wgg-FgX0!rAKDT&Q?^|4vNVJx8wn|&J>wI zD$aXgu0o!SS#oaguBaR;rds>A$(bTXP@bq6+yB8Xu(N{@S_|J>Bg(O}3g#^pEFOAz?A_z_Rnh?pxI{#N zMy$3r#_Z0desg%jox4GSm0M=*?c)VQ#O}!{g>S1PXXH(}{t|kDqo@&dk|EhwDA0b$ z?+WBV0peb+t@B{AFO5}N3Q0a8hVs=QFf2h`TOmE(dQ{Wm;@=wd((%<(e z{U0 zG}SdXINd#%c|XA}`(F2B(Wxrvs0o2})MNU8Fpfam78_gfr)A5?-}o#gWY0hA4_CF?7)q7nqx$uo;8JafIQ=I!x`|sJK{xW| znl?RWw07xQ&EtEfH|~W@p#LRAa;Jp=IWd85Zi+pKv16A3<&MtY#UkLh>9Q-x86^m} zqTY0@DPDsfzJdUjm>0e#UKgm|V^mNW9DFlw>6Fj)k1qH2;-RDY%NRibp}~lmXK6&) z1Ubg^gU)#|jHx+={y)#^5oWGZ^Pd6}V4HA?5L#TdX%eGK##jH=4?<@XJ(Vjy%s%(P z+=Pi%;XJFp8vrXma9;7$(g`$+-7~l@r3LsS0zWQ*nBs24euvd~c8tc+fC|YP;PxjX zWyk99u^h2hM|DEm3yMZcy#ULk_&2@Vq#}R*b#STkOuAX%qN!*mC_F0N1=-K=iG&*j zJSg*k_HFwe_G_L7O$!*bH})I#ciRXum#){8b`~l#-pgosxFDoy z{gW;}9%h|xG&)4;n_;g#C$F@Wu4%oBvuh$~hu_u-=jWUEXduy)z?M?RevEM|#<)Q* z$_-Zk(y#U#RV8thS(@wuj3jkEu22l+%jWKCaK4!3f|gk#%f@5QR6@7^&{#U*JaO)f zeR$>4D4bX5WNtCVQXGcLcl!BhqAJe}Z_qfE=5B0lU}1CaG9kz2+%B^3SVZTVo^~CO z8y<|On0>4ff1Q4fdxr_>Q0G*c+Uz3S!V5UvdUkeKTjWqFy_RCKM0-$Uy9ImpeBGREPnSl5@gx*Lo=q3B8aoi$)YQD3na~R^H3gjXo?M08S*^R}Jx{VXVDzi?| z3u9XW|6=kDg}*$sFacy2THoG2S&lUOwv*Esq3jE4@5B+DZxtR1=5O(~7(+U~Kkqxl z2q~MvNM_s6X}>6}%ekhlbz@;Ybl++4aNf<=ywV&y$(oJu_h)`ueA&-){{l=Uj*P4z z7vl9%-0eC-#7B`8_}&Ut5b51g5oH}PXGbnirCQ@e(0P9yM*@{8E-%*k`ejNge)!AW zZ~OLhQq0NXp#NV}_;*p73wTL}|Ff>Twf;yJS3_b+taRc8Skm;V-ks~o&;TQAiHeU$ zXzgCdj)vHksAYdgr(R;*!R~LO+r{#1%U(7ko4~JCTT#LimNZQ|o^b!$12O2b6w>M# zM1|6h;<1dz2$9qZ6xc#0RGuY@q|=(a^N3!+j`)Y&OrlC)pJvBuajn!Dmu{7;G7jAnM;k?1!P@`W zPjSA4eIynN;m0)3%4m;&Blvaz5U&2)KC#|Z-W?& zt6bL6ZMpEy6iz(G;bu&fB(@q-H4Ay6Vo?ET_VUCgNeG=4&3Da>uP1LFAL8A-mP@8N zX({x(whv#uK6Z25b4!Y`d7wl@2Shy{!zYKX~T0# zDQ6p17KNe)jSE#S>sqPZYrPqIuQFaf1{F!TRRw8=gQV<1rez${kU?7YFNd@{X6h@- zQPM~a>>x=t>_4KW3>aFSKaQBiD?nL6qVw*+p5+ZSa4k(QHC^H7Ifa4Sms-zrQ`b{C zQs5xFiXJ2f_1MW`xS@(&Tfi8tAh*cZ<9S$G5QAJL4BGA8DDmQyMLlKNVa@?aMC)D6 zL{S3AcdTTpyyDJ5r1iYDY_9re7+dJSQxV(upTrg@#YOAjGyTjFb`8o~$$^sbmb_~x z%>r9!t0%)~>=M25Ipq*3(hSC)PwofG%Usjctn0+9=S zuGdJt6{|cP+7flRZceq^o_A>2m4Ehf0%&~C#US0-89)weM4~0{o?>|es`7J$O*A9O zl>G+D3kVhdMbiUb@YsT>eu2sWxH!v!s$}a}@Y8Qzls;F{UL9U~c5=`+P{`&)zP1WK zj6lB%eAmeqtA}Xp~F-@ku4q zS;zh$e%Z{j{!bF72%XZce(ox4ezW-aUs}cj%w#v1sPTG9BQ2Pgw!LC1Xo1aVKU|CS zDz+q^Tgii6jJ1=e{>$LIn6~r4WGL8_k7Kg(ry%BKa9mu0fYL4u7P*vRIV1{9DOt7I zv0N*re1?1t#!GnOurJ^KO=KM`YX4)ajwKMV&ev#btJ4Vr#+PObP7f$m6J57f(tOh- zU}$pU%L_-iLVzWz!Hr1>Q$D3ED#sALW!9R3Bm@Vm zwTD%vRS(?a{i}G)+#JT*{Rke~M%E4JI(-Rv^!yLzg@1(N!FeemNnJZKfMo#!I9TGv z%O3k#VbK9^oLrWj*9eS|-W;Rc0 z#U5Y~@fe#*h}efbI$6ZvOU7KKq+eKaox+@}ut62p0pOphnlrnY9Xpk{`;(Grh*}zTB-)-UTQ^7rb9%vrkMLW z&8MGkoEVDWf&IMku)s)ocPIGatZKpBdl*$2BJvq-f)Ql_uiZuNsRN*-^GF0;+!(@=1VX}@MV~IBa(%I`3H1Xo z_J3u_cWf8DLB9k%u015#I^siR)Fe`r#mmlZ-@`Iu5c>x3@fEFxRr45K$iKu1qV^t9 zUudkh;)Q$%`*xaD^7Uk#+Yng)nOKnXPT6KezyLG#o z{wJ<+M=?X$EnwVZhzeT&e$}2{)oH@Vv}Hdh_@Y3X8e=LA5yTL3XaZsZgMRFB##$!D zt88PUq;JpA5(pmx zum{ekx-ul>s0L#F@-f_?n;VP94(XGsJ<^c!VDGfT`4^_5_5c)8LJp6s4@$)eU={re zWm@H_+NmBg$b+}%B_@EK77Tr;n+$Mr)tX+L;NtAH3D4YKfzk+W^So~(BaV`$@J9;8 zMG+Otkp+$Kai1-mCc;d7FS)%C1S2G|EbQ~Z!CuWJu7O=d$BtDE0>)iYhXQl zgKlo9)RMOAD+Oz!o?iK^+#0fyw$<+%H?SBZIqpSJ^WKEBgF8PjY#oXbW}h$$8z7b- z3WHxRRUe9x`iUQ>+m%vZFY9DXbDJdbt$|w?h!SLhjx2B^&|_y{Wze&>i4^UA!i+V= zAenRecjT?oV+AV194Yi|s`aN4U6I=p2h%G&8JsG1re2EO(SEJOL_Pi&-_n0N!)Dp_ z2%lYU&ItTt?kb*f&Xirma(TrW21ulTwGo85ubg9N(vWbMwmfL?ECX3qf!^)|QO=9c z^a|oC&XxYSFg1z9j9HXSzi<*+pQ(5aW9bant<1SNkJLpSR$ALKaUQ%_pZ4ZhLR8DF zqIKxuG)CET4Z-X3k0s7#y#iCJ|N0_3w{x;_o#w1_i+HYpx zQ+C_i<#zdS--h-$(=pW-7EMuBf%DUEWRcf@b$#R;AOWVO6$8cWgnBDjx1b9BY?=CK z&k}U_8Txj@=CrPz09w0qMkMb~yAnOF}45Y!oFxxdDs;b7($$Q2X=C*#2uFtV#*syjVmZAKGi=pX1FnE`x@iO|fQ=b42!km`z>;6Dy=FuG1C3 zaF4`D9%X)@;dDV|kVFFQElK2lnkd;ul$w}mFBC8SL+w8FRmpkWCTO=f!1!m=!;?&Z zudwuS?y*I7DPiH6N8MHS3+7{XJi>Zc2l~Lgthe0@M#dC7merdkKVg5Z?T-J!k|P5b z<8|-<2y7a+eM;dMJO8)a@8XKYX`X(LKrUU8`bZ>dR}8$k=RmX2zP-sM;tod&J3N`q z86?PEOjktYVKgbm--0^wYR!c&TSWU)Y$S|ZUSYS7O;kE*@C_w3l%uYgXba?T6 zr!>2$ukgKv$Xwd+kFMOE=!tLX+fP*#TjUpO3%@mFzBvHD`UkvLQwhjp7~Y^{!k#dj zsOwNXop`>riT#=*0;R4tM*b>DR{9Z3qS|%+)1^w2l92L35jV#W0px#_7d>Dz)YCju z{zJc>9PKWD#}v81&b46ahO;EPnLUuy$#81Wv^K>7(3NIg$WQa`Hn)GG>({>Tq_0r( z+aB(dn(^N`UL~6n^utWCZ2AxWDA$Pt+*^{&1Z}!8)q^EXWqxmY~0up3gHGg`O@Df7Ge=cM>xZd5uSc__E^w!|Qp#QB-ZxzFm~V7OzreY(WGZs^9& znZ(`!mj7$k!M3S+ZRwQ)A^J~+s$>_t`mvuAVIuHu|FUJHyAOlQB7*#jKljn_39I$z z8A`tjij;GX9fEDzfzX79lSsqc?aHI&ttcP*HUNM5Mt{3K1hC?M23Nk3Nr^LizxCu^ zsXdn7*n>1=XdZL!2WBCq{nGz~rEc-?ZQEa-?Udsq8p0%>o?$b?(jBV44h^dGKOCYe zat&R2XsSFwk7J^t^?;X}e z_q7X8L<}er1*Jn!P*g;ucY+N>)ToG3gs7kt#fX3e2nm7|X;FCWlAs{i0Ff#!^oWHh zReB1Flt_XQ2NL49eSYVC&pFq5|M~v%l9|cO>{)B?z1F(deQ$plzQM0++I~H}>S&)C zlZUSgX4lttaf~r{{15IwzU+hmDPBtx;`b%(<>Pope{-XPZzX3J`!dY3O@!X?+K^E7 zi*XmotA12S^x>rFbNKDsI|qW)lD*(D`As+OH^rVO#7kD`iJSwf!`rTj3&R1^3V)e%XY}CKn-S)#aXGFxmS07ghm{9*?e8yfjw2SJ)}2QbC^OJzXpeK zA-P*q)T$iW!lE79iyqg%sj4ciuN(-Tx%fS^Q#gEkxQF+XH(-HIq@=0$TjcOyNv8bLCWJg+gj-x{kLw z+e{|kCHec=O?Oo9FT05vG?XSc29h)+pX^r)P))e6xk4$}gf9Y7MvIFTUb2G_%DtOr zdsHyqLSHMsAotpt{mEylHV&DgCC7#jhyYDv?@AYm`!=-sGSBOrYMf~=gf3dm?Y6(kBpt;d3lz77R|rCu+F zlkJPGu=Y-$=DqOo98`H_+mkWa@%_Ax3N(c|kAd}Oc-KUJfj-*>hSOZWO&fZxtqrR) z$*0dip>xnF?z@X)_nFQx(Wl+PN~>{hUXG~H+=&s=lP{6mv%Ip(8|B{-wPfT z?IR3%)gA#TUgC3ifIR^sU|SgoGMiy!dYw={6opF}vXaQoi3j%X(&q3Xe&;l-Cv;s(C$p%Svf?IOBFM87F}4-;tLy9Iz4$+A9L$;^ zHM`1Nc_8$is<*A{Et)nwUv`J^D$K>jGS8tc7L(z1bmfraxlFGB-7S^J8m}39mG1Ad zE*{R@-y1o+zVXxtLmT24H0Oalg>$D^~d&(GQb$&+ln@FIV64cpFo zzEB5c)|Q?-^*ZRum!CwYe*3v?xz{AqX{s1IrfB;Q-S`aT6ZaDQGT6 zG@n884hdWU3qY2wG8eBosUI-IcNLNa#p$+NIrwBo^YgIGY-9^=-8@ENYCyBMYIDLx zQ%#Gb)ptKuI%dRTwqRJ#Ft2fK*o4tCe_n{_tT?PkA}2QFf6o>CIxZkLG84#a!oT&^ zdWa4Nbi##2d01I^Pji+my7`zVwjkT}XwT2N3$oJ>kCart&E(4EXwcg*@RmXbPXHJj zK7qm}Q;Mj$L?iq2qJ4l!q%6_xWuT8dpV?d-FHz6(7D(5?n`fHr4SUPIqrLfVkM3+I zvoXCpbQ8I!>e4*htDBs?9qgY0rC~j=y_KjDqFI%Vs1&-Q0mc{gwT6XC4acD2hg^%K zdR>A}BXw0z+0o)XMTEO8{w?*W?(Op+g4C`rjf*jCRC*&--*&om@0USsP$8KE{n7*g z>dOW8`UIwNG4jn3xsp$wjMj8lE~?c~WpZt{u?@Dri`5%n@PUg}A;e!+8rWPcK)Ud| z@0^|Jm$bhiJR0#{$n3}5Lo9%1+Gt6a0trBO${B4GO32_t6pHY&wa>gB>bzRLdKG&n zvDZ$jVe#1Bex(n=Y@TtWB|7rNV^y^*UmzObjVBr$!cxk z1g|4T5GxJEkVObm7sytQR^X?3qC-CJK@TdD_*opyJC&Sl<$Y_+3kwNON4vsNa&Yd! zPNm8H!RLA|nQES&3q0!WqM;BJ(w}6z5&|gI3!?+eehUkho{_9cpC3Fr**hyf4&UGy zuKPMMkmb0Tka$0abVs7mX{e^~Ii}z+pX2T|lqdJuX`69Nh=b-xfhbsfuA6=qze3Uk zdd|Lj>NPTD#HR%oZFL8J@24FDg_+75^P&rWU5^FZ+guAYb($3bh&0I4!_&*`+tSdR zg^rgRn;h&*zK~gkBO5#EzVwx(PbYDuPL1ewpvy9p&j9r>X?qPx*NA>;buTfTH5zYk z+oOI=V3;(AmYu9P`Cxa@@4d;vX1(1Qj)z>_S)Fd%sv)WGxUY)BHi~c?%zi=m<<8$m zbcb;QJ$6Bc1St-LFIK0*B(P`r1IgBJoR|%eZ0ig&Tw{D-(LY5eMW(3iLvO0ca`M>l z0M>Rb&3Gs3USFP@$f4f(?93#>Jn+XN90^x$dl;I&ZPLylX;uBJ}Im z!9(dC2o@Sm4x)kB8~lS;lk`^e7AHXB*G+J!e$&LP$FS{7cFmwZ9FATcs21FvJlm|C zbl3UI9DK)_Xfv9szxR{9|}{n_e23mD7ij4=aNsbH%El&-IY?RLdeGJ}iCr zswE5qqfzS;cp*ff&qbx+=R}&;`Y(&ujk*`!{zuWH#y!w=c-4oux(`ljqpc*2znlx; zy}st+1e9cZg+Q8I`3H*Hz>l>-y`$@a_mVkg)a>W5qmX-y-&U7j@hU4_s>OQK{Krc3 z&_GVW_&bZEmJoz`0cxwPGIB*?jmT2$3rll`v5g}}S={3^g=sT%DoO8)=Wi}5v38Rn z`j~ZPOL3A!QE=d5c~|Ww?yBCK*bAWr2P&*DGV|SIG7UEeK$s)+b4+-vas#eELMRVb zxywH=^5rP$DNMz7oD!%8oA%hcyhy&pobKoQG61i#{!0Lc+|-(wt1z)hddt-`!SZ_n z%IE%=N5Pl7-c@mu0U3XVSdtOB(VaJU?=S6X#ibH9MG=UZL|G)0YzY*`h=#!8n(eML z9L*KuhXlD3^%k9*y269i9J>la>OYzocsF}Ymy4P?nc;790RW^id9$iZ2X3cc}Es9EE3hX*T-9y|6mPgZt+`>A6%yC&nM_ zGKGELbOkG^du`z)Zt0E1>;~&KD2N=sM-ROB9sVNPMcfVQf3l*DutlpxlkcBwuQ%-B z6v06!*IV?JyBE$;5^yWSJbJhSWa|9ZqVwh5;VWEH4U`7HT~f0$(-dq$4&eBg{ds|G zvKpxe&`AA({6LenZ|w!y4FCbOg3xiR)Lr<>PP6o$nQUJR>->*34{cEQJd}ttIN^A_|LtwZnc>Vf zoto|ZQlo@ls=lv3FXIik$h};@>gLB44Jr35F~j>@8-{YEs3&6AuHATI?Tfae8|ukz zZG>Em?m@>+{_{<>cMaA^*pm(l=M?Vm$-8uZQ{JeWz3{~DRV(D=mz@r|Cf45B04?ob z%BHiP;S|W{@L!#{%;^Yz(!RHU3HmWaH=GionS1C?)vcgvT=PV&{snfOOXCBj#Ox$)9T7=t+$z2U;6{S8`A1zD1n`$WdXGfN2)lvrM%H+^Zbk9UrrD}UhR5z&~ld`LOqtGCDP z{oaEMtaWZq{@X7vOVG_6CcIlQBi=Z&&9{o++j1~-qqC|Ox;HO9H{>4z)y~@VbaQ3{ z-A+a7-7-L1VZbT*0~ynkqyIo3Yl?mYav5s~l~8*k3dXhA(`I)}(ZxxUD=`5O==|_% zCC`CaA1p4zVPt)42*Xwq;FeZOHh|xW08TxIowb~BfcTXuN!V5YIWK~~^j35P0f5Vu zDU0AVLk=3dpd3E>O?dDKvYaTjlnHvPM+ZVV=C}E!ELsyzoxp5eeC>RwEwBy3?JC5r_{A-@7`scC?IkhIg}c8E7{@D#}o zxJwjN^teEszMLq<*A^ff3=^*2{G>2BV3AS3IjGYKlV*FwH#6Xi7#)cB4UAZ}aC3i# zF=F-~AbZHehwyD&mL6c(ws2hm5&5}X)AvM|Wc|bmu#6Lp?+(;`JgIj5!IQw(gY6ud zW9Qs`OX{MNwF6rURb?H}IRJR`3F%!4-x39x2~%JFh1|zokPMjj;v;f`dIfn&>^bVb z+d21xE819f?*SRQjX!G!3&ho=x&YAe{r5Yp()#c3Z~(%KH`Z_a13kvI*n!;5 zMqv$T6-R9O&zWd}3!Ice6CD2gZ5I}Q=SWG`;z3&KGe8}R^}Zf}fO_ zKtNvWlsZ?4#a1QGV7o!pwxvQXJ)x&Ib-S|p$=(Wqx96!n8i@4MGJF7854S8CDFue2fNsu^h z02S274y)MSF^)TT!=b0V8%)gm2G85J4ozRJ|7Mf14>A;Qq;4H}XKTP&^NOM>)m;5! zK~v3qSE_^C0UJnJ-Fc|$v<#%)RKG34*djr^*U#q7c=Ld8R?TGX?oq!X-`T;lZP+_6 z)@)0^QHF?bl<2zN=4Vu-e_Pm<)vWg3`tk8&FDCq6J-i-R`&|0{w;Q^d&^*5&#cg+S zLaoD#70Zf~%o?CJ>s34L>zJ;eAe;Yd$VMf5{(BHh{vE`>e_8&o%i#XYWg_PPKzqg~ z|J#$c6H5$m8)Z1 zJL8b<&%n3j?*J>y9;3mhDFbBP;D1j#FqWOSP@7h&5%vKMGJ*x@H@~Ov3tqDNuGOO_ zX=^d81R!nSf*;TWDW)4c#79~98|GNfs2st!Z4vJI^gHnz+{&6LEcXh=9p)2LC}*oz zkEG#GSC+fPZ<{%S;8(T*Ha>=UeKYrWD<&DEX1|TpGFon!TpJ05S-_->Xm0fnwnB07alLjIUXh7Rgt$#9($%l4zS&H9; zrkzGO9kaX9Vbu|-s(Cq*tHE6dMTg!>3muF_*^;uIinc)W4eeuL^Lj-wqNpn|&@%se zJ4A<7qqZM_^!r^Vtb<}72{qQzE^&1*hE-FaeRCi;Xuasg*QU}oKdFq9%PbsZe)t94 ziux%h?_l2u11yF#VyP6sdt?i>;Z36bVmJQPbk${KB$MMsQ4(8KrRWZ>uYbxiT(_?n zJ|Kjk0r@Q58ghtc%jxA1`aHS$14wy~X1sbW86xwsfcT7asWPf2+%O$iT6+>??VZ{J zH(%Mi*z*h4nI2ITD81aVm+DhmU{=I>ZdSFS8#+PmPnF2xL=Mt_Nz+C^B96{oeZayT zvzi=CK@RU@`@OCDXa9q9^tKrYa8(6sgCD&)euEb#GAOw<=GeI%lFAk+ixs$%mChnd zroZ4WiQxuv zyiEElredz+w*aXhp!B6fX)e@n#+03G9*!~lbm3@;psu*Zu%9mMfmL4yykyhX19ejz zBRQX+lpxvK%w}sCaOsJQm<&8oHI!?B6>ai^FBCc6tuB1!qtRWK?eK4cC=x(^bsC@Z zr&uR)xn4|&C~I$k8^{&r!}qGgGYUY{-?8QfstQ0FVc2@Ha~~G&pwz?%1c%f4gDsWz zD*U`zv!Yb|dS+@lbDAbM=yT`_LOf?NT_Xx5japYG@i2hCAH~7OfiKB~TBDy3H-0wL zd)P>qk)VHA>?p-upl*cA1hqj_&hXV}9v z^7qJ%)?jmizNWm19889{R5nsj@EvKdY#ive&Gio#xeyPoe3}`9Iod@F z+4yPp(?AS@D19zJG9jlcozgt$ZFsC_ty>s=#x|{v;yVCN3-&FSj*)PAiKpZ>R0=bF@McPt*9dqsy@=L zO3b$o$$E!Q&XFD)Up}Tmrb0i+hfOjZ*35=7pz(Sm8iSgYqaB?BSqrEPThVOv`~=nL z>s4skwTzu6gHlvRJ~F4xhoEaY=E}CL8dUjFKDaRJk5|UDEK#W;pUO zGZ_DhAQEDip4los9UA_HzwnGu%+7!O*nRlpqdSQsLZPg#rsgintCTHLKSx(d25AZs z88C_mF}@Ud^XRmqV2>*tEi(IC+psO0wvi34Z`6a9g+%3F#{rqpYD`gnP} z!%q%HwCtQ3@S%{-nJ)$xzMMLtsGj)76B4gpqSx-qC*%r8mtKVzYMyoh{8RC8iCNwW z&ib@{$h2bZtLfVNKWhhM-`L;1y5%((w)xSoFNN*=h$fW$1u#QY@>T>lC6rF$e@ zIBs=lYrSmN{v;Q;94rKiF8TF00=dQUqici9;POJf);mfk1ilAq>MpoEyi5LT<{Rk<%WQD91p^OLk|ty(HBA07DJSg*~eCEMaSqe-|D zr?%o0fL56^Ts=Cer1(LQ>l zFR$z~B|&^B1|M8}$^+AXsr>%Al+M5%yq_DBr52ErKV`Mf7%T0n{pX@-=4Ov0ualQq zJe&16QdgsX1(;ivoB#I5ZjTf^0~!LFO5?YyJ}^!>8M@o{M>Em^cgX?Nt#?Y^W8}g> zRuey6Cpt#l!FS@Jo)NN63H&Lx>VPD0ufXE^&sX)!CP*8*%0(}+D*~u1`*gbu#5~ud zZ{Ad-x~4PW<|87C8@d7zX6vZjdkhH;>Ww@GfxXZQT^UqhdQu&?( zKeTj`+$wttFWH{Z?F@I3Cbs3E6@cv@2NN6aYXobwiP_)9MNNA=CmyW)QojN2!N(A|r@bg(rFLic01 z7S|2#Bk!JhCO*&w=mCDW#L3Et7OB~6{S{UF^rY`zmQHlQXzn_9=Jtl~sMoJ;jydg) zt#@31GIE%z1n3p0o0oFQtdtYMerlpNu4uc108bktL`KrB|Y_8aqi(Z1jR7k zg+NfdOWR!KCo)xVAKKk1t8?qP)Grj!&jsQW5++^QcD10GeW`dwl0b(|zL6uqTrc-3 zG#@-!R?@?JT-(+9A#gEc*j|v?`~CRGv`(*s9gQw>kUXOS2*b7&C-&jAX2gb^XB+F5 zp4y@~m?WBVQMsM^q22-kGJ$x8qfTBA$BlGe7&L$AT=&rHwW5KLIy3LJOMT$iDW zuV{`4)o}j)-0cWtW^Ya!qE#63wz6&T%SfNmWutAQX(OvkdJ4_%+3E{S7@ox`Vhg3x zpvtZImyGJA$)P*d{;PA+3`+wK}&$a9rMIqUT(LN z?v_(h{64FN#D=6LpLf^wU-MQT^pHynxvPNK%xP^nyARSBOM8^@l}B@Jiw?IhpvKBv z2HV}nMM<9TuZbVh5VES{^6_68J9nurYVBgYcXV<{>NA)y7j-2iz5}b*#IM)HL2TY3zWQMCq=v)_QAKa;-*}AK9+0 z&<*8;HE^ZAvg8nbpKqnb8izi56s!Fs$}iSxxnGn$ z(fkzX)r-1fK5KF_>bVM6|50(wS6(cXn}H0e%D8XWVf1KT;-qn8nY`0kqgWZ%xay*O zq{`Hz2N2X~KX$dDY3a)&RJBP)uBzq_sNG68Eh{rp_fbY|<`tuk$aflJe$ccdf}tM5 z(96iBof7oXPcPR=^V>4rfGCU5u|>Hd?U!UB^uXq~Alp{(w|k2w4JNYfw!tDZIF&t` zLOx52ACX?7i33Pl1BB|qrt0#f)bebnW-SAFP+;THwW*TQ?4vK(T=J~(mB*^F`e=Cj zoY@x)Gt(Y^)llWO(P$^LoGx@zaLg-;`+1M?Xn z+Hb48?R&X9zvgvXniM`;cH#!H!}k=V9Chb0S@V^0Uti@-562B}5`TV=PA)vN5VE=~ z)TXdDa-aRF{WoTdHRNKCy+i;0qG}h~r#JK~iyE5ModU^9YS*?!rTv1JOm42Y4Aq9) zQS_cMHQ#b8Ep;YAjAbo+CpE=TokaQ6=VWYwI!rQ+7JpS1R6=7~mfUaapm1JCC{yEC zri|^Z^H`c<-z|p~4r9c+6;3%Ce)IE;64m7jTM&`$<6nI%ZNnsW&0(4jJ0_yCf2Kd1 zSggNqGBEkTdnv!*sn53DCCag`pBw8-YK=ZpmebF_jk7volbo06Wllnwb_3$%~-ZW9;K~7Zk{Ex zp3~>T)e0nPv}p?Ab{;xlh*$Zg?@tH;IbXQQ~#WqkR;d4-u zWEF_ar4JqPpa*3ia%giQZiD5a_LauiCcf2KXkKp}1Xd(zP%nrpc=Hl9Ksz2V;Lh0v zYI4Oa>%6<$$MP;s@M}LF-?e(WSi|Ax8l#&CEeGTmElY4g8oBW*Nu9Wnp9B_nAny>u z!d<7WIk%OCQ-Yh>*y%6qm_2Pf582-G946qK`bg4mK+|9T$91-z(^?ueojGMjL+azs zD%!~Seqg{?>8w|3r!p(DMW?2O$Ep`hTnfZH;SFcc?X~>+9hK1gY2&h;F51p2s>jzf zCQIvphkpO}p-ax=2MXNIU+MTrT2GYO)c^MUS(5U%RK+{KnW62bP!~Aw9XKyvPEm9` zD1Lbu7;7|4u_Gz}wu=er2#}hak;sOElRTE{BvznVptw__Px>51xlu@z1!&G5oT7NE z92J6HbF+0ATC#kD-U9Y;#7F+`*f~Z`jQP9*_VuLWWZT#+^}(9;>ktdql*P(Al9dm1 zc4LCFkHl+up!qNArEb;kP1RXOQob(}xt?ZR^R|6cI^+znPwjsn?GN;&JBSPi)h{Ri zDh5d!^uJ&150v|?`!7Ye(o2>TdLi`($c+8{>ZAYn2L|SWVLc!HAI#(LiCAC>C;wv! zbiNq0%tb6|Iz9){SOLrF4`jUv@}vKQ(fki42Xd=PW8_9x`j1I7tk?{*471Spm*esO zwax$GDJ6{Azpr%*3^VBL%zf5_;dhHcT|35!1#N396CSX|HL`T8Ew-LiZ0XnmWJ1W6N9|14qM(W1Rh0rt! zT*-kM3V9*M0c`*O_sO88;@@+RAoIjK;CRqlAKqsONTkUeG|x@o(xNor)>aVnn4rv{ zE7tCN(U`a640VCG-R$lygWRiMJ1f?nN;n-_7E2uDG9kh!`)*1(WyN1<`(*KI?rH_z z>0Yr3r%vbS=lPd5HDxL84}|ZDYaI|v;9wUGevMXlTIwXDBUHvEx|%gntjk6Y zNU_ODo;tunHpRljV7wEs+@Wyh-@c#Z-#;UPZK%E!rg0_Pk!?Rme>sa|1~oC9UA)mg zt9b|upD!M;kR*dc;QH1A!4D>G@vg}4v_Dv3Y5m2*bSoqa93)TvL->7-8cS}%WMnUn z|JVYdOTj%y!7=V1D2&D!Ef(*$4u3bUXZ!1~)yegU9QkCn1M!l|6y>_aZpkD^%f`0= zcj^d)0e3X@M?>}&{NsX;&u(_V9_mb4(mBIwU(jlk_Kyy&E^*k1$ATLoFxlaS+Hv z7aC8BuvJN3<=tU}`StEQPS@T@yM$N-$#-c2n1>!p`i1ne+cUBC3X*0mMJhRA1kLvw z(SSdJEYgSxJXkz1Niv5c*)-*;U+{@$_b(gxD-t5}ho_8SA_!utE8b-O*rr>3be;Lt z26P9VF`7aH>(M|WJB$0TIOy4qTd0*o@E(qNI@``(hU4f$AR0U&M15%H?+QjZUK@0( z+s_#)>TPuUAvM;z^biwp-D3l`{}}9R1gbC<}KN#4-%q2hbjneEE;>MfvtHuc&bV&T>Wcpkqp^!`kR zW4GC%&7EJgzWZhBhXvj)xMQVYhTqPT`V?d1+qDqF+--6#9uT3tM)UV#1}xa1%eg=l zFgdD5+6dR|=KI`b7x#~HS)a|YYrIo!R?Vn6v%K1Cx3YiuZ`p$bl(26++?&%r*lbI? zy$&&3FpUz7H{qnncx*pImyrXii|=*FyFb6=BZOGU)5*0RW@{fGzh6iTec>so|E} z%Yfy8^#=s12DS&O2x7mxH(c5F+h8(C210H%}4tuA}T$REi@HR+~k9cO1#Jw zk9WkgW7NI6cN86+*C{ed-7I%^_%0)=qR_sm;u|vS#Sb;Nx|-x8McF|sbD#xfiL8YvMr;q9m;ks| z^Xg54tT)$$zJ#-cz5u?6iSs;ErgIFKA%NO1ZZ2SDo3wg9+T6o?#W`d>Y!-aUCQv)` zg6vPov3B3-)qqO_&#Muchz$gKylBwYv;~!9zZEdZc!_)o`c7YnF&tcy#oY_s%|Pds zs}8U;rwo1#1bO-tFC}sGc_*;t*x;8~(@gh?spx{dqc-mlGZ^Wg?e^9&;(V771jDprt$X;7?cExnZnZZM4|~{msAB zWb-0smCPa0RLfrs5fk2o6)W)Uje;abo%X8(=@k*Yq#Utp4@up9!kU&AcD$R-A0B=M zn#BtB-El#822N9dG-@k;$}emvKVw>wiI~SsyNNE48x6^0KGH>yuIJ!WAOaI*DU z1a9m3uH#9PwPeL0&pU;jnPiKsAhqD>%!$4`YNTvoq>|$rC-s{sXCPR`0?GsQ^UP}i ze&^IiQ6XN2BUTiPEa9XUwTk=v6I_?-a`--2kz;#tDrolMYl8RR0}N-v=&Q+12Ra8Y zR2wD~vFHy^eo;Jux{y*WEH&ck_>u4 zmLKof3(RuxC1iZ*SSNPvslkrPmU{L> z*FeD$;A;+0G9=645W#3GT8d;b;v6iplJsjKy6OJtw3GHhc|NMx^+OFnV>_7|5RfTq z7%@wAp(pw#rgTmn#TqFDIzw=cFcU^_DgFLjMk@wHG90^y9`H*xC9Rw;&%7!qQ>vEG z){m9`pl<2B;qN?$T{~goJ~^PlZ`Zvtaq{Xxjd~%Zr8AT^Yy8UOeYLr&SNt!1;;WsP z2k+V<9FPYuTT0On(mqsL;V=Grw+t~F7y5NvGxRH0mHBP-*|ITz3USs+V>aX~s)k7# zI z@D{^X%qlNnMabH%xlH*=K1~Q@KD}%QiH?HcDT8i6arajDq>*_7``5Qd_KoYm4kxfv zFC)2?8MS!M=nWqeffGd>!7=b^l^6>$%E=s*A|co3-q4hqu=2&>-%%L_&pvuLGb2Lv zKMo!YRD09B&RJvsKZy5ngocS*-O&+^+Hw)XF@XaT%g8^xYZ|lLgf|%)E2U}D_t+LR z30t+RGA{mNI3P;9RP-#zLI)p-!ZakymqE*Rb9Ln&Nl|63EE#>4GYzU!u`9-{xVIjq z&B_y&Up}5T7N$2DfEZ1x)bP!kh|jpEI&LRB7K^mc{zXMfS!&gVbWx)KO)jb$tKxQE zm&Mxn^Wy4)M^dqeof$#U9^KQ~GwTTQxR1A7gBZcWCmyF>u3vMrv0>ZcHAcP_u|N=m zfpAY7i+iu)2RUrok3W6IkpRiW?C-E_unpI~y0}Op9$6-eT4ra(l2iO?L#>nD7^_H) z*MtH5eG)Yek5dVd&)Wtrb%A+t)@V@gIny@&9dyM$+nT!6fk1ZYSn16|R`VJfrxlE^ zST?9kFvy|T8l?6hN1p95{y7xC=?3q1qpk+V!SM`DXRSGe@g(U=K9l*!Xz=Fo1Ou|~ z@uhrQXEse526cVCYxO$LEDDlvzg*Lz*+&JgVe1Ljkjl~?$LkTW}E)b zi14Ge*Bwvi=3HEcFg*2-{pk;e?_72U2p8g2RJLi9%BQ^CAXxz+F!D~Pv>qu=^j*uS zgY>!)p&6Det#La?&S6W<$ax+Wf_8Oa4~9yt+Ki2(PiH_@8)MfWy&v~N!JV5X$b`=H!U8FqZ&rLd=8h>pU0R7gM zzdx=Km0>JvUr$Si&_^)x*^nl}j=wXsdln>)-bqP4w~MA)15mu-Veqa~eZ|YKSWfk2 z^lR>7;9MOLywBXCotE}3b1c8H)%8zeHEu0*?$i1vJ+$pk+WS2_2lCBa*HKTOm$Q9a z*#D2QY!+D?{hC(Vdsn0)X+=xdqC_W&f!#z+kteB}wsw17A7F!fzj$tONNf!s3MkZ7 zLa)UH3_Gg$TiQL^Ro%;?oaQ%|>q=NQ0Z}5y= z`?`U*t?_E+`@z`p3tP39sUDU}v_h(C#5!QkVp!>zSrDXOk{=bkq6~UbuyAGOXCyZd zJ(>Flx*Tr#GbVWgUX0LQb`;3gwnZ`zYCVjiU9y>Pg*0wP#TY%hX&&=6J>1ut-i;~Kf2au$G7C_#I0~T9Ky8X#47(Vc{DEg-N@@2RMp&qFb-u_ zID#9biLNn;oLD)oYQvnZhJXG}XdKO*aFCJp_Z#SUJJ}{X^6Id{krYvaT78Ux}uo zw12fZnsgA_TOlN0YKY|VjbS~OhqatMM-6&s1p&-RfHy{;&^`Ejq;8^j{$ua_19W_Jxz=V;}f zr~dqr@94D?WJE10k0>k9qvTcvj>Ab@oXk*=In}UFK3DPX+ltq{>7TOGCqG$Fo-{wt zWd>+pa{wfSasWP!t3)?qmj{qn+oL4CbS;T0(T?8|ahX7F)I!3kMTF3%#qB2t(!ak9 ztG;@xpySrAO}&rF5uxGww&|vCq-J@I&JNm|(fkXTN_q)aNtZv+4{#Z|nTt$CNrwMGv=%>*t(`*i z!S;I*Y|sa2bTq$;c#>)~5joRW0Ql@&!2kdVOWa)$zQLq!aLxDEaB&65Ng7hWitGOS zrw?L&h(NIrK<9tC6HkK711XHV5cL;Zh5A1(2XGYv@gxJkAf8_y<5czT?_u}<{p;LE zp@O7V-%$w}v*dkP8bb5uLCitM2&m}tos8l(^m@ut?H>r_$iNX0Ly`Vw%Ge^;&`0D} zwg7jDKD%%Y<`z8tuM2T79yRa~+yj-u_ys8b|JUFD>oP8S4m|4unK6$0Ho6Y6cu2I4 z)JHD=onF6=7`*iA-)$$nh0LpaBKe3x%!u__?Y2hzW==(+a!YX06KHX)3-(8q(O=7 zYj~!BxVu_lF)b3Hg}H#(o6H(kY_#~&ABa|*U+vZOoq3c$9#3@9sVf>)oLKbEPG~VU zSqMOAqwfg8=J5F6zW0Bu3lj=|^O&OJZJTRKmZ=w~j%>+rK2sHaUeTbe?_lS?Z--00 zPFgO7X&Iu@oE8Q zitR}}4yP7KK8>spY`($oY9?*znRhWjH+us95Jm4pAy+Dug?HZg)JgJ6_k8yTqDFMh z+DK72OaTouk0mYSQmlBLmoTRjG#O=EG)XoK0(vk3-d+q>A|i#^aF>dKhuOgqZN177 z56G#8_G^C(*uSnkYp~mW-sRns^7|?|ZqS^aqGTn4MIO+?W?)KxH=+i9ga7k$TLCg* z4e=y=gGCz{dn!5>37}_9_V~E=aS!B^s^vCeM|gMyY)V~>wy1I06bF}ILryy5VUCl$gUVK zij;hDY9pg;cRd48;^*vQ$|5txbjy73j^lCjrabcg?(<#qHh$DS%THeIIQt+v8qRvS z)fxH)zHJm;ypB$RnW%w#f zQnIobJit!$uNV;GSp9*j|D7O!&m{lrZ2yNZj?ib_Yu%O%ZI1l6lhPLGz+dvoW5C5{ z|3H+EIRLh0PXo-)5*QlK!=Jz>So;nCYR96d9urnkGImM-V=V0wf(^9 z;|+=e#6*4qqj@xav>eSqr_?Wh`}Pg}VUv~(oKy&>?gofWY?bIx`nOLBLr?Tyrfxl$ z<=$@3zM$K&b|p6e`F@Z6JxMOLE>d+o9~%WBcNc6@RaM*flY*}ac0o6i!~KTWoe zS_3`3pS=3!vz;l$8yv99dAIWX&h9*rmnx&$udou@-Aa(!e))wgWa4-3$`$3Yp}`vx z1*^@mx&i(JzqbZk{G*$(Jo8+D0^ZbSucDl9ipOo;=&-j0=_Z}X9p5~TOHIrG5#<`d zEUYSa>>+83O-e!U!FHgn3#R*Je|O4#E$Pb!+i8BTCWady`S>wiMBR;u%3K@P1v|5a^eF`dc0&E z=)Sn}UN378A9sD)D7YY|h`o0)@u}X!XF+nyhC3#cA!Pe(oNw^k<-`T#a+%~SPFME( zAc@zy0B8Dn^0`!ZdLSO&bdvX2WaQaVO=)5FXK|9^>A|B3o(kJm@j{zs13JZ)@SxU1 zv~mQd|0?o858fb{*!=BaA9hWK>?4R|TK(>ok?FDQlrs^t=Qgjs_c;650CpS3+|;Or z{V`qU+*un==HG4e#HX7U%fi5dj#2jX7?#~p3>5Z7 zC|yyXa=$o}x$0|v#m$dxGCbYkAoDe-sxS(?1x{Xo5V~Nr49h#o4;7XzQAlgZP5ntN6{D4R8&c^gu}+y_DgW-K-dlnx zA@NORroqNv9OY!c%QxZQ?BVjd{QKu3HczF8V()!5QPDa8i5thMYj^HubE!cW5xFSnh$1;?{1k^Qq(D|;vYb%7!p@EXOj_}o%FI-Iqn(XLgso6} z>*9q&(mvkZ&9XcW8AJ5J*+ z?!nnTE^#XyG^{!tY-!CvCgKtGu-&X9D78b46EoeuS)+DhzNowvqi;0?7+3y?T>`!A)VV*m7bRtBfAfeT<^oY*- z=WjD{iv1c=2T)7(0~(e<+#N_>{Uw>LHxZn;7Q_BGO)V{vovdHOFHWv?i?GWE>|&mj zR8xZKu#*16nKd|XkA8;j%fN_>O8Wo?bBADG(XBIp4mfaNMb1D)LIIrQRrSokflzcY!`{gT$#h3pgR^lHD-@-gLO+)NSMXiv$qstO;>|iG@xy3sq zzbnl9_^AsgJP+UHhwiG|xoxuE1g1Uno~Tz2-qt=&U9Ikxxum6jziEvOnN>kDPXc=Y z{&DiaVl)yvO_L|?tKe%m*)0~7XA{E-{E1MXj6unl+@HFs=WyDj0!)wDGMQnSre zgoL(-8p$j^daL^ZWc9AYq;dS&#G`0Y$k$m=Y55$@&A7)c7+Lgx-e@NaN$z0$FzK`s znpkn_uTEYPm)VN@TEJ%Vu5l>?GhF31G6Jsg5DZfiitjc?={32Rh4`~&@a3P7r)W;J zW2HUgJv@WC#{5&>&KDQcd&UYcAkw7xsY0s2zX6XXh6~11>F4o@l8x{_{sr9K1KmUu zKBL963BH4u88N>V2zoFK(kp78tAEOK=3sbsPJ2`D8@#%7rnJg`?M(+;xsYps-cBoL zjSY0zl8Nj~i#0qb?64Fsj?H2ne^~#z86fqTPfM`f2Vj;zwB-1tThScAe3|RsWs70+ zl$vxljl3+Jil|t;O35Ek8k9QaOxMs8cKtE9=?ud`JWV^ijxUB16pzjJ=)_h0?NvAy4~=j-`= zJRcVyzK!;izKC4mF~BU>1Kx6~3#1drEMjPW7Qa_cZ-e#1ga>ciI87|%m)1XyUVOi+ z0#^9=_uSw*Syb)xG2{$H6H>bfZ0LHXyh5MBK5k9->qQA_sS_kEFbQ&mqa zB&!UCHo+tZPWC;4Dh`ckTxDuiuCoh4p`vt zCRpm$?rSEEwc97Q{*j()n5j+saUZXBb?Cb$m_>Bo6$O4O3cRi1D`v87pY?TUF7lwY z%qh{wF*+(`4xfc#@bNlG)17y(U@BGc^)A>M{Aw=&xib|%_VaJ*qaDenf3$NE->qaB zY>JPKrnX4A)3oG6Rd$ldN5*tM6h?B5xNVk??c~9$E^(B;!L$pO$Z8F z&L-`u6#Bgu4&uQ^c26EQ`S5mG?E8Vj7e|r39&%y7A9~0ax?_qac1AuKJ<+Kod0%`GTUuxXMfEe$mtTOb)-78 z@Ja1Pku}~mM?=*2vekGOCY{s%PzX)MTYJBM9&=Yzcr$q_EFw}hJ?UWL2&tY7niCn` z8T2hssmr?`QD9IL!y(mkDCL)OV%yWu$-SQ`VcY7Tgic!ajy`-jp77Kl^oH*{*$MA{ zOlXXV{`IsZJ@9wrC7R-_`DIA$#E}cFKB@E0$>rkD8>#y37{;8`B=$79uS}<{gFrwF z;|_b<$pdtp8?^H5-LPS;)_)gaEXNz#UAN))*CcI4UrT=Xzy5yy{{5tbpSnZ!HLq)G zMOw*BdHT#Iyz`vAbgAi~@XW7m+h^x>uF>x(&HiB0E-K7!pNx;RSd(wgw5|AZ>Ar?^ zl+qUu=58~!HoL+CQ?%}+hr?^YRO; zJZF9>msLiG1y(I?M%QfSKZ%Kn(X`o0cwK%)t?x;6^o?JB4b6nccXxhW{@PdZ?6UB* z;BaC;qLH$^UJ1E-LmMn%hG=K0c;D!Y)vifOPdN;Gx690_Mfud;;jh7)MfF#LeK&58 zKlA2i$LGXXYwTZ?p58Cq89~*qYWAB_Ysl|7PK&O9=}^9BYQ{b+vF$o&Zu`4Hrx&v` zB5U8fJHB*N-b;`ENDJc+MBfHeQ6`xw&OIjOSejg)!yvmpxYb0?LqZ%@~kg@W}G1v2f&g-jpT{5RN2{3^Eq{v;G z`iUmM$td+j7ji6HaUQ=Mz;(fH(3mZ5!R2&hA!}uSDSwe{`O8&KSo-1#jag)yiUF3V|RTu#^`&uy{c<_v%w|X-%@%frg!1`(_-8% zi!Qu?QK*e;nM2NIoDKK-gfk-&_01{vDFffg`$peyi|Ik9lcW-*Ac{K_&$j8D z1S>ooW`6yx^4u++{T02^+Ta`Li{1DTy|8b>A>J1+d0l#6b&cro^`Y;D%R$0f+Eo_p z#Sb4gXsXdP)#MV`xxqEOUTLhL{qZ+{*+MJH>;qNrvS(*^RTag$?!TOJAY4!Dw(aFR zqo!eul3=~P?_70@l>B{@3J$u%q{N20T@=H7qn7@zFz(O~=rWn%5)<+J9_n>AdW47X z_L!Zd9eH^6=~}_$o~)RoN#}%8dM>>PI6*sb`u*u6_xGxw|66)-Oy~LT+W`m!lW+k9@Q=i?kBE9r+i)sr0?sMsm(-H3!t)XJaAqoo zU2p-t1y&y{YiHJK7Viq}UqqecH6sbJS^`=GdYy1|>>`Uunjbi`F3#dBul)FMujz_aY2r27qGW;e$=8Rv`W2Xy{CdUs9FA55zA0L_ab7i8cY@yrEw z0dTNWQ8T8@t0E5%T8Oc#8}rQ6%pF@;irtSS47$`2?e$qAH~9GivP1Uf$8e4OD*+K< zTD_Hp6G@il0&!a(U?4;3U8LXf2*~L|34O8`Z@d8r3TdLOkI5HZYBJsKd&mVGJ4;d0 zKpYhSM2Fx_S@z?e1DR1zzOlNY4SUukw)q6RF4dDA7T)eobIvT2WQ%sssd!HOYS!1f z>Ddzp&stgeGsR^V)(+G*?u!t%mrY7I2A~}@%Z9Q8>qpo5A51t=Mlg=vCD6il+l|R~ zD&SSxuh8J>czEJYpQQ_L`zwDnEwR|q2UF4+@u?Ayouv%a?u>s<>7-CN@Bas2+m9V^L07BtD}ym#Li zFaKD=k5kZAqB}1~12WknFLV8iS@ahx$+C&fKkq+~MeF;HsY*V4FZ9J(?Pwyg)L^Wc ze&oI!EMiO*0L>M z?~RQ0P_+D_@{6jcAwKGN`eF$Q<`;&-e>9K?WnYFJsM8bEQ_h&k;A>osLPWw(4YHNe z9j(G5yWM4C3g`2B*tgam&n6Uk_-ER#@&B86o;2-wKi^}cDIz&(TmGDr=u1Y~(WB=0 zbqKT>a7{~coodv#$KP(7E)*sSD=CpR*}R4i(^$YT^m{3sBzo+!q|Lch$Zu@ywb|en z)xZ|fe6dbu!NL?(%VkT{#s^h42*`}mR~{>I19il z5gdfLo%o&Y;?HM8iASMSdY)|Otm4;ex_KkBRj)%^(>g$UgxtYTy@7?ihf0@w-u`{f z?p4i&p)R0k6-Xlp&DF&n83NQI6fDfbq!!<`^;M8Sk6({8%duBkmBL+s_`zG-S7{W9 z5J%C;C`<-=dmsPcer6qPG}=zz_u+GFxcX%{*(FA%^ZKrD(W(cwFJCm-^E^~VE+di%Ky>(vEFO?Y_(bF5}(rYQzmzLTHC+iveFy&S|SyeB*&&=F8j>PTU7 zML)ZI9^T`eei-}u7`i(KX~Ie;%#RyYHrsms3}U{VBCFTxmP}h4eJ`N|#A3piObRuw zZWm$%Sogfrv1{8pYw=J%pmWV%c)J*k9DD6c@>l=d(g|NH)on15?{9>(b~gC^@P)nC z&m-3#Gmp!L`H+|LTUJgV6CG2Nj4X(QNt(q)FWfDpj+rjZ375TtIXzTe*S z_)b%cU~UZV1Vp7d!?BD`;w~1dM-9yh$vN|Gxc%G{tT?$O{bArUXF5eD&*#HWAN_VJ zAqOQJEKou!wdVh}|E=@Ly}4t0Fey<%>U?2_h^q9S<=dD46biqyQB6OtmF(RA;3O9N~@0V9pUo_I!f=?*!qONYHJx%iT9AVor{PtcrI0~Vfgydg9Cn6>JQls zwB{bn!~z|kO_bz$v3OEV~k-few_NINC%Er*DzK$-z);G-Bp{>2K$WEuOR;y_I zf!?)?le@l-f48!z2|L!mUsJQJ+yOuk4ZW}!w75WRiKIZ>3lbc>=98ty06mf1J-g%U z_uFC&$8#Wl;r`Q$A1zxy3g>vm5{!V*{EYCaowqA&cE1<>X~L7o2*i3td9nxwB?T#7 z*IUo_Wc1?2qEMNhrM!}@b?GZn3BJF$u0!jk#%f0`1J9Q&IAkw5hcXp@;NM;fF?JO{ zQ6AhgEc7>FIN#iF?r}c$1Mjfew=&uenDFDsTmdekfRcRY7zehg$+2Sh&UK?E`#Hhv z_1xc6m$2>(e7B?RK={H9UX0wrqb&d!gk*R_Xb^uY~FyG0fvh@KT-u zAbt*5x6jcq*qY8le}HbvaUHP>8d2S^cR4>hAd#P*gt`JYG51{S z(h}8x{(z_xJa`=7A6(b%_^KJ)@qC0>h*@PG^?-5n1y5Ys=w%a= zgWs}L4~sIsj8pk8duEO(n@VbmOG*DnmTfp{qr8W#oZ9$by34l1u-!IFlDlqxOn2+M zsSz$jNrbBk93MO&xa-F-uf%} zc4pe>EAxx};(L4U7}zULRjgIsyb1dan4P&BW*N*?uK(;M+nmGvN~BJo(~kh(GJVo5 z9RpK|2+b+4llwAHDK!gA91w94B_~$pCcvHkQsxK0G1Tg<#3`oCq&SY&X5eOIJ_u{T ztJ`ak>X;J68jyj(RBzh|tphlukl%*fK4)M@FQ2fcg+zRyC0KWwpQXV8k9jIJGBmt@Gfr$JHdw3!>l0 zf6*l7-{(CG_)w^>L)xA0hd*5HYO4K1L!svNs^QED$~o%ml_)M(j9I6HsrgYf`%Apd z>w(X#S7>JG-saX6_dMq7D^W$dzgtZcYLXB{6oHdNuS-R-PNE@g_vAm0hkQ!rz!A)1 z$8(-T{R~aD(U&FU+=>x`n(X^Hr`noaieA+-2pw?bP$r^n>_52ET(!5|G~tY59vwZTKQfWn=bRnPMu7{WU2lt44Lklz7Ee*0%;lB&fBd4Zkw&V(_v zQe|>?;;|9a5msWFheh_0ao3OSSWfg9yRLoJ-`|zI`?%j!68w#Y7*_>a?ZelT$?z#( zMvoahO`yoz#Rg86>8&@*Ur&Wpzv8)qY>U?p=dhSgs;Ac_&nYzI^SSfCrB^S-6qH_j zezd>j;e7^%wBCLD?OVBnpWkUoS@{YdT3E2|-*)h!MBb|u+bz*#wOEQ~zU_jja21X# z34+Z};GFnR@PU|}V4xNu?&GQO?*d1pT7&lh@N2Z`bWR6YbtxlMTo>MlRu=5{hN6*& za31V!4KF-qW-8N@pUqALsl!=)d0Gud#D`B>+iDc0K`2mv|J~dpv-X%{u&dUEKdv8n zx6ef=xw$D8_|+Cbn$G_p)r(pL3WYTva4f!yZod2fhn@Ivy(Edl}DSC#vi%Rsg#iK;(du;5t4#aKp+5zV&1!fw)y`g#9l$Le`7 zf?&Ci%V1>nX%|s2*4GCv$7?gs)?pdIJ1~ElD3SPOP6TJ9*aI&B*|kPqvi}{Ym2b0n zVz>l{FCu=jmq=3W?kL_Fp8kBb9wT#TpQjhcR5|wZ$ zejQN==gDUNQJ03CL1>UI8>Atf0~-!0EW`C;1*?H)1x1I|;jae8*j_p>93 z`{u^&G`|#mn3Q}NZIrjvn5Fi)@Nb{T+ASeG?J1gI2s{!MncfV2C9I$09r8RnJ^U$T z&k~&%6wB2w&dLMJ;)|yXXyOu@w%LKUB*LDzj&{st1+T@rc5Cw|Yd`Mn5!$$Uw-R@W ze~U!RGpoz>YXBbf<_r_&U9OKO_BdgeEEtIjkEQ_=F z*qge`-}9-ij>w@!aixbhrw3 z?p(Qh&OH2m?Us3qb8H)QvFotNhWqhzpXwVk+*;q6v=XZ;Yh+JwKB#V~`lsL}6?1B# ztO}eWw5wtcYtA_XDHS6Cws#e1gq2e0^x2KmLylicfnT3%Q=<%#0M?@hf6Ml?4bM5m zG!(d`I!R^eUVYaB)@vN*s$lJYVIcqEe&PH$1w?pR6~oWr-lBC$5zh%oXoF!h7P;4U_ERkA+yF1NqEY& zYxCB-;GiyN{AmA`+yF6)Z-ZMSr)exp=kUMleMeI7-Vyr)<7WWGZf~%eLIL+Y&w>9G z=Qz*T?p1bP4(1-hMe(b2qAU{+<&MN6_Mioe6mAzeOCp{h_$*-WY&^DXdp))Lp zqRxPV>oESm{b3HiGcr*Ye5z9@S}}0uxyV4IHz8H|bw+hLtw!LQOKIO%LTs5Q?HI=z z|Gab&nzJ5@?l`zF9j#ou_R|=taA)J(@Se7!U*=P74~h(dXU7*PjeCse4|#*EL9`q% znAQ)f$(*W%+Dq*3`5$%ipH2K+SY;1uXCPmO=ZKERQ6DJLc$E2+kQUukpMO!$zHip+ zWW|<$z_bJ^XncKjJI)!hVa3zQpni)lDeLF(4$z$?acXSm0vt-c_UYW@uhuYTd*Qzc zQIRc?*YwO!gyOe*_|L~O#by%WVcus4q0N5;n_c`DsGz#Z24JmI`hs@_3gFs!0#oKA zz=O+pIcUI@CKoMH#8$<5Qpv|O)wC*c>b&5spX&w1ePFp1IUD~)Q*jwL3jS4mQQZu*7 z*EL>S84_H$4DUuxFc*`TrlmxW9kbXOw_dUBf3Wwlb1^FT@X zPDJt@>=#qc>SWSR1J!-Bg{>D=LQ|9TY90h&R_<%8e7o!-x(=XM5r3#~UROmR*1`41 zoq;M@WefCw>7P^dgmNaDi!^@q0?1w8de;ncWiw_52vW6o{sEwBi+hFrkcamcp`x1|L>-XLorp3SkGf$k@U zDGvrgTPO^XMpS{=0@xB>?oZto-WllY{FJUt0_uHSBqdloQ~yQkg^}8wYl}B!dv%a* z?Kzr02J%tKZ-iFx4Kd55|FSa#;)l2(oCBJl_qOi3?I{@Y**U^9G0AbtyDrdS>69I1 z?<)sGh5fej_?N*&10aKB=a;C7QoZ?-Me7sUqaP{|HKc~(Wdu8va~+xik2?q)uZREA zbapqR{hm5^ml-K z(~}ygvS|3=M7InKzguY5apDbvlO>^-70vD^-_(w5 zNO-Yz?}IyErrhM(WTwt)PP<8;7hVTn9?)$ckcE2UfL>9PrP9oI(N8bsMs2FnHU~w& zJ8h#)xjhvAum1eBbf4BKEd}Y1yol=U-?n`6*aB08EnnE1qHs`YRp>$b)y}J*TU2c$ z_oTqC^ulgl02^|AEX^{IURs>OI~T#~?@ru1WFbw5XBx*@E`bzx!J*MVORty+tKs#Z zm+c+3y+&)UHTS;wB}hKR*7}pntFDkl{|Z z$!!rowg$o^#;>P2UBr*i`rS!mqBBGhd?)+0R&??I@ga zd(xa@vwewRQg~akjb8U2Omn0yxYHI0_OtoG9-si$w)en@W`KxS?Vd!jvC8@*559MB zMKhyHwV4WQ-UC}|dy`6K56RV^{Cv^M;&WhGX#(`q^4R7do@oVqupH7Nd>$?Tm*=;S zt|x*g=yz#7V8JI=Eu1kg98xVz*Xm}bjZYlh<|?LRg^B=uVFasyCk4bp>aXa4cCpZ{ zpND961`_gOdSIWkx;0~5PhS%{z%^b%NJYOwWgxeG^`{C?Lzo`CltoT$T$oeP_$sQtTbSWWEp&7jz9d@*gsoNKS8&TBu$7n>J(4=JXy_PpuPu%TQ^z;g; zo}a%`5z@*WSpV^G>p*)$&Gqqv@AAKEZVB>tpnH4}T?*vMwh6?%S@51_)YkSeZrmJp zP1l_B5^70cg>t;0y{sh{GtmO12H0(mHj8E-}AxSlfVBUC{69>vM< zN3n*3Tbs+=o13gQj6dC;Zhgo&Y2)SsKS1Mi_O~lu$ybCjKm3CIbTU`nfNd^xliuDc zSy*;spr;O^bJ08ggkrYgg5E-=q-{o};ikm>*gKBa)+|9aA}7l6=ny^l`5<%dqW_)* zq+!t^%hMaoWL+XCqM&MaYKHv*@=0qBkxyFBJ1AizP%7tyj~bEEeZNg!I{81CI?(_) zX^0V(FIme^hly->Mzwxlc@_Lt*62CaI!303HTh9W^f*@?4~TTKf|9Aw^Ya?6X-v?zuh05H8jkip z{axqJ46;wlJ-80DTA4C5s+N1T#4T@!fv5}oCwiXDlRAn}hwLx0v`Ebb%gxUdXW|kw zi~G3>ATl*>qpuZ{b(QMdhJr^A*x`<=eJABG^|V7JT<09W8U-0zp4g44zMb|BhA5hQ zi|Wx=2Zp50U8RAR-^Kq zIARtpt5FlX357)$WF5W;xeIm}u)J*fz$`p`ZjwYBORrGIq*?4@H}~1ktGLLYb7VPZ z);;R;ks+Mj5Q)RwB6J-5c{H*> zo-W}Bz=hn}Z_R8(&!ZIPA*?8>NCW9_f4+7z-2qiC;zl ze5Z|$W+rd}M%PVB+!2i0NjwY*bB+RMBFwm~GgM07WnoXJWFZ>2W&L={b0s#hONVDw z-asJ4!D>2GU7b~VPwMBS+Bd~o`P@+XP8axn60HKEj-u6q2+jlZsEJ6y4#zMMynd2h zmwlJs+-<&Epo~c|!eo~?f)+u2!EWR-Ya`dt($m|%(SG>&nxp8o90; zKAXS)pAl#-z9BFhxs#{Ke?Zh<@|=BG9gHR*x>sd+r&uqF0|Kf&ou35zD6wZ>(SUiw z365a@veBW3k4`6eiW!DJ8u!co{js!j?Tygy3wa3=g1m^TfQYz%WRxX zwRYZ#oU?S^g?y2OIz^)00r**10n-n3KVtkxNU6})kPsg^yvhVL)@NKG^N~|74oRbX z%?OVa(fiEJslev$mpblE?I*{uAnoL<_Z);aC_II8913fJr&X?q_tQ_F)lVhI#bhH5 zLh!QqP+)4H9R2B1Fz}c`>W9gF)3aQZi8|bnhT>f9R|fk;*%eSdwiq>G)yz zg`bvkA_1oR%zHMNb78YEm+*L~@BN@?E51_)xg9k~%y+1NcEgXz;ey$&SSJ6NXe0vJ^mn0Bt1LrIg zs(*p+yn-Q0gesR6($QNIyUS%POh9&Kxm?t%LoY21bZ4GUED3h7DpNytHVgO6YPaVr zJsa%4p!tecgJ9evw|MHD9)R`%6Z2xBPu|L?1l;6&OHeEc3Gpp)UbW6VFuls z?pcj_=X^%PANql7KhCK*)c8q+TKCM5Bd@h-kV23*2(c+|D9&hijI8DD9&NXr*F8#hU~BnkvVzm~ z#i3lD^@uvdZ4?8T1|2uWna34I1v7UnV2NMBw|JREdnJui1OE)AK>PY0vWmM~W4bAs za}CYQOQKz+;@zI2>AWN0rn;p}E{GaH(M@_r+Nia=9i5K9^!in_p{_~@-Je(1boDkR zyndIQFf^TYrk5fV0qB71%SR|aAXw`!c=Rm(%XU$&Ar8$ZWb~D4UKqm%0;mTp6?)*n zvtXb+gJ$BhX9)0|RqTp}oOk`aa3k@@Z&%iZ7e$lb6!Po#j-MtZS-jN>7y5;7Qe3uS zWpMUEWbXd%K(^mG{tKcS?j~#F7V?NUr9*wxNibtsJn z@zBpgTHPp_Jb8NP#n+7pxy8e?HN}d`P7OQVBgf~a44c+^RmWn^Zht(WBr2p)EOP35 zP4(J??{^S;wR8^eYs(MT2%z-;NL9+qoNWQ}EnD7EaFb~9gxH8~lrU7@TP}g8y%+^q zbtf6y6-hV_jU-m%xyP>95hzyf}GbLXCco0EIg>h?K@2_W0>u#4(YY4xQc8 z>J=OPuw+GSva9Qu*0~=ib>HXS5!#dng<6EfB6Sb=Ij|*b!KngQLrNG{7CpFbiEBt8DpTmuL)?Fo0?oDQ|i}4t_QM1y=Z)Q6ULs*|4Y9w74#s7;K zX(-Q%_O|zCqyAkLyGLVSA%!9?yGx#a*~v#|6SU;a&f;_35P`ve?@ql1b>JM!iWh5h^m7~As+#Ss0&rm z^W1ypcF>23#@%eiPRe|(emA%vbfd+VqnV=bTa%lba=clbvcU)K!zMSTkljXqetehF zfIoOZ)7+;0$^BjF`TQ$c0eiKsDkb|LL?smMc)5H|{DbI9c@|MF7|jA2p7A<5Cx=1G zR##$IG5@Irn0l9*n8^=aMp|hdF-b!v;LDEV?4*$stK%@ zHMLQt<&iT{vFpdFyWuSG8GON4KoSSqG!=ZO6Y&7_yYqtj0d`;~KBZjTAF_MpP1b2H z+sOuQO-Fw!gIb>XXYpC-p7@+lg?4ISXm4Bd;WMAU-rfc?TOH5fTx4Zrp{}68T*nM! zN6z;Z19xB$ZDP}KjlfQ%)!_~hzBHVQ4-1QreG(KowK(?mv%apf z;|0?d-O`eyFqF2K>ze$y^Gf$o*oPXa1E(x+Z2OvG_cu<)PDAJ;qwaGRb*nY(w3+G= z3w5l8)B~8{0p1^!2Y040iz3*b!#$5Pze9bG)S?olfQ%8b1-T_;R*Th}-b_bqUj%Cq z@8TXy_r4?VCm#>^;Uk||i)Lcl4Ri-y*;Ro8Yln#%^Ng;ot#scdn1Vu?5O1G@^txS&mi+UU>hhqKbNc=m`!Ip70}hh!N9qih7ZX>N%{Qz z?KyBI`tqdX$Y@K2^S;}G-fpW41x}A#A3mef(+}BQ7h2y)!s&y9#^kqv*_bl1Wfd4v zx8roV2XLqPIqeb*_${J26kW4e)H;^i9!WD{zo=w*BhEpX=Z5To>QF_A1k2|~f0l&T z`(bp+#byVxw7%t(Q%DdKlNO0{Y#X}EMXpVeOJ#CPc^20Wj;FRt@$vV z*sC_s+cN~i7G#3uBcwH4qF0Lox9bydKF=jjM8kkH~#*n@PjIDRw$F+8xyB z{^3T&hNqr6MKwYy_al3}*7o7Fd)J*ge_dB(q4$F|rdG7zmgy0nxoiJUbam zveB&4M4*oYB@HhUuxCv$?z_OMZzQCu|x$A%*RGKRDIZsJ<1*Mwb&(4UQEYjzw; zEpeE{x&WW@`7)Tp7w5R}-zM&dTz>>HOu*OA_H%7xEu_~~+|y9l%xhQIAY|`z*%{lF z#4r^4`S(hj7vzlqD5%Tm1=Bir;k2+qZgwLrANyQR2`3M<7ujq)1q`u?68J; znzkv2Jak5MYt=R26oB8HJ6lYH^JD@!0)D08bA>W9ensLQo*F+Br&dcgu&G`|O0&1V zsASvI@0T_{40bSU%6rmT`Rl5Np8K-#mxf-g5Snlg>B`(B<^5d!TsLV_m|P~t51!688pJ79`Cwzt+i8~_ zE^PR)1%^M3V!0Gd;k)r~w6jsZ7^gyXXdu@b zX9C3svdU?SDeV|`GpDqjl~7uP_Jz=_xvoSwdt1w%Pn=u9B|+(y_Gd~MpY{!3vX2ZC z{rt#EbiSMwT;$Kg zBrsc`GYm4(gcXy_Bv2%wtC(HbM<-a%)iv1R)xcc^ZQa6_tO~81`X$&i_WGMbD>9?X z&a`EC^+$SC_Jo&ok-vuUW-qjhdl27w41}wo$Gtd}^>6vjMqXi&oIW-t9e)J_1?P}N z7AdR%oc9yESb*3J@=Ok+y+h(wJE5uuv?SC|4;vY7Dw6Ifst%KHykUKvhR8;+8~|&i zyIh(C9EM>;83=Sz&3P8GG<EeI`$3%r>OZ@u zUX-wxnnFf3Uo#?A+m*-k? z)7K{>IJfy%rQ#Rjf7nq`Amd}-$J)uYy`?qUN&a8l9@x z7#qUZ=yGY#iMtJ_s7SS+3Y$(DG;tN~lGm%B%M0;29Y*cVNy0yXDuVN`8AugOBkBQz zDGPLOLTvQ%U+&=vY(tNydnV4UCdAl;gdl2+PcJIhf`D4o$#O~~eA$LOZfQJ#xeE_hp)_z$|L7=YhAw;d7@Ff-jeK9UwSM5T7Yz{Fjz(t@>hZ@~qBPg=?9` zL83OS<}T3RE--c3-+m!PYZ5hi4iI}Ur|=l%f@4HwY2{KCZ~!EHE#=ZNWFLjDx^*U+~4?6{G=w(i!?Lll(7t8N-)Cm;#`5O z5k2PII;H{?57L41*aq}dv&l~qGVHyNehH3ht3{j(^)^o!9vyu%J2ODEfVuKP-J;z_ z3lWV3eTe#UW-L3Dp4h_50KmINBs=m1v`Dw#^$b{m(x*nCX927Q8o3*vhTPfO^q@Jr z-3?bPxY<{S_;s_m`0m7|kC-cgB5%7NJW(AK{U{8zkUyUx$l*c4>)QWsliK^`$UGv6OhT|P6P0fqwpm)TLfkKn&s9$yJGSt0Q<_hA zGIJ5OTMx%t&}2-mi6UlZ3h(yO+;?qDLcKsts&KGv{Bja)N`GQ##lJb#i~j_SGiXs< zD0?J@8QPDM4E2$R@_=i4GxMzUkDff>6=~a^z3cv2JYm=WIVPI6(1JCp9EoqT8YIGl+^eJBOd}XzE(emCJ zCgYO}yf5YAkHU=oy`|d`CdCFFJ-Z7mbPV=XYv;XFzb7;lq%OgZ$&tNK+%+rD*3R-9 z)VX_Z%zx0(^4vv7MW@)b))3`};L&Dbr?>Ix_EBx^N35ua)6k#IZAA~)+}U#WE64(G zJ4iE=dHp!>hwF2*$5QW8Q+r|fvnXMF@Dc{nhV!Huob%9q7G}O6HGtnj3SP~}E3YGN z;icA586}7uT;LO|%1O3y@iQpn{7~)drv+}qMF}XUuk8n(21m?_$DXc@NRX*%HZV+5 zE#N6W;?#qATajhX5dvmXbLl?2fVcuZ0XdV)ECBadZO2ORy1pbQ}vS{ljxuG>YYl&RfI9|KCI4*lGS$L6p_XS2RR-d#ur^(XflaKZTT!7z#~mo9 zyTolSTW_7)#&_->T726X7#6jnc|!>5;HreqH58?>WnD0n!obb-5%;3qf5C-|<>Evc zo-vdf&DwOdChPwve}G<0jdG0l9x==rI96nC?zjz3=s)zKSo)}(s}TR-o)eZfqk~2Z76D#SQhHx+YzDDQ1LFaZn7{)co{y?#$kiz^)$XSR14~4!|4dH=*B4w zC93cuAYtzACFH*BQD+gHB5N5BC!jLG<+&Qr7`%LK`AREQ87)IGtPQ8s{3SdiTa|`GRW8nn5O?7Wuvgg< zE!0}Y3!_&@crj0Z#_4Ag@G^FaSrtRJ!%^6tr$H+zNs;JA_;NW{1pAxg!GD&B5_EZP zFvVVY16`>MPMAi@QkYX-wKz}%raZ-VaW}^RPq-4HM z9x|qqHxC9+_N1#Iq(hq3rCYo`e>=T7j(_GwZe3Ji8dEexBWq{yol2Tez)Ss%Mm)6t zJEafAh^IQ44XnP8*N0X0oc<~|X|ynL^^qr1##uTxfA+HXsmAk>sWETOV1NBM2&fj! z%E<`A*zy27m9CpfWMR4~xjv$7_nbYZ9fzSrvPZLQNDhW`vkdzAQLE%ZL)zuw3)HyJ z`|{;4!&lFN{NyX(WO={kKly{2q|TFj@shXzu#oj)=x3q`@MNQbK*r-)UABnekYHdm zaet^8PhEmq?vy?J)69GQ`=)VoAnfZ0TfME!-Un&#QMVAc66M)|;L5#;HL)HHTOlB% zd6E1VwFI%N&0D)XAS(U%sQ!ycCB0GhPxh~kwJ%lIT}X$8HH80A=><$R?shOAZY}56 z2}}eNXt-cEm^KklaK3DHps(M7JKFSB|4+GeC~BwEeby$$ogGGOpzFi!YOo#`v3eVx zZ@<>+v0u-5R}ui>k12Oj_%*-P2r)qs>_B1Y_x!%VA1syR2i_6% zyu;jAnslRTI62Ep;>vWz0T61qn=wy zT@*!L7y&Bfo|u_%7&|&0veByd=w+x2Ht_h6p4OGvE7f!3ABmghkL+WME?#MX@jhBr zD4e*c_OQr7)j{}ZW6rzP03D?*pYqoDpZzp;phmSF{O&G|l{=M!c7?>@jWz1Ja{wTi zVj8@8v*GzK&$2A89_~Wp*tKf}74v@no=pZLBMbAaX{5HI#4AU$yZ?3IC@V&xnlo_iIw=Sa30giarJ_W6a_hb90xdN`AfLQ-SMFT1#|9ElHVxXa7Dg8#@ z=c+tUNt{ytqD7kZ_`Ps7hnH7#{<_4bX*JCe4Vs&8{eVrfn-PJCxf_gcbkzh9*Z5LY z5a|MWG9A_m#f^#u4 zqZhxkjTM|-@XrE7^Q-TbCnhR2tB##KI~DJYK}B#T5c|M@0Um@5+Al+`3sI}JVL>5r zLt1X5?We#HR0tGA`8AUinrPC(Jxyba*GMCpdtu<4Unm`byqIj~XTA`_H$NK=5h<9N zxcr7;Cfau;eB?@GUo=&*nbjf^UBuh+-08t1+Ov?yK0L1@t|xyJs{VN<=+f7Lr_Od) zNBK6bSQ0jj)oEjHikkhbliO(P5i%(e(O|OKOv`6d?GspMOQA?cGUQvyy6LxVqUJox zIpIRQy}CF3#P_}1%34{J{S8-}OZ8jzl`p;r>-J#~3Z}|Ys@EWII-(yh)M}^7AwUrz zz^zcUVMXK3>6$hgHZ9m#r@CJ3oAc4n+0;vuK^o(--h|znE8l8g--K}zBhYgd9R3a3 zAjKDqg~VL&pwE%W%2&!Vw=TYYWZmnN#dL$St$=kLCbbnWdX z-`m(~HGdzNB1v!|G&F#K<3y`F``Mp302P9x0}BbKnA6@hDsmvVey2Y@2 z+AM8tl+b<2)f*8sCgUr;VJN8|pR&kDx5cJxqRC;m-J$&MUf(OU?L~O)_)gxFkMnc0 z{)~S{qfo!QrlZ51gEqQc>vI+ZF{pHa>F`jY1t2(A)S79Hr5#?vri1i7Sx5n;xQedz zW>;>b_pR5@vd(K|P^a%5ms$zM=@#|`hn26Q)Cz5V@zK3miOTgwd}DCh<&o5wWtOBT z=fGd^^+>YoNb#j1ZZmeisiv?_x1nuDGZahvDCVJDu^~AfUtjDEk%K3mmMj3@%8Z!Vh1!`-F9pPVB@VY*+P{ z3(vh8SOMZz_2OC5Dt~WjTUo!ij#YJ#z(>!Ccl znr#D9IJXC#>?mb9>msOzkb^!F?K3B>ez4P>T?60n)!r)TKcH)UxaS-0x~3yWAN%|R z?m^~{OY?``9Me-M_#2uhLQzhau(Qt8VF_YHi;z~}fD58kCn0>`9S7$KesVYU3}4}@0qA<@ZDV=K?LLod7R>85_YNQ6vZBtc z6G7^a3q7N>dy61Q#-?#3p{`x?TKci7XH-GH@d&ky101U)yVj4cv27RneiDA(`kuOB zOgEP3qc=M_!nk`YHFgA!LFrMld*7XBJ_qaQ8L zvUnTSV5;D;U{SD!k{kh2bsh{6(TeqW_OXZ8P|% z8JtP|f4dI&xlSPh0H5%ZBT&se6#&=`RdO)FXi2pj6A-Zebyp;-`;t=L4AG9!#7oQ;F06KbXS*U_rrYAX^Uq z$bJ80tACur!ju2OQX9k@C%P6zX(U4BaPbF#oJyWM66|-Fw>pI!xVhdi08q zZ#p+)izAp&j_x!M!=;G8IwWWR zBd-Pv1-b++mLIRB_;XIIeRjlGkOFCLZ*IBVCfL6_?;Y?t>>Tt) z<0QPLNc1?3)Vr1JE(V2A(T_Qbbj#AWEIlaeLzIU*eI`MdE+6BOde(#cESvOXtKk;s z+i#s;Gt%R2Z*yj@O8WCRr~X8ol;j?%BuD$y)nS7P_l=Y95F3&RL-X1$>QOgg)*2Hh zZ#rDZDA|6x%EjCT+!p^!|0R4nTH?NDyUq%_Ze%@-=F#E7>jRu-msru6LEbxv6rdHc zA_v2v{!WhrEa}&SR1@x&n;jOKitcdL^{!` zG20$#`F#EMFi@l9wnR!J2?Hl)qsPQ$M@Q zt<-;gg_*tw`-CnGFOI)|=vo6{57Zdb{}l1?By|s9RD^^(By`f#3qf>ccmn%o9BOF= zdt;u4P%Ji%f!TDhtZJ0FCT5Id7^VdnEty{ieQ`Bs&ggV3g^Xjwc2-8imqc_j?WyVE zOUt8NvKYUDcx3Z6NFJ`9-T9eS;_!$s;nM*ar|3AlrG9JH35@A13EV_qA-^sf3ydq8 zK-kW5SgOt-@$RIRtxxos9-jN*PSOd{urpvBR>12R}Q)0Hj;I9m@q$3#5z?av75#rPu)JPBU$) z28jreA5Li8*u0o^k@l(m<4bcaxrAL1DSggKZ!h#*UQ{lqSutFDX52UTVPY2p2ZT)L z1u(P3AwauMS*tHSsmsuol-Mr6lFyfbl4ika@7}EvauR4wWOULUT6KaF{+MEGjWO^{2whQi1oyop3ES>DK{ z5j(RWCwt)K^rf~rxzHSbHhI5ab=vdjDQg!qFYfju9b7MMP8XFR1>7JwkGyl?e;u`6 zIRuLXQOleIDNISum|RlV79osL!p=o$A!FE0?T3+PM^UmK{)VOLEY%XuS0I{7i2pd)bdTdJ;(mtlK4oC9nM9$>Bq zez0GnmtG zPhVPy{}q+Mfa_G1cw=6(S!W(=lBVV%w zX}et`d`RIL;^kJRH1JmQrJXfsb)%qR&_CX$0VL>Ww+(IiiIxgRfLb&CGumBE1?Pt` zsqS4g{SPKyDo4<4f-?tbk>qPrx;E8BWimdgm5-*}p@)(iZHJ=<5p^S-paks*J)?4l zWmmmTJZzL(3Ng7pK-jUd|pOlw^!W^hSuN+O`LTz?^_fZ3$Jk2I$LSghms%P z9yE%OK686DG%9=0K9wLY9-bZCbmjs~h4)m?W20Z(+f0t&oBZ48|MVtm%;QnVn@o*H zg*)p_y|h;~me9hDZxv=j1$uwsPI_vz9ZQ+YRT|~{-mzT1x4m4d@Mnfk=KBL)KiG0% z={;W-uZY=lEhwly6ZTgmCPLx&g9lOa%+O0kJe?cqxW~S0nXSH6VVUi%nB)A*xYBnb zAq64*?kon(vxaZAkR_hRcnldaPofsx4}m#SFY1{sjr0VxH5`HH6k+~V$K9E%sHFb< zr0xvYu0;u2fmGu;r_Gs|spaHjQZo|WLQ)?I>d)MWmezMf+uMbaZjd?HBM-9E@%J=z z82#))Ljf(A7lmIHVttMA{itEB*hxo1Y z6G+gXp=CJTr|;7`(G&I(L_qwcL^uFX&T4pjIP=;w%lyCjKK~A?fP2RV0Z0*& zsCG|i)eI%*sO|zU8eHUR5xq=&`e(i*&p_6H4O*nk+R{oOJ;l zKim}4&k;Z}3l;XYYd#lfFp4!Iro5&tW_P|BubFtY?({oj#vT%!tX3Zx(DL~ix=da@X$s@sOg0x@|pEqf1UfU zo-r|8-oNwx*}$-wN!UOi%=FQJJXX=i9a6Xq(7Uj#hNY~7uS@ky&9+p+*;OBPbET&s1dToi8 zfrA3yKi$N`UJq=|1v*sP8WS~mPs*l5=OS8&@r;MwGqT>0oiZRi5ytsD&b!Q76ydFD+0|$m%9G+Y_ z+r$WBC$|c}1ξ$?aMY2Tf3=ir_y$)t!d>8)w8J5XB^H?DSz(F1Tk)YBm2GL--%P)d}aKiS8aWoecMYrT=YiOa_gDPV=%On`*AvZ1k=uo8uF+` ztp6xt>CtvKUXu#}h?!_=!Wl$}M-*~Jx%T4DrS;^-v<*O6w3&{OsO+v-HG~Az-HkhB zsx(@)t`Qu;JG1e_;rx+?6u@X8l5ywgyFJ13P-&lsq%lYml8OTTJ2SysEVpW?Ywyo> zy2&*DaD7a=;rwo|Ph;ChoMRJb;FdUsYYDSsi@*)904Y&7YcaKUaUow!BPv_u_jFCH7@bcJxo&^@+=W2Gl=n zx>zgxzQmK(=5!O^RNsilx8mcwl!a4o-%fJnhyAA3O-y!JG*l>gm;@0Pon=tJoim*= zh9ij{5*#I@JiR}$T@|TD&uOADDe#2g(kbJ!lr53lZJ8}^S9J)7Z$!UawEQr=d-Nz> zK>vu>KL(Oqg;cCOM5K5mM~EH*+$TDvHH_kBxk{lhdr(@_)kuq~0B|G?sW*kK#yqdP zeX;4d{b}OG0kMt+8Lz7_gR4E5!v zfvR|B!xHRW{TnBj(3zs-6G(sZ#;NXJQz?3NyzSV8eRAtTs=ep}dOwo|T` zrX!RCBNV%yS$+EUwpLm+@u=MHI(Nh)mUdYAO31OsXG>V_I#)Np81L3+Cfjt2b8oI+{Z<6eF)*>e=N0zF z#7ZeNE(v)j62=Bv!T1m&JOlrcj|Hwxh>scfSs_%!BSmp`G3THkHd?rKuRj&+Ira+9 z5Yc0-zU5&=;nnH0aQ-%ei;K(V!=#PtpMJmo7mSquX1O0Dq|*$2xKkvp`89t0jdu)@ z6pkM!5>9D!=h$R_D7ne)Yt`Hk=V6H!yG^n!J`eiV$}3xEq@7Otbx{Nw_K$09Re3Wd zDMOk;>y?#TN=3 z=icWE{eE&)P~Jv6M@ce_S-HV$Qn&lSjL&}wn*vYn;=cIfGVM|1Y|oG% zO1?1p-SxvD!ZK<722FZ!923o~w#mX_8~yKQeg zUN@+|VW6Wse(Brq`kukIlz(pL!#>RRq0lc9vfx4w8_iY=)eaFL?_+{6Yb7z^A?_aB zQCbfck<_r@*_|6o{8y!g#;APnAE_bu?axc)W3Q$v27J<&-a2W89DVUE%(5sEHrzI; zwPVxVAndredjnsP5E555KM)yfI{!k)-XNQDW#7ZRDqp3#pjpBDu!?6prfp6*qu8C5 z;Y-^}0tH!3vccC#Gp0RZ#0g!lU#4n6I-=95C*r=#`RKMw8z5=gMt`!6?N1J8OE(H2!D=NIRT6)zT24R=tt`l$3K$B9) z{h3DT+LVpse~z>^HrM0k=RDjlPDeWKx;%U>wEG}HD*Hvf7i>?O0gmhVM9wU!Nu-#} zBoREQj?L6Hd%=&FI(NMEYx*qjC%I+g$VpUq4C-cXg zbto4da{#kn#^JDo88`nYZmtLV)F(2_uq7XiJlEXMpq_s$(g)I&9_qusm^sKQNok6Q z{Rsk9eA-9uX?Hb_JLCq|D1Ud{ul_jukb|JKzCd&42`)@e<}aQPR}#)7&w0jmewBQt zf_|&EPl#`~TS?}xEG%liM^y7dZ`a1}w_lrz%WuEaiC;ft{^Gj%q5WKgBxeI6kIe5Y zQ%M|4UL3@G@(jw031NwUZr08q?gtZf+{d4S?P(S8MfMv!WiI>@RyQO`QpnpvW$)8n zWEWg0C*z~TEKb=TZoQw`0aq@v@I2Zt>+^kczqaWI5q@R)??PNdFlF|%{yz^cKURJi zQlM$4GaCOuU|Anb5}Ty}zFn#yNXMf+?IG2sb^kZjX-I9d7T_=IJy;qfLW^@lT}>jf z&-CypM|B0IMaxaBznQpG-JoF#@{>G!2*A)tw2CP6&$oEf|KP*uDDU)ZrdAZ5bKn@b zNw@)g*#b2bW1ngq$vhf`WqhoKYN>dFZ;#@EX=7ltB9jt^tKYG%(0{(7Dg^eb+Sg5J zvXo0qOG~T*Lkt59Kh#*n2^WyxJcO~ngEf{8^hTus5+&9#K0pIYfm4&M3RF@$M`H@9 zS602(W^*ZiMR(BTI**Z;)ZSaiqkFSJ<7HA$Z?V(^3k+p-f3Y87C(4&Dztdj=+g~>P zW4=h?3(Vl!l;ve?!QGI)7zypCT+@&hMLLtCh)7nYA<~f6Wam~J^3aX0_M>Q_!2GHd z)f8j3O6T+5)loMbGRHTMHIFtBFIiOXq}fQ!-zm11!{-+%rUWrhQ6 zY>8dCWo)6Es);(Jm8=Q`%oldu&NNH}gsxN{9^_t_H-1MMSF0->?h|ksKe=?~6JO%y ze(4kC%C)`+TAUuPKUu2j;Daqi7V%tlt}#%*$%QMiw0EDC9WchCir_QDSc<&>XAn=j z1g^P-EJ;W5cBKs!(S@DcrHXcF%>j{FTY5~YpNG7uGAoR}6rB2-V7Fg7gSOkbC(>L2 z(it>D4g_S++6Y&)W~sqY{@^fj-Za%>1%V|S?RQYMCpZIts8t)_R68_mk9|ta76a8r zEmT;nPKy{(0GABKHcHY&P*qi~9gUL`>sub^@?z&jd*SPq2WG!-n|1GfepN-~sMhSEIxJz^&Nw5)_iy z)n}A`o(#VK&L-CF+#*3~V%~k#Nutx&SNOg4gWs2lVz5gvb_-ZiAEFeO8D{JP06dFA zs;Jj0s=Xu;lmb-8$b*(W55728c*xXve#OdeEac>$$DD(-#;lOifdiWlpE))EZtcyy zj)8+eOQ$yaE&VqatIWxjnLTWl`k>5n?wv?RBD*d;IMss&P{DKr?VE=Pn4;!9m#h;1 z2MhK{^x$D}Emd<39M7P<3M986)uxSa#D&qVdZcn4_o`Ekohd}dZw5`YCL3cSvITQbbaq)Eq^CQnsMD*6=l1mv?6&dzD z2e}WD0^H1fT42w$C?eS>gCdwAHvwiE z4jc-k@Z_5pL5sU)K)vc+vG%zsiHCvpuk;8oRBtN2HnCax2F7~Ip5+eMa6eZOG+!~R zP<&VEuIrtCKE6k02VRrDKe72{9&J<8?$t@eNT%m<3#)F5*ZeutRkg~+HbeT;Cktz~ z`mnQ2905W{d>!cx_C*nPBl4p6AMk`X4Cok(T&paNZNRR3 ztehJ&EI%ft7H25)M)f{){oJW+0Y7dKJxpT8LCYKgDqW2-(u+tU$ji;7fLEEDKGGR? zF0zd&80CAFZdXdK46~(s-RU%3w0=#3+?>__h~a*24$L*W)@Na2uo%SB`$F;7gG_0# zydByh=Othjnz;3hs`-j6N>J{dxV?%pfMWmq_YKZCG+69!sPIkC@YrS(CBwGCh!}im zk#{QmMuozKujk9l#WAQ`ULg}@N}i1&^C;7{kbYg5NM6UzFFwzwQG3)$@qqzQ1ofz2 z!CqN`Po8UDK5S$3D%V=3PKOK!H~%|oKv`c_NpXm25M{VV;oIP%nRh;%zV6nP2EbWE z+OL$#lAM#g6_g>VsYg#=Cmz$T%gOf5JlnaTr2148Odya};YJoHVG09fFOO(&%jpO% zkmch&Ls|ln00REIvEf1HeVVWXqZ0&sRbQ-8xsk<8QcscBzejitWuD41dVN^OGiSFQire2a8L-YJJ2>#p`MDu}4bc zEj@RALocFiE){tzp8uon&HY$I_W6ECS;x4tw=HFWbH+kyF)onDWw4d`vW{MTx`q(K zFTF(XE2Ue$YIcl+JJU(YU{iQ@H};+gGqBEUdiO)->$mZ%i&qbPd5l4sWL9pj$LXtw zBxHt)R>gX@!D5p-u>@%fqQebA<$XnCEL)^bPz$?6C0LSlmfUob|9cHme5>sPmDc!9 zg=g!Az4fbF;TwLs*Z^6foQcb=uqbrMOtY%|vsHbi%zhr3Bg(%q2HxCX7mjxE!^rYg z{01W5V0;cs2)RJK2r`HqoUC{#n=sRBdSGT7VD_jz&~aHeeecdxcx*}Q#POF`CYAHQ z{?*v@box|-Z^$2y`oiqVGE=bXE6hMkf}B2d3UH+3>Db=Sg;>rJXz}lksCeMsw{IWa zz9whQIp6BO_fO+^c$&U4*ebAZT~`<`_k4qDxL*C@-mzSxlE6jFdka!&tYw~t^r-Lh z=#doC#sFnY0BHqfEg^)Vlm?O@Y6LE546o>Z@8b8l_qF(}KU@}xuK2~c8kIWf+*_vk zq;w=K^1vC_`A>X5eS!FZ;LAv67X~4(dDB#>*xu^Lj*DkMy6N>!<3(E|tD;OvJrA|> zD48=3X%;Q1M)j`NtlWaW-20}Neqstw!=m2K=tbu0IYrjlM@B@3ogEt5ng8SFL9#c_ zP}$XMaym2bc6;7htd=)L zH`vAP2Z%$({uz~DaY`WWaL<;zJIxfgn-kJ`mt8J|4s1(9*2cZQ)pJdugKxRiRGLGJbm?#T_)b4Z6;=xjM^gn`qYJ^~rNN1jI#?eILb`F`HnS zcvddn)@aQ=MzWjOSogN(nLIbm7#j)-KGyS55>`ld%0#L7S0cKC1z8@`?b0B@Y_sA` zRB#_1P`YAK9>D)wcy>NVJV4Q?c64!dan17M!MtksHT!ARruyXG{r7o+9Fs`~Tu`ST zekpJ9NCWi6)W8{hdhtKw2e;iWNf0Rtw5v0VQtTWQJc=_UTOTYLHQ3*-A{OW&Mlyfp z`KhL6Y&$$dHSt`o=M7FqUa|-joJhOl2mrnYFSj8k6Dqa_Ko+-ZCeb4+Y zF!IZ)@dVcg@_8TFsh!YOX=NzX{;I$iSms;#(R~73KC^;vYIrR=QXW~F-tolTyd))m z2l47Cti;ZY`}b*7>7z+B|H@y{E15+;7`blM$iov_iqN>jJU>iq^Z!Glim*oRdyVJ0pBA$`jo$uSr_^NXS zF`3#pKbQ)=lR95OL;X7(wN-#Qqix0;z4do{M0Oa^#9=Gh;J4$qi~41-8AdAFUNunA z2d(GQ#DeeC>MIY2Lx{@R2K@oBhSD0joV-^=3wE*2-G7CjsWcxsZxGrdo$%-LJ~I`J z=p-7_P%HlLnYL?vmT&<4ros!%5wMANB|>2$4z)>#$sx^xC%jKxy3+weIUP(O73mM` z^iRq!#ZWe=CQPDie7rV`~p2fn-$(gu-J~r{O*L#x zrek@9z`eL3O8j_txr>dO9a-Y;gNtiQFU&Eb+Bw!fIOlf{v#nFI zt@DT*2QS}hyzka@cEC5A;+t`-KAI_`$s$|dx3oCx9r!%AmE<$#zc?-X_(neTx%2I}QBLGBqR=c1 z9)Y|9PMRm2QG__k81Omu5${U;m$Xtyb-rC&PvC+mh1EsLPxzxVmmvrCNA(@S=Tq*R zvTx~J)As@RIgPnRy71fel8}a(B9#fWww?A!1k({|Oh+gE!v;~VS_$xj9MyR!>>5-@ zpCL0+J`7PuqyyUF)Ax2rjee&C!qC2ua6<)=|3(j=H8Y&>s$iF~FEG%kSqTPz{v_|E ziYoFB)j+usw$@?#i8{&^o7G>vKR(WY^@ms(X2yTUN~WJ;Fw&^sx!bwZS;SpqI@qB) zL&~&5mkq%b_6iDO(nZ}xLU8@1sxwN0JKV&wwGO__Fwnv!e|6u3kck@o5z{>Vky6wr zzG;k5!5!)ww$bvVl}yW!M>t?t!zZ~@GYt#ac2Og!2Yy^!@*@9I+YQ5Tc}fX0+ zFwuK4JE}#>Ym?i{u-{z9m^e}93Dh`FkNp(wXP*eK_wD|5Ppj3OHgo$^`!ma_@1JZ1 zd-fDdiv#lYl3B`C?8p64;n;(}a(6##{2^Y!1_wX9pJSAAB=soQ9B0e}*6K?RPQ>$~ z17F$O(GZC!dCr{TQD~WBpM*L<%PIcgX1{ccqXRjN7 zA2~`Q!JGqV<~>fQ3+Yj-C-V%t^9)LMl_LhaxI8JSLV5z@3H0p4Tp+=Tr9L7F#Q+^2 ztvbc#cDjRMN*q1K5?qF%*g(L}!0M$#&%Copn%B8Z(2C_vw@`WKpm2k#NW&uWjWE^ASxzyiLosOut zPoE%@I<}HbPeH|0p9!sXRf6+vtV|<`p}NZC#&T_`)|0 zSia68jL(v&XFa!paQ8N!wWOTqHvx4#yexn-#8eUR0m1!+a|b**~^v zWG;4~HS_OM1hr`AJNNvb^e~bVmt81tlxY&KaBzi$`+k+Lg)>}O=hP<(oC7YU-S(iz z%i|CqtKb5@;he0g|PbOxUQ zuK^y=W&26AKROYoPP6L6r=X64ZxIF^>;wru8}yiNO5%$*NF#PpE1HsVR+sHWo=K_? zGiiUd)OKp7-{rZSBIvy+bHN081#b96ssCs%;QqS~zCsF18V7!3a9&X6tq*Xqe#s{# zHEapbyUs^6d&8C7u+svQNfAORHNeySTuEsN5qXVHnNzdr~ z*GrrC9D$tr`5^3f+QqM)VX|{=b^_!wE_?S^_Iz3Ww|!;d)dnk%F#8RZFiuY^{+x(7 z)x+y#$^2>=TJpRABWwKywyBFxK}+Zc^9KOpXd9hudTy*$fTEVvO6Pm5TYV@hzQ5b0 z_j~f_P-KPQIhR*Cc@@UfwM63O+?vho`LnIr^#NzNy0!8KLf@hfymLG;!F}&)&-Y{v zD)+qy6{j;`Zawj4+%bE&HDWh(EaUZu&D2MP{xemurCY)s?5kfj7ZU*qQSs- z>V6UXa?n_n*}<?DoAHs zbw6G|RWepGu|XO28q*!RPPNIpzID^^ZT5)iJt-eEt{v%Q#L`C(sR)m`SQ|*Zm?a7f z>xfjxLY{i`9NwAzK0Gkaw*5>Xo@+a5Oq#rBWB5wP+;!@S`i_F#kfzs5+k8I_^>s}> zx6546deL{C8h_*Zw9vA2Ni#0RTPi`PYM)$v@J{LPu`e*e7LzzD1N}}D`pN`Or_kI? zLG*E}ZQK$`3bN{xi)gD{*iB$v14~3|j^M!?t47=s4s?0gQe860KKumJh~Xb|By z!LmRzjX5(2USp6`kMzaU!mT5y9{f?hy@o!e#VaTS0J-ThMgw!(73A867_Z;x=8U@$ zOr{$0+M9nLDd7HvAT+sw3!N_sI^LMVR6-5H8+f)h{DD2QX~T&;LJ;i7ok=23SI{Od z(>xN9hiHF1Xo`-i{ZmB*fKc+q-p?F1hMcKzI+-G4H5 zX10z~VlW3@b<1`lzQF`Rs~wM8?eo5m0;fZz%$x;Jj=;_o{CEO8XZHJr=C`6dr}9*u zaYey&F{k%)HHeB=DLbAjddkP3DkkW!;7SX^MjuUXgK&Ys?-iOd0$dlARW4urH$9nV zcBaPjM@6AV!TT>4q(BZ!5W5htC9gqlTrgIzkOyT5P5=v(BJ$2C;$GogVArDH#I}8M zp@{I$mZcHwHRREFgH{;w+TBvEymwIcT;)bM+v2QEtcj^(s(|;5v!oJ?^APZFX3r?q zz4%lQi68NSj?g3e!M??F=bR+?Vs$|4T7dfzsj5`j>$m=y^mI9bbYd%AzQYzMK}j|V{A{GUEySR`7|R2dM? z%RMA+)-9TC?~Cd*Qw1pZB8Z)-3a)3yU7l75yf|42DnT^Y@0*{GhQ)?F&dM$D;yH%l zMh4U2eQ>e=xO*yJN`yr-lt3ZLt$K(>WcVx+j4qLSUEbA2Ru7bQdg*mwgUjtD8LM-* zd$_Ac&v0q=dZc=Yq4bdF=}x`DsqLzh#{P7{ez_h3DV25Y!+?C#UR8-nTS>+3f$!&o z#PXDC8X7ak^?|wH@5v{hgD}wvG^9M-@MZ-rfe1ouyz;y}W?2G&x&?M^xSPRF+=Ust zROZboQ}}?wizjcJLVKPq+V$U)%F6lN9P)!`uKVOF24#M+<)7r889K;pe$r|7!#m5LKw&yL4?w_8n^KhG+buG4z zK%>3y2FVBiusJrg0bmMx?@6y)CH^JpR_fU*;s(ygvyNJM!#K&o=%qz8;n`AQUt#bx>cQd#pkeM*?qReTNui5@h>-&^fN83 z55f1p2+b64K0#G8gZUemO%90hC!7?njaT|s+nR;>Gq0l(_IE?R!^$}Ah=`|U6hF7M zswuARvsm=f`Gey%`+jv_O!z?O_Xw3c|ASotO;n91xEld0 z14?@#U|`08!jutl00F?sg$JYhyWQb4tecFWXz8Tz%3y~^a08JO`&tLoghZ1eV!uk( zYw;7&hvaKuJ0zkgBWxP)6-+RvY^}4aR}sW)q%K{T;b_QuI23?Wp}pc1`ce@5&_}|2 z-HDHM=_e5yMvj<%6+rF2mCUpjHy0DlUqs|R-;m*w(dW6T^i0V}{)7(VWsCCtc?mu< zk-_s&AYH8&j7GUVaFHm{K2v3yefOvEyw*qx8*Vjq)?NtGobYIlwcSVUSyXD!arT?u z+nEeCFPROHjx9+3aW3mE?UI=RkNgptzY`%Fne&Q#*NA@~fu7a1iR#;rGELhI6DAZ+ zaIc8YC6Tcm=v3UfX^Bk|nqV#98af|K?OC z^Fhb#Ve5z!1U9+BPh)O>>N*b$Re)4x;}A>fpFzs)))TZqy?v%fv%2+}|$ zWCw|8!$^m;nWnd)Y>)z`)rE(9AWdmMw$dWh++y9D+i4&DAjEc{Jj?Z}qidgSCQ$nQT~Tx_II}jPlot&rkbLefw%^dEmg) z-tN26L$DJ!n$4_%B%x3#G~MCSh1%A~)-9)M3Kqn;4@YFY@rCUbGp@0d5dm0YSnu9) zJleEsFJ`tl(JF6e@ys?=&8wL3=^8uOb2C{2? zqDr1DrVmC^Xq&I_MtZyosgBc;48#ZpK}@(}tEe{bw-W8bL9y3{HFm4PX-bzVrLmQd>=uG_Rx361rJzZ)I+FCiAE@!G9`ufA}&)J;7 zd>s1p;YLu(W4m8tDJlmojz(YGWUp|4yE#9`TH41&M4lf-9f5*-o%VctO%C_PNN4a= zWkr^)w42sU)o&f__b>CykJwBz`7XE2vXxt8nitx&OdQY3sLv1TEu9@x7Lwsk<(|#_ zYtlSycqdfn3!}3L=BcvLhbWY%7Ii3bd~qGTXRPi~fe)W+Jx&ys4jMW2JHKqr%;4aWX1FM#SOva_ZY^@G6S+}xPRB9R{AAD zvNJbIG^zzCum@<}OTL(wSJL_lrC32lQQTd>*JkWX{j}E>k5As22=AZTu}EF=X9$L! zW;%46>pd?Ckp&T9FX$r)JB3L&JvtU+Y|a1`Ztz|gATu~y?5%U#`PGf?{D$FYXogYZ zEzVfL{mNsf?SH(^SEy3Hy#bSN3LCLo_`EryFWVQry`!uGQ+&Jw%f!G8Iu)tUHlFHB z!hEx!*n;8##T8Cs}_@_WZv>E1=@kFNc z2OVcM7;G9^I-@Vke0yKPkDH;)n3~1YF5@BPXVHG0RvWFCp34dd(*ry3nl><)fzpPG zuL7({w7=0PZ`fXi)vO?*FCUB1-`FzakBg!A@o)9xjYBe9>8miHHe>Gj8vn_HgoJSO z;V_ZVx7sj;9$s%@TyL@W1FpZ&KO+2hz8rYUqxw^|$oNi6Ew%l2U0A2m>5TJ6*POup zii&?{hbH`rJ1wfevwoU?A5+xKlD{9-!;CiqB782cU$ML2B`hx~7*b6O3PkE*c?z-L z(aJNQmT=25eGjzF_#^91S-!Kn({qA3>>%4%dYE@C&pR^WQjxW{#$$llbfUh%Yqs4* zpuiFo5eR-reOuC~FX%dx@6yvKu4bC!>#7xaz5RwN6Jg8f7WQow~v&cO&Gc=8cJ7XN&8YtJIf;M_k)Y7DN4bGGmMjZY89;2tCi-m*ucEV z8O5omcxCI(G$15ccGJc--0^juMpjW1>L@@qeO2nk)K=VPv;x z8s`-IgQ?LkklEDzcGXk|@~8}4l!rVkAoKpk!($C8xL*9CX$#(qDH!F z);=?m=tkt=UIh_R8I}mdUV8bm&8EXcu2s40{h)7!zg$*iT@8Kz8RHZRTExv-v8g2B zqwOJry=-bU%v49=);WjgPpMD8JP^2%6>n8~&+FhU*DN=p3{fttTJ*Pt0iM)vbmu`< zWCV~j1v9{p75~)3dDi0=Y8bIE2ql9ws+|;60k|ub-hV3-F61w@cI1!N-LY+~sL6D- zcW~HoVqn?xE?k&QhYn)t1-jlh&5QjWm^?j*fmGGLAUhHT{}|zEPspWQh6;O3k2lp$ zA*1QaJ#xk_9!a`i+?iM7P0lBlYvzY5aN9ppZoqZhxVm1<=8ZNpb_EiOq!N}U`#EB% z3|-Hcj*KgD1#M^^+{St!Yk;ro^K#0DhOsvlA^R{rUd?`*$9XGIP#-&VAk2^?JRYy*s~3rF`Jo&H{&N zk}neh+mP|!m6#jDrpB-ZHMA30%!!p1t@v1cmFwz-!l@5M^TaMR0%LtwcP;^afb5ji zXXH-y7k%a0(O@-}Uz@pgIo{&a$&q%k8jw7}-tS@$9Bc)Qp9I<_88z+SF00A_IRhT^ zM{SDwrNB|o*lax_+jowx37n4rlZPv zL>b1Oxe4xr8*Tu#s~Hr}t)`~ZuoD&E1?m~*Q+hv#NaJNWw)*>myOkz;h*2}S{h(le zdTs|s{%Uf!QcTnCyFZ=EP#0wHR0pX4IrLcQ*H_lJ!^FJIT`k!iSaceC?yRb;ihFt^Cj13D5ey z*84xH^ZZ5ulaCM_+d`~&1(z>A?g2Ucnb?9~#4BpPV+}m;QNp%VIxps?d z^vEySDbj`^ZwXN$5ii#EbRImF^{HC)hWO_&U`Pmh`lIIu|Cz7iR_}N|OKJ{eoe6sJ zq^v3(N%~02pXgWry9-2F563IOd;>(BlLmuXp0?$ z76~%C-w*L-1_TaPoTDCCHZb@dv0=uujGUBdkvApwO}WjV=b0WgSPnDxkd7dfTb|2> zv||VC?-{fV@-n&v27;8K1HFqfFxB*7crs2}H25XpgB3xtVLd%d(Xm}-iwAg(A3#Wj z;QydWsmfG@UASij#S1d&ijcK8&+1`YQ5DinG-DuQY5=sDoRNv&7d=;`Brgf^%)4I_ zEw50Hhz_`T=HksOPSLx~$Y*UN(r4ZDciIC|Z5-cSYWmw#3F|MTF07Xq@GiKu#$8Hj zIuF^|^^^dC?vB@;)^3=h7 z5?ACl`XLP{t9HOlQtnU7`PZg5lv8$46fT{Q%E<1K;|XphbKn(jc;MvA4Ku-X?R6qH zrHC(Ma5AlZS_gWVW=Wm;3q(nqL6P=4kf?#e@qvg44@r)h>C4c3aFQN34jf{vO|A?& zkfOvqKeY~i8aQJZdB&bEAjrf_qC6rp2w7p;k}-HM;vi(e^^|9QUqBAZbYZ%JcJW&E zEF*$&&rOWtx-X;~bc;cpZ(5fgDew~1mG<;)@k*TW5VzLq*j>1+*Y`y~pErN^E*>)o z+Y%6Dy)!YUW8>rDm-J8he47Igwi&$#Q00KM@od10CtTQyy0X4lgqzXsmgx@QYf<>K ztaXXHQ!*^|l7CaC#gXqfbG0(tk114IOP5F|@-Wfh6$2}~(pnj%!aSt&&Ze|@on@XC>{+i6~iu6DVc z;OyJy``1f?qTXjpB(F{t7N0-6$a^G8W+^MrPGHqAYv~8i%hZPffo9NwZ4r~mc)k#& z=vJ9eV3bd=7+RE<;SU@{=(l#`JED~I^^F4%^7Z=NGU36Ke0gnA5{RNPk29QX5cbN&_ z4Y|6U(We>F^ z0QJ}w@$d^nqPqfC(r7{L!!&Qo-v-m+y_gGZ5Bk!kq(Y9?xqCVlNUx8No+0+OBvKQh zfAUN%P8?SgN5W}NsIm`khsNL}VA&D82fQ+4ovtpH&IhV%SHsw9Dr?#IQ z@3=M8r0f362PgOW)gh&G^+_~`DBMfTyKXa12~N@%poW}nAnO2*O7*tCmh637SB|UxaImf`8k~Y zk~?Si#J0rg#&=hrjj7g>eox&GS{(v$=aD7l#QA5g5Z;HTafy+}7F}e%sDmphMt-JS zF5^Ux!8Gv)cwxM^$&OjR&kdhOjWr)s6XS4lM_%pm74367-0@Dx)XEl`uC?4g^W+d5C^-hv1S5uen_yLTH+*)Ty!0b(yQrFd;np`#r?^HF2W+ zUWjez<#GQci+4*`)|W?@V+Y3m6g+wx=;@(e@F1~;7sZU`0&i1 zwA2LP=4N0rIG;|ankIYv(s4DaB~g*t@e*_>XU_OmOw$r=BLj++|+n1p+|be z&0aStdQ50{G^u~6%lppxnR~+|iDK}uTJiwzdMwkJli94rodLh;J_z@2-aGcI_G5KI zh{j{07~4H8B7%&Lo#k^%dcMYXr`o!3IpWY5MI+h(;6=)7OEAICBlmU- zdYEwu6}(oK&1VZW=F{4?E18IGmxUJ}m^2 zTp$GlPX4>0@-2WbAd7LbYaBN=r&JtpCMo9$TKEf8kCf30oxWJ!$g!_}tg7|HDjyXD zs>efRI%|S#ET<+f@=9L^4oS~h6=dD@o^+8Pp?sWu$7#LDt_I&C-*VHPW6%4`I9C9oJtFR_5K3=yZkH?*EHE zRVB{n06m88e3Gje_7ag(h`+1x=z%!tzN2~;zRV^#xCrhV8N4!~EAaRCg5Co?2b03k zdY0K>W{-3MbR7h8&p$@hbrk@@k|Kc_11o*N>cPl2m_b_xy31Z_d3!Fp79g2XNUYK# zoYhf(;z6!|lKLNYzfCEbgMNF(4rl3=2x`AC;p5#XV$xwYu~veHXMmXOM>S`JB)SM9 zwf^vFHNK$JGVw83({$Li^+7;GET?uX#V+WQAzm(#Ge;(D24F#aNz>^=mAWZ zCF|POsu{#@7lu&yOHISUCP8Nmqz^(4YlGzJ=D-2d=dG>!#vc*Y+(8r%Hj-t}_M{_P zT;l{^gVt*pL&M3Lx81XYv{@flqrYZb+qt6m`om0n;LevS{@!{5_0LHsGYao^?BH43 zu3+zTVt%jXyyePqd$u)N(j+h-A+cRxKC5$93XbqbB>qi7+Y;jhQ?&krxT4DCYC0u4 zv@0XbS=PC6fAtzdjfHeZ$`R|QEzrMEN`2e^GHckuAJ{s;T7nkC>NA*848Q|-=p9@{1o{VihDLfC-#ciO&YFM7`?rrPP!IM;H%{_m_F+DG@=m=!>-^*S)nWf*Pp}OhVx`Ow1B) z?^1(OT*W3~ueXS|s$g_ zk`9eWdGep=-HCr|$NV<09sZJH%!)`sJ5<%h&0QYdR1vsaYjWaZs*vgIpRb#Grx#lR zglysY;Ef=I1&UTxeMe)ryMH4pj0{+l}1Ls9n7 zGVGRkw*KCBuX}aaA@L|>YwEwiUTVHvy0GQFG3ZvF(dObxMn*b@sv!IYP**=2 zFuTGI$Zzr*i(=Sa0nl?vTAo4oy?^rS1A#9g!y#Tpsi+(22fn_}*1aPt6?XlEwPMEw zbN+F>>6KJ))==XOfjAF2HQzGvt;dM-sgKSZS`q5Cx}$ZRUYV-G!ly<0j4yPIsD^1O zycVY#oRJp;Dav^~dnB*jyP~OMTiJ5|y1Zvl=KKBE&#b>YeY4TQD}U#J!InuE60m`T zkTTFbbg0|ppF(oz0DsuR!hNbq;LBS)VcYaw(jI7stX}DGBra-~=2*sL>T}k2_8Gi+ zMvPTH7^a7Gg-NdbRf0Ij`4N2QA*3k>5oEA7fGaQ#*@Sq4N9TUyTsYQ`IM5B@yc4X^vHCQ&gr^-RLIa zfcM2uAHTpq8R}bG&Ajk_az9Pfw!}C0q244;o#`JwcZ=ZRY0h;Xa^nvE9!23=6dFVhhEdt z@jLhP?)TjLzVYJ^;rZi$e*m87ImQIoZ`lB>0z!wWS6(M+B;wM_;=kyHTQE2j?CcTT+mMm>l~T z)(li!NdMLL25wH!hn#~C>fg8yzFV%-K0|g2dAs?6UYfq=eawY)axt$&+Ks4);L*`c z1Xg$#ugPiX-ZX$wOyH!(jb*$kwhyM;!vB=M^KIyka~t>f7}8Y>woIf5sa*duJQ7$m z`2L!Ci1CA*DyJHTLb6zD%Yj4kbB(ubH>!UYB*YI~_<2q%ozc^iqHyoSqb@c-6vcNd zrT4+^fv=3`XG`sRIwWKRXO0W~P~k~-Q{EA!80u}MrI}uBoH`=9e)f-)@7HfOy$ji# zua!q-$K|l4QT$n-pv8d8x6XG|J?dG#7H3NP9Nl}dB3o|HRDnun@o!@k@rWQ z@M;uqsuh*an`g*7YV2teWWWzTK%QHW<&W&`mZ|)$4Oqgu3#rQN-e0Nv5x&OJ`Br^+ z!4L6)uM#*Dl74-D8@EpPL4p&xSIUZuo_iF{xpY~5SA10V^Xtb!qk^P!&rzn8vr)o! z(n5v24+dNEOvcK0*1ejq5TCPZeV_=cCfg8sak)LJjwee!>$3-fTX+rfk!MHHx7SFX zuJ@eY8iSy0i@=$(knt)M!n>pJwLd|;8=QN`wsQ+ ze%VUu<2$E&yXB#bE@j>#(CI?qrn9faY;nk^+0)AQuw+FcJEFyS$ zqxg41Xn7?AY9-rnArCYEGQ!g!pI_7HTHN}WrE1l=p?{^jPfKU{-`|*5y#CbvYjygq zJ$oCWNcZ`tHkOXby_0TpLcCdJyc{cc^a zR1Xh*_Gy97tlY%vUm@PJT@GL0y0D#boL@vG+f ziHZ3HA=8WPng&3#@hdHxq`p#fw@fl`9YNJ!&rN) zbS~(30dQIJX&IK=UA($S7?L~^r3?g}e0`kUnj$=H4OT_2O7I2VveJ=Jp=u=j1g@So zcRDw_D)^;_!lp2m8NTx63Zm$T_IYS4#es=GGI>_+-=(_6~|$993$SBPtC);}HPmRYY7MzzeaZn?bP`Y&;T5_OEY3L5Wx zck7;<0xSDs8p&?ez4`wtra2f)GbV@F^@))j;PKOe(f0eq|Y z+YEVdLS%9(U(dE;zY}k%iJg|kr`k3vsnvw&^9=zY>A)QQOo`-4ur#gF-lc&GUlZ%` zmG`T6Jy~4LwsGCxVO+c1X?^$pMGPb%&v}BMPpj9i0z#IZPH?09fvbUmCAfB+GE3qs z|KTt<#;4~$WRXu0E`gnpD##XzXNx}gBbEd9;FwOAYrV%(GJNMgu5}9DpUdRGRmbQ( zEFT?DPEgf+B*yavWXv5YTGeG-?ZWMcBLS~+_a;5-1|uzHlcv%xyMI`>;(I9NPr{K3 zirT%fyEV?s-Up>C%bz2m-3Nnk-(O&q666or+?zChS?>+2=H!x+9 zjOuWez#2C)@Ns1oe)sJ~ZOP6!mQB|6;Z^@xn-vv<4$XFj@Q0b`=llWqT#iClHS^|^fZfO2A|$- z zao0XVFJSs@T2za4$QD}!oZWWs<4exH3SQ=zvK%SPSOAKRc2U>}Ze739O_Az7z7(^A z_i3BPjwpjA@NP^X(l1%{iW92PsUcAUJb%@ETMsVx7$Q!MHl1(xfObvmyf-ILNe>|m zioxZP0{^1e@RHq@Yo%qkryf;(2zHcuTzFpi_G6OkjU4Fy<-1wk{80uGgRml!EtX@K zY_hiM#4?wQY&kFif~@o*sSDuy{GYmCq?@V2$=4%1 z51cppsP~bo_^!vu>PJyJrZrzUuaCQ{7d%4N#EdH%Qv~Wu-(&u%)=eioZ|}Kr`&3G3 z$st8p^=(!;&qW^SUF?h4EtQwoqxf_)?^#~HzciFIW!l=IyY&d(&uCR zD(Q29Q96gfeO17>+T3!>afKw#D3z|b?&XSk2X?2HYR$sS5>YSRME_XGUPZ`(s)f5N z5PSrZnU{hWDX^HuK%`@@8ko!}s@k)JBiVD>Ny#Mg(5Ck?fZ^1P9D51+ASYos=EHJd| z0{7~UD?e^RRuk@}UNd%>AASu7x1%LZ72x*s^sel>ZQ^5a+`l|7ZA;ldw2D;%?F3(Q zw~$}ivjdu{$!11qSe2bbPD^&JBCJl0ngH3H3-QfiKGkDU8j(qHHxR!XQ{!?Sqr*ZXytG=Akj*lPhRe4{> za(9QYjr~BeJbR6cgXx&{g5*{|+{4HI=0Q#gk<97?Vltiv^aWwPMu`wRBoZ zlXIN4dz_h|ivm~~jd1LItSh{aD^N?fA){7r)938#i6c4{QqvfzcEEY?(bzxr!^y@| z{R^cck*k$-ZSq5tO>qxyct0^-wM=n%!837$tY9wm5dr29RckaPOAV`7%c%Sx}ff8hCIc;(>{6Dl>VTIB82P*J(R(R#uD_{GZQ-fHg?qqzqz1ln|O1QytYym#ulR~YJs92$g=`O{8N*Fd&|#TqZxk5iAn;~xiH+di%-IMZtS)i!67 zrV~2{lZ@5oX;*%AcTDD~(`SfDW$K{diYEc80cziFzXRTC1%eWdhv?e{VI}bBX?P=&>Qr`du~Sx$ z-W{iQkL?-w&L`L8=NE2Tv04*J{b|7uC!OMSTX6_|gy`!Ug{EKn50Zv??Ks8gdqg4b zNCP2;ty%0f%Rd4X;^%yG4u&RC;ydc??*4T}z#3iQFHN-?MmMGYxh(Co8hTT6LWQ^M z`Y#L>VxnQ?x?G@cl~{evLU3D1$rrw}?dXI@A#~2o3vE^HKIQ$>PC_1m@Ju3p^=l_EsSA3gu@IBIi_#p~vzw^?S z2LHGk6I$uH`b_5@T&LJqGgDnoakKI@01qrMHNZL#=dafcKK>8dZyZW>J~qp@_mrwG zL2TtB7deny$Vc{N( zoxEslicS8*^N-3>@$P1Nf(O|NH%+WXJ)V!Z{U0MJ@(r#N&C^Uewa=F%t!)8bzG8-M~gkV|@^Ue^_-DU+Tj+KQ8zN}lA7ZmPu3Ofi@i?Ys z+scKndq z;hvMtI*f~6_4Pu5Ma%g8M`DB|Y%+hIo_L+#jeKnJ{IIH->HjW|Bd#kC!D-w?lhWp6 zm-hOgQ3{;ALdMnA;rGs%0*4_Tdm(t`6k{uX)UkO&&9CsFOPM<=Uf0BN*~a$X=!K7I zwrT-&XUwYGcs5SZKGRE?4`39dVT&xWy2#M(KuPEj;}1Gl!~&|)*)iJRLOLLzt9xxZ zc<1SUhJ#d_CYtFXPiMZ&61)S;oPG8Vh5`td?U&0Q5V-%}sbuN@AXKv^&`e)-7XNgV|6 z+L@v{(SZbowv(R5>}B+EdkH+o_KeV0V%^o0=0os_b5)FrcIRO7o;7!qqV`u(cGaV+ z#;)j;&F}8?wYY=6obVCK?62v|AIC$CUhn>d=pjhrcyV2G^mByiMe@6Xi1+X)vkz7Y zquaC)YS693J-_R(Iv&gG=zTuXuM0!!%+<{t+Bv%ogwcXfAP6PHf-J*rCS3|uilrh(p ztD9E|D9;aM73n*Z8?vwlFk=2w4SqN5**>(aQsm?7?qBAcG2=2b7$}F>$YLZ>h+Xzc z%_=oxAY+Ew?P9`+kFAkGlWA|aIwEGyvN|2rqsb3Gr*^lWD7tc?#PO9Z<=4N1UG(Q= zcXGX^79l9Y6&z!xQOw3#SeFbMqGUps1`2bPthm*2G$h5S#Z_42UXykt1$h9bo*ycc zUTLU!NUjBF#??cgMm}ih&=A*}z6{r}J{1xZ`JQA$s3J7oagBtaShl-P7csr)rl6nI z0WiUqt9$A4%pm&P?AH{TldSWfM!6clCsY_xT?d;#sNb`2^|`u5ijUW!U8H!6Q!iAH z`lLW1I%pw70TygJ!-;4jVC+0eq> z)Bo+1+DuQ9P36eyUmDuJCK)hi$|m$LA5K<=7%dT^_T;p|?fcTw>&m=NX5*V{!Yd^7 zW2{Q)a)U&oj`Nw3{~)so9ksJtHWPnocBIR>0~=?XMi=7NXSW5$WoILY5p>H@0WLldPRBp^>TIf#ZQ^m_ijHY zU48oeRg$RHPUF1D0oa+$ ziNE&382@X4Yrwih-+Czg9((llG78zOJhNmfWk?&Z{Zq>UIUldTlt1N^g?W016mt@l zao+>|_SDOIbR|Q_@njper=$?wm%2|=p%T5j`S3!fH7L+J4F-bEqygzyn}Xl{quYZ* z+ykezk&-Jv3}3?#%*Tz5immx6Z@gA(P4qe!;sb7r(p0VFUVU%j7g%Cj8qfyOIc^cs zk%m#yd1mq}gdm-*hm&U8w+@Xpf*qU=J{g(1=-0+8jFA6sT#9)wn#WiTrpu6vB}^Y{eW zePee<*ba17XPyj$7+0a@a;wocxUFG4{;{fj{qZW}n_)G>(xdziqNe7+pYfN#6+AjW zCv*)*$}Q?7@&XL77Sgrk9Hg74g)S2vtIUCVs~HLmoN3nA3|1wQ!~2{ zzZ3x^>1S@q8Oc%hGQqKFHx@tsF<$_tn2kx|s3dii;1ZE6VmP$p)<(B9BgW|OApWA} z;JJ{N9ceDa3q5jylO_hM_9nKF5}DAF#~&H3WRe#6GC2{t@B_GMP-ek@2jyj8Frs^= zfGT?Z*MQ19SuDNr7ut2%>ZiHXSrK#AGnIPThth5_np5}XcAMRqZC4d5R=6FLDd4gw z6=$b!ENjIFX&6X*&ZVvEuS_RHTFN{~^X>&Cw2t0GKVx)XdI!&)fq{(+O;sO9Y4>Uc zr~i>h|E|~?Y8Y-)xK2FPJJ{GwFcF{% z7SJI?5q6<{G+^ag@F825s*!>cY1Y49kyXJ4E)9K~`I{6kRsuHeZ~~K63o(8Xei* zpq{bMnpJ$^5bPQm>FY0Y$Sse0`Ab;N?veD4bYbwswtAuGix^#pTuB>U%>Ki{CiSs6 zp=(w__9BCRbdo_v>!POC$K_pJUc9+u==pLq-J#!Bx|bD#8ws#HtYsX2BJ~!XeOs7v z5-H>zemMW%n&*JK$jDY#kbW zh*Vcsr{^Mv4A%L+2az4%3PH-egKTF;JTt$#G3Oj(s}q%W@ed~%~G-Yew*iNnB0PjTRKlcv6tbmoSne@`DS@L zGKH>oJEZtGDfNYKc>bQ+g4yZL{a;JuHF&KG0($?&nV~NO>5h!21#H0*2A9#n)X4mU z_Ywnn5x!23WnQCSjq@CLtI1rnf>~5w*4#l~-?zGK6iI7|6dF^e24Li&dT0J@^=0gA zUxV0Bl=cK&%NWa)Ox-1U^TV?Jji(O=(VCDb6tyYITRuGZ>9??*07N&}U$%whMPA(Q zOG!rdJR+!{yK(cV=C}NXVqVD~%DmeNAUyvj2EmjEi)uex6qvjKesTisJC^{iIo|7) z#C%BK4aeo&?zSY<-oYX-l&I;kvZtlhze#_dlQFAJM)|(Mmi~CZ5Xzm@wh+;oM2HT~ zO7^WQ_+&i5DZ-#%OC!7oEVVtis_0i&@c_)-8UILwZ7W>Ku%{?>I-j5*65~@<|Ezj^ zk_ga<)#-~UbLiM|OmuzW`@-++jC(?`LcXjDc$bd)qQ0Y<_c$NG$Ub)>!J`Ia;R(GJ6N2UDIBXabCrQl&GlUCfDw586}9julZ&JU6!)Q z%lTdt+&tqD(&BaVs`qs%FEO6&T9)ts6f}1#rq-B9;4ZCZ7;aGTG>|L_aG;U&7gY6v z*UMX1!0^`?^g#uA+euqVhhTSdlGj}Cr14eq9HJt=)_>pfR$1VCZ_-3@gLk%0-4ui- zTnhaz8zkNJmYIKQ&ySxE=8u#5He?G?#gIdC-om<>QZa9yQFNZw;OJZE0ZO<@CX@Cl4B^__1)M_jqw7VPV>1cvsdu}+MPH+VX6o-GW+7N6 zVr*lOAhnx2Yihu-qzOcn!&}Tqkgy7l29mH%(dmMU#5GgN6`VGGF=ijw1;uwJJ6OIw z^0~|flQ`X)yZ_4mkrA0SM6<}qnVJdmXsXOn;NKK=Ef~@Gc@zD~v``oChw1RJG znkbyR(w=Mbr@}N{{qB*GP)px3A=zIi>XUibzZ!^NN(_LmzcF^(xkp%_`=$VtCv1Pg zOf>&yyO-kN%d?+XtE3`F?TO= z^*R0F?$7=o!~8LN5UP|=xyIg?MM+zA0Da5=(A+_>{lCyX6AS4e5?hozm77{bR%uLe zZMy(e3Tz(AE1y3d%Q-t_4A-6ab#B{uP15e-1_QMYU3g@ckBTOWHhDw>d%IE^J21Ro zcQKiLo02BxTvL84lmP1ZG2I;D)o%lu0b60J=6P2^cv&Dso~oPBj2RGnHon8Car0E_ ziMWtVYpLBC&!)d&d>D8&_j*= zLGSIH?!ldrL^FmSA+Q7!v}ixNcH~RrmoZgZ>xsZ6!mau-Y>tq(WWBblibWKJE+V9$ z>MB|ZGMFN+;C_|~b{Bk<{>J}%5*e{N%sPXaqpYR9XbXQ(Mp;WH*36}Z?7|v(Q=Vb1 zh%O-oJZG~bmBImCJ-U4m%@e+GEBDp-xBis^2Z)u&4& zazV~{>y_NN=TxGQudKx6U1Jmlb7xYK5z<8{16BD%?kODd;~4fNDW&ag8LhNc=fax? zh-cZA?vO^h4aRz|q^vH(6Oa6JyXTc!+EFdw=sel79c&1@wgJTnXiJHW{<$Ba5PVza zKPYL_PvsgwkB0akwAKBwd?!q%$_#Csz$=~l$P(A|GALpX?eFH28LD`WZbh7$B;#q7 ztjo)2VeW~QdY65$_I%nylql#{S};zRf8t7|nUAt{-*(FqN?exdOkos$@{a{0=@0lBnG88N~qV^yD>hdf>mA#A-_X4tSw5w)By@cWr>c9iXK z!k%ONJ#tw1Qs>VV-;ay0Pa6CNiw8MHy~L>i9YB%;=wFP^50Xl$_Aw#Eq-JGPnJR{Q z0mFv!^w*1m8pAuUOZy{2Z{Ma^mHW$E>GYmA7fbqO)l-F1`Ie6y8BD`huF)jG-M$1b zQTs!=yOkJ?E}RPdofr5=vs&;L%fJh@pL^)W5=z)q278$i16x`oooDzFQ-zwis@3k| z#IkEnmeLKUUv)f?u+Ua^Lv|E{20xxuEL6%m#~rHN_5JB7*Mp`<&YkRI^-NrqmB-$& z6zujp{==r}$VuxUD}HZI5#IZ7=L=>Oyd+fvL3Xv4>ps1 zM3?BClj0g+gJ2E1=XsE8gSV{)S!7qqTkVZ04qS0$YKS;E?UD@`xD0a6cyx)i+3>ES=d?kofEb$HPlHaU^P zwk!bJqC_fy4Emu8WxT-HQ_Fu<9V>=}{Gi_)q2Ll^i|dH2qyrI!m#? zJ$w4cpY(+^7b|L!+wR%BjmZP+HYpvazO5=`Yo)2)7BG4k;~N;em0ZyZyrkPaoMar( zD7b?9II$K*0@G9p)=BIcIEf{RYq=6`--#FZn$>2A%sySMR&kh{j0B#R0 zqrksQJ%i?Be|Zs-@_c#9Jdx{Yj>|ujKz8-Zn^F64{pB`(NXfth!Fq$o1%w(ED*c{x zT|;U7TO;^>J~&mEkEChhEh&Kh70Xj&0 zEZ*}8=M7HZXM6hFJpyHohAJ7Ps6-y7h~WCXCE}WdZWovrEHS-uaj)Jj<8RiTUN?Z) zcVTx%H&v;Hb92Q7)&_MaqK#&MKK5%v5hF{Evr^X&*=i;lY}cHn|)u7c&q=N zst7STv*qY(-ro%pZiABWDkA)8D==fCw7+HH-L)j<@;9cAAD_0(NB z&OU{(&u_Ee4or5Cv`DH#{YOS5H&69%*#)d-DSX&TW#+`mJQaEVC-ny-D@SWKM`&bjW>d_NLystac4!b#+(cgn4;85K;ZUZs=E*J zCjb*=7SoqrleeE<^jLSs{z`RBoB!UXbK6@1pF_}oeTszz>&c46_hm)JmNM;DkLw5Gr z=pslpIgke*%Syw(EG@RKmwW7$abs;iU&_k=(41}evX6H(OANbxtRK@A^69p^a%VJ~ zQwXu|fQz5ZodDFh*X&^W3b1+DisN_TdI(i(F;Zm9)eW-u&uqbzD{WE<>e#CcjgD;Q zH;?!ZqG-sG`ftFX$H29Yt*W&E@c{@K&6*8CnDQdb0x>Zyw zNpnu5DIL`#b;;-XuHufu?Vv>T=75wu3fReNLR&!8<=)ygLFeV!gA=BDzTt$Z=6&Kr zi#4$);j8mTuQco!s-_YtYHPTeH`B{?z21jLYQr{FI+}bBDFa3Asx8HT>MB=2f+1)( z)T{{*$Osw=5Lr`d#&JmYB}Pt*v&&WoaX0ZcGAd7JSl!s+kPNalrB-hPKjBnbaQNqB(;WB5PHytv=7`cxY@;>31o~NK}hJJ}B zA(>&5cy~f7em7Q|>JJ8OdvHqdWPA23FHE2kgaA3ofRRn0v#yUCpBm|<7j?=WPI*15 ztFh0~i!1xFsUh}c-lAQUj~@NVcL4?7ZBx#dCRcE?(>eL3y5HGXInf}y7Xvr4q;Nyw z_Ho9}^arf@f!cy;uUVeh8+7Kg?d~XcNZ{KaKu03r(=0tU@<(coR+_Q1DbQ}c$;kN` zFa0V-V>RIob{FR(P8h4isU~{)?{xJB-NjnxeI0W(_DQB9a25J7P60MHzDyAc8ofS( z^{04WTsh@0b)Oo>m;E2K&r}PO=YaQ*pt$b#x+;-lcp^?m`Qo4h zL)6Jw&@D48TTY5tU(eo`MxCTJvTkF~aLU1A8Wdxr82+HdKR4qA3`05^sd@5nJsBm4 zQ6e*+mtZ37YR0j;jD4&#TU=2Yv@88xfqR5>!#P6-=l*vVe$vP8q{~6<^PE!dezpLt z!LZ?qAfxdzmNcQ)uy1>VE+N@eVG488e-CVN9**Cm-ANTh$N!2bebFD<8h)F>l1(XJ zPQO00tgM(fbISJTaYX+>-vL+cg}Imac3P7hlwOKx^~PZCrWB@{A?675vzW4bps9iW z__TI4;D%E_>jL&r97QTttZrSkS-EDf0yD)AotF&DTv7=EE=W3_G~5;sc_sa4vbd=UhjgFo(&|eNrok5NlxYm- z2}syp>_XSrw>cx2Zhk(fGzn9U(HG?A*RQSA{O>#CHk{Mxb9Cgq15| z(C#Ss$O#$`8BIc-AHp5M-pyftctNo%@+bH2ShQq>V@^=6KDv+Rck&76n+R%g8?ISl zl{#6enu2_Le7#HG-sC7@!h9%OoC-=@wiwE0810ElB-AJ+;cxnr;y8+{a z9CK(=6&QKuDkDA_kR3|N#_}yGffU-wX$NuA4i{fXS{BCRVw{lP>{c7;Y5%tGl~BUX zgg;lvPG7CFuN6RWa>m?t>U`P~VNwi=SDKX3|08T9yI4OdgOP&{K%NAL)`etNMCd|l zSCsT|!-&YR?f;;>nMyMu=1L4eJq;Kry-Zz<9z4%em(inWrZz<7B@Er7j@F-rI2K^{ zJb^tt#65;@HRB4eG^^E#0#dY45!}ao4X1B(x%6N{=RMdFjmnG_>Cwp(~ zg8vNnD_Z<1*MU_Bc7zp5M#6F7_Dp^${wTC;03s-O^z$YsuobxaHO=<+yyzOaq?#rc zv{+K_{YLAN%);)*s4jC!5K_kp06#WqITI!_L8N-jX_!e3o^j=NadqLaK%+CF*7ERm zM(M?m)!I~p3^n7kUxRzDbzg`!LuH<|fjUIKo!FbI)g8#U9b*@`j4_UZiX^*DzaUJa z260n=HxczhjDlMh^ZTLWTIs2U8AtwuG)gooFBI-B$X=f~<$YWL z`k5vawR4Z4R_unlAn%m^p3l+&<~!|;1FV4i+43jhBGRRu-fOTaUj~u`O`C3cMt9r6 zJu>|E?wyeSj)nJm7Kc-(^Ugc!k0v)|Dl5vT-A^@LuERhRO`sst=~LD=XmkGOxkq{TZP9uxWEr1-==RGLOhgjjQ2fk=9DHw83wkj($b(Yg3D{l9N~i&_UP zpsgD_rLsm?E9@VtU2*dtdSF$ z=YA7Y?Ld(86izZkD+AFuO2nnM(SP}^CnKdW;Z)A*(9;*xI+F2uwC*q3d55oUl6eaG zrTxiOxQ6PVd&{7funi=~pFiCOW?b& z6FESZU~iwqmCtf-PjMY{9CEf@A8$BCT4xBe{Vqx7xX4eIIZR!N%zJ-C;O|xbUMJ46 z6g>N=FtHag>$Rv^)wV@M?ZzP3t&6zm1$aw3SVz6s)9e;TyDf7o86*70CuY}YpoSls z{_*ruGi%#6??Kwf%%h45|9Qfp;GOd1j}aKQ9AK7OXF6S&()uqVY7g3$KTN!jez<;t z@hvf}sdiskQV8(MNs_jt^!NMDZN*S@oK{*+-I4apl5rXqQ}+;9CCqw&L?F;3EOP{( z8sj6-o5`PYZM5jGK1*3&YpP(sAzDp0)h1w}jQi8#RNNqZ_pj~vPG)Ssf9;Lo#VD=* zD_^v|*UO4m)f1EVm>DjDv>0bzCzA4<++vZAy!XumV+V%cE2yYPvZS>wX5T=BONtLI z-)O*96mz!92ed_8x)dj?+-!o*PX(Lkpr(HZAAM!gt{dia@N#27A!6S$f^pGxKGu$x z58}1#tQk&&5o$49NxlGKqS3~>*Hi^RuwXJ~d|!6Snu(09I^OvC%;?o|T{+#@%Oovi z zAhb`eS8G*oj4PYFtAB&wboZZfi+}mZ-7ir6-Cwg~<2$c(`TW)9-&8vIGFs3utN01L zo!}1tqE5}ACsv|!`tSR;>GuZCt}mwUd5a0)6&d4nzOZ47Wr@|dbzJ&B4t5=3+Du(; zST>cK3~;#o-XZQHG&2kmjBj2aB+nw90VkJZ$REV-sS0+DbKR|{P6tUIqiMkdDKr8d zN^Ktwc{jzqNtMC8DLBZ`7_tgMz|n{IJ^opD@{0@k+^61eppO`<{(NtDef-^hspa?R z-$u-Dzw$zue}$UK>>E0*2FberV(9w8Cwq3j6PbAM8zQTqpLxCSqS80rOvO)mqJj-h z@vB`c5`Sb+ifu8w?n98C1WMrvJXmrp&>@To<_|nDoM%7h^tc@2wV3MOY;>@%>e6G+ zYLq@>kgTVwueTT%fJs>YD(Q~t3!Ws3Tb`9J^j|0E6hIY868sRUd_nd;hDBdC@NmJ> zHzT98zjly9b2T%ek!NwhD$k$M=YM3dVm*V~n(N-yU8|p5%=NgNb>b;>`!DB`FK!o( zl164+MM8YY4NH&KuxemX^%nbRA=+6(HOrj0({7UByrb5bw{o{=ZUwe(CJF6BDq@3SM7v=SKY*210DZQY{L zuq9F=RssYoJAfKao_mEsis?98jXBMBLjLC@cxFOE4ziTAn-f&NR}E=G!Usd7(EV9u8Mgd#+{_LF_Kr}_D&Ni?H<;UW2l?Vnhe ztTy-mB4wUey{qvj)=b8A5AHueSF?UEVNQxhq#r9d%mCO475@0#_0kNcBQnWl|0LIa z(InDlQme3)@jbLaqT}-IyT+2b?do`;FI~akOw}YWE1ep80)MrMdPHu*5m6%CGrS5< zwgI3LDacNDV6xex#0c1<0F?^G+OkOU$|FaH3EU)KB`=KQ}0$IS{n+zjUnv%#{@97XLGD z5C15CBzJTGnNa|^RApdx`x(<2dtx3<>cNUK;@Y0KGgYi-=FJ(jmth$(OIk3lk_GhpTpX&l8Oifu+uTD$a65(>KGP@ zxyH-ykf0)k09G+_u~v?qw`g0z6qEBmgL~Xo6HkrWJA9sP;zi|zHzh|atA_^;6{Qr%mBW5na7AGQ`1;ypf+s2iu@kiR zck%~u!lvTrx7K7RJB2nnVB_J;#WKAH@Q>Sqv$@C#W0}SII;9C?bqcuDXk5hYpL1R= z52Q}o!?O75)zOGtDN{pUA%T*NYa-Ea18+Dv8H1n(KDMFktBpNHmaG!L((D8m!dQ$O z>(f5(Bl1O>lC5XyA*b#@Y2!^!&)=+Ao1PULi8D)18Rj%S+oBjdj2p zvcqVIXha=YtY8KJnS@nu>!^1cm77CX4$78R*3K=H>`5e2Tf(DCbfi6hl#p#7dvUr( z^LDORYJx~NLJEkfDxi}Q!{nzhDk){r7z?Hxv2CO7d2WzPmRi9tHG`U$UC6GiL>7((r;IR#PCU8ZT0h`8^~GIA})!( zZn}%BG;>blfgoO0?UAgZW>s$nn{=(q;2zjV!+}l|A*t&G+smJlAu+@Z>znC9FsuLbKJB%j zIZIu())MQXena-FnRj?QuDgtLQNp#4{jZ$sO!pB}Yfn~d2!Y#)UTQVGK~_!U(^QfN%Y&2FdtOne~rb@q%goG~YV$4CQt zDP}?4NZ2+|SuPSCqX-)Y(-81pT;g0y4{+(^&{deRy=_`)8>mOPdVWa}`tWeSkI6tg z*PlIFJI~hYZu|Ml-GZa_*jW4C5>qdJ5gx?K^Lr_3dg%!q^a0r_#I-Vx5q6wVCq=gk z4C3rZSJE8PxLS-vYW#XCxLaoGuT(Ov{Qdn+MKBt?ykL`K{&#J>uKDlwT~J|?lh!pi zoQADaoK9Ne8;3&V@$V-*$Hw5?5?n>9k zFygjFVoir$@q%Bj*W)F9sW-zHnZolmouxiGVSw_ukXI5X-kPE`TW5waWM}kY4>6jn zBNs=-(P1}0;j!`WiX1^?)zF`nS2^4Gu>SL-f|?bLfEEH)(K<4#vfj=hyB+ z_;Fo=pouPGDvNGmwY2kIm}+84M;E_^oo9zJtglRUC{Z#52f~?mIe<@2-ZwyPAKmq< zK>qzyh$Vmg2 z4a~FM*uB$y6*SIqY>i$z)yT6j1A($Wjys|Xev%y%QBx{28-7O@Y$SQ#J?*oMqU{8v zSf0G*yKI{vj_+5D=Q!7z8UH`XJ=`n_7$No=U&m>--_4x_qAM&(H&F7I0&==$)ADA3 z;zU!F(ba-AMd%-RrdrU$IEEOdoWtXWyY*Kgv(@{ zB2eL9eb$1zA|A{|d7CnI-b#jz8)Wpt43Vd6w=3mlY~ z(L;%N%aR^-pm0>=&aXF?XO!gnO{EyT=?UN&G_uFrHs|Yh>f61t-eKjh2>UswS<5V< zc0Ox<914clm_`#eACZN3L+j4tyP)O;;Wa~ zsnd;Y*dl5PkkNT|V;he*@|}oS$V5{H96k7Ni*n? zqC=@U_Hfl7cW0H~M0@VKYj;cjJ7A7MQUzMvq*HMa_m0B?S)4S8rh$nGj)&Pa2L8zs z2u$Xe2`6=1q0fKh?lLP8dQnw?h;&+*hB#s{v~bAJdEg(ANx*0HZW*Q;&`u#7P+I z!;_Xe3XJ^x0{41{`P8D@s8lP}G)%iZ%%-cr`|r<-JBX0X7j6Px{Jkp7X`Byo-kp`j zZg*Q-kSFmEujabOzTS1eE$k8IphgFrQf)X--y7dp-l!!qKjeA9su`}Uuk8O_fV$?v z_d+sMdg$poBNw791fdq2qvgNK1mum}Yi)T)Ao7Nx-vLb*ZR`;h6m5P065&;Yt6;5@ z3e^N;Ody5?D0^!VoNE}L#r4NHVXi|hTZ(9UVCw31#%M~1##?9aZGLP;f4!>y_TAT> zdL0aubS4c=WKHf4>BI!yXmg9Td%v$!=L*U21mwzp<+EF_c z{o-g@U~DpxEu8)dp}wK8%k2jU7C_u>ygP8hkG>3i>c=3zgT7*uTs6KLXOj z;vF!d@cDW5gzMPVB*DzT&(HK+j8-|~8FIVRIrK$AWPIvTrI(op9{Uf!yRsT6WPNns zQh46}Bd{sgIfsxL<>`B;*zH4M37>bAL^K$2$Ge}UI$&(IWOAQKt=}tG@kCF#x9}d zyAOobeX#$jzRx;o&P2+p(eLs1KaKQ#=Nrd$^9=`JvlQ7N)VsWxBLW|Tabn=4$-Z?9 zjdPw3mib*MIcfc7rETT~NzzZ|5vA>vY;|>i=5WBLf-?(IdBROaZkCGD4INoBdv2?I zxchQ0)TiYu?6#k7Jn_ZkF@GOTnK?K1r>spyg;+3_M;7PG6uYqh^X7_m;1=^?b{0|A zR*bDw%;iO%*46j^<*wHzgKYv)0IPd$YmeH?o$rIcWXpuzhc5kIKc0PS4|I`t?{aKgH9im^-M_n%19@(+t zo7?D!$vx*7&vm?1a5LLB@$!`9h`DsAnB3Licf&!QF&{~#)Y(PeSo7>tMG5^AA$L9r zXj95FKF7K%t1>>WoDppCF>A0KsJE)?P2B-Ox*3ZDg=072)XT*vwQIaY!o5x$QObVX z^kumvNrrY>9y5?{vmgry!qxR3e#cVr4&aJfsV z?Gd$Lwa1ne@k+<|e$!j!;#4TbRoEEx#&*$x5m>_bHZZ9z*3vB)79dF&F1A{5xY0Gq zRbERMk{@_`&N;ddGw%==mG> zRd_q{0p>sZtq%#rgW3K({}_VFXZBHnBIxAAcAx^fHIA1J?AaiOrmLaF6^3pYBs5+@ zxAeB1D%Pa=rF44rk!P9tj$dvjDLL6Kzbv-e<9Z|p`Y$&90?GhU8W9!#VM)Lgke(=^ zBM{mnqH)LvMGkaB7XoKcXtnYYAX3th`IQAZ3nZg@KCJwpSG3Mx#L@xCYxg^n=55;7 zEL*+%)T1ajWn|C>$7_m3iuVgcumfy0$2RbjL;1aPxc;bPYs<)H1or^!45@K6aSs0+ zKkq7lbI%ht-$)A^%r$x{HRq7TH((c#t^qe#_1P2c0EXpsp5T2rSw+d{s=j)>s?-JNFC$|s z!|ktxjvRV{Z-Y`O5Gdz7XNN9%Hkok9!K|tGl>KD zUC6!Yk&Nm*7PtCD|4~$*IL|1CZD_%H}*?$>RUIaOKUPkO~fpyTb znaJv+Uua|8BYBZ|;9+(G@!ES*Y!HVCHE<|kD-JsRjLWvmvrB(>fp7*|HKei9Yu=F>H z)%b*i61+J-?Qg*wd=c`Ezx+L521*MbvG;dP3yNn)!#~yyUN9J$Ea2EqvhP2b+8Tgp z25rix+ZtvNg}(+bcYlTm21*-?=^wr8VMF$SIGIhPEOva0+%OO~=B>@f4@4Nd_ei=N8M|k!9-yb8W6P-C zXwIN&?RQ02cr+8l&HcQDvQ_UcfC6fgl~T_ZEyI8;FMp1&cb1Jw!>U@c~?~LTBum%gc*7Fc%V8v$75K=QDDv0=`Ll7ypf^>!C>jg)fF2*h zJ=c^4DYiZ%kH2StDMt-RW4hB0C%J)#oK!_f&y@6`}v+zw8v$v9*-@lyy{FuO%RSo1 z-YUP=78row&LA1KNVY0)CeNQ`A(06FAvWO}+q~okc1kuUro){XvvmDtFdPUcjP)40 z#(MPzd5=rtM98Q-8c#(e8>_!YUv$dnUs`|oW+C!XVBnFRgfCkx%ml&Rm3_p4dmv_J z@@62oS+L-|7X)7(%7UrKYdodnvXaY+f5SCk6cmLvPec`vZ%Vmn=gOho9>^^?E;K@R z_wI);jiRRPC$GMn_Dr?a2v}R6+w=6qOIo}25G~VL)&AygPq!iEX4~&>p-Mc$}4%V%8I&NSQafuN1}SDk1Wb_(BAJ4l${ zM;%S(Yht{adV?iz(e3^2{SE9O%PGB8t)1+<+a{Wm7U^(7w1eGDQZ?-WDrzu$)^H!ZE*<{-FAJg}Ym;0|Cx=U9?(XC9U0A|CF~mq2GzB_A=p z7QrP4SI)gqH)DLbP28qz8851#XFXbHqtNW%`0Y$&c=6Qvfu>KF`=Ngw=zSay?lC5* zuXYEPh^Whm3&}{p*9GwRpKIzZAob_l5UNTT_^gZ6r?{1_G%&Z_X z#d+7qy9uXT8g_RZ6M6u;OKVo zqQSLfmDRy;jSWT|V3QWVmDE(wG7?-)1Fw*c^&kR{@hbMt(KeS#3ZD13wkO#5S!B+w zFzqLKL3)pVNnz=*m6(Bx7F==Qkaa!KrN{LI`i*=!1~ghYwxEw6lGwM3 z6T~QyzP~i~1oF{h?IVBBT(q8+v`@Xfz7tb^DO>kM;}fUicVbVWY(mSACo1>;N(-vW z$nNwmee^x`EJw{Y1wvn;(7T_FY!u4Ss2zvuX7k-KTb+htj)j1IrBNcF}AwdD1ioRIH1jtTk4 zPo`KXBJg;5oIC!@UYc_!I}+H4adcog7mQ;6K+bD68fm3H^h3zvJOKkO^{sg+*UVB6 zF23xjH!NLEH#RR@X)Ddo8J(kbe}!_nzB8dL*)(39e)IOzQ~IRl=Ggw%xRb?(l!|(B)zhNUDaL!|-n|Q^I->k2(r7AFH z41OB()chq<0M{r49fPs{+aZh`x?xAoXPu){ z{#tKt?n!Nephvj}h^+X=|9~F>T66oTVTomNXbEF~EL_Mf0oA{dJAlg_8=Iy$YA5n6TD_5Cgi|XwM8og1E+8L?v!TL>1(hd*{p@`rPsu6IPD?tUEuP zA?gJwZT_&ri4+Q{PbM7P$rMj<;(0EQT2}k-sF#x_d8U9GuDKLf<5L`QpW2uidf`Bw!ul=3O*VZ=s%-B|p?G@TN_GQXH_cY6FN1&yVdWS2 zvq!Py8GJoC7I_!G7vs+i1a`%v)hm-N85pe)cXt{uvRBGQ1|{^%RXFf)-MGr6R9}S9 zF000WijyN18WVH37F=u&wX8!-KQqffHVxjj0S8oi+J)B~UF^6cF0D%g`xU>#6bf2t zVcnn4{Qi5Lox=D_4UEI!U}>hO(X|8FEBpL4lZsYy<0OZRy&aeX;=^*Ez5hD7G@OA_ zagC#c_%8yqtu-0Wv0q5Wd%hOifks*z;24v+8jSU|)kSiGWZWW~e zP==Ar4 z(WY_itAOP_M&8Mn<=$jlGghg#4IuN0<`u8V!1Zehi%ZbcSZN}4KGYq$N+EJB@gYVj z?LmPRt1j~VPg9#tW-9G%o{A7S_CF)9H!i_-2VWXX&PAp*Ren<^BIOD0>j$_>?CbIc znL*|9?gOUBn4x)5k!2(PW$5!YB}Iq6i;>CZA-ZBxf5R#aAX}rec-9l}fE>BvY&AyT z09h32v7U^OK;8?CcToX3ZF}08{m;oeyAJzGwUxdsM=Q{Xj=?Qk7vONT@S1=VXiYe( zCC=YyjeI)29%6*6AMl0#Yh}`CsC0riAr&v$Wx%zWa=6A^8h{`0RZwS2e6>5&)EK5; zqf|e6&wufHI{D&wV|%2qoMp@P9VF|?H9FzP<3hNqV9=yM#V-S{Ej4nB8jS?Ka!Mxu zII!7=^>t}8m)3eXIDVgN1HK=xunIJk16(Oyxl`}9X$4gt(WkBIc(=4mnK=@@ec@%9 zpN_B83CFo#1&15_=FwtHm#u18x^4(}{!+M|P`WZ@gh;G!t< zb$*BjPs2e%5U?BHzI{W-UT>PL-c(O>U9`wp!<>I`xU9m}se5AHyJ1wQXz{Z*p6PKuk;OK1t#>? z=kTn8y2;`zhXDKcqF)-K$36m|lbE~S!-u3?uYcF@@~SU7<`46Y!@I}tVWQ=}p{|0U zrJ-BtpJ4wy62FiWc7U9wsWRYFK^bvF6`9|F~n$ynDV~{5MWlHToZLns1hW z3>$NhZQ)9!&@zZf07&ZN#T!ddKRfk+7$PF#F?rv{DA+rMgAu|jV0c%H;u?=VYN+|L z6fc`%K{DNq#$6pFpU=KIGsm%v?LU=!o{(GBtmy&yr_SGZo+)>Z=|!xKCJQ5|5{ZMg z$%F>4wK~Qm1qa4E>Pc6dm_mg8U8jEO&+%)iUp$y5Z6CIQy;JGoYR2MxzN2+sMa{ZL zflNS|SMsXTTU=dxlqPELFz;iWI&B<~w<~{spn7W@bPIQ@p=xH-p$ayQUXg zu>u04Y@aPoc`-}VnNCh~Hm?`pDrs5jb1p6yx)H8uL)({ zVArx9O<-JU-*y^bgP+2H>RR5d_=5$mOW{7 zX8aH{SrLAphc{K?&qWaFL(CTbUIC5~zhBa|5nNY&@Zf-#(hZzVq`@*k_*p+0&KLGo zkY(mM`e6@GQO0Voj9!p+js8*B+{%Qd>X}FmTfMU`VT=^%{`7buK*g~&64QaAYsPEm zUa#irxI*&%Y}nz9rE_d@HCNe^ZmQsiYrLBuSm}y!n&C>QKD;GUMVs05(>T0tZx!et zvTs@_AV5{z1QHJZTyex8fq^3Iz}&;z@d#l2qxa51jA_udJY=!Cm@j(DEUB=EPl!!_N>#qPiO>`>6!hx$j3gj zq&cVLR!QZL>OqJhc^1A%pbs-sXvB0Jr3^S~wz@|l*hz(>%oaMxOTRGQh?KDMt8&Fi zGD)dGAC1HJA!NXYFXHEzkX?ihCG{C<{Iyt&a(8_0hibl+bhjl-&hMxh>?C;Bwe02s zV_;aXW3*7={R9|)4Bv$8UXN+ZNdVeAJPj*^66s2$Y9%dbF!jifHB9B@QN1&ccVlZ8 zxQ2{(3W)d9_li9JYAVQBy?-`^ET8oY2~7gWF)_sVXE zm+HflB>8{e-2gwv@H-y@-T25{O@Ap(d$cXm9We&YGAt;3Vhh6>48m(^V=Tz6rPng7 zr49aBl7PuW$k!*9BM(hZZk{F{^r{IO)ER?l-Ez+;y$~TzPF*70N`18fDVX{F7uA-q zg=&31va;(1@qxcCi*)R8viC!^rhTOOR2ond#}xuMAQ2J zbi8)(ed%~AX4hTer_t}AC8r@UwfB^5uNWNu!5H=$#H-8{)Fy2~S=tY#A^{^spzIs5 z1>0(zZo-B=%7p;fwb`bsd+ICZ$b-q%f{5=>gI^+OgXIFNQ8n)U2c##KJ7{GjHJK&k zH=@h?5j1O5akogstB`{aMIuEaL<}FAcQ5Bc43?gpK%J(kl`o}hM|`Ke0#QUnX}8Fw zv8C;YXd1|NG8nNn+M-D>m0=$Iwz$QBxTR~KmN!g&Fhwg{*=6K01~GIumQdw){SShF zzj-)%$6HG^qsNFaZj{cKf~X@9dNdkr!^aKxZBA1N{}75eW4lx0(s1VfnTbnK!&YeG$feWcB`|dL1Hx7 zqvc9gxTB@zf|_S*s=#;r?BDK)b^0XdKKCdn7ta!9KNFI0zJVErsKky)<_=3GlHQFF zFrcNFW#Z(*Q7BQH4XkNgov-)TDDi%Fyt;~FHMvha8jTyoL!G!<>`tcZVw_>xOp_0| z|Ei^OMg8C2Qg;0Owr4RdV(Uyg|M$Y-;FkBD87QkC)&_TO42i1v+uab@tn6lvg4HAe zmi$BRB}U8R;vA>{a1nV6qVE9byU(d@$CfKQgQFzxvM@Y|hpz`X>Q3oF88JQ=nke6O zzRDZU2GixU*Xk|6#VYB=$HXfN+R`wXV%x>&P(*vyE-6vux2L^6taI%|UO7;*--uT!KOoIlW8~WYICZH%`lJ3e*xvOu8B3 zEdK)ms)w7)0~2E^9He*J_pjW`?ASN#2-d&8a7TxPbJZ0fLQ(vg;68siSm!mh7~AQ1 zjb_~!{uBWcxI=;#^KUblXcRYH22*9K^!|fPN=V#zlN~gG6&6n6A%_)A-@1v`3V}QHh)Rr8>lub8tn4DEu9xSAPeilD>bZk)-KnyH0-D1WRo5e-Pu; z0+2+!%g77%;xyf& zP)@q-S&DfG(sEcWwEeLZfB{}-p%@P9_b_KUPW)-)0dQ**b{rEpV3-lhmspNnB*g`E zbY!v0E@#_4-*K_f1_BQ@G6*A z@d0cAr1Xwg(TYa-7fu!|35Ki^gC3(1XpWcu|qw{<9J z@k;nRxo6plOjD*ZbSl2GVeS5tuo>XM^9zk55{)SeDk+q3KK6BCFAtqmjmK z@PLz33KX!V8$OtC)-e+b?FKN8=nl**^g-s6(_h0!znK_8u(h}zD-GKeh6;br*~Jj3 zD~k6>k*WI2)Kznph?6x{_C5E4ai4FWAt0skGA{Xx*^B$|k7tMVAK)1(_G+XYiUMlZ!bd{-coa{QChkw@l2f5HoH`G!_x=6phvf(l`4 z|AVak5RwJ|@)CdV$=lU+2?zn~47uJbg?o7N_~<>ob{p;8G6`#+*AC}j(=~VTS%^y6H5}1CABeu zNJJL`LtEBnhuiutp)oShXYPDu6F3JgYxq7~M!c19v# zE7pF+uFlKWeqO3lCztPpLy$X@e?xvG&_8l)Wr~75J!L+?4&5+-dEbKmf^Ii3i2S|J zXs!vk22ewCRF@SXF_Vj4-RpCCR=nb_V>KDIJp45_J?u5t4j@m_LA|+et=^0jASN5* z$E)e3tzjgL@pUYiDlB8L&y%rpZ3)kU7X4jWFWI#mBry3u zPdhZ!@qW@Ds{$O8iTy|kbg1)Xw+8;0VQAXz!m8|&w8GpY=LUC6Jl6vwTY)e37uO0N zb*UsBvXZwj3zz9unS!pWeuRj9l0M6FZOQSXOH5dxuoZ!jC(F`&G+Fkpm6Sp%HXs`%A_d8MTF*qEf>xPY@^b zU!L0e1ajve`bmHN4sk8HcR3?L2k&Y(Z|$`DGLn6fRd=kuI6b>KGa3#a?_IEZ6M+GO z?nDigXIoQdVtuOp$Z%6>c02WFyL&j+zaS9C*0wn|c`d`jvif&T6KelgU)h21}MM{XUrQcm41?T<3dTY1?^jm$=L}>h zLWcNWU)%&p1%|1AvnQhWOF<=GJWaT2d`$MdsO^YlYl@ge%!+A6NI=~s)4xWiY#>nv zHP~5v6D*l8*s?ZGT~7lVZcg{+OylQJV)w$yshob+_IXx@@{kuVsbW>sd1d?hyJy%) zb-6E93pzh|ukL#b{M3V*@4Oap_jb(Z1TQbuKp*M-r_M{;?>Q)1g;w;47Q}qSkTtY> zgQ6`TO+Wh{_vFksbzLi&Ba77@?HK}}M4}PXL{Q1MQ$`;HsTO;f*NO3|ouB1@`_JCH zHbg^wfnR87e=Z3oLcEyH)f@0w{DJv~a7^a}P5D0(jQ$y&*+LH*nsp(E|d(?SvbTQhn{F?6g8 z^HfvVp`Q&@n>+u5xJ>{4{xL7#;d*Q9lwM@#+v|YBsei@!fYO^dvCP5hW;1B+L+G|B zAIA6aUx|-xu96C1rRJl8J7)6AeI3v9y2?t(&OR-yhe>OC!lzdL+D-*? z(xL5joec$h{MRMUR5@5By}5p@$_#3zIvwKg>P%>zZt8TpHlvp^LB2QLjCARLDHnejQCG3nI9RTzJ z7yO|PS$rzS;tl%hrRN);j*HtdHB0>8%njjKeojL(y+;G>t%nqwbJ6}2R+NOBMd|tw zkig{dUE0 znVj`}a8ii6rDjn+;#fuF30PwOfM9r3x$%2`VD1sCAP%;NB-fe$*IvPOaX>#Nw}{W_LjETi&0#3zhN7hweVCs|>uq@8qYn);#6t0|CY-hkA$yp<_Zp zE4Y(S&<7yD^WPhXnZTsiCu288qRjS(r^pSeKCDrN?)#wP++FN@rrv1a;Q6QLeqSzD z8Xf+#ABqfa5j|sE>G#|scTSmkZmsZkz8iYR-hOWNXQExV#|c3@=wGpcbn3{&$@!Sy zU*yQAW#%BgbU}D~F2n^MSQ@k)Dj*PJ$jpM3&wf+SK&kJs?j$)%>S{-}6_Vmaz2v z&+4sr{fWn|c77Y(WubViG0UUlw*+Rn1*NEcQWe_wHcR^Hq46-cJ;u!`pHBZvx>1sU zAYfmc*>AZ!$&gJ%EI}5lhS|ZKuI-AIh!0cZmSu$5SeYIHRnMlI9${D{ zV%)2jP31|yOIswj<8ssDpM8Zs=j%RpRuyyVoKPMJKQv{g2x008C4I1udeWOHSarfo z`s~j4sxNMvs}}bk?Ehmq`JwiN14M*)U1;^$u}~wWW8H~YuEo|(J;P!>@3j7mHmXfT z957YpDv~^z`k>e1*R_*W3L;a4O{*94000xg$Y7_>P>C)R&wo!}32B-Rp5^z2{e9`E z7q}q&#ZCwZM-vCJ@VY78_1)~@fiCI#84ZKu5XX}VhZk^h^R?Ro3YrUa&Y zyyOXz&;1`z`%txG(9ih%*%D~xPoMnD#r8pZ{SUucv|Y%&?zQhQ>zq@z?!?8vt#^u{ z_xmYcBCzLYYHlRmP69c6u7+j-cW}pFkcI*K(%cNXSvycW3nRJ;0b&7a_AjesQtvY$ zvGtHMLbnHB6{s@2=GNb^!~H+>3Nkl#I^uYz8^N~cYQ4WdlEp~_+xrRjThw!R;G zT~akFQA)Yxrv$Yk{17z8$)8sTv;ou>9BWSO{f1Y+796gwEAMdyhn6>hhe~u*pyiCocJSb7qC)HPXNbZpuNJJq(sJxqN-A{444O_63Z-PB zO*?g#KP;IHnu#m)P2IObTuzT;36KyA5EwT0^}&oww9_EAHT3)U`{yAF`KMP-*y^kj z54xWw8f*+ej#$Z9okytmS2_K#zv=NuQqO(Mx$HUgeT?L(tP6q{9zwS?$ZOZj*=Jd5 zO}x4;4fYCgv7DVmZAS_7N056$oDYYCLHl{sJ;LlFUqH6qjH!klF`X_xA|A`)&3?+O)M$pa#9MW*DKchgK#~zTBe=>16CO zSj}+3SJ2)M?mn0cK^_lL@mXbW(B_T5@31Tsl$L+rJw#mY9yJajB5pkq$s+bFXijzL z*P;jedgnH`;ySd#ioUQj22SOg7U?JIs*j4M%#2ls_(vGsooM8Iw)Ba3J3H&_Y!M!I z7_U5kRw;8wbkNJyK{SOHo0&~q`bT+HV75OsmiQ24Wsgd4i`+L~TZrB_)U%w<9y~ea z4kJk~XAXu#pyA8QU)@CTFAPi)1r7#%*b!m^H;^}{kIG|PVTIDaAUkO9)J}65upc9E z(=t-@u6FH8%eT&N95Sz+zJht@S_3!UAtiHf=LZclM4e%A5hankpVJe1Jg{n$0EUN1 z%&^C`*+Vz7T|n(|3T-%fcJhwJ7d1&gn)4O@&u{&yS=I)-5BwH_0(3Z>8PtntSf@8| zjsvGvN>F(p1s{))H-^&IrJa2o{RbP~swH+9Rn5Ds9Ue%B(M;e}E#hce0g}~ehUXM1y`I&mTdj7s?C))bIJwX-#z zjW@NW&?U1)kBbE+qh6vRoLc_wCg#K8MI;~)QY1>-9~O1v3u?EAln`o z61ZNJV(aoxmsq9<(x#k^%E<#~4s#AR|4{9AX#Vp*NLXe^G8iZEYlANFehx*I6v+_zaHlBGEYnO&#|Fk=Rf;oX76MzzM$*xPfT*(nb|En z`fdkADeloDjRaR^`?>j>R?Z(XrVhR>Fc1`A$+2dT%(t~{l}hCLjRDSmb=_Cb!PD&) znicSkePQQ$CQDoToiPM!Dkfk6`LSRn_v)wBjbRnHEUed^ev!B*g+hBFbgUfa49Qd+ zbB=)=&=N*HfIT)jV|sS{wIhGUnq(-d>iEhhX0U zF*L+f;a9E&eL#yy-gR>T-bbEypltQx=GuEnlAk7=oP81E#!4mD)LH12YtE>e8#ZrO zglfh$em5MGzG7h$jYu;X1;MHdWd%_vfppw)h|K>WtL`L|2%3RB{Nb*L56oaPG}!4S zW7_HUS1g6H7L{nhQJLieeUFG&oTZHqG=Xw~teL-dStdxEwBLG0|CH+{-5x*E7%T}_ zY@VdQS!~c7`1-C#>}}SBVU6i+Gei@P;mD~(fyV!Sf(P;o;bI;S&eG+jVTIBA`k$AN zKc1wzrUk19%AtQ=u$=d`-O8Ey`>DkKdSrcrPMshGdo;#SL}z(kMP+$)bJPV88465# z{JCE_U0(+5^lR(fF;h~D|!sYf#G^&mCv6X_2(SpI$FQ~ znk2h$nMvw@82->gW+CoE{RQJi(QLZp){!p9`u8x}tVzSy7&Po3YiL8I#>7mv|DNYg zGr4a`2OrZC?)2DNE|4|%e4v%0Ac(hk#x1*88Gvr53nPsFh_*gg!~#)A)I5-h<0^#oz{TRlgV*PC22Lp) z9@FfSV_Q`P-}X|-q0%k-9TT1vIN6D zd;DP`L%r#^`!*O&w}uAMXsq%75%#9>P_}>E|7oeQHd)GU6s59-vW^*AEQx8kBDOLvfGb?l-aF`RUvFA_iYU=<>Xkf1cf~j9q)n!R~8^v+a&+KC}Db3ja_bSsPP9 zi}|d0;+kSA($Fw-Hs$`48#f-^z1}}s!KFPj+ashkxNklS=Z6~9wsV^4iEJDW-q$-F z{ppcxc!2$|m;(lRfJO=sXcW*Ifvv-mejtD`iyZ$y2v&bpU}hc<4hpGPG#;Ul*;`9q z8aN}W!@`ZUc!gxcgYNZpUw4>p#}|$;9kR`mj-243Eto@*z7TC3z}nW*E^`|sOG9$8 z>VIO(v)!IRT!TuPFrMsAwUjAvdOeHyPu)YL*RtQ@_R-8JX5B6iPnDqRfqGjwA#nERo=G+0unq&FhlP3gWZ4&**W~dNOc)c*bOjxR@hHL z{Xa?H0mN=JX9N$Y<>io@!7Cu@g0rDI7a<-Y`6*EZeuhf7!R})ufN~%D+VP~y^YW2V zA3td?U*9|5-YKQiqiON(Q0VRvH8X`3qoOC^;!1{$eaNZ73@O z=IKD&MeC;Qf&iu5^qX<~2d~Me_Y2_n;xZ1m(>fJxQ3z3EP)+B#2|Fo$IgEnhm8Ky2|Yi*2S#swLHwey6-jF zYeN+e;iZgt zGz3RdqvWBz-C@NtZrZ%8e6y5XU|uI}Yg;bnAU8}&9WJLiZ1f+GM(#J{9=vgX6$bkw zXC@QG-)|w#wT6IL!whHQzOWzf0OV%=2c&rmpyaY2^YMi%W~8sm>q0SPzl#A}bTsQQ z^cGB|7$BZh65PcbuHAXqoecPmo(J#$d_5JnFFw4bzxGmJ^6(Kpah^Xh2Q{Y-9Ejw& zIR7QVMq>_`;Uf1AyfhzHmK=5tIvv0Vj@qJOvhdgZoS3~3J;>0Fx?DZ7D7&2O&aFRK z0A=L=2qumEna>{x2m}cr$&pliX>w3>6nA74h=`EqJSF*=$pyf{(W#y!1pSIkrcJpg z!<{X_8FHjFD9#PO1F;7eHhx=l+Q3Rx=-!<5^m>janug}ujuep}agNzBz!`zM($x1cw zh=F4OBX>M_YO(bFFL;5R-rC4E#0G(61TxrvHchTaQ5`T08qIrQ{y``@5gzXCU9}1< zmBDvt+`CF_FXFD4X_K_RthqPt?aG?(<<+3J_!L& z+^}RiGgdL*!s*&y{EkV;BPZ62&>|>}?{T9 zZ>p<1bNBLWfTlfwHNx2`1UJlK&a6wVS*IWdoHL2z{F%3~fa7q#&4kw48HNz6{tr|b z8+kZhow$%xErM0C(6zsjlh$#xJ6k8me*NNN!dYU|=;o{DmdE7=h$`%!5FhYW1`O8Qudwd@({dzT*h)79)^6EMn#ssMfai z3Y*lIZE&|*Hu&5t@2E!=$1+tS7SvtCw6(z4LEEChRNQssj6c|Pkv7acDE=uZDfSZ( zF@K*>k4{Aic>(<%>ZtkiMEc|-RowYd--YVLJW+z^Jrts8l4W)vFCEn!5fLif@lu9+ zL-7i%$Ax+gG3C~apkfHqaeauoxst_p1Vui}tI!jrMa87muj%{TreW^09K>d^k$Vn& zRT^qXzkXM9PuRBqZu$Cc_oPj@n0g`i3D{#C7dx9{h#$A*oEb%NWB2b;bn3_$zFEea z6Lm8u;x$499lyRMI=B^*a+vaa<{C|IxUa9H^j5U*@sCMH=O#SWtSY~^F6sOBRj3KZR!MaQaS2zJd}n{s zIUFz?8iDT?wDi`?+x$%l67rI3sL-?uR9&*5c;0GAi13pW@VSwE7@}9lfYOXaoK7MI zv-H{d9b$|{usFV|YbLf(w;pU4e=QwdzZI_KxLP-?o^hv;FKtsP(Y86}@4&=6DZeL~ zF-THB3Qn_)LE2F?c(B74h}7lBtRq#c(-M#C)vir_TW!&4F8}~S+%=pSd2u+YQ0rLx`t*U^z`|ddmp&vq zO!yp~Elz;sNE^dU69$vcWS1cwDDw$(!Zgzo>g2-QlRH!`P@?Le#Sx zLodm{Tt~NWh8}86vN@b%zx+_+P9nw@WNsy4ko%ls&_9)PV&a}A8q?>qw~DjFFC>60YoBKN|F z0y9!Qxgx{l_QeEZUv6c$`Lz-6Mt-kK`^Nphwj8(nA!J^U<@nk1 zH6vy7+Sc&~t4o`6!`5RVGY;i5Oyeo8-%ou&0 zbBNG&jDuWtsfWjQ9OQhleQvwHQsBtqqY*YNX)aFVo2qSg6kw=BbGJ!&ot>x#bYosG zU3^&lqRh;*kPUPXUiNDhYmK3?zGJr(ds65GR~^8Z9tBN^OZQyrQozrjilfIe7A5-8 zTp-qK6lX>)0>POT+bI~z-TQByj6hHv8Zs$1iE!f^U5ei3>s2dG(?vfY*l1yu(TV^lLc8+|t|kx+kwl{5qFZRa9K>wI~__ZZfVmM9m+NBJjRq_%eAp zANNXG;#qXSL;@SskId6Ev#XJ2y?mVGYV*l$7PQvhYmDDp6jZ!NF*XdO=4exlP#^k2 zJa;70ZeV>&qn;|-3K}Ew(FX|R^nCsfHQ;mc?|~-*Q|SEE`98cJZH{)2$t9|8GXMkiaaa>d^EPhQ*b7xtK$@P>(Ta@5J773yaevTd*bOfaRi-^!9BfMLqwfOt9nuNyfnMMqwk#zKfD zygM@wqGPOst8acT@U(65nU-%aTM2K;ex&t5{cc0<$F@}dqQ5|nf?k*cXGx{(V}I!T z{|6D#l7zoG1w0(NG^PPz)CKM8&-_D6{2%m;5d3$93urF~*e?ky8u{;Z*?<3oVSj3e z5Y%K0g`F`ZxkxnM?t}Oo%u%FDMP4f;YlL>g8``IsH7>3I7y72Bf!dpf!{)z<&CAp)NHNJa1Muhm^x7KB;oCJfagVXup z^Dbh2Uh>7su4@kRlH^0jER0g1$kJrA8=4#YX90xAJdjrf|8*SSwO8FUA4~i%AEo;p zb$!AGgut_ocIb~+?0^shy}kv$O+oMHFXKyoZcUdq_okcuc$yWI4p&Em$jfhEQV9!N z`R|ndn_>OOocC+-VBHjrFwNk*lv#$#^e~ra1afu$JN2J`b2M?ccmjU7=L-x#!`Ivb zS5rnnY0vUyC!&P#UlCNRv~O>G3FKGvwKwQROlUPvfpK=RI`tiirn9@SjSxm1@wO<# zX;6o%T&dI8%#V&W?cTn}r!Cu&o(}Rd-rhw4)FKFfRf8p_NDoy4Vjl2xLI8*sA(iN| zx3X{pk!=N8fagMDzeNB4K2C^0euk(kCH&vJo#u%@y}YWufO8KU7U!Z9Sz^Euph@L% zV3)RWEv|v5d=kDMM=M>}?KJ#_Lb9$>n|jJI4~f+D>ia0kd)a$^8ylZNq@NnPr_uiQ zYR%2_W#8|KmdkjSSW1?9Tr)=I(<|AT9ent&`TyZbr z{OSKee>9Bw|DUVY`2Pkm4~HL&c%jArd!vMQ*pvMloIAgG##s?ysD)Df<8W|lRDTDU zi()JBy>$8gjM4?rZW3c5#?KbgYPO?;yY26mhrV)IMTr7{#;_u$aq=2P5M=9G##*Sh zeXD$Zd)*k>^#8n9E<0$+$i4qT1J%X{X9-gTHhgF8z$3-@Omy<#K@*ub zt|JW72;ej`&>l}Eu@?Cgw(WRU|=M!JiE8D zl>&DDFxFv)8MV3}s8_%%yYDzV6|cL-kr*o&KT{2wJO^wkIA?8IVBY4O_}H0a6K$IN zpKI9g+lMf%K9oM;sZj`m&^kjSpRVr<$CW&8f_|Ei-#}dN-GqlFI|~F`ZuEW5E^&+S zIMcXm^Jk{LZR8RSDQb2KdF_HotWr&jiqfM*r-J>YEAWXehI(?C8AkxN==J4B=zply zPAAL$B*|2`ZeI(r%YVNWMLX(r3JuxxCS3Kmq`UVU%42<7zlmg?bjAw!;ql-s3W*<< z7Xtjgi`9(l-5mnd{6yp-@`jrOBVxh4?c`#bJi~wCU0qJ+bKRu2^S*)fXCc+_S^H_1 z^TjJ(-X)Gs-O4QuqOu1MZ1obT4ge;=?Bz(`ihhL?ZeV=XjdfXADq&u?!=0dVUNX<$ z%RA(4;-%J#GdWH%Kigb+!`H>7wju4x;h34hQv+M+;ELi9;JiS80X>z(;oD`N$B#Sj zqXOJ1;vrF#&+;v5(>_RBYA)T2cKn&Tcytkl z5Yxf)<@U2w{o{)lo4a_I_4UElOg2D5RI4Ih>Vj8;-GkKCns>K#%_M_rJ}S?e|1|88 z1DPscaS9{b#F>2HL;{VkxR%h=eC3TJ;8j{n>&=(JOP8b+AH>E19)2t3`^7Dk8rB1M zD#w6cLnS)4wlVJe3?1|Seq5~hBGNxvAm%_g9eAf%65!9l`x1*{DWyxQErXuj1y;m! zM9wF#4pr98k^0<|7_K|1mM*ZCzl%RJb!KB=1#psI#9~JB{nXQNYa`=Dxx{0hTtEv} zbnNjPzK=wqu%;(%r$v30J*0-Xe`^~f3iH7dNG@|}H0vaO%LQ=6`5tt#yfU>Zw&DBr z9)rZhhR+&(9cp83X4ICCCgVFsZda+hF*DYgf@*@Mw6K!0;j*>Q()W?Oq6J6MnK4Fa zx&Zs#I_?njGH~`aAOODZ67xDEw~F&2zg{P00V9efZA{aNi^L?nE;S*SEXQM^I~`qK zp_}Oy*U5BF)2BhbUHKuFqFO#7KrPHK&59f2DZ6%907fr3_bemBBOWiG{9^b ztm0u`aV86BNvxZh!)NrxR)$##_MSsE;}VMyRg`pdsGVY1ZfGR%`*##(UCMV|VWJvB z-2yi>9{-gi!>&T15fV6k>dX09lpqU9m5A$r;Y6tuZ)=V>7XcA^s9Rt$HpnQtAk)2R zBD8EveW!Kn6y;`RgG_wA=)fT>2&d6t;`_Et_Rw*yqk5-HS`U6%KXv(Pe{t-AAMpIs zTLrL8OP+a+P9+s>-q}y(@m`Wgm&~ponc3~QJHU-^hcziy1r8fW$_4}{k8lTYEy{|F z|A|XAJb)?1$Am(Tu8Xqn!C6`TQQ?xqdh|78t|3&Zk+rE26_V4;%X>hDgm!nAVLTlM zsit}Za#~RWke*yZQ1DRHi@>-SKWJ&yIEa_tuTRdj=2M21ItH(#mWo62fg{1=@)p1J zdB%Ad2lVT<`>tLAv@EAN*}NFb&(a(G5rk=nc{2N7a=%vJLKE9Z~7?<}9;@OMg zWk$-TIVLBIKeWD(qLFgu=41K9eNMno z5>@~&1G%5Us(`WkKWJv4oL^dF+KqB7OP{pAdv&7kw)}`0cYhHxY4CUq+JG6D`eXTP zG}kY*y5;~SMj{U*r`ffvSj9E`b6wk*&)_!<#Wq2aL!?-zPe8Kt9b{hf0O_(@{1rHx2l;% zA}R!LLSX-E9fZ|WOimk!jy}N^LBbR)^JL6Gw8_%3c1T%Y&R{QVG8Xo3EiT!ugTDZh zpV|w45w;jNQ+k^VDKx^B+7_*4(Zyq=nOC|tImrv@^B4sVuR;To!{;KTq^-XEZ~KQbNesucw(Bhz8BB6)hlay6uCsf}3>p}YMnIBSpgLTFnFUJ9yo zTvND>wT}f~v`q5V?s_h_o61q}v=?gOq%2_13Y&6VMDO{(nX~VEJ69im zy^HtDM=?vG%re-YXW%n0+z)|IFOufPcPxH3PzhZc->w6zMk3|o%tp=B%*wP!kCB6A zs>)zMjb!!m#F^F;PQ_P|Y2E_bR!a9e<(65Z-sE|fn1R3W8d?g_AA!x2yph(A5&n#N zN%f$+GkTG33xo%{c}g{Lm^4$GOJABeHmYaR@-5b7BU~cr8^bG9(8%awb2E*{Q7apg zEoB`tTkzlbSv~t}z5%u&NK;-wlQ1z&drydQ95L)X{T^-i0n(O<#0vC#BCgc{WUq!G z*4{ybasF1*Sf-?KN9pTf0`{Mz&;JZv0Cia|2C%b;~>Cc&d z-#WV(ye$43LYUy=!pFLjL{Z+TThMrzCj&wfC#?{o4%Zk(UcRrfJ$*K!R_uq*t162e ztl)IDamCAz$C&DC#?+QvFb8vDSPn?^17Icx^K&WjNSUX`5UCHg>T5wqxNu zN^Z-3CAEC?G#6U{KZXz@fWF3KBE+&KfHCaGpfQyI-L(U!Ur>yKF*29{R(|=vb32!Z zITJdW6(|G%a35muSdHV9->zB)y6w*kEo~cX=vJ>f%ePf&qmuY95A8gMU+lI0DuMbB zV(M)#>2Q%JaBPRzxu+VQkae2r@~W~F+A1Z$|&N&&BHDd<7HBA>gXFWBg<~-^=5o#W(d3z>z_{ zMz4yMHbpk?5HYUH7Oysb*2%2bi~RoVzqP(77m+DO@RhL!sd}t(u5GGRdMm_6x^JVx zzig$&`F)Gg_jS#?DqiVNr+9YIbY5^U55m}}mk^ATa@F*v^?9)8?E>9Z0G*780??$A zC+(!()-4eRraKc0oD3k*Y->FwWh?QzP3q@cqd1d*z4)*6>Pn|yjgh|Jo1c3cBg;Ab zL!YTcIs)9k2F{ljpL{nOJ_-ED6AXUx-w8eIBG3GK+deCq9_J&LS5oN{lUl!iWhYNu zJukG1EEiL_5f1GT8OJ;}N@wHl2n=`SOLLTg{U4M{tDvt_>^1HitD9Dt=@Xg)33FKLLyb{d#c-)B6e{JHY+cp&0w>$mgV#;BfJ za2KNRbK!G`6G;aC>u$`;^NhpGJtK@p#NqQ)4~il-$Y0p;Ch5%#BX;Jx&Xwso&p#`* zuZ#Q%JVwQIp{5#h`r;=Ad0nOvOF0*q!?)tl``pI=2h}Js8fb=HXgIN<@k6HP#M}v* zT0&Cz8wW+FszJ3^@6#`y)~@}SX!39_2HgyPLkA!DCyBvBPnkfwV`5#Xmx=f7;VjX` zp8SLADE@gToHdn`Q@|+gCkod)l!a(W;BI|oU5}zn($Fb82(eOnY8`L&LiuMEn}nr_ z(W8ol`d%*9Rs6)XZ5)!H=?eTNur4QIqy(VCutLFEJ$3n9?YPEHVA0ARJlU^)gZsnu z|Dda_<7ZXMDiU&Kue9yvh$(sT-nOcF4WVy>JSaphrhg0c)dWpwMdYS+8g=tef4*qj z^#Yb64e?lX=n&wz3q{NeBB41|T4K_Uu6anc&@Za7Q9FQqL^fa?RHOK}JB2z8t>$TY`)2X8Jht#F*+o&sCQ zPwtSx&vY0v>LM!F466Gz_*uR+P48Udo#DQx9Q;M+S*En(O|XRJksOP0wOp&`G7+2I z$9lu)1qd$A%zwHkSE4kdsQbCcx&(O2=RsEOhQOMO^zK8{2^@x|oQgD~9bUVZe>ukt z5ae%+@8lx*XFR#5>8pJTe6`kscGMqW1 znykUy#jDv?n?=DK_IfTiG(fQlKGyQxx_6&ywV(?051p{oCJ10g=7W^Mf1FnUf*2Kx9-^G)7BlmATd8*=QxRag@{Q-H$ps|QF z1R`q2iRE<@3gPhDv_^cJW+7M;1a_>#dr^vCRYR;`regrH#5-w_P+op7czR_o`bqtz z-=8BQkf%+g>)Ia2JcnqA>C)?CXNl8M^yG;%zUO*uN6%tX35P<3wQr26Z|n#gULD(H z1y*9bqHaZdD>Y7rg-c?SsUvq$zuJ9|D^#@G(xW@(r!LRmf{sPiqlaWxV~ItC06bTCbYF?%XoUcZ?b!RUdcTs%8VsG zPtTxkZ74EtVjGti?mr$QYjjmAUts4GJu0;%g>143Fcf?W!O!rxQg87@PNo-qw|(NZ zUEHwNn*8m$zgOQwtUZ#C5Y;fWpU?usROwM-4wfDJCr2RMPt*MEryEkAd}Rhj4K6)g z{q$QSJ1pJA)g1O~enTIqo>Um&ObDzkx`5d8Czl&AXx;6t|ukKb|nRJ-Il#OJnz|D#@V z^wc-Er}Pun*V79hLJ~?sP!tsQCxhkaORAO}HqtUiUnP*U5%uVp4g~HhBZYZ}GoqUb zG&!oA8PvX302A|zYA#Ae@Aq!$eZ^cz?DH&{mvNY7z)95}s!g7K>poIDYRNu}ov{A- z2Kf3sDC;c&if!xu{Gl_;?`ZJY>Q6R%sg^51zg|zV92Ua%n08F7HMqbTiI-nC7-={* zW!?-uqPC!|-*5y%3RUs%3oz;8b&cDN(f=h-o6&!%Ht&Hu7bM3gKkK@v*kS0tC;x)m z6XLTgx@e=J?epk{UCEQ({`jLQ;%!yZp8qJ_bF8l^kSV^QfM`aD+(J+6X#tP?Tg|)|fj~A9 zzDBZ+yux+=muA*VW+d0T_G}cw9x_SO#6saKX=5|hE>)U<=s!P`Mak)@mg4jgLCgk> zVc6dxLsNdtk-&c@ih`Ux<(^21hiV!-MD8{F{Ch4T{JmmpTasG!R|MBQX6-Uy^=h)JMxK^U=b*aVJY& zzURpfCEw2okP+#sZ`$8|>Usy5IUCGpN`C_X_}hsT-HzC2XDyHpiAfCwU}pv-k?yH4 z_qG*f$#jUROVQZ)4#3NI{i+udi3u;vVdQ7Ey&1+MHZyMm*~*j?y& zMvYpyF`oqVFpK$*SCQ9IS+B6DvHqq>-|{tiI5oTO%^ahsuMzD1@2=`jKF3|Ho$(Hz zxVu4a7+0`?X2E@4!IBn+AEEs}J9N7UA${F|b7Y>oLkK6s&JWd)`!+i>Z$ha)4BW!w zjq)d2tW4)D6xv(H{DIkL$4>3rhQheo(%AZo9@iOTw7fMJ3pn9pO?ISYc}M=s zrmaD57*PmQTl9R!RoWk8c=&6qrSoGFDNlM1)wh@yzxTT_jV9409bY9EDriOMp0tmD zp)>lnp`#)s<=i|6S#pYNPYqo24@M%>2-^w}jQhyMZX3GdMmV(Fu# zCVg2KOMdOyR91sbqHByG3TJL>PeCuml4MMh9Dk= zz>KkYm=x*icxwB=@Hwa;y|wutb=u~ygbO#qtfruMrZ#sFT16_SdrlVcDhg(G`egi? z@)0#*cIpi%G|IdV(1JpS^p-nd>_7P?8N+@IfC}cc!>mEyDS~CDw-h6-uavx;wYHJ2 zWwIzgyI3d*X$d~y)vnFvgVZ}cRK@Q+PIEjh_lxWN{-4%CGBV~$@k5t32KBnW5AeW% z$a%j$*-c@G)a0i(1jBLk6}VO=3l@_-#(wcT{&3}5=_-Ko%#}p{i42*`a8K22v2#n7 zkNlZQ8*uak$TNYi-r zQ+a#FO}~G<6qB{9g2V?^1KpD)G9(PYn?P5>(f=lfH6+N^Am+0J>o>IspM8U4XShC3v;u3@6Wj{vS6<#qI}<8PRE@E;mQ z-gvaK3sIuo2Tz3ga`e937Z4Z4ypz(~%jB9NbX^J!2E>AM1m6_5jO@M-TwwSi1s zTNuhSXX| zJmiZFd2IA0`zC?!t0hsM>X_ zN!-OZ*m6ov(l&vIq0*BlMXju`-Q$#(drwonC0R-OxL@^5p#l((^T+8$~A=tarLH2 zlGk+JmO&7#16|s@ILRxv*7{zmU|)Z9>oy4)Ypw_ESVkn}N=TBt4YE0j0R(6Lok0a? zYb?`9E80Xhq1-38M$9yW>vyj}4BnW~ zb)IuF4X78ITsFLD_MpfRE`?<826QtR^>M?D_3U8E`Z<*Cxr zQiwh-Go?L^C#$jZeUbAIL**PElxWpiN?sk{Jc z#yJO#3~$O-lKj!fYd6Nve6+ay&ePUL`o_A&vD{-2{`mgeZ!f3leh^qXLw5VA=cB_# z3h4CEE8s6|ZhQP~y|5+9n}1pN+Rdcg*G-u_|Mu^OAFlapL;2PF=}3+)=4bv<(~ChL z06n7{I5(O$R;UsS_PE0gG>|yiP$jw<1?H-4$6l@J7=~DeH=AV-_l!CI+Ld@N&UGQn z!vFPJdN;-9cd=-}HLm~e5i)v__$cof5x^%{ZH!I_ku$Gz+4grt>fMuYRFH+4Go4Vo2w)QvrTfL|~6x*1NBN(8Y z?YkB676exvT5(or87*vi-_#r{b@oKa&seFKyn(0pJu^MQQw*@v26OmPP_S2%G^QnQ z&SN8&7Xfta5bJ!LReO8dKjR@GlUshT$Ld?ldjVGB?&yY1|KU8oQ3*dcE>P$0l`}zu zvdcSh!;7lL>{4OMuwYbWQuRZ`fqMT07a5{gZ$kqO)(e+n_+QMr^gdfds4sM3&6N6c zs5!jI-Xek!@#daERs7xlc~j!ks!lG(4mIdcCpa_}ZPW0lo$mt=6UY~u?7qHw$0)4C zX}L5oYH%{Rk(YjB=3}qN&&*ieTUxSo&$QCMneO@f@3R@j{g)uNMGio~FcmpNQT2!^ z1YL*`=&SbNH~28YVR~z9&!TdloY(C?x2+tks6qjVgbq>N5ASAE4CXhMG3jNg;isk` zMn2fgHU1CkcM-%-aU|I9L-|0mkxMd4f>!l%vA4mO`(uyuLE?DGcgbIk(}5#Mo$2Vj zTu;@$7yE^^WOy~F6fHr*i&E<8{VWh`$wA?y*)^rcJ$N(t9UR2<-I*vv`hrPq`=hK)s3{aFvvD#6CI*yIi5t+bgDBEZ%;6v(|BB za(fHz)r3Db;n$!-JoRB7^LXxz&uLgGE3_HxI)ynv5=Fg^5e4v`&}vD~Iem65{uEZG z+of)+bN-=y_v-Wmc>`eZ-lh3JLY=}Q&gD7~(|YC(P|Y4}zaHKG9A)}$?9A3dLBX3o z_%8>F1TJy0KAi2PoM{r%ueb9pe*X3NQwNMv(Co)0GubqH1gkLMsF#M%r!Vm*eKr6@ zV9I66#-DMCKz@#(+7c7?aJRSmA_=jCyXX|nm2`)aPvLpujA&Z9g^MII0aG`8wkoIO zgT|Jdh+1urh@E@F$r}Uvedk$+)+n&}HvY0-`E^a}3rtCAsX5JMhf*8m&OJN6SC|65POeiUL& z0xvoOjx1W!_OlLs4aG>a?$Ugy_q3e7IV!aI$J^~k&N!&hd~?|bd#g9`18dq4;Y1?6 zm0jUFa*TQ#KSo64biXMUh8Te;b(V4cSkA@z25(=qb*@F4y+K5*<(c~WyupIh@qYCy zryrhhh4YPoI#>h{Bc(wjUzecsl7fJ**@r(uqThV8QVpMo+)Rl+w3$ZmSGn+NC`zW) z(xJHJ7ItH4SOk06!VW}P5g5m-SrYT~33dfupP@qb?{(0KxA%`>@l&%FHTn?AXz?>d z;ngqL)?Rt3|MWRdUwzzL5~A@;lab}GI6HWv+Sc;B8tz-6N;N&LpZ|e4Q3Nt5Nx3l~ zvFsiR^!&1-A9trw(~gm-gffGZOrz?@0swg4Lnt(Zt>?L0`00 z-S5n`8wPPDDfRvO+z_^OgNu0&=gfGPNCmCp?K*Yqp`nL^^Z8TwLeUN0h{>P@Ibvd{roF^WFK9 zD~N*}dDZjA6FzNLucy2lM^|<|Fz=o-F_~T0WE%b)5iQp=7D7xL(#IJRG{J?u5~?LV zfSuR@$A9kNnFy6rzA}OHkAAxfz-Wfv=gnIZDE&iQ6x)Yt+s$X93W?E8`5x|OM!oI&D= zBPmwvLu{}N#2&$(x0(bq6q>Iss_LT}F@flj&&|5UE=F3EDc!!QDz;$OohH5-d2xe$ z>YVA^*{^M=ABVJ5S1Tk!+{<7~8fmy&Ww>*+t&}#x+d-x2{M0sEiiX5Jt;=)n)#24^ z63@#A?pS^CAvB3k$>47>i6;vVbgBY$l>`v!E|9vDP_DE1C?JM_KyjjvXZU+uX=qmuXe3S{$4CwxmX ztfkT=1jiIMlwvhgrxXGVuE5rGfr4?-r90Zqj4{UK;*7~FG*I(9G1I@*M36>{Y>RxR zHhF%7XTjqo=)#-jS!8@p&3S7T*&$B5ko}?B==n6V_q?ci0YU_1p4XQ4gHju4t{eMM_dz9L zq!R$iIs@#>=j3U-8-vVqj7T+QdK1tb%9J%|2;yLjJKB{WI77~SG_8__37kuH>+B#+ z`DR1HMEB^NNU2!_JuwyHV&L=@TvK~eVXl^4FA`yTT!(%pYm3Ee3_`$@fbKOgKb;%jL4??Lz{ z--M&LaS`ct=Z{OCwmXV)e3S9y!|6MEN_VU=;!xoBIw4zcWkzm*E)Ic>hVAcfQ%6y{ zK4Gq9A3r?t?G?XcSmdLHTP=ve&)^QTO9*%8XIjam7{ILn(V(#8V+;%6iAu?05a@R7 zCSsj~UDL>!=AR8w@OVP~u~;X&d%Jj{rnaci5#$YzkjY57R&)8)FW>EqQFqy|h~zA= z;fJp;oF<>ltgO^VX}L&g8TdUFqAYlF$Sb1McxwOBy(xlHFeQsEnLW{}s!mwDC+ zc3PWE#Wc2OVZO^>e2_ou9O(C4!nSm6g#gDlwWK$|1r&P+wvSX~y-g6cG&EB=2X=w9 zxJ96zO(Tc_{sw_k{cs0qO$PfkYNkneuENxN0j~i4z`C1=lUNLr5t-IWK?Fd9A{u?IWBbzrr zT)4IUBL>!}r{Afpq;7p6G866x;E)&&zgtuxi$GQ0xC#d9H0{Pt$Mg9H5O$A?pb9W= zKC?=F0!j<%)Gd2zNZ3r>4**J}pH3w2W7tc(@?e9##AvVyKIlckDX)-MpMl03+N<%6lY z-o6CwxPdO@{dMiU6Ff!#5shw<;UeF#t)0#%K)Q9?jn|f;>dz{jeitBeJJ9odBR0kNQUGRnAdJ*#MT^^g7MTX};^}o;(Idi;9dO zR^9$|F_oGtgYwE2a(T-ex(EDTE*&NN4?v;-Mu}$oe?;cBy#S;-Ma@lxt(kY5Cb4** zA3Q6YM*0OXzuUhrpNklJ9}#Y$^b^Labz1yzv7c``Y)^KCn`-=HF%2L2`*cu%AuaVH zB)Q7Bu`(1HI0Lzjc;uZ9fW%gf^^D8_5+^b*(gST>a9~&=auYbO!msmsuUrQa^Mzo^jjZlcL-h5le5r!-c(uUpvydI~%zv({rh6&( z$*U0a%zb>9GATdo%i}LnS#r@kQ=xz-FXG+g-O@fG6Zr&9`C^jcJe0 zy!1jH z#V$Q~54*$U>Cx>*Ov!KrN|~3OE;VLg6Bi_UG^TdDw8cTwmne?ALwi|8q?B#1CHKjF znuaGJ6}CbZvXT3koY zj$tZwi_41_%cD?%OUCPMqlRuVYF^ae`cE(E7Vd{|YA~AD{Cj$(;aZVSp)?>Xiig^a? zOCNF+tQO(lKzwx$)&EgwzR-Ey$20vpdp9QIHR^K(f%@RU+Pu-%zRHiQ!lj=}U3|dl zTZ+~i=6B&~5;ugG{CV>y_L{E3hYmAp`s$?)n5mQy>K2fhUc)LI=|jmOFrbpn3)pY* zi|Y*69*~*vMzEiP#A82DS;JV?RR0Ghlfd98$Y4_y|6$Zrnnaae;7HffTvA?-vGl`2 zW|GQw4e3P-ryZ|uy-E#^qiFqg3 zr5$=$f;H_^ixUOweU)ZJOlvc%zktS{HnKS%GJ&tsqpmk@jsTj zbY1QLaCGM3P=45^=%aWZ@NM*~u8!fiPXt!i3goJDv3}Y#>hfkEU6=|{W``AY* zgb-rPDC;m|oy?f;@AUh}b#Yx>Jf3sTeeV1He!UafWgEOji*t>U0tF4>NlzcF5kl9M z<;|t*>g<#~GQ5foi3pt{q~2`GonblBW&8DOdMll^oaN^)j5;ml4l!4yQ+STI)!Ut9 z^~jQ41Ht}_cI$RMRk_LN{Yp-do>shJa#-pIrPa?O4VL0wdw;)c|4Qe8vWfCtK!$-} zXPC%8Zac$81b9&e2n9EqCXV|_;Q}6x#StI#4 zgs8{WV>X%n`Tv5dCTn(2R`Yt1{HDyT+y3p*c6Y#L_0+&QFyswGno_q)HWk{JkJ%1y zd0c<*kl{VfPe1A#=$B4I3@h=cUt{^TE=Qe)bboh<5ZL-V`ut-@1);XV2sq5JW@5z} z4mSd1T8KwyF!1wgbSesF7;o?_LhTd_7G>QUgolY3Y_Cq_K-N0=QTPW*yzWNwzU%u% zhz)MwsxuhIqrUI<(0F4y3UZN}{kd3zM3qP^yxw{pr6Q@iG3n8;I806T&rml$YArqx zVthG?we!|%#96#Vyv~%}UA1GEq)s5>kN^t2CyW{1{lcI{3@5p%N0l{#;*#H8&(LiR0p8!X{V`Sx2OWL5~dMR8toY z^xV=WrlOFbSR)kYOGHAeX%xh+#X&m8;UgGIzreX64}+X=BT-s<*~Mp$7+Wenm2qBk zDdbss3NajvhVE8aTnf!mRCtFKe_g@v$X@0&}5kmBJs~} z<`AownEb0ijV&|IuBs`y(yN@_A_bPOL`aP}_n~C^Lp_#MGl>wqB*n0qD+7y~8 zEd=~ipO5v{reA7cx3ywTN6Gb|JimN@6{=L?p22Z9=(#ykErS~hc2cD!fbNgn@4P}P z3@@chZ7+1apQY8%u1d9TBDS>XvnPp^DsE`Dw0tOA6EX{p9DtsRUNrf|D5z!W=L#7z z!lw_(zeZs3soO03cfj_WOIs$S#p3FUOCB?o@1F(Njzd_lwD?HFgY0%~RCqv^UU2qf zqqANYOh$sI>C2-0Q={GO%4rwJ*57UM`K*l)EkqRVMHvauaD>COYt&mrWcS7K?R)_f zg0na}v#S!Fc7E(ZU4*bx-FHogPvVXqR%5k47pq%q#>(as*H4pwd*;gZcLJVo9$v9h zK(4joaM#l*4x&%>IgNk>nH+7=?!I|D5w}lJxEQp1pbpaD3b&n;vgV|=!N9BC$T>wT z$8BH3Sx`l@f~+FSt-N}{NW7`q$I_<0@$uBRg8rbu`}J>T6=ak8Qx&2{E8%efyUXf0 zcjA+IC*!X{<+NL#$upBcgXknH1&6o3M78&qhT&ucgU^h~kwD<3oyOz+5@unvd@9nB z7E0K7g5z>ba!_V*KQ`zpv=kea*_gOE}0hLyt&W zTew2|?etwm8Xi1sFB%J2*vEjT$k@SCjJJn3r5$QINX5C)TMWZZ$CTS;mOyjgz;%&E z$A{fZa}N#-fA@0=qc>c-v2!CsM{ab7vuS(jL@J^t9%2|$u;d%m%=wM~ml$2AI8_w` zSv%;Hi7&Tn0I;Xt;W}_#xtC*F3ceAVlHebdnZbB?u zX~U!^_&al{jgmAGi1h3q$Dd-R+douM1@49tuLF~#V2$v4`0VZ=i~KWvp0IEN%goj) z9<#L+=9(Nto*6#GW9AzX3Qy*VNE>5@O2gLAOv*}@vMilr3gn+#LE~j@;A6{H{D=ld=_|es%FtWnBEDLEr%&(L0 zI{xIUr8@e}_eUS!d)#H)JAR#f-GHYE)`|2cVCPG*UgdbE@Lq2IcKM$TQF@$8h$; zoswTm?ZWlVA{Kg1rO*O zb)(ODQoPHEcbq90+2P!PTg#-gVYI2-MN_DR)*F<)-AF)@x4#-1>*!5}C9?-vU&J?BY%CdaBoKtB$6>tjlnB~srq zMdwJ{;V-X0mvSFWS8DckVf#zWkQ;O}y$#~RRZ8t9mvic?O1#2~^XlGB-JeP+e_1v) zW$|tr=tkUVb1qGf1 zodbOhY!41QXvyCGcuYjWSkknKFeE|?GzRpLx->3Y+IdPvigi8aI8ou3uuXio(2@h*yx=H&c;kKNheUu~C+ku(h2{oaaHPAXeI zI<+j!*=|SvdI+WN_AQ`G?J;aepMYgZm{GlV?NMECs~U6opA3^Ge~Vvl`PrwNrY%OSD8nODdj6f?@t>vIRrx> zjJO^k+xoIy0wkUM3IF@6sBuiWgX;WJ;})Aq3;J~n-@`bc^A;EMAwcZ3WQ0)m$eg4$ z^}a~agSqI3YNofz) zVvl-GWtGia`O-!z7LMu!m`*y;vU1oq6DP7PoJqEG99;%e|Pgl<_7P z@-b$5xtFlbikQ<|4^`#LZ$6=Z;@azu9tlN>cL(Yq_}UP8f-(dO_y%~#v|~8!zUF^$ z3K5d@M)*AS>PxOG^)avBKD%7pc(!lWCX;{g)j~ojH@~R1kMuCC9h6bY2;c_&yIsJ# zv7+aixwG1-yE?{W0`E$Z#w9^`d;` zMjj}VjG@6mpc;A|wC^JETI#@Bw*ca4|D`$yTTX2vB_D+wpJE9+2tD9|&jbAhsRCj* z`@5(+B#Hmf*-wrFu1~ie5&YQB-i}n0VJAR%?MKf@1P(PAhU$_GfZtFrdeSwu37$HO z$Kl>tK({Ganbg;LKR1S)4$S*bv78MkqSosC4{A^3n-*a)-(EIGj5x3D5%VA8xFb2i z`+o@dDo8%CP8`n~(wle6Z1*`Wwv?6{(9D))aF}YpiQvJPYN*b-DCI>s-9Jk#&bplpA z)5&$_`Y{_88=V5)6T0h2#T|Hb9>?lm=5@$*IUgvgHlp*LOiw}PQ=xAfQgwf)m(<=% z9OqCjmp`;FmpZFp%bI=b9+80O{aIF+6>7;MR&|vfJ|m@1#gpZ!v7Lx&=QX}khVr<9 z*c`as5eInVnEX3G^YRaWq~tO>KN>8l)ur#HDD=lp zS(dLKl$1cW2|w&5C3BDyd79aUtO4$y)^lY`H<4^u7ZjYKlVn3-R(x8{7K?lN;eLaW z1l8%K;b68-=r-`ocOj2>-=yb>zp`_$Yf=ye>FC-b*Uy}9j5+&V*gVhO)Yksq!xOx+ z&tmS}%n?LC_mxTi={f-Menj5!W2JQyoyiGw9kki}L@xr45ux_X2^+|l(_D!gAem2! zmj|Og-YO6xoEEepd|%-HHEyY$+MVGO&Gp0OLut17SYB=s<~ps$+w~2=tX3q}om7&_kz6x0B8|JAJ$PSVFX7(P=@Jy@oaL zcd#A@X{XtJ|I)tc7nswCnnz_M6~Uz}7@;Qe^IUCO*G8h0Yz_6~i-|3CLhCik!t$Qp zT!KbZt!2NQuwLYc;&UFmr%Rf%G;egk7d-xvfMh8RYXZ~}Y^;+ofIx$B<;e7uRmcZN z8jr}^AFZOH9zpp(n-OVciWQpN>XzfajxNpK;@l2b(9)^baSjUk+xH=6I`+Qvq%a{n z9$Y(f=PIo>$Yyk7S~uXnrLCnR=kOv>Amb5C@a)c0LnTx<^Io$FI;RLHmvzgqe@y`Lo}qrim_+X-HL(|pPU zVLXy{)0tv3$A?jC%D0Xi*nL8iBjVV-NLMxs>$34uhgz#>@dWRzR8vp6cF)>0fn`;} z^`psZBBtomr(0`v{i|Kl*9maU<{gZ<_CF4B_3C;RFM#jO$__B< z#deU4YFSq2*o`2}qk{s?q#<6$5t=4dpTtvDd=M9Lbrht%V=PE^+nTs62g5TYhX}ZJ zTmA6b#(I{gbo=gtd}n4fdv7V{hx-enLh# z#PmS=4be`ao+ABJxFbLNqbH1U*r(25m+d^w6rpVqotnbm-IqE^tEm_$?T(y1gKJ5! z>KG-&u#?9J#gm@vJs;b<)_N%&DC}PwhDGBa2<%7b>`0oPD_~zt8%+8l7gKe-Lq7!H zQz(*_Yc_SfPWByp^5s_FLs7`egMZjs{#VQ+SLl4SPrFYJJ^?`vB2mw*Z)ozut~*5~ zh7m&AV5NP!LW#>KDOsA8_f@iA(X?XzDM<`C?^Qj3G?WCjOv4 zmX%boKU7i@+0+ktYh9|pD`)GfHdSfU?UpQrQng3P{k}MJDCV!>sY5^7c6J)3vQb&y z$3seP+d7J$o;{Y+_2<`{_Nv^q;s!g#cRYm2OHQ-4FR!A%fE*d;ua5>aTxu9Yzexj% zbLxW7vyT=!Ci+(%COxq^1h+^N(d};#fN~J>-cEc#3>d}EX+Vep_;kK?lSIA~gE|t$6=2PSUe{$}fQjZu zx-pv`C6qL_HMY;P!e%CURu|#{ksC5nq|NPx9SLeO0l;6VV|lUgSp4t2Mp+3u$j+FM znXbh9zLmhVrs`-wO?}wH)z{l1eW4QJF@e1hkob@0=`AegsZX?uQ4bA))A;iQl5@9NTA$|yXZDs{=GkO`P5CG_w(H<*-aM=(A@-BdM~7T{mhf= zCY;}|$fO8{a0q(7FF17enDO#nOOy)s3chh;LU+v+rNgL%f|+|pFSZxUE%4{VpXq#q zP8nq3Wpv=l>xaW0$s|iHR#d`EF?=*K6$`TT5(*i%G;u1d2gk)cj*gJxLf8FhcRWAB&)-Yux(Y{d%73sjpm6X;OkLQmqgv;aP*WF3@MI$; zTD(Oyb8tkSs=1(Yr`Wc+V#;MC$l;WdMiMMo&CFnH`lsC2=!@K^8@|}U2LY!$!U@MkQ2k@4~$;IL(oUz=yii?(h1A?77GBKm#+=UT#Bh{e-b-1VLr2YJ z9*xEhAx_J%#;$OPz(;W|JR<`{`816o7H=Rxt(_C5#SG0aCbnKEqhOZJyOnx@1_VgliMvvNjl0Ncj#NXm7=?j-Ev0584Z@{|HE3e8S4X&T2R zoW*^YsdB_SM9(jNXLWTts}|QX_FQA9AzVDA1Cb)O0i>;J$r4}#K!?Y){GHwXWkP5Pj#m3^~5v%qrtg_x&wVC|&^Mu9~7W3EFFAF9W& z4^e&YbY2-dJ0GVy9wj9hb9TY?9dxsQVg5f~)}Sx1WAg&XSbplihEOeTR=3mq^7>WDj%7mkQ3e zGoWt$)RcS<>C0Wo=AK|Ml;p%}XHW}gA2IzTnEX~n7* ziBX8Jr|gtN^m>1r+rHI!gOc;*NLEzSxU^aLC{Nm;QEu|<71C4g4j5U5c5-iRHotCy zj02W$_OVjAW>Drc!ad3u=O!iN8D};iTU?sa@s3RpYJhx#Q4cOLZFeGs&IQTm1}ZO4im9_PIG^)1T*bSwckWV-rvDXAIuuPJ}}?aU04LsN&ZnNkp&6} zE1C>j#67Tq1!m^k=xxDMxN zMAcm19L+jDF`kU2RP@XVAuiSrJDX;9D$EZ_SP=q?X78tbnC&DcXo{9`&$rlae~_-r zn?FAKo8`DPLYkh+&Kk}$X#m7Zd1qViMnPth;r^?sk3A&9={#r+C~RvrHNs3pf*ze5 zwB!j}*_b5lwgK&{D8?b4c&qzQ<&SMilywu(tLDeG%b2uMRC+7obEfLzS_sY}UYAFD zBYZwBkNoN``K$y=EqABCA>ohJ2CJcn32+wz-4sr&=LU2?zaGgDV88FALUZuu9xS0| z?^?G6hEg|n?~Iqn3}CYCm70V>o7Be5F57rwJFcm@D%wexhajYnG^c+;y~j zoZXw}a;YXS@6`t$9^>!2adXze0-vcbn|-M|?Zh3{1!>6l93g1YkhW)74G-)KcX5M6 zrNdmO)A16Eb5d_bb-zK957B+PhVVkxkfCh}|9+O~a%=TL9-TfkNyJbWSWgQ$`qW=9 zO8DnHwdu%arVm|Mht|U~^q3bVJQ$Hnq<1|33U+Gx#0p3^hqx9tRR^u0+nS#i0ki{#Xa*cw@zF@emUP!@h9Eq_nQs6^`E*7 zM(^)ZAMt@sD9`p!1D~kNi-dI9grZw@S!lQCg>h&fsnbapf`(@-B+?B0%_K~&I+Q$k z4=-m3fv%ojjn!+$hcV+n##NBGqnx*679MS37VIzxqqfn z_9xo+oyq#-ny1@M3(zrDQNAYtM?{moEGB5B-3z=dU~Hj3PdX>zrTpLjASnZ9Qk%8N z*AM=hiOaJRG?(28H6v&2Pfe(BVe7=p;#6JZ!*XD22#)N#shAG#{hV@i854gICB;Zl zngcbzS6Wr{Ii@GIbKzmg;k)nMAE=+ntcA{FU7RORY>HX~)#BmwgS=VK$-c!~0CTPF zb+$C+`O{CLzG7~6c9v(jzrM8THxGgswba?@w`f!!Q-M+u&+vG|IQQMauxv%oDUDRY zJH4VMrVWhIma8w6>d5k_A3pPAN~J13ZyR1*2&U6y|1(n#IXTRU;K6u*#AtvUByWPf zTwThaDhjEO$xr>0=Kb-!Z_sp#Lzmv~td=`|@23>4v);Tvx0w)R&)(0hqwjb@9nj=e+aEW3Sy>YbQViVVsxrQ4Z8;4hf zgvRrED^l9N{b+kVVfEh8PWBJ_m$It`#D)M1iKc*93nXs*be$jS#7!^m^5^f?(mqr4 z&$UwKy0FQQA~k9$^%X>3-d>_O5a=&80~gIv$IAw0$CAT}x0d&drml+>+4toiZEUQx zZ2J8?%IMQUhQT%ZapreKbv$_#JYSB38l*?slYAFiAl-t&JIWLRl$mSuW%6BGfffa{ z^F!to=T6C!Ce;o3D9UWbd+#ubcGa9I@or=dbSJ=8}Q z=$z%I;Gzw*dBMGrzN&Hrt%komQmb^VP9+M-vlF`ZwrT|Eoo*86Y??Aq^&&!>L=($T zg?|`pESqy{#{u^ce(;>B_?g;51P2|G593Sq=x#H@JJLkw%aFZd9#2}~ZyDwzC3#lz zJ=mjUWuM|#PT#jfoxZGE53kM6!ds))-4C|4z7`-HBRl3O*A3wygR6~x6}`2aOS_{@ zhR;__*-v4g=A=F53cFeCfdm3;;F{~@jcZpNm!8m6IFdK*o5C_OIN!JR#zxN%UID9pMG|Oe zJ;C=XT_co#Dh3K|FmGUr%`p^zU_7LfN(P8?-01M@fThEIbFov8biN@{a9mSI@ztf) zvD(ivc|K~JD{~L|*CB-VSGPR7b-!}u^}Ih0uTo42_}!L)tQN1lUFOYzcdOTx^gkc# z`Tedj`_)VLyS7o-t2p4EBb4OVl$U^4;_z^3t7W+zVnq64;iYc+_u0YEV zQ%hM`O3J%3qo!hG>a^_a2^xxcVD5=cVVAlSrCTQV$8UTlF#C$d3deEqBOfJPiXM`z19$v(CF*f(R%iF ztIH{uk*6okapQV|H#^|Vh>o<$3g!_`f^&>s2tyELDS#!(hYkd*>6?>^&=OLp_SZGD z>Iom&k$V{;_@$uU-1FvKI9sL=y<# z$0HhhRQ*|tVCLYDzpu^Pcy^?X0t7)~8C<3ug*vakEg`|Lw?k?Yge3N*ti!nIut^@g zE;70R_Ijr;++&8+#7a%O)5lH(!V(xDyDfma?W_`tsyy(o<68cz~3>@ z-E36iUvjdBP$6~u^L$H$y=YlUeb%Tm#45h}s!4f0SiwiZIrl+b=!P58D}#Q$HucN?AEqhY z>g&j7gXD|kXS`JaNX=FS$&X=5s<2) zpF(Hu>WuGzeNk=|qk`nwMn!!qU(%iv`a?ZURbj~O$M@w6d1HO6MuL%&-Z^8*DRDTk*|6-xCJZq-``jAmw{e8)9qcej<18#qI{ueCD}NY;`w3kdmf8-%Bx) zqwU#@@4WJ^onW4z<#{J;iziJ7d~$d7D4v0p*!2%+G&tGagVtO1!P=-lXfg1h&JmGo zzX)^DVhu@=3K;qC7-TtZ3q{$jE4*(cG+b(Mjc_2`mnV^h60;0-eK+&DY-x&_AMEPE zUDKRxYk@YRDNezn8^PR?ap(U^ky*4!&fpCA+&RbMDmDl@a{$ zYqd`xM2(bKDwZAF$=WWva?>cdkoE&Ssbk+;HOm|?$bOv4N}x;=X0tyo*N;H+Tv}r9 zs^0&)bL)`ZjiIlJ;4!4Xs8LrxeowYWnuSw!(eLtU-7{>Eq<^Zh7|QAG_WC)|K^>T3 z1nRsA1ka0y2tnW|dkv{>r0PSg^On9GF>1~auNI|0t&dYd{!$PKy;AwB9o;VqIhKEX zLV5h@bG8}F=okMQ=s`mcXBBG*`^YyAJJadMuFX*zL54EdsLFM3Xn_}M8r#QPO-j9_ zq(&5lJ%Op8ma@~2Rl3LZW1nnn6ehh=f zU2vxvGsZDMDTjDzLE3B3L$)LK7+W_yU9nviiXNuuvPJRI!pzo|&cT@+oOBkK zA!xqn)FX_i7Qp=S@9Urd# z{uc1~sA#OaF?`&)GL1~4D>1*%)n<2X{G8a01HYerWWQKNY@{Bo#)9T3R6_~r&})w; zA-Gqaj{i_93(AeS;4MrJE?TO-ZyEY=(m|Z?_v*TsIM@lVCAdVWz{<52+ebhfWT3Z2x)g4UH@m9?@PE%myR9+(3 zoSabXsDCb8+U*;JiKgF)?9UfN-q}deo9N_C4U;U-Ch8jsT&5bufAXA4^fnTs`aKDL zzpku-;-`gN6RM9}wf=F1i!0+|=HyJb;eQ0jbE?t^wPR=OY|DQix45>s@bD3xU&qiK z+yfxCJO+Tsx-6zA2YwdW`3#`^0tyUNmE|c9rgZ5_(Z{^v>sd(Uv+?vM2vG`Xqac=n zkWEHmsaO)697-(&&Ekrm;I$ZKsD@B45Lwly>;YU|hMHOu;y#}S^VDvb@~uvHUODR( z($Jv9mj3Df)gJx<>pII%FMN+f1T5k0{~#FFW$1CF21N_Jz@X!Q@J=%=4asTXSG0QJ z`Y^`J)!&1(5x>8l6nw1wmy;bsi9p+^HS`naTiQzDpG9x%awgoXP3eUq{ZhMG!=*%A zqPJXA2EJEyN|QTu3fNu~c2_t$We>fTP2TjMZjyQzr6D^~r?cG>;cKOEJ;|iGHS<#4 zxO-Kz)rq~U)lsNmZ;ub+cU%oAF=*d)jmdTHAOE>KT{M4v1hP`G_z`^IaCvy!QoPLa z*0{>r_BXYzFTd7(H|iAV?}7Z})Sa8ZPPeu-Hq_dZI8|~D)L*(b#Lc@wV2b??R3=aU zKN_l#ZaBpp!SXRA=MBv|?VVov00+Q#j3}{(dMlABzTALx_YWn0J27O@@M^7|?VM}r zZO?C!4>UQho^05JPeE`}A3jahJVW;Li$50WZ(e#N1qI4h9esB?y`Ui3UW0LmLcA< z-<5{e$BhHn4;#$|rM_I0YUN7DzaI?npLhmIF(fhq&W3I;{n9nq)&U{Vx$Fc2Pl;^E zwyZ0>0f<;7)FP^qr4D$VJLvOmW)O@27#4ZkRoAFnbTK3F8F(I~(d3=QxvdV^#DoZR zJ81tkb%l3xNq4!=lSNl|@%^Bw`ZTQm++xL**(f6>*d(q1%SwDsWEWyp#_ZsIJ73>A z+ckdlo8hbgK4bF?Z7YT|My0wWdBb)GzKBmOgz@ zITavi#~apkDX)loy_SAcEu{QDmH1|r$Jl!<|3bHHzcbmUS8r;ALa|-uKH~6p5%tOJ zhtj2oR{p^S7`5-pz1lK=CtJA(;bf6_w~qYIgj@r6HM+@ z!W{Y#HA7@k>Z{#RS_oLzJ>T24@l+t@n^)-IRSMceYblW)NW5!5T9HDTo zGA#*7$is>Oy`@awtR7_|4dWh8sp-V=F|MgRF2KcJrt~f*F0{t@@tjkBlRQ-fd{w1< z7s`&C*C~j<%h5*{Z!Q?~#&(d0pXr5%EV$8Q^ICVlL8{r0#$bvICSu9z9PSGIRSD+l zql@T;Q)+p4t3;CDHEo=+W#dlO6SSr2^fOs=*biWUtu6urj9Q+KC9;#UUpYGz< zl)&F#oRITVkZ>kYg-7}C9Oy-a#pXclI=e%t2?N;jsKnOmOk>0YY^S~(jR*U{FFFiM z=A-1_qWvVGj?(tJjl^nb5hbhXOZHABMBGvDu*l5$PdK6e;mi9akAuog*A2}ahKX5+ ze~`&fo{yruGXOZu)omLJV`PFzJ zP;|m@QsX;tvH6FXAkqb_nn=D#Fdb$O%0J|(g%xQ(POIo_Ub&wyg!D!UdSC5TNx5Y? z=DZeWl&duU$*e3vXsW5}l ztfeW?jqdMB2{%!93+I8Xz>)l`BEd+en_zBI8}=t{{^oa$7p*1_$+CK31DYSt1-i;s z4zg_KljZ0qFlbsX{n|R(Lz< zCtc0yTI^^pf5rBkYXp7mR-9bjD52i*#MH3XIK=3?a!OI$E!yqW*LX#(!;YIKwHj;| ztRIg*O2_|a%{VK2$SDH5KjYfqgu)~T>n=oz>{Ihz>@KiOf9PbSi9p84uy{szeq^p5 z0MbwL(oR|2IEcX#l2(d2@pJ`@tp)pe@)jnTdiwftIW_wyXFjHXd(k0&=-6R)_yVua zYV0sa)ysvi&RcyddZh1oCtTAHdWo_!ur zN(|SUF@Yj0KGSWnokmE%Q=E~%^4aQ~f2Qp|ZQuL`kO_hGGoGOPz+RGE$3A~n`)Rup z2A1(Vh%@0&_8(BA6N>29nV31Z#`^3pqy6wWOc>b5J#N!r^YIBx0|xBiBkmsWR6e01 z&gbq}aLTO?B!AEl4+b?H^!JZ>nHv?}g%BEghc2+RelEcIVn?AdXQ|MPI{CcrqBdNpp+K3*zKe+ zg006OKio#!X*O!o`NB&E`~n4q^+JRB|6cv@+|C$2#8f_r!uqBiQm+8XNkdcd-5hBiNWKw$?`5+WAKq zA>KNKL!urOhR0;uDTMz>lx!u9n>umz#kpdJ0PS0;Rg>Az+M&lxBiw8C1v7dJidX&g z-LhIE?dLvC$17Pjoc^hU10v^`MxH~Dad*LQaLoP0^?t;~k6V8QX`Xpb)9~wR50SM`O(>ert5P zt^EpukKrx&M+Go-Y2lQH4*nSs_C{f=;qHMNdWXbUnGz#Al9tw+>&m5m6~!Y6R7-v8 z`dQvw&>D0)Nij=_G>RNxDu8dD=$_*+P$K3}h7*v9LHVrcfbk*gky=g!ABy{cpIpt=}Cp3b{OHQKOFX6Z~ z()gCdatEV{ry?aW9F!xp%-(VP;GN|9{oM96`6NyA2jvU(St_+NUTVDt)rxx-XF)aC zM%>`1z1epdI!a4&rXHTI3Vk=RlWS(5`Sp>PsNIhKt9Osmu-UD@8~-L~uzdR?ePrVZ zhuqr^&WwP70-qI>-<>|-R`p|YzyV!bhIvGQN+-s5t0d&fj~Vb1P~bjPgX`Hy5}Qc& zsne-_h2?fy)}9p@ZWS-Pzv*oM4R^lwN2}R;r~8ir{o>b;`M%{h7-NAmj)m;1 z9poYJ;u4Ths}fxY1ViC^=C~Nw5As_zsfl|xsFfq+ifJal)vc?w-*?K42Fyn_wnw+~ z`Mq{sZNIPGOYx|!LUU!`@;Pz21M$2>HpAlXw7F7d<;N#o!J9p=HM$Jnc-+Xa**x_S z8jh;G%W}&4*W6jSE>A&BVc$auXkc($A`v>WuFw)lj+l2$D!9wz;LYFUM+DQ8Khngg z!;T+Uiq@Xz7NsP3zG%Ix{$l!+>G0Vj^S6Z|bBhAr8zYfg7>D6Vf%{mP1`csQ3&EiY zx~+)bTvR8!k7-gUmA6|g&?`(h!Qb`?)&p6v?Z<-EfbeAI@fO8hKeQdbqNs#WRTITy?|V(dZT$J?t?ifu6> zK7JJo;R>Fmrv=0&bd(l1QfnxDkp}V-lzyrVJ%I;8K)pmomXCLFauDh^6 zQ!TDv&wR8I!3A`Ri>EuOl?qQDRUangKxa)MmYHG0OT>)`V!%U7KE%s z(oWtYlPUS2^ow!B1hoBgSbAHg@nQ7FgFrLf2=w~p89LQ8pEtCPF7PNX#amWVD`GV) z^+gS?MGz$&E{$ux1e#Zeeq5!1#c=i&s)Sx#>xn#L-skObPzmT@?_q!H@8S=dS}o|e zmed*OgIBmaDiB5tl$hV(S1muGTf^UA!rLhV7{f7!<6F;XNhx=oxoF|N+G6c(^E6+X z(Yk}MF0vDzZZ!YkKiJkPlarf@*kz;T)34g2YnU9r2?M}XDW5&R7%u}%#?kYZ`7CJr zMIzT#bfTm6UUwv4W8Cm^R#H*!5lQ#swU1FB95~7#y2er8f`1#PoMW_vB}bf9%`-M?Jcword*ej0KG>hm^9}Jn z@@+V(HXap6X?^smeO^&xPGGrtwkk+XOy|yMkfX}gALYKqQn%0A8U*l3l$W!N07Rsb ztXIA;xU3d9*Q0vA@tfcK&Tyb#z+7q?@rZ&lmve67{O8F=fgPW7hRos|59D|k4_%CY z3h&VughiQ>e@;I$-Qv`~%-1;GBM$%5dG%2i7gPvIzn`v)^?^#=;-7b(j!OQScRPb8 z1Pb_AXspT5_44y4qIexF}mJ<*OcE2(EPLN zXT0mKASU+;;)pknVhc2amcWgeQ>Vh~h2&IfMa@_6Gk-Yq8>oc4EK>)FOh8**agrL| zj&SaIq@5v&>dNP2=oZmjFDH&0o}V_!?73$mlzno-9Rf3O1IG)ufg+(N_ZDiTj)R<* z%+Rim43_d`& zn4lwRnW&cUOdwOEwNV>(8^$rGKk`GYUPz>K1vnWQ`e~N~ATYU= zpzavPTN@kd7y^Da9&ky(#3SZpWFH73u4A5w{DI`bSb*S6?>(*j>RV~96yF+>txTpV zErR6Czp}3!U5qh_Bf9?f+OLQum-}h98Bu>!DWzpfXK>t0=`czMKdgT`^2-&_E2N-o z{{fWQy_!|J^4Ty~0Q!n=SxPfUTp<@rEXDEcAmi<&DyRzf<5#tsf1IRd)la=^@X`F% z)29&LI6m>E{%EfVd{LgkFF*gNw~5iE(h>77O=m(0X_K2W9*pu;+4OSjojbl&VvDU)`)*qy=wHne5BL^d;6U~{3U$mB6ROLcangGU! z*6?nL(i=#$f_GoJujg1+A0704%JP$?-1E;RlI19HfsB$2+V|jwN-aC-?b{_HuM=RQ*ErXp+)7KK_NOyLOiLCce9- z5>C7ll=;WxRNB@FRq^MD=Wt{X@J>rv5Tke&7E@Uk9u2{P6}&wx2>AJ*`9cyxH|>OV zWcYa1Zc`Z#i#{w(@9v$5Lf$cgIh-RRXfncN(jCu%lT`$jDL`wWuD|t_^(@Kk*2BkR zF3kV;S`E{uvr)Ew`O9xt@1**{Wx2Z@h%}pzPp=$$p4=D3)+b@u>LvSFYBTEMOHtno zvELwbxFPd4%XY^G9ya<+m-GxA&Kad&@V|e>*&w{m-`H%L>N|O@%|hf%M|83cd_9Ex z;cZ0+jwZ9n5D*~9Kiv)|ixbUy^QHbV-Za1IU#WHQp7!HHm)y+mZs!{-H$FaM<=Wk# z$7zp0u_2W8TXt&ry)+`Opz_QTPUAsFY2Jf$0~Gz+edAlxJv_jEZD-Z zw`tKTeb}i$JQyKOJoTU{5&R~Hb`z*(cn992-cz$R^K(bNLS|XV4Akchg>P4mDSK45 zw2m8?lp1U3x4bB7tZf?g5c-;?Esc0uza&#%m$cX;RA1kXl}!o{dMqd-|E235&x@x8 z1`7~}`H&435Qs<=N#1tVsVGQy-+v}Vb}`u_9dF%uB1&LOE^uRz4PG+lRv+e32Kk^v zx^sq-Fy{)<4&BX;>opa$S9YTs+ydsKG4+L?R>j)Zw8VU0x+QBRlvgMhSjawPeHecy zZiVIG4C4f79tGB-lBV22^d_irB`G|JY5gv>Nn|hb=lFzq#Ca8%S%`b_eZBfLzaMXTj=)F3=daP{V$;*3+AND#7IYCqb zM=28sC2AKt{C@rqGBe9qCC>HWj?f%z8k5cYy1OT)T6pdpDb#c-EO@4T7!P~hx;Q7ZQCE%`$y$L>7$>_AD;In zibrcv=CJ+c%)TGuFPlhrqbybXMRb>+v263M6;t3h4nAVNu+g}h$oR5z^WPju9zFTo z4BdF=rpi~h8Ynxsrz$4xB}C4)6C7*r+Umb+Ojh{4<6C*8hF5C6bGb10|0p{5c&6X? zk56coSPnU}%Bh1xPC3n{lS7iu&Q?iIv4qTNa|k)E_{0Y}eIh!roJ$ynk!nsk#%!x7 zWGm*K&3?c8`&WPU_Skmc_xrxC>-BoR&WmpT1Brspr3ZK!o()9qAsi5PAr**0SZSz* zCma$Eg)|vR1uCfT=A3)<@@L*gejz%E!?z#1{cX4n`x3hQ>4*wh5eixL0)gIUI8~+7 z#}(nf)z}Eh!Fh%vMD1!|%mj`=cD)_ioQ9v{QSz;I9N}AHh2+)Q$bOBCi+qKdi>AS^ z(*%u|uK@Jk>hTG50UZS+-2byW0nQNZu$83=)K{q(^(Q`y&UeJ=P~<8@$pSeaBV!|8 zB=K@d5}d~)t7MbebA8_>YJgM3v9+!K^X=8^kJYDHecyZ8t%?2%(v8N)57_74_Mb#g z!?M3S7F4|{`Rx{Bps%l8@-ymNUQofWTaPMAVhXDCP%TFNbj{yqJ|21RR-x;t40|9u zYJJxZq%f%Mpd;yjU5Qn3?_NR`-|;~?V{_ds4@T0`IsE_h}aMr&3LO!SQO zTzchen1~(r%3xbXegZ1B>o)Vq5~2vWNmnz}(L3-yyveqakafhK*qE)R{|%xSOLVsW zYr@C|2CymH{hM2a2ghQs`&4>Zx8!SWWajonU%qE19fh6bEKUcvNvhtv_s&M2*mayZ zU#>2@)rPX>9Nt{w1#@UrOg09Ak}bnKvxt4g$e!qux|afMD%Db|l)0jIa~ONJrG9uV zuO+$7cG$}}a9kn!{?%MBhuyj`_;<`5TPfJ`T_Yo-D4N*a=mju0nb3N~)N2ay@!uJ@ zBw-(_8~X0>8yDY{lRF+pZNIYp?B0to^mk~BuMash4L}%;1Yx+$^k$6=k9_YQ@ zFQQHyxo(y!FlX#&3ua<=``6>6{iQVv$SV;Ae&g5MFZ$r~{9PMh@klAO*&ADNv78+< zy1i_fqghnU2jJ4`Y<5k``}zc?eHZb}?lyQCpas#$kTMaAF7R4Vg14M&Q;!Z`v|kpv zOLd`egioI>O`j`|st9aVf9mrpz`g1IGl`GJmqu~>+ku;I;UbVsyBs|o5pc&^6RgiN{;~FT?{U7j@}1tgx+aSJuF4cdut{l_Exe5$qzXi)6^>oq6Od7ZAMH`5 zrJ3Fwsc|EG^yRI9HsksB>~8pdMZL4E9n<>;WQ<^C>NK)IWC&%(H^i&(67$)mR27CS zeuGxXNNHrmN9z{Of@-8j$By&C<;I`B65Y4i!zKtsfN`LFYhXaLLVl!zWt;5x$!;y_ z^Y8=^MN4MBn<6&`d$?|vJ|NYO|j`m@F-)2huwCA*^Ly%tMsK?xlMRUott>Kr> z?@H2aD|>OtzgiB1hxXM*UB%|_u1MUKxorn9B*(+xLkUpsQt3IcCQ91i{W`dEbHI~*P4Sg=hBjUf#|p`%T~k>11M$M5*K#U_Cxw(!QM>p>Zp%yL3{{u3`b_K!|pMZ+$bFttM$6!Y6w ziNMLnXi?~;WMfrVu5`d+&BkJVITBQFwZ_qK1DRNt5tDyo&3p}d4CL|mgmKXeRJ4Rw zjC<*6dyO%l_KvspG1awdn!jz}T@=n~Q1ReZ`jDJU zY_vvvudBlQ^|aw`aKC8ql-1=K?cQ6k4WuHm>pI~eG{JJoY@|<{ z@y={LBnH&%!2j{C1Jo-vC870d601hWJR)r^##wT4&J2pvD}6BJ-n+LwxpannGwPPK z?m@Ft@uE*~Gf#y(YhKVbJHWL+@g!v6JSJm@KqL_Zu<0F2qcP+X~ z|75c+XNbf9g=va5L3Vr`p>KKvi11PGfa@h z8&JJ>ap3obhXd93awbyS4we6y*3YV>GPf{iMsU|g)`|{}jE^XhUxkF^jch+1OAlZ6 z8@PiQ!cNkJJN|NVd8u1~m*dBn5oIm(Ev+Vbi*b=O5U-1uWpzBQ_6pd{SSsAlC8Yw= z_XKmJ8z-@2dsPe<0Zb;QRx8_MR!%Is`sdvWK3y19&Nq%H+#`IZ)RU70K*}Pn3V1B) zf}L?88{HuGNxqol(e<*HQS5jnPtES)a2rbb&qYTM2Lq>?@u&C{mWS*2%f5eH^zvXE zL`c`5uhYrR$Uzqs`sr1XfaZT7)JI9}rv=T_T8Cm`Lp zZrkq1+QHoOS|;iO^SB}mc@P3gb4q+X-f~U?yIVL}+W0INcl&+wIF|DU6&8b22{~9- z+5D3DyierL`#lbdx;rd1WP@1tlSeXED*<(;^h%8$YYX>Uq zxxczH-_#t4*mBd2a$Lwi4&0+UU2$i8;u8vUJI_%2O&8f&^xpLDfX(p{P zr+((xh4}=SwsF5q52~nynZ&F6@#TaQEy((f72`6X{x%+N15hCs4yaKzeiWW^>Cn@A zIf>8j6>cAEvtWX94mA4k*-Mx7wz8fxw7%pSwoK#p%bdoAhJ-zo3cIq^NH7ef)Qgjr z&xe?N#4y#!eOW&2=2zh}9*gU%^7vA(AzhQ0gKdt{qozeui&uG77e`;Q65UpAy=ip& zS`d6HP~;M?eye(&KAm(GvU{~8M2xm+=yc=a!JKVspikNFan zg`;Sl&#^Z!O?#L#mq;gl8{f!0TW^-O%iVbf#xa1jHbA!F1dp^;*~E~Hy*c`99IXtg z%p+-x%o45oME?2Fo*K>)bME=`MM_A=>)swi2UgK4LU-FCc=3TRr|+Vk*>wJInhQhk zj+NVbN(sb0z`#3ng@c7iT04*w?-E?90RYttwSk+j7;PWv6x$_u_FhMngJb$a^Pl zJJdk!*f(kd;ZUKR&^Cw2>K|#9Q2O+*zxv@z`HnG8wod3@NB*rybc|^@{N>@i8geBJ zY(^cZ1vJ8u&=C%iMi2qY>Y*d0ll=STJ#U4986WRm+$ zJbIFuTk_-{ol?EAFCJDksyZLDFebY?sxw;B7r!U|{pH$ks6DpFk=bFnL)&Z(b3in; zu4m@kD<^1L&D!_0e;Zw~Jp3(Qwfg4{k(m;6o#&mutkmtuAuzRrFpK5w&faki35)jh zv+JmheQ<9^_`1?psW0na33#;ufTXy$QZf&2kwZEeXOeS&Q7 z6d72WOzNlIzV|a)jgk@r6MeX|&&V1N*I{ICSS-v=vw5A2yuY^>Xi@>jp}>DhIy!K61}_unf#;DxJWn~%-x)e++OiV#>L+8nPy)Po zk9X`{q)6zZd@B}PN{%66sojGgWw43)8$zYJn79+G)DDG_HA$mE*C3^=skW4rl?eqW z4Opxf|4PfUG&t2UDQGdmpo?%nfDLm`I&(kF8G-h>H^^JIq)_9IxU{s4ER4P)sK)9- zG1cC)o|qKWRIjk6b}6(AAEKwh-*>q80em5H=$8^|!76QGeJM9P)qFq`2D&`!8alZo zZ@Y7#OgS$??>N4+v}n_3K|lB9<^~K|LF709;S9hCEW7dV3XTC1^UP_kb{&WA#F^;9 zZePlQYs`ECVx<-#}A>>$iHr8jqF2EYtH|mzX;6n=WX0;MYWccqaOtjQ7-PuwfUxBe*YM6 z**M1TFBHxV%qRY`50&uLe1^=uUbYK+VcB8rlTDw9g_x9pM?Wv7G|yI4+Mtt9_S4|G zO9LNa4#B$37fa{fgdqt_pj*6aVI26rNDL6}4ng+`pD zveZ_4R;QNwG`dO-jti_p=U4Z}lkG_*C1^E43yoC(n}3*VVPB zA1Q6<(`Pzd_2O2a-%Xx&2B~RFd~EE6F|-k3h`47Ri-dTqPEEslrjjS-W}HVFCAFzv zn}=eJD+AuV9$9-CYtV6!(>!&etd5C zsL87!_m7$5~7m{T`~Mn{-pJC*s1a5dYo+INdnUbDJrBPVH3Q6f#eJ({A4)i^@B&* zOVEUzJKggteHvEAqp$EPRijt7o;pbO5dA?@p)@KiNcZ)i>ZR33|0M0(hok^ymO%O` z&z@7dBFm&C{cq5NNS9z6Osmi+(v6rY(!5_hxp4-E^hb&dv<%W#P3WZ3Mk#VwfKL#! z*vfOb&15>|ae06+`14lHn9FdF#={v~kZv8A*L+O5_Nh-5L(O(%^n80=eY-IgGxKVB zAx}c3+FQk6EF;Qyw@b|6`wt`b9%AzQ5Fi(@X!h`iXTyl8RlCC3w@nS;;0XdWymrf9 z0?cHv!73}QCOE|4s)b~QnXH{wr|s`6pFP0d#d7RQlfC62Ah!`0g^bIf3F@>#25K*} z_y`Nc%I}4UJ*^QPP0V)3Q;-|D!VwJdrWq1Y;rmtR*7o|w^dN`vAh(2!50@UxiEM&G zOYG`aG36$520l+&A#+yUw~88pFHCTktC!k4IOCjNV&!B;#Y%QaQ)}^r5eARIsiJ1# z8Q#YVSnXl7$jub;%l_aija~-MX&PpEhLr^{p?`;gKpjfsV`49J^Sf{vmZFd!3rn57 z%cl~Ka@KO1zPg8$U%UN*aqAsv(Cft9cZm#-B9FS))*f{!$ zdC_gsPO%KpPVootUEE3x{{(?fklHFV-6D1q5KwR#-k{4{g!=-5Q;%&*wR4F!P{gz2 z|3T?tRIng6^>dD=xLr3{iE^0$oF^8g5ReXRY#IFBs3q` z%ur8#*A|TptiECcUv?2lV^44vfCp6&xb+(48zc_kD#YuEsZy5Tifomgxr1IULC5p~ z2ix6$;G(-qLHT$SfA;Q(&Ijig{?R@u2$^)4AfM}(u?`|2)DLqFNT@5-A76qv#8x30 z$kKjB?q0DLHn1fAu6x=3_L?`rS}7|`bUBfA(lU-~ z%=3m$a^bE}F8_=`4fNTIu)FXPO_XGeWFt;`(qCO_qC<;XO$)8+;<~du0-A#DI**0s zMo$xRcpJZGpU!+4M2USBT{9L+4D!#q3r9*D7nWB;16T-=AiqvzopWb9;*xVpboOK7 zd>p1v-MF+DjpiOp`>*%_?`QY>I2C@-nW3tY>$gg}8Fr27-w!hG4#AmF@Q+qZfz}1( z*J6kMY&)uCE|OTEQ}8~=NG>-XG3`86jib>E%yLOiJ?Sjg&@7yDL%Q3@J}2<(Ha^tFr#1BO)QKl|P;IzFwf!^Ad^l?2eje_$0eJfYpAJ=gY;Vgl`$?`8bcA zepVcym%L6&)h_(`%e>tAsOS)JKCN&rf%CuVn+QJ&&py}KI)u52OfpJQtdvn|4M@zmTN#E`sj1LEeL%n(t8w(i_R~*>b^nTEt zxdnEX_u_Jw(aH|9(aA56|HjdjAhLDO7#^K%sEy9C}~M zjsmYVn4o)9s3dR%|NEtI4AV&NvnS{a>BOC;04CqO!IE5OfWMa6p#Wj7Z{@Pu@BU_- z7!nv7q+NB43t^ugEpKdjmp<*!9A#Z~42S&|SqAM7Iiaroh-PUsQ6L*=x-0fAe5HZI zEYc{ylD$6px^v%Q*@v)8Bk>qVp*FFr9a}%Ux}0An9HB^H=F2D&ZKlYSSSz0@!R0ZO z58D|gI1L?%V84^thqC(*ie$BPuX73k=kgrm#_+Fj1KFkV55wJm3ph3y4A?Dm^u*VK zOW?B<#iF^^-|Kv~KLTVkIBS(wv=trGm+O#vk+YC$AV4x#&IFZRjWT(MRfJa0P52bw z#~mwv_BJAIR8y2e3@BnGVmW6>*+f|ssL=Pcx$)0~Yq(Qe#8v~l6d|x3H?abu#odqu zyid}S?K(<0XU zg{wqVJL%0UTy%_km_1QnkTIL9T=ymbef}IwEH0{(H=r@Oph4cJ`a5*!M!eiY5~CRb z*OanX1Tvcq8M{%p37&9{9kn0nr37q!mB3mAsbaJ;Jy!ylyZZ z$u!NYtel-W9u-jS=EOS4Dw+O0{@dbJZH=e>GY^k=X)62@>y*RxuJl^Y`Q95c+PWn{!$bd=ufa;zjvELo0 z+LAqpYx!iydS*3es;z|}SwzcuS-VX+a*u%e51ad$QP|gC2 z-jOxYiHUFVHfe@v+#)uDN=JZME{8;2k^jx}I?~h1NYfr}EZqSG0>haEL)mLi#icD} z^I*-l!h<0T*~8)AjP^(^@m9A=(1rqJ`&I^VzBIpNFJ#8TWdS#Ps2|HZl1dRnA&g=# zJEu`po03dHAa^EOtNl$!YTukbPy!j-wl)W8jovmXa`+ai_Qt^#<_>`wi`-@8*zzwAgWjh(KSMpdG5>to9(`LQQCuiOXO&gJHo-W=jUBZ~-*W7nPr9_|tNG z9kFr=;TNi$Wc_4gyt#hzi|4*m^?LH$!~HKr7Jj{7d__{JRZFrvVE;<9sbH8|4ipCO zA9bld2fZ`?UVhO3!QLMx&Sz785oRcx%>oG^Zb-FYzg_xd+}YKo8W-0I6oKYqk4oAA zdGeE$P&EQIe+@=Mdb7#jNY{4sSv|BoFP@=?=3nio8)PzGvy@_noB35e&g{1~0Ug5mf&%R^&a^AG(2R?n&*@rX(xrdaSgU(YeJSsP&2w?AY zxH_pqae+LS)J0A@CPaXRg z%e2sA1;i++zXwIXF}y8#Uwx8RCS%(WXMg*P*q!xRdIBahcaK_?hM1R$dEZoAfaj9P ze+392s7eCoR+1XywVyybw))l-P#Rt}XmU&6iMGU7=6--74espA;@`0_hYi80$U48I zHD`6<1hK&`8Rfu|lZh-2$77yi(y@)WY+^_aTqwZ=umFcw_*Lf#6KnzzKs;xd*j-bzK^rcm z2-a4a31hdvBr{ll(t!t|e5-zJ5DY1A8t-cjF3Aj_>ZSu07ezP<0$CSuyCn1fHy-h$ z+++X^jY{~e#wqT`{aCw7!6~kH+Eb|NG}jTjF(6dMr}hR{;M4iGGuw-61DdnsqC0Z$ zzKnR*8g;i|k7ljNHz0aAL*Euq3PL%nTZz3+H{NSQSNJF>l}%jn568IXo#B%Ll}c}a?eS@exEAt@*LQZG_JKi| zt`bA84wbHmcc#tl9@mexW+TL)(i2?gC7hB<7EKfp2xGC39Tz!o*MpQ1m_R^v z1^(eOlF2!v7dTZ};RS0o_qvgW96ui1-_Y7nvg+^-INzM>#`JwezQVRf zKssZsx;zKD{J-kHYwI5&VyRVt+=Ui(n#VrCHm%PUKVt+;FJwmm0-|^aC5SVjCmgv_?J>`GmR}j+dBS4BgTx&m&{7wh;{td6fgnu1eg;>9wc3l# zEL9X-=hA4{MnBSe=^l^(%xOcT#$%6@ojIx9NC}CPPr)km2yEv5EHRi^);E<{`kd!q z;rr+j45NN0N<3r!QQQ5iw+&SiX_9S$6(UVn6(76ok`pO}?L2IDoa2+wxjSSFPLski=3yA`m23`RaQ|@{SxXX0CrMGejW>fx_w@QQLleyG+ z-b>~x$Zea*Q$C&ikF*yw$mBhhZ~Bh9;~1fnx_)TR%&0(B^eY+C)#+JsPRDRAfHik_ z$u4giyGvlf^?uCJ=t{d;ep?SPXpc|}ovG#;uV?<@>-AMJx7^o^m^X92+uHh4_aL0X zndvczxQ_2Ko|TMXm!Bm z!;5<3VBNQ{Rg5nNGz+>Mw83*Cp;9WNj=ntb%3&TmjpQY$i*Pq+tH};8ewG5)mxsS) za3ffMJ8Q3xRFN)lNT4TzSh_c%;Ju%#m;5~;__-%G(B1b_%WaXj4@CWh5}{mj58+_z zRwm^ZE*&Yo2I3OcF*nKaG7yGM?{U(PGdu4Lxe*MOXh}gYOjW(zkfvrlDZR9OF6?jo{spr6 z&!cwm4$Fm6GFhn~=DAz_sgCuzFnS$Lv^i|Z4`eT>9W~$33@w4!N~=kmH(?8-qqRh; zXkEWp#OBbpaE$~ST*-Vhl;*wlMpep2;n5^$$|+_?Ex-~;gSD0bewHaR(6w0h)u+SA zeqYI7O&$us4u2$-%=1f9#Y&vjD7#>mJXAOxnP)u7enyn;6iJA;@*4uv)_*rQ|Jdxn zFZIDa)y=KBpNtdwzm`C9us5o^)jLwNi>-|}!-j@L)|F`0I;7{HoU6t7(dV~`BxJX@ zza3c0OiHwkMv`H$_4z+f*w%fPKJfL0$y@PKa7`a_tr4MM*+L_G@+S=5CXn7LCLD+b zv|ZCcp@`F=R~;e%;Ebc;z;resFyPs_SXm~bp|(aoN2}po#)UCIvx5rf3a)4Sv+v2i zd@%%1dVKhfo7cdtdslY0|Bf`25S*pzIxtBVtay<{5!##1b=1v>g)-W0@Lgo%_o-U} zxZyLn=h_lo^`Us@4?5+d`H!XSCoy-T!35*@-vu?gs$XeOc5&E1nP|zs>p#y~(7oTX z$spxN^`?kn=Dd{6a@0J}I<9k}_AbDljb=w@zfrH6u$Gd-pbB9KvUfZf^@JvsSljMI zWw+ZP&RNHWe!^=Y3-(3ymx;EVd$Sn&B%Ed)|H+ti!e3+uTXBA&ZXNz)PuNPfkH&WHI2(WA^@Ulr3-9wtO8Z2AI%g3ZwzqI{=UtZHa+w0>r%--3 z$jG=a2w*uGG0u*XgS9!e?ej!AT3HZUnk(0`VKn`-u;$ezx4@BOC3-icJ#FXFq-)>@q=J~Hs;<(hjpMCYefc&3m!}|fdXcg}_zK%tv+H2GtcNZsvwt}lzPFLQiolpLI4SxrG)h*Zr!eO8|Kn3%! z3TH~?Nb%w1Ebl$g6f=XE7M<4CIx1B7$Y#Lj@Yv3y0ps^xdv;YUcE}DrQ@)iwDAtU< zofGtqlNfdo@TYh^9D`*oPTpu21x&_nGSUE&2#tXlU4_}taokJnJ~FNXk7XjWZ$IAO zG`zN&(y=}JL(nl`FgV&h?71{y#+~{}@qaiy)Ag<27xpJhg4YEO(b6dmT4|ka22cb9 zle>s%NZB=Fy72Hr=;|k|QlR>QVUyy+<9%JRmtC#(qY~pUn%7x>%EXlVCz)1Q`LX|roRQv8P}k;F z67zvy%*i;icR*ntS+_7a`E7Vh)UH!{<@zvm`Bn;LzL2OQ)DeEf?k4&l!OSQy!LK*@ zj)vTGsh85!mjJ+a$CkP+Nz_Z6pk9UX%>XVu_3Pu1woB|{XX?HDmJ&IS82-!IV_tXX zM$cq_|33ZR-sXRz*Z;7M&H4OmfD+zlx3ZTxYh8mn#Fd~;-!GApq1F4$Go(Cz2<15+ zm-`rd;+|frA1T@N*Oc6!{e3Q=(o9;e43Y6#WZhze{r3;43f>$AJL~hBJ-NAMfs%^3 z8||}iyCwHaLe4CvbOP)vkN}vrmH5F7$#KF#x}Qi@ioZ}Em~OVJjw@@zJwqL4 zr((UL9$ZPEvfKU0dd*+iRMw_GS4AF%%>KG(ME=Hyn~$dcJ*e4WsFSj9o9kmSPTMt~ zcvyDKI=g&jwV#HHi&2M3NyV^7ak0@tuX^*ZLm~-FW5^TdLg}NVs;}ZO0tLK^b{gyl zh?MJ`z)!+&WkaT*Zo4(qcHW3UH%m1|B+qFJ$Fu(^I@i;1(uK!$-kp&ovdDznQ}NHNI_PhsY)k>;Ky>KV*uH82D@0gm5EON{=Zo1X2kSP{NG zmhho9wdcy4tsH?h_bHXK%jNjU@NG0IutR1X6^eXt!%Xy|>uM}AraEBtzZWVOt}l2> z{CDadi}bA@4sK4Ivjz`{QYfFN!V%xJkEoVYFchiQ zIGd)bO6>gSF>2%DakZUd%CH^q>K~(=^_4Z`Rx~VN@yz38KO{Ei4cp<@IW)!X1$n9Nr#y zdU*Y2MBV2#dc(9rW6t-}Zy#6ktzprGWN@o|3~CrmCRRir4HLjP&Ry6kg?Xv=SAp)h z<)=)Lvq>2iYIpz`O)F<2%M_VTFzRSjm^%t8B<9T2%C&La+JMQdZpEEvX<fFe&NooJB`U`Vwc z)i<@avuPO9j#(r}DAfhuy7i{;8sa(77m=VN$@!ptJn08YtmTjM#MD+&VuxOk`ZmaK zo;0JuTILmAsh}vEVB_y}+_?P;iuQ$ z_lxZ~^C+>raQup@`pzssQ!+Xz`V#^48uUH}xT#4Bcf_i2@3ma7jMdsI73{6TAL&F& zwdq_#VVFCFy{3W|g8t_E+OHw>gP*@c4G)q;(;;WRiEs{6Cwze3yGYQ3)?awxeV7Y0 z`qY4S(It*o$8%POM6k=1MlOXqg^MUUmE2f-O2sT^KkucUze8m_!5p*}cgE^)tUwzy z11%33uv4k4(<-){jY++4Pyy$?b`i9 zd1BKqXAsKalv8dGZE+(H_qrycqH5O-8=kNB-9>n+9~51P-QZapZ8sdPLWR*#Qs&^s zmuSAxiK#`m8~=<_k`r0WUyfkWlQ0OtNF>I0V7AK3#%nuvK?oK$RS?YGMWYrTUfTTh zyb^!WWn$UH(d*E!S2_6g*=VZ$FmA!@f{L1@@YmL<5%*Uz@dT15vCChmJ0$?MT4I_Y zWMnPwmU9-e8JDs~zE6=NXxPej>1%tX^eZ_~neqIxYhf0$^uw za7csbX3UBbUkhrv1ij@u2qzed1oe|_AWxDr8T_cj#4eX|gntE?N@zINs6*#tr1KOFuf~Lq!ElkUU^q{iuc5%v2E?vw9pX?-cL$~!Xk8@)yFtqG zaW+mC5_UTJF&XJtl~*9QaLS+$r>bV;6UxGfkx8Y;N5X^nFQ3mhaOVDvhzf^ZeRXZ* zAU+vP8#56C*(39SfT1La;yUx*bNyBZfVbz|n826iECmJeu5%VavB}ttFX|~wdH3>1 z>uQ!z3##=39hUM4#tC`91iWDEVDf?b{M^v z6x8ts=|=vnbW{2NE;l_e33>w~bPK0%Yl0u)3Lvp1@|ZD^!0| ztDb?VQ@h(Ge+Bj+Gf6|on&b(14@|z1;4){TlQDf{KDCdifWHbDv1*VD*Qy(h9upi1 zh`{+o)KHmm?vfXDSALQAy=_S+TFzPgXKYvZENySezVBCHn}aNxKyCxn&e#k_3+NYp zR0}WOHNgX)V5v)JUEkWozt_OTCO=#>{aR0LGsU8W^Ynv z_17f|t~Vy^OS+WDQ!&y!#+UwJBl?3p{ZzQii|-Sx0K<7ldr4B>XIUN<{Tl`ow}+8hTXNv^jYM|Ag&C&-E5+0{!nx|!BjHhO7 zc8KgB)oq-7Dr&4Z*t`b&{fkhxfQfl#vA#MwQ{@gjDrKo2C#<5xCXp7hVPb#X(vW8( z;g|)bwK7CK#UHe6$c{j0g%*NY&#PEuwSQ!3yQDE{p_ION;PiPQZNpOLy;L5G( zcwYH~$4k>r+g=(>6i;i;VSW1+YMHio+u3=fV!!Nl8BHStoxy*5zZsr)JGt`{HGkK! z*xrWwwzs#5@D!E@`L4oAVm&gYLmSHA-Qq5ggo=0^YdZ-yzly{cQHAP{8cSc#RC}0f zP8gWH1R9{t_J1NbuC_+3xhNAVx6Rt}hBXck*dPOe7~Qm}Mgr(Us`xkYGZ$MYWO;YD z9(EkVkAnqOvcH}3c5EG&R;sWJxKZx-{k3FV*2mdjE!B)5?|uK8v^0F6`)KC9T0D?o zV~(9P6Yhd@Jb34!{d~t*1?V~dAchu0oW48gz#(S}?0mSvR0JsAg7F)KyBzta1sKp$Otd3(l}Zw}gL99vT=Tzn zEQUqv!(~KIbkoIy%;xqx1$7hQ6{ElriQR*z_gFSHJ*%t@6dvSCHQ}nao_#yf6Yls= zi_rrg7ns0uD_)IY-+!j5v&%n zy0$29MUw5vkbi!}`veEr`aVhr_?)_i(BwYpP%iBi>Jvq&FQ%abe|o0ZL5gBTU&QC- zw!@n>i=5@?ZY6!6IJfX}+}_!qNd0H_4^AE^o)=xB-^bKT6>~4~4sTUa0^?-+l%J^U zZvER9%t@j3%!xaxkN^C%%B1(yc!S1yiJ8|~2d(ucdP>AVcxC$Rjg&e25PhvYrn0F6 z4A2HLE3gTr97n)Q#;zRF?HlbaRf@nSVe9;;FRSM~zf`5KhM@Ngt~NB3*o^4U?xNb- z#su6dD4hGQlHsp6V&MyW558W)j#zR24sfUT5Slt3p~gkQrq6Y7wHz2^xZ?|{a$eGWq?m%2i((ktz; zj_gF644JqJw5mWzn8EF36LKcO;B%2JMyhf4T=1pypOvj;do1s_*5)geMg|8&$V?xF z?*zA-+fc^5mroz10ny2j5}X?)Sb^c)g{=aC2B74)(A3obt`{rNN@c<`X4=p3orYr_ z?|!VtzjNT6uKcR5)WFr+?9wl5&SvfH>W6f-B*somK=7^iR2wK*ymX}r&}xIG3#Z)!FzSTeiv*3K2q%TJ3SpHs z<75_DB52cV?M|x8nBQn^u#KC{RILrBm?95l^ZUH@xMmqm4j-ml(oMI*dZNo8zP=t) z8#vKg^Xc=$SEHv4$T-2f=Cl3Dr~b1@NVBo_U@m~%Q3Gab7lTx~txYSe!UHeMBQ-&h z!CX_FQH@gfcg@iC+936PoZe^oFBN2d9HU(x5BWCM_HU(GnyR$rjiWa3Dgme>5|atj zl!p9ngS`StNQrMBdkPBW(0VrtlZBN`hUCwq9v-ynz&cI$n(vRQou2`|ocQ-=!=x5; zYPBoP(8&IZ{+weK3=;BKEaA_%6Fj%obrt^G)h3C-rL@|R5bahUtuHyVXa9J4e);m@ z?4McBFHWH*uf98|*4-6RQrn#8^TXA+T^kzDoMCWOiFNBAM((Iu>I5vwMns*eL%vY5 z@DSs@bx?G+W2f)~rV+DpAl4PV4fr9FxnaURPL`5BeSV!C%3~4VICJ@T_0K94rZpya zI5<^Qzn zFP^bCD=hvh-8>NcNW-=thW-F8fchYQ>p8}cG4-dfw2_tx@?Z6ru#!S#kUKG@&66T2 zL&$1`vdVRW1*wDL!}E778bwkL^W^4?~bGqLnnTG;q3gn+fVH@O!fCpwR7g0 zrL~47R?#uLV7ACZVu)DTepH-8|1b3YLfF_@84T_NX*%ub|%`jq0=yD(WDC7miy(NBhB z_fWqvWe4Idy-s#@908sxCPm?AhsyXGLjqu%U4+ukd*%XM#-|O^wkKH^*S3XQU-g%v zUJ909I$xJ`G5u52v*wrw16jFY)!X;3KNjYP=ULVye=msUjZc)`Y;P)@i%9ut7(CJX{3lY~t%ay6{WF57->?wBbIXesUeK4xbUt5(sw0m;F z)Y!=zM%(vF4V_!?d@lAFc5D5x;rrE*ZclK7rPfjaceP2It_~m|H|0*_3cyDw?UYQ1$gX~SCK(IIvCvMD|$w?PN;YwEdSgpWqp!-|$+k$u%V_oR*jsM8eX ze;zZ&%1f~BpK1xqtuypcc^|;vO7(PH*n-i}y(%i%FE^Jf~ zcLCT->GRG_?tu#43KM8ed~2|Am-J5;_sbjiGM-Q2-9JownizYwKMnY9{pp=+JV7zO z!_C&dz^kA#jcL^R$@4`^t}b0xGxX+(B6$z6NL}s#6zh|$uMcbQ9b30~zw5+-)7dhy z)$9j{Z`--~Slr+OQtb(^Pk%ka?;kF?*>^|unAu-@w?Vbn)LlMeh+pxJyJbFn(#ZN^ zr{ZL_oIBW(plq7Q3OMsObcWi7`u*;H^3f4F2J&B#DZz2bhl^l(zbyAJm1&b@A;gxH zSCiL}%8P}O=_XSq{2y%wtVz9JCVJ5$D!ImvKDF*&di5TEYr9a^;^q~yt-rQY^_Zsg zV92fvkF=+6^M+pKx~at1bX;vHVfjzryobL0v;Xd#Pr2_do9(+t_RlE4NbUQ)54OYp zde)v-=TFt;-`ek)x2x=W#zE1#wWVLy7^uW5d|?c>U?JGdHh#1jAppiy!%RMRh+{2< zC}&&aX~%7lPm!Fnm}$GtQjkF{{Q@GgFPfA#fDC!gXfLC|g_dlk1m2zEf>Y_Ux+s_7 zX#?m)+XcV>@QG1F8a2i3Jr25rpx3b9*coS@A@r0lBk*bi&j_|y$B8Y)2T^vQ4=+BV z6X%`N83^*E=c4*y4*5a3L*tg+O5sf&J@UF_PkwWL_xBP1o`)7?2-0kc&hPJ7z29X| z;ffKuqfafinZ&q0s(S07w1e_+Keb*xZI`%q;R+^}apM^dFFJyb1i17lbodj>8n@`s!rl zd@8N<0ahlZ3b5%AeecLAMClHh#fAWYzCg?TAFLO&vUm29lbbRYE8o1!Wtrv`hgLH1 z^|A-tryO6}U2A?l5aRofD7Pr|PTQ|bPmYaUczB1v(mw~&K#8u7iP&EGOs#*sFG)pl z7N*spQagV2kNhF&ejk0-j)UPTu_9^igm+=6btzLb3s)@iN= zP}I>+fMzszX>3UJoa60FN}r5&S1YR9+J8FEF7E^($mz_1Y3@<_oWs*=W$KkDWukt> zyYg1z;qt!b_%FYY)dy>+&4kQDYC`W+fG9olVaMeElXUL!OuzsCf2}#>IETvlP)NOn zMCCM7y>rN6Iw{8~gyeiaZ9|gtsZ>r=DV3aZmW&)m2NObw&D3(<9M&-Ve6RQK_m_Xz z?KQi0UC-o|%+QH`QLk2el zKfMw>c>PmvspwX1?l~P9)xfubC(c%XUJnk`2M)a>u)vT$#VLAyU!UBX`HK!zdh6POnQ8q`P)cz4_EyAGH64Hu+z@A6_D(|o%~?@ zX&YPl(#5RX2{Zm##gb>*xzN({pL2mQLZz=^`Td5B%b2l%l?ImprxC49Kldvi@~zT- zT3s1YkEvKbyB4ALvbC)>*D5XLde(_+!TUq-uqRnqJT24AXlK2j;)M+AgpeTWZ%y9- zzM#9B`BOj-5}U%C$|pWBS3Pa_cHx&mA1_UCC!*9-*am{cq*ROF?Bf-X_pZo&z0dC1 zp=T?`jq}%YUClpIV4ocJdZ-?E>q51iHB~{##}DEKI?u>?RY>qbDxNdV7PW@;KC%_( z#3MVST&*qZ^(s{Q3fSBD{$5UfeCZX7T1cY4on9@37N7Ark(se0rE_iW9+iJ9Y9H1_lrUe9Zh*UftP*=-LrTlUbX_;}QK zcs^X+ozeL$T+riU!;>rj^xa)4MOZU2z35a$19lXniHc5Z^L*n=KwyPPG-<9h`vO(- zvNi%LG|63Fqx%?htz$8|bnma{rGNp@xi3g$>}=wM*yJ!(VyrGb5Q8nEBI zZm!1J4gx}l1)KPRlJ+^+=}xn$Z+vI!b$0RXhZlrqfiH|jWZ-%*d)R5z(K;W$IO-xM zMal&Jz3mJYA?;(o>Z{gEn7#Vl`1Jm}QjKF`|AVUVch(Q3*YF}T6-D0f;!%{x(U^|7 ze{F6s$5`9!%_=ZO=&iw+)!uG3U!v}uTx>Gg&kpl?-OK9NOn231*0*N1+|W5>G#>i% z!tuH^vur$8k>@<@pz1lk37m(LMM*TDe~&cBa&B#^18RlsbLuL#oC2O+Yc^9COWo_EI(^) zuD1%!R50LiOG)?$TUVai>lWTeiF5$oj z^0jGn<12a=(UlT=W5=31A|-o%BRW6Z1R3j;@Y$Ntbk}4IvRJ9&h)~k3D0xTLTxVS3 ze<)9igOe}MmKl*+pC!6DjTqusR{db3yZ;v0(C^2d=Bs$sRk}F*C*;Y7Ps0aMqq1+C zMwJI&wAg4Ztn^x?U3qfO_^a5%4o#nSBy%*Y)S2yW$t0xKlSN&v$4+AvYMFd>KGeyU zvMtYi+W978>ZnibGnGuU&+7I*=YTHV@-oce&a1$Vz297le|&kHV=f`^_|y-V{q|?! zQMZ4n!X`~o=$l|2$j70}AweMF+6@+s-XKq}ZrMRPIqaOG%(75$+WqZ>OP50OVpp`z z@bJlvFplrNr`y?UdQ&GY1oj&Y)(%4HRK29s3N=CxCL zFkIN6&p1pEwgaP8v(-Gg#k6q_NMK4~kzcVda|MchKou#-{@AWLq+VOYV3FZ2Y}z%p z&od{V2*~0*?8`+Q4%^9RsdBXzv7ov;e2mh*@G?RYwLZ?y3r4_Qy<&ZWyma{&C zW)MID9l`lkrz_P)0P!DJPzekxA_h4^r+WhKp%igmUvKR=FiY}6-a0u(8qFp>m=-DF z+8plvA7qhd@29xf)x>X=y0*1kskORNl^$i94gRjPpO2G&vVwAkn>)4vy;46(E%*eVKK&Kf#iUKKZIHh%mcO3;VRq(zG@>&rvE(Qd zl?+lpuqq5PWce9 zV@O*nNM@PeZ?h_&Z+$c#AfvXUHZxvcZI(C%XMZpwd(3`}Er7RtRi zg7tj&ZGC{)ZNM=-10w3%4M0W=|0HNTh0uIg_c^nHmqRqxd^SKbGUDIz8HJuGF~ zJh{AA%j_1Z}=vSE0A*HM=X$B zCDdV#UHB`8>qHv+cuD47c^h4x-iC5NN|9f7;Cu zyj7~ZmW377TwY)|b+zwJ19)z<0%m^QXZ;xBxv!!+<5YS1`55 zM}~CE{(kdK?PZRSC*^ZjSdgLd{rq{g4rragKDEkH0P+i<*1u1bKZ%%#uo(ooV(Zjj z<%ug;1&$0eiO#WTP>{MTn?XteX&ThUsJF=}FNd_2HA5Y!-tAw#kRyIEJ|`;7C`wVW z=DWaqL3so;Ke*dj8PofA0jYg_r%S2n%wSa*@h%Q@;=M8QGYh~?nB1kYjJlj>T|ack zYg9|qSQE!2_k>GaXTL#+M&`3<=NL>}?R5=jr;i|d5ey|UJ-Q$u;2%Pl62p!t3BC?M z50j7t2xWH_mDvi^>bs-=g9MZBxA8J^C*lrG(Ch%6mroxNe^O{~K$XRJ_egr`iV#n8 z36%#A-$Gft3l^%}%AuaiU!&%lc~jMZbZ_dlAUyX|*->?WZFdZh2dmS^(vO!;)j6o2rn1zIkWV{tltr4X&%V0(Rpo=#B=#(n^8N%`%a{6&DO zHvzbcIXnI#rFUb?K|B9qdx(;9v#U$PO!X^lFIv=wZ#Ck!VNyoP!wLsqS+%K}DU&Ca z-dN?@?|Mkb0lE8^R)&Oj4XK&|AAff9Btr09_SVjs*js=RpdOh>Qzwz{%g%HIfsFc& zHVecHpJQEpFLs@@PZs;=<#nIYMgabJ>J!dlo_{efd+U*3fZxv6;nRLb4gF3yzZv@Y40o*c`AaK1 z-Wj%5+iVwCP>l(vekjWp(vAmLqjvsSrH0h~VfUuTIx&J9LXnqmUr@78O1)1>r7GY5 zB#>$5+{{m-z#-A%mRqW#Pw48^s(%+p>|j}A26F`4vR%W>GbA3kDIcPgZ1N$hlWfA7 z6EFGMAsj~WNT<-pOAJ;s6B`W-Iv~+t{e_^Bkn|X_C>i5*lFUBnj&<%uOVrURm|V!; zKwH5kGyc_36=?{*@t39Z-@4+0=+i?$t_fxqY}0lVjv8|{Lk}&)j-a5IB`E-LOvdX~ z0@d%&mno9$Tf>e`Z6n4SVlCelcVw+=%{52H1Z$3Dw?Lp&7SuY4kE->K$1W)$@Y zdB!OI3SPTDMN*i%egBIq@5dV9z5L6dJD}TxVf-Y%Lo6}DQ|V;u5oOv=&m!3FxQXen z^oP_DP5@Ozw5yvSj8&ZNH{Q=@s;ahdmigj@wa~B`M%rj$uxj2XC|w!FPSc{nEOYXJb>-~`(ONWJ))C1PUr5e ztH2JnYt;>LT$p(mc568|-eNzO@Q=s}rNNoa93335w~5K<#w7?o`NfxRk-`UO#`61%YkU`D>DnI_&!KwgqAY{ zy0h+=D6k6|w2u5%Kl0uBb5%Zdi|vQLYiX`7Z#=3qeE1-E?{d{$jF)!@KNU?PurG>p zt}xG6FAWm*p{Ul5UwsjYD<(SX$(r1Q)7|;sUhHx|cv69oQw4oG;`1?9{_&|VELPUO zkfwqvh(9}6^tu><`;`28dt-pPqSV-yBwL+dP*fH zpafPW`yTd*s1;Xz00Qsu`w|%^62gni@K{vVgCnk2AQ(L-K0Pz2 zWKA40_LQ&C*kCzj&wgUZ}-=MRLQ>WR~4qnxYq=Og?HU|9%< zW=U^dGcC^W*#2lqyp1UQG7b~l1^W#pPf7H^4=x}Tbzb3|V`^MJfAJMNhIMQAajZ|Z z;}dss;eq1W85w6sA1RBk4^w?4B)WX{F39bJsc!$}h+XSgbGKpyaQ2aCcfVCR`q{5i zvOX4SVI}j@%QJFu65xU)|dCy4;jPs(_l*7n0iu=#$rPA z*4;FBxA--~Vv2B4qNxO8vWqM5J|sx9_XW7~-DSH=dtUwb8_GvJ-7w^J|Fk=h2ORmF zZ6EOWTJ$P004=e88d8W#;WbA)iUkT?D`%KV3B$lQW>)~#Go4I`Wn!!i#)G5Xy z;d*zoj_ogZC`&Q*xtmTffV6OwRKTj|_(%bnyXXI)Nhy}d?(1j8Ha+=S6mr(_NTJ?5 z$83*OgSOL5CCi~>m8;Xo9hMRIJ&wT3H1(#hT!LaBG6`IPCp*zx@;5mE+);Wr)3d8| z#$w`M=BZ-p7!#9Jb<<_CN|R$w`OXwAV*Tk~>$_*e+M$o{4>mQCYfk+=y5KAHs$gj0 zdx1vt%8k~{-$$=xe=9utw4QjZ{Gh{<8_%9_#d?^c&QDtb_ZUQ_-(R+d+2pw$B z6GI=Qjbwe6fe%8+<_(p^OeAD5KVZ@ZJY;m_PQ?7!(}fekNRmF!AdGZ56VIG6pqH-1 zgv_T?@bITdn6djX;}6<;EhJNaA0)qf`~)&6Gg5u5d@vzu4kF6~TMbrcW+Y6tNSM`G z3fCVn{bOs;4}M}xxKkmfe9^S_ki|jD+L`5CFQ&5~QAn5$&on>Brq|NP zjDym{O5w<;`4#jgjc5yUTnG7wP_<2@LiCRqRvn1!@GdsD9oei@qL6wq@YsLH^lW5i zZJZnjW04iAT;-wAx=^O0!ezTbr^}7Qx!c#@dc4ooU-1y^+spqNd~?iEe%j*L8B9Y& zH(`R`ez-c;fLa-Dg%SnPa4w&_&F=#IZ=j14JLF&Q$|uY2?=7`9HSM>SEfZXvu8sI# z{3WT8jkx#Miz&ljWvkmKKeuU8UQ!R;WfJBr;ft9F+c+cw;F)cUTZ2|?k;l)F+T;Gf z_;B)@opW&^>A^wcK=E<7vAMavx#>fFfv?kh%hLDtt&KU9mio4ahhFtO}b84 zCpcL(gXl1aT4a6gmy|=0Ph?`Zi8Aw6(<=Z@bX;kg#fjfuKRNKanw@1o686>iVsnn{ z(ckUgPdcSlEVfh~TzghxNzc7@Z)o&*U-m0y8J@Xa(Lmi^VAw)$@!=(ny;O38@-lTY zThxIrJ}xlCRA|tB_$j?sK6-sPJ_~&TIvY|XP8tbaE_pJ5(j5tdbXWMX=%#9A_i7n3 zR-kWB0xsVqS^sZ8SY6N48K#m&S(v9LUFyS$oBxAIuV7K!xPE;e+%ZB(hw|DHXuup9 zNl|hwdCWo2+Ccu1=&m|;Q?H2>qw<|z9h-ppW{l*yAXGQy8$C)W`&UBevy^Q5eN&mA zwOgJt5VsEJt%?1H^Th$mNXkZ%uZs!=&5=ZZ#MG4()NI{uB%kg=-BX0#zjL^HM(1Ty zOofw`{}OCQ)%xx)Z}`^!`M*K&e*@oNgCeeWxEdaTnM%W7LSn&&%;lm)vPn5VvJV;A z0~x90%unVlE|cIY|Jeb&dw)T(9sXM#R~{TOdS>Z*JrSMW+ELnnsKPB>+*GNKSZiI& zR^kd+fm-P{b2aU7X4l6V3G=1*45yCEGa32q>g)_D5MeNzR_G%A>v>B%VsVF<^5j%DPrT_|97QK+&RKI z2rdPpKE(RJ-OCu(!tP@|vH&~LF0aSFAUv6Zy1yPd>Ch;Pif=nuYOdKc@cQk9cf?UG zA~K)6bW}#W8nblir75cD@RrL1uRUfMSIAzm+`yeYj075S)GcC0Yc+8GgR^**^Z{X8 zKiYu1ixD^0pe>3>Y<&rD?QU26zO1celWQ+piyXPl=sc2A)>J&(%&U^>(xLPL9AB>B z-Kw)LvgfIoA5OE6;!&Q(x>WQsYUd(H0pQAzObt@*?fjJ44BZaHCC3-Z3H2B;bg89c z%B*HYkA1Y&?bC+I<4GLZxDRh!oykEeTW(4PGw=A9FpMs6Itmzj?%d*zPwO%HyD&e5 zs52=js?e~eu?R0XC40~l{IkBDYD*Xq?Ed+Q2Q`jP$jY``v~I5Ii2tGC4PQ%@9EG`c z>=)Q@>?JlR(|I!3)xhRqL+HxFgY@8~7fj5V!LZ`XRK-TDiW4V_LF)bxViBV~SG@M`xNZ?Rwul<)-GBMbjW?{aJ; z;{|8>#=!#$0xw1nhwC5X@x66@|3pW6vNazdpaIXk2|xD>$Qjn2)Saa->+JY(JisWA z&Prlr^$-=^4J06>8&zXH+EQyZ8 z*|o+mn~=joSr;oHAju16_s^0iBxy?BtC=Q-%}uS$tRA$S@AOHgki(-LVZ+R($iLd< z^X<3PxL<~7IzFsDo)T$Rk==N)^H7w7aZJd--anBbbG8@RtTKd`GWHrPpB}JSota-w z4=?56&)HfZt3lb8E1ZZ98i8frny|i@BJYu;^~c0?;MG}@YHm~&kE&Tqy2UN^JJx&o zH|7~xEF;G1(#(z+_xt}K1rG15Ur5i45b-vStt!Y~`0O~Q>E~8x&aUucUNdJ9yK#FD z&3}tR0GYaS!hn-cQQ@nNt!mGf^t8j%Fm@1gl4SsTGzJ_UwkF7x+f7c6RqYzmsv~E4MCQ?d> z4Wy4UFWBlk2dt<86V%xE@??66tpNqkL%{g6H;v~) zGR=*Hb^G+7konC8RL<%!IBWUiO@&H+Hj^%vd=QO78l}@0O-1O5izF<&lRUDklYv{YETBwxpLq z-;Qk(rTBujVr1?f_UAJ!m++%n`~0`A3TXjk_<9ysu#k3u4f-y;vb*&hYk%yR!VJ)S zN?mAbmJr}+#P0vD%IJ2}R}b}WJ=NDiRe?d$dN77b%=+TiHmznkr&r&2e<{B(T>-zm z*lR#5`GA*EdH;`~7j!-hJ;wTky_p-$1y!=zU3M|IWAM@{U_r39_W&xbTQKo{x1L; z5UUHwm*!kSFXov_Kq`wNP8vBgm3mhvgnDIHJNCvpyV{AULC2`JC14Ml* zsWCg10kgPxV}%gDX1(`*j=tWZM6pGCsWEZ2;vbJKSlly5fB*YMybs#;!%>oOj7w5- zW7Om>H@O@_=8S-)a-e|5Y{R9%6XAGWph0^FW*XpkMjJAF+7&n`-y@zSvvP;oyNs5~ zgUu~sn%#^Eg0%FZlCZ{j&jS5X%LA(RH=pl=T?e120FhG0Gsn?l-?EkV;DoYmvgN!F zqB&}WBSun59dc$msSt)8WX{=_mc?#FiM)d#@Bo4W9e)H&H2XOrrot%3SpxegAA`Vn z+b&mU(ld@Vj#ayAf2{XVeA}hp*B(?nKVNkG>PNaB>?d}HfM|1P@N)%#-Ir$rxa=bQ ze9I%NyJnE`V=F}2;4Lob?Wk^4Yo=EUFj5ie^lp8Z_E#T?m3H+}39nqI~*_CUz zSFS%yi>kZutxQ9r>TLbHGlAcXGs7!7?%h&7v<8;{b7EZ+MGMuQ+kL^)MkqKDWYA2J39c|Bbw3 z4F0DFiMK!(Tn4stCDwYkffI96$cjN`84WMBsW3h?K#%>5?ASB2e;i$czj{q+0s3Ob z+eYezt{6Q}7yQxxD%wy+R;R_J)ovon4fEqN%d;u@Kuzp9=IXr_ocByd#Q8%yYG1fV z*opu6gg**CIPrKRh$`Jagt1m4|~|n2L&>tVYb#6vkCYt!njF z{>uE5EVU~2D0OwbEnPkHc_n0TToa59G-0Y1MRXz~k+$aUrhCsm954)rU0}JjdO--Fa@|P2RIX%ZeAi+l2B; zQAQ`C8q|9U{7zabvd0v)K8zV_95WT}jG*EqwHm0CpS`YJpPpE^%{`FykD2HZ_=-FD z(-d_2pwk~hQ}xot9d+(#tCM+v@9W5cVW4eglyXm)DcfYbIL4l12kPybXw)56cxiKD z$iU-^bs@NxUohdtbt(cwB5xE{0Kh|~a5uE?KVrU41}cA>vL1lp+40SjcsXMZLWv9Y z8;W(UbYTH7rpU{3{~Y2(mDbaLMqulYmVv;V=f!Rv(JM=Y*qvgc;1KK1R`w%Ys%Ovd z82Q$4JlQTGFN`$Kz5eNoJs|smTYM+GU*N&Q0YjHJdU&~}UJHJ@xLojFm0JGva7y@M zrUhvTFXrB^PtQzLj-c!!k&t_Q30q7o$+!e#R-UPMMBEuK{HOnL5kv;E?_dvKYC{^O#muC;ok?z)>FWzV_g z`GJu6_~A`dwLGMZx))PKS64}oFG;LMmz(k~_Qpl6&X<&Or77t_V2k5jowniL{C^q|mr(xoW%G4&GNO$yclGDO&h%nX;bl^|N~9@{w%FVTHy$6y0}%e+;~qA5OXXS zK>8%<-I#EN|3S%3w))5p1^G*4)#D!(en;IDynoLgH(h1-udJ>XLr{9uqPT0n<;|O_ zc8EWVvw&Nj#zO7pnTT)>5_cO=l&9IT;=tW}{Ew_xiXJY>Q0XI$_Eoa$fZ>bc`fudf zzWU;ue|ye#Oh|{e4ksHk zw4qKmVlOfdO(Q{)@dHt`MgyK;i#;1METwZ>!is#6G97fFcLZ-;kwPnpq$59cEcZds zHh=>ygrw8s*|cVePiKZ_J8z!tF&D8+vONNE&nY0IX!E_t0jA~%OuzTc>grgvHEZ?# z24uH~o<8RRTj2kiiC*6ZD#IeBV3byQl|D({J-%PT27=3gXu-syoDGV5jWlquEUvQE z$w{HBGTfi2cj)8>r$QkkAz&dQh=^>mUWbX~7yG%p&sy2oQkltA#D2sX5S$u{UeI4|DG0mdD3Mv)`k6^vEFIu@~bT6n4}L@ZTI!`?;pwq2A&R@ z{(4HHB=3cH^_M2a9*a}=!MD0DSqT5M5Vka)OE2kGEH-53eT9xnlhh3;McZ?9d7SpJ z9@L_4g5T&=uZm%Ac}+H&n$+&5;k;4Fs5&o_rX&d?Wde&mzFTG@wP3$U4N!_jvY(e` zgij*`7Fv^1#==(l0$LkgG)RWwkM_^7w8|DkbNKUhZ~t0@Om;xQGZ8sSl2u=Zm2t@> zvL*tI<1_6iH+(X2M|9%RQZz~XO0C!WM-*_N=$CWEMG1A<0?HW zv4desbuVizZ>}a1c$Z3>n>1#%_xZ|xF^sr>(#S@>q4CM7;Qqsqxc5b$(XkVKCC=B}GRoi#Uzgnn?K-N82&OsSSZP)%RSCh>j=jiEV$@MUj;2Z( za%twkCSdU_dN*a1MmkQf+l^MLk5t;+d&+Zd+w;-Hp;n2BwpTmj@Q>Vsk&?YA5o4=a z%%0eDjIsN%;xvp9S?kd(LNKzzmS8@%VPkeO)$(Thu(oF9s-N0%qh($H>C;iyVNDPg zRP>!U-)e;hOz>08lL;q$xH7wVqPv1dS5k|^ol35l`N1Q^FrI?8ri;y{aH(uczA8Ri ziL)=2bP*K$H)hzXW+{|chcSt#iFFa(O8d>j+?IDAzT$+Bny4(~a4wc`dOvDZOU$<_ z5boZ|7f3|&W>%c2*!x=?- ziI_|WO%mQ%2IK{smZ)m6swu=oMg8wdQ@)?xX4`l_5?sU1tEO8X9S1q7&aobhdIG}% z{lhEU0lyE3tP)LO3_ zvsTy7kBmu9TE2AJZ*go8!V0sKsR7-f{dTR()sShPU<}wudm~8~$2P!^YG+1O^$TDy z07;v_93lxejM4JfvLnou+ zWy?SQ{CARrWssBY-CO7NmcgXkzp^aD@PL!i{9XHy-p@LH{LJbdTkH@KR0+Pg3@!@b zdws7HE-}#nQ(8FwHcc-&B54EqrKhy_kNu4TO(S#0%45xq&oTP;`lD7pY93P@C{lWp z98`j)7_1Ht&TR*pioGdmD|$UgpeR=d<(B6ZWm_P7Ah_Hq{8`5PVABrqS2C3BXVMJa zr19p?J$R@T;6ZESPXjf7J zWR2Gu50n2(mtQ9b2O?&@VIc%yy)pKn1P!2c1_HfdaY#}#rL-RvhPKK>S2!#pih>+iu!exp4#tm>r?JT^x)N zLl8k|dn+9ojKr2QGfm3U2VrBcX$pL2C|m`@pt09WM_}bXph7Xwv6RhKyk2I;bJk+k zGu!OjyWQ`I%zB_~+|k5!T5x!G^kW5fGfiZsVs)72X(y}kgeDv)(tQaEU7{~u{@IR5 zr#K@M+-YX|e2}L?^$$&EK&YBe{&B>o8AUb_7=mmr;{Yb& z%8@W_rO3}cSi=uGpVkjeN0Kok#^&L#b=&S;l~}mb?iQ@tdxiI(H?aOpqA4Fro0R@n zdcN;BUa}_WkO^%B07S?Srj1Evvj(BnlI6 zVc|H-L7>=i|V9(fBNvdNrlbI&62X}6CJ%4&lRuB;dxC< z5&J(d)l*I$IWTsasg$zm6b8FAbCs!B_r|Fk#=mk~_U^MY*`mUCU@{#l%5d>wJ3f`- zPZfRoC*F}t+C(x*S!biepxvU~LY1^G4C6eCeN^ebzq1fgkvjxLQCVFg20t2;v(7g$ zl~Pv^osEjf)I2*BaaHwgu*&g?Gs;S7L+AWgdZ=fYb)RYm@FbPx{`&aJM|9R<{GV!T z(3#rfa8BifxwmSVV6XmLzqveL%VQTl>Ip9ib?U2~f%j+U?w7&CO}={DS-RyofZMQME5kBN-#=B#%8^}W_>r{%qj?f5TKs*|~g*?F_Y zlAR_nM*eG>l2l|awGu#UhzRfMH@=2FpC-|fgYzue*x)JWzWn^A(^}$k9%5EUKdRI~ zboF8Tq2Ip;7CMz%8VanO3ZvdW&GgoCYsoYy%{9WSTHXxv$gzM~s8o5wS?a8BO%6E9 zV*I|y!{IOy1bQ^*kKX@ryEoE_?DYTT{H-ANPDmf@VSn!I_l zdCWwSqimtYk)+5ydf;yJzWsE;lGy(I-xEWoZgz;TXpkq^zWV}TT*aF}kpQgiNcQVj zn%#Nx(OxMhx5~hN9RE=$4&G%?zVUQE@SisiQpc>~<&0*&MEFR<>BsVA41V?-z-(=! z4xVUPUi)Qk6p{*EW>Z9?Na(I0$~K@on!@N=mm)W%=fc3XT&Y}{FKFAHxh)zpmY!gK z#|}nvCRAB0oQPBq<%uu*?g|L&_<2)kV{YRQ z8oNa^rswS1LZU|8<^hDl7QH=Y6QGv|Ob8h3kPhj;OZ-))%7QVh4SIZH{EXrgY0iK* zM8+BgN6*PCmb&v#CJu-K??W?EXTV&4lQy`{R&s{aU+Rt(<#qHU*~^Jz!^l67vHXVq zX5XJ6f^E#k#wwn7KFhzoOL_sIIv;Zk`uGR`a|HSV0!{z427n(Fx9hMMsi&T;oBd@? zOMm~B+_gRK(Opr+zQu6ymfdbqAFy$r_Qw*9gl+2CVy9Vjr_bT6&JvT9m9sx@JHE`;TfRMe>XGMtPirVj zyLXkXhU!rgW%FMFX{tv#<_z=YuRHnp{l>vI zI_FDbfCWo*Pw`h}>4RT_Pgp#ce-O(1jUh@hb#dYi4)lg^uO@Q^(`ab6ItUvF8A!V~ zKt!V^!`P7An;|v(D7pPO$Ci0G_oSZ&n%zl7i@E+_&)DYudwqWcQH-6kGIO~huSY^H zgK$suiA<(hL*^-nd$Ck`h-}5GlW>4 zfXAcF6N~)TE)a`mp9`T~q-ma`RcC26?D%PPhTP_=efMe|t>!HuJX?75L)U@ya{Abs zdE*r~7-!_S+0hRN4&8^vd#vZGN7ABrbkkQeJSCpd49?r?6hY?$gZ2&W5vL?rj=I&> zoKfU?|9yE*f~1N*RIjNgR;zLBMca#U*zQHJA)Vu>f%7pBT)sTBl2Nt0t4l`g&fyUE+mO%xNY1yK z-z^v#>ijjLdDEf;mA~I@pYHxk&RXB?GtRYc#Y-IB)A#0Te7fOQ((&v01Cac62?Oer zoanVrRoC6+vsLmEOb^<^Xv&bMu|HX;H_#$1{BpYJb(NgVPoQyT$HVIP(PJKds`_x3C!XoIOhqMOyn8i@qn&w*s$_CE*dm2 zynnPdHP`!23MJmTw5;Za=nSrEzBcg|m2_R-4(rL=NU3q=GrkeOnqw+x=(V`LLkVKr z!E@#uRJ^aNQoofxe4w9j)ASzQeUCungEDY|Z#{mQ-UDI#!N1=f_99=ooc<*sWucjD z&NKxIUz`&nP%>ism3TtW^FC?*IL9~-X<=`+bSXk}ZctK-#$nMO31tnQ=rIH6b#1}1*_P8dQ^JFUU@>sbY-Zl!&ds$+VZ z4zWj{{%KU3GgSDx^H1BO){>H%vzewQA>c}q7wOB2U<54x+_7R`V_ix=&rmXD=S7sV zMK!GQn}>33J-Pks-M%~Hz zVNOgAxP{Z@moLpli~LD03OwB}yPfZ;WVe@fb_q_82@phlLO;q=-s<>eNpCUU)PueE z0&8Tx;6NQU&ZygPIV&ml{Hcc_)sZ*Jm)mXlT|w5jBJJOiuwCif2cqqt9)1#HO#E@# zc@i@ic+lG9NzuTtx8`Z_fO1r;ep4h`nxNsx@_e&T!?90%p|cxx={y^{2*;9(vyG z0)I7i1iohIg1Pizc&Yw%gQoqW*>Fa8fo)z?RFGCqxlQPAtDgGd-|hKw`kMkmRzbM( z$)3QaUnM7R+T%6NXVGG$XZ_Ucn`Chk^~nnD{d>l8du@#9-GIMPYcfj%^D-;U*ki?s zd)VrQgZ%$NUC%pPVMA5Rhf||$MjRg|6&c>ORlaY@%0K%1#wRHH5$8aK zO(PLRafyMF%q!CglS*LNkbhirIIP>^-aHfW zXg>;={lehtRC0(N!+6BX-Ysg_bq&~O19((;Mt!@+V5pM#Pq!)G5}xmr78m!AO=I`1 zBl^o-=C?AQRPt@99)p!L$y`D8m!L)CFblvFqNtnX@f}t;Al{GJ+IU@KhcTLVF3&;< zpD)#ZugtS%{!QIfz;zqSkwbjbT0-Y%TTbmP1-Kq5s)5+&Ixu3KFb2`tM3o&zLx{1J zx|1zOH&Dx~!nbwj9Si0aRfeiIY;ZQd?_)3A2`g6I-mw~~jPLpRaAfd!_gyiO>I8y3 zRuemb6XzH)6De6)pzTARHBYL#L^lJ&#j;$1a$Zk- zF2~$RqIK=_Q-)Yz7yHXv251@otbLIYJ;eH^?uu1~8Mdy?`yr0e^T{9cT%|!hjhIr^ zk;1cs^{+1fn{5C6R|a28rCeG?hQp!MBWhM2_x3?slE3?>7Upz#M^01}kFTGyIIV(` zI6=StUJUsvn3RT82_@#i@;M}$qkLb}!G68NhPXjSIOd+1@AQeH3$Mtm+ec4LkiBkp+MmC4Q}aqHxA z>#SHi+ zl?y8F*o1;iMW&i>0nWs+?(9?Xb@!yDw>&axOx@n8q*w5j@+*q&?fjH3-upXBN~Ra4 zDVgpScqF(qVK4mjE7%kKWtm6E0)=m=K4|HaJ|k~x=AXSXrNYm9`BfrJgHHq+iGeq@7De$= zhLKbz7lJaUi+AYyNTfH6nC042gA?5vFn8_B8;1r2=z};r+cV3Z#>|nR0_ZXc3z`SE zIK&-7A0GPK2_zN8wmM_4c3}XPuD(tqsw*_pX3~Wf4Ja_(^tGrz1{@=K$!xT9Uf)+p zW^o{>WG=MBKm|@LR|uKL??Iy%2M~&n;i_)wgfU2HK5xvJo=33dpC7I*g^%S7OOpd! zi^8_C51^2R$i{AyBhL>zo<E?QOmCRl5IWY*y&$`KcJ0ptn3HY%%D^g6#(N4e` z#;P%Oot0Hd2l8lfd~3HZuB`~|vDR1Ci+>>AsjLJ*k*>y|rzZG6=s$?5zqgOfeKd2u zXIeS)0M%Yw#?pQE=@H5e)k4?SK+|5Q*L--CVO+C;N)2G+TZ^LlpuJa}j2L?=3=_W> z`uqctk15G;&|nk2@el+04QILunqWM%4@W7FZn#%UB!zi0{<3h4t37o|+#^)t$S>0$ z$~FFv=iqCPIvn>hy|Q8v)PN%c)9OtBL@Lwoz;7SV(FP;G)1hg_l;iJ(#2Y;x^vxx< zoGC@DNU|b#@i-5q|B!tKzGZ-aemFxt19h7$=ZoXVc?96=rXASN$Q=a1`mCgy>l{lq zSQ#_z4V&IEMyP0`|rbJ zmm{CE#CH>nPySa3f5+;Z?>NSuVwU$jA$t3{em=`CrXuc{7`KgRH+&9AbUf8m#{D9B zzT&G;+#AjCH5dQ&o=f9B&p!qT+t<0&3#Y=KgO}kO#oa={W6jx%OzqxQT>8n?iKfn( zJw$KqeVil1YVT_5*#V5j&S^QeOW5iAoI4XAdb%C+D`tPX#yJ?Py!2sqvRov3X_wl} zuC)Rrn#k`91G)J=vnLF~QvXfQt-YtL zR12&$rLr>|abK=Lf0cOq;zwdVZaIi@Ef)8tVT5$o*xXOdpE_iftlv9&saYQc5y@VE zvwAD;RpH~>e!;}to-4X;N9G{Q&i~`++T)r2{{Q5j%ZhSmM5)}$HMdO{x5O%yTdYEE zvE(+)CPMC|qFgIU$|d(goBJh+kUL|mlyoY0DUi~c2s)H`6)x<8~9{rG^WmGs)ZA61%i zwra{Z{k(7b;NQ2lKIdI>D++qH4clBDEUkXdw|GLm(546b1s42oW8?$OgM(Jv$UMg= z3bNVcUK8FB^-&+^E#2r{?jiHujpaAzFyepjn(X|q+jHH=u{7DNl-d5&G&%XHL{)r#yuW0E@Cy&Y;`>dPAyX$1@ zzc;*-NPg~6kv!@qQNfvQ4Z%VGMi#|i_dI%Rv_+>J?VrOeIYlZcpExUz9@BsDRKnM_ zzM-V_bNuIxGY99N@^1N0Qx{h(!3|2^a0j(|{1s%xRN++rs$&^~h*CWATaG;^kE`eZ z+E}GNl~&Xd5QbyJ%RRn&x|sPa_?1`ibDw)S-y>YT03+NYj26|PdoU9l)YEuT$QDViiNaq8p8sW??_P z9>7)UK%}q~0!0j}G494^(DEOFIP9-Hi%3rN1HDl|LHS2hjfEAf=g${mFIu3^`S4~0 z&kXJ2-IxC3nWuJJTG=O8n}th0+2P4DHNsQ_0o)-{FOl z^qh`&wv~8vaD}_Sz~L2aU}@tc+r;+8{tr1Wj-~rEuC>RS?J|~d&bk!XcR(%Ww1$^S zGVHU|>9&sR_XSne#9>OOb$_1kbzZ%+sQra{(4&;6AjwQTU%VC3K%_e+)^f5@KA95~ zz@3_Z1FRp1kaE9K^^{e`vggP9LGmcJdsa!LN?dkUo}=0URGckxH%B(<9V_WERVPKg z*FuL6^D>TJagKi`v#=}YmPCr|6P_Ca`|m6_n~DC#{oolsrF16lipVbsIW5LB_0;1% z_i17z=d_yctbb=&f=|Efs&wXM+HrT=QUM~JjH*@#T?7@L_8{}V9=E?pf*`-t^msZ0 z=pn2ZjJn*dTa|bj``gLuu0Z{ZuCCkTK(OWiF(7;Wkp)5!r_Q{}z=Gb}8=Xn1&n~F^ z&44gW0`!QUvyhWwrW1)RE4{xZUzH|Y)nB}9a<#gPDXjl#e`tM8a~#!Ce_^s*=}>wy zY)6MHT*VUd!!$^zBDJ9&YC~fs4$gGt-S#41;Qx+BJ~#FwDhGKM!_K#6;} zCu?cd13HwyYgpb)36vY)F7=seDn$Fx(x@9txHV(`YGlT?RQAxwFcrJ<@8?v|Kw9k= zW^&n+%L0E-h@M^D{kAc8hNlZ=@i|SBAfR?-!?0GU*J)<@@t)_ER`PMDXrlZu!U$oy zP?CXN833ozG-#f~@QU=yO)hv%&Re!Ib$zOcI-a4qsxAfI2SIms`8F=1ZBBHw{zNVI z2y|h-x4g)X&ix-u_ZMdWAmN6tOTu$_*Zsj~iRY!H9>!heO?5fYp&y}Ps+gvp)?@nQ z#`0of_K9R~3&CT@8+!gZ!>(daRhR8P9r@+)yo2N?WYm3ap%x{E{aR|&kzxw2ALw;KjctY1yqttKm47I*hw@9Fvo(I29R2e+-y z!U7)2{jy*v*$F?Ss`?AteEEF$zGLj0tJyT*VrhP%?+p0)3-0}o0r0n;6#xG@j5vjBSs8>3v#=sYH$%(pf;NQ>4ogP=hMGU3cjzU zlc|*jFP3seBWh)aSG^kR$0L?+I{&cA`YO>ZrG67e*GU6mx0b+|f4vJ4ucX~PqGNWM z4hH_IASW;@Q2aN=H_4Um(HyhDv11+h4LdYaxa^Sa<~MTZbIr3v(S%F9>u^YQeASX( z1?W0tcL8T)DQTt)r- zqtwvBq&xmm_$98g95U&x?{)x@=ok2IYA#@JTUpF=tJYXT3}CV$x^#FzXb|!QqX%RM zLg+T~{?Pte&xNWKv(wN0w^I9SEQVVylvA#rF4a8~cOvjmlUgU@ zn&3A~qY)ce+_e#_t5(SL4eDAR%f=ZfJ*wZaZ!s`!h!m16-QkfaS)46G%M`;Mg-b4j z-0AW3#Vak0E6odmr!UrgW7|hcPw5L8sC}$!Jkyl?!~zAV;!=N#p2o=%8RI;OdcD_V zZFnBu&1qk3R^JYUK;9ZxI!$fr_|hbo*o?=Bg;~D!z+H|ye0b( zAz!p4To^A_3pLLdcv49pZmoyGt0BFPrk{nM-dNE6#ABZ6(xWzLTbK{X^e#F|E^1-|Own)zPHyTW;^Pz`fexTZN9!dX;9V zo3MvV4tn_3+InSn`@mvXS65=u!Gx@tj~^OsW{xW(LbzQf|$TCbsenR$zWpg6@ zNvtUuI=a1D0D#b8PRS;EA6|ftr@%oQ1E81PARAV79+9YBPOjbJqpRPwAI)wem0Vw( zF^Km!N&ntDviVQVZ?(=wzdC#n=Mw>A_rqW?`F*1NVyz(Lc6n4XBZjBj5~edUURs4@ zGrwBn3!`Is3m2EckZp%)whf~nhOLisP-XMsKjMcTu0G&fF0|R&q2-5uHJJ}iqrGY6 zp~!wU$(W4}?l%5?M~FB7Jb#|NUAooMaQC+}IT7RG*;lG-);EH%-crnOPy2PA)_iS0 zb~h*F*txhFSPYgGM)-W2(2W2*8<4h_ik8voOdSLn4GZCAAXT8)^-qot6|Kb9rw?|@ zni(r-m=EVw>{{-+@Vh$lnE>w6*EH2oWyEs~U$CN*DYvy8U$k`F9Uj>LIzP}XAZqWm-A+&PMvfexP~4Usw} zy%Gu7V@0*^PU%czYN1Rp?ey4FGg{H2g602m?CMezDYbF{8^S;EWF^b(ChoDCgA8vA z*c&;&k-IpLNV42P6y2dn__vbc2S7ec?HD7ZqNmiaaqi$4;RM^6zLpmw7%P7@ zVDH0PI^acY#$*sw=sp>OQefy@NbiY%_wC zg{FI4BQm*%9cUd*+xN{3#1dwZlJ#04{P&`OQ_aBE7%JO}`^t8czEjEz?RxYkyqXjJp+$v;7Wqv;(m$ooWZCHrhL!pf4pon= z1r=IFm#umgwAUqA2zH$=RX1+i@liaX`~cT;g83ZwK!9q)iwKqM`}6m0^$D3jg~oqI zd3tttMZlKct|}3{tXY}LKVH_9sT9F@CTPk+uRhEPsLCfjJpjg@jWts;v(lITNPo_RUw5|kxHbF(yKh9@Uy^zx>uBO$m*Bwps0qzwb&6ov|7MV z)jm=0#1&gjjmAYs3$YBG*OMEz#S#4)d=I@QE;lSR1gW1e4NsbV_|+OFuKD?EiK2`6 z$1Gr0=VqjzjbDi+0Y-!Llyn1bm#5e$GC#etf*gDd$!suiDx*3pJ3zDZE;~ zj@4W<(9qCe%^Hq&YuC$bi1ZH_elOAFX&SAl`1!t$r=cwR-F?-Yx%;eq!dzi*%4IKq z&yev_NE|%J+lXw5Pvl!jw1WMYiUum=D&Ujw^FtlQ0Tf_a?86wR8Kvk~@OH@suHPrf zuBl$0WJ`?g{X^iwt7X<47krd~$`C6;jNVPOYD;qPrU;_l!t>uY@1I(6S%oV!|b{%-aG zqZ49mKA6odrF(m`4)bn-JgB=V7!fe$168|VM|Mtd+F*NnV0Nq3{1GV zs^5SQM4BZ&rU|lwFY`gLCKc~{UsG$oe@R;qm|wTl?Lfn zwymbvX-|NDx}rdr)NO)FBl?mjtC)%3$2+8?!;zBo%tVqPuBV-}n?ALq=XP}|M@r7% z^%FVU7^1<)=mNV7Erm<=KfQ|jMbmix^-<9i9n5a}E6J*4(5z1(Y#YT?Kj-GO4=VJC zo;jrUWU?--CLuo261SGtJrY=F^SO?v11H^sQ6S%}i0+gl$5*mlfR@!mnQooTl@Ay$ z5T~8XT5nS_sIx;$;#>pVLmZNC)P8?acW>s$^vo~Vw;3q~7_PnGvpn(X)z1A-QiH7J zZ^&0Sg0KI~E~#_MJdL|71Xmin)Ntha1XRu7^s$>CIb@QFVd9tD2#3rC@4Bdm9v?1| zq^`vZ`FQcgx53zfiAIGdMJ@V|d#48$-_d^k?1%HPd`xkF}Gc+)DpU=J# zTuW+Oj|1TCd=6Q(1;uhEOwP%@hwiej;b^vxI*V#@Rst*Bgjd;;0k=~^L2^W$LpG!A z+h#vXZ@EjOW5(S9u#45{|zT-~0-p?CSZAY+MrAZEP^3ElVjhVe4o;|3mY*o4r(zOar z>EX%qnG(3hcGvC1O`T+d5{UU~Y?XUmT{__HL#2r1mEE%X_+n6bufJT>+-dVyIfm72bFM)k6tqi5nO0-ak9c+jnu_JnuK8G+q^Y z0-$AGoX&zze0`<@R469WNjUoPn?4WU_Iq4FyPW;pN2>X?iW%nle{#s2Qq1gJ=eV#* z&+fCgt@?K;E@BpG`=n*8e2LsyA<`D!r#R$ z3;)*j`j26D^((OAXVx_4ruqMo_OYASiJ!9vtkm(YsA|HLN-Jqr4| zeR*ER8io8U9Z(g;KTKP{qbnPCcl*4QXDrW;H-fecRf14+w|#l(y_VxuP&(WE#T>5K z-5p^w8m9UZ7u4a`JYIkIdSK{O_(T7i*72zPrG`DS50zFY3}7K=r$+<5uP6NQsT^{# z@w&p}V4?Z$lf^?GCs){~Qop9e>Nn1RqAow_E-qVgy26x_l6ue=5*-m45xl>mr2bRO z5w%BG=iJd^6={1TYiAv}uHDbE*We%OP#~da&s!1Q;fh*C&>efo0$T`~$vF(} z=17l;X49FSPlW|XrzrzBYouHf9{;uGpG^Oea^!W2LAKS6w>%qTcT6^U%^narUR(2F z-44J;#nh8*Zr=LmE04m)16^@no>|jPH6@GX!RL!3GRxx?F#T(b8Bbf)l(_R-wEPEm zU}mB)?cj{tCbWj*`HPiC!P6taOJ%u@sm3+$TTDw!+nEP)$bPJqlSciMR$kL*if(-G zyT^DQ4yEaSQZJ5^&o_-OLh^&rIhwV?`J5Mh|}!^9ssw9$FkjwrTEUXNL)`Rwy?m z{_Q?2Gy6sm1x{`3ccR}vK$NbM^ed|Cu3B*SVIex^lu*1sAUD2nB0q*n9M$w0CCK?> z1Vit;1cp-h{d1M=IBDLa^99-~m9IX1s89BSL2mPz-N1rokSvRoVF%Kme0jf27BLH0 z7*z7BnVQY>44+f_p2Z9_4^k1GHc&*4McWmW^Bt$Wb;$gp3uNjPx>F})%#vZlpm-w7 zxWGd(ge#&&C5jQgan;GqLobFZkxGQGO{AiQ!TaEN7S;R_`6_!3vyXk2AzoD;kR5#A zt7gRTPtLKKH_>rC>|(FIt(R2a9%+9ky6f#H)#(h?Qx{Pm%G*LLc9%R%w(>@LjC$Ru zP{@ma^@dig3hR#*x3UrdnsYQGXEnWyFjqvkT%R&>;k@KZ0aTixLFme*taQ;m7vSB! zp`vW-k3(FQ<)+1H#|z&p)pFfz14GFI&Xaxn^B+!m^qDo*S~Q+pn@7a{55}uQ+WUtr zMMC&vlr~W^xSR#PsGrTyh80At^y4ZpO1`G6NxCI(ieWA`G{?;aYVWw&d!9ejko${f zIkIIMjr&sb4hH|02_5d~BtnNYQFVju9BZZuWJwisT!43k)4QKBQa=9=cCludB*zXJ zZc+DMI6RLPp7#6ZhHD!1VDL#zaf;H}mhYzhD$4!s1GN&$tKDYjd0&Dn2W`${jxaGN zJF2gEYFOv4$>PdSx=1&%7M`F#aCFe}8&Z+In=u|6XfP-aA?$|BPj!+fcT)YOrE~99 zZo592o_3#kq=r)lkn8+oa1+F?9A7{rlIt-mkHFFr!7BYG@_?Llk<54xy&6Vi$0xnZ zqgzBVFbSRKjHlilS>3wu!ey!HhepK&lrk4%xZyBd^}=?uPit)qUN4x&yII9-rf2k}ja<#HxO!2hs;#5%uZ1aUVM7P>-SPrYai@Nr6oM_1Q~(f51~(yG z@j_!KarZJofU$kIzP|_#E`Uh&k&==6b@An$Th#{Ld2d}#;8xcH^Hps{&>x=^XX$6n zc9!#Sgu#B6CkBFh)dHoBJk>qkg*2*VWUdUd{}j!6!Y(Kd=VJHINvIlz2M_~a&xlt2UPdNWaj+W>9sHS; z%yz0el=B7;xQ=lNzaxyCc8Vq!^W2nxe*dkR&RZAXy%_(gPixQ0(B5GI!I8oK?{!mg z4`p}y_Mgl6eDzm_$-Zb=iE9g_ZNm3|;Ja z)-!&Opfm*dQ`}t%y{~(sM2J@Zf@tJ zX0~Y!r+BRe^{U}h2`+|vbSalmR$S+CfZX`VzNoz^WXd1FB7K4Sx)ZZs-tvHo#kz(M zf+kr`gps)u402b2z_LA5P>zcmZw*I@)a8lvk{5((>GAdpWD&M5eJx%AbW(FtcTv?V zE2}pG@>2V>JI~M9A1H3@v^`;TAytPC18nk>-ISXx)3B7bDg1gx{1C5 zY+`8)r%OZHTye-m(bPjOGRHk-OViBWuXOtJPVuydTenK0pJ6P#9R+4Ewd$)83#_Vi zq9Vb7tpnXH@tacbp2fDK_7V=@nuY^+e_f?96G2_qb+tTwrMC7CTcYA{6+a zCQ^=Pv54KXEg}BV<9pzluQPBrv!|$I0nzt&r>44GxJaEXG%WIZZ?$nWGP`?tx)8&}HJYnpe48RMVIi6gF!$z~28F zx7|ikgaQ{RqWFXKe54E%Fh8wjdY+C*ZqbO;Dow`L)5N;AlV3eTkDLm?>PtkfnGNx9 zWbpd|N*;u_`fvSBf(LH33CwqaK>=5r!X4Nq=^N^c5HqVMQk{~HPS0uDdA6B`3IEPqhl=mGFK z3J9}fJ3lgmiC`?K!#GNv%>No^F8*uvAbq&g>fi7%7VbV-l8+(zu4-*=#unAawJJSh zU4yK$ityUXA^dg0Y9`VrvPd&_-($*yNgpC+M@8>1Ib_s5<Z_{(P$eU{ySo^kz={iO z_xDA+@j5I`v!dOE!@#90lFmDy*_JNdA+6yJcsk3SChPH9l~r{bvThEB4Te$s7c;as zCw&?c?Ghrm-wONAzDnWQQKVugi@Cxt7>)DW9au0yZJqD_Jhq+vr_?)9nWrK{Gn;x9Qfr zcz!O-l(h-V--K-x-of?Bb_-s;18cs|x63K2$!PC;)5og|-#TPwlFpyBZ}q-&c-OI% zJuiQF^?q6Usrc$4Utq9u+EMd1^y6B%j-Rp>47kF;VDq60M`?n0>%@vlyJ7{73GRb= zZY+-%^{6q}s#P>#nj96{r#M(z#ivl1lmRWi;r&RiLb3>QmLgQ!sSV*to54eXHlg=5O=Eu*D1# z`!9~!Z7Z{=2Csmrd-io3t1VU|3nEtwjgj|Y>*HCbG&;y(%Fp=sBlJ1{{N|sX@cAgE zr#+-&kuU+c%%+(V><_fnEnw3fJQa(4&xzwo4={=Wf}c&0we-qiK+UxAb`CPyMd%3f z8J6n2B@fN%rT0E`zFL=bTfo2=b~Dxl1sw*_roXwz2^3U=VCR7O*$F6z84sm2nY4f1 zur7u|E-US!Z`D6;rg`UzAjDm#U9MV|S(;8QM=mBjM8I<4P)cQ%rv%6YU^8n^+gmq`e9oL=&1F|@p&Ck zvKF@6z$}0wO!gj}LNGms@U~_Cg`!aUJUkP#mkT6MAKx?VyIz3+_Q52qnJRVG+MVQ% zu^KjD4VL~=0=Fk{yrz>hQi{@(YWBF|{290hQ(Ng<#Bx&Ty%Li{@+!XSB zDN2lF%!Yx&q{1B`@%|?Z7=#l>Q8JQ~&2(GE zSlgcxnf0-+ej2iWm7ATDeI@XL)Gz0P9~sBiVi8`1xh7EK;qv&vq#jDG(FzUx$H_;` z7zQ!EPNOOjxrc$z8JWX1^w?wsN?%~GjU+w%T`v{UA9zgE%%sCFe_v7kSQi?w{XwEb zNU+<)?a%U?Kc$Y+wy!zO=toy_yAPL)dX7ji5siwvvQOKB-*(;5akI||J&#r5Ppiua zk0m;iYsb4m;pt&Zx-nbP)1LE!tEMz^s)x*|*^yERJdl-HdVjl;22zLqDJfN7OVS>U zaqlBfKt0{n)6R3&B|X@%3RlSS)W*-4z+G-#M-QBhMHEoz7f`XN8tcT(q7-sH;vN~h zE<*SYkbNXRyoIsef#!t{b^RJICn@17H?EAE*_Y!(rcvm(R=mIscO}h z3es>(?;$k^U#~>Pf-5cA2fhnuUqvv%RRwnD&ZUB(m$3JntA>w)vh091VRpthD)8$I zxjyEz8!QdG6M|Ho)j-xMf51?3T0O#e#EcuO`@QLwnM^YLG*%uelW~I%E(0s86`MLD zzxUcW+{`>f|MjfIugv(8-|*hPV@eYzC*a!eDjdwR-#7UAkyNis{^e`m?N0DNCs^23iVxZIu_NHBPsvqh{5YBQd1Tklz@-aIS>BryiH&wCKRmzNO zKQ-N1bL2VvEr|Ug*H*1mF+l6kP_(MT!uV45n1#_FG0%X-lqu6irz{lR4>UVS#Y#i9 zJs#R0hU&6SR)@0eG(@~c)&t$2I=m+aw1w1j+i!e;{jO*MeG%Vz=nL6zn1JF=AgVL( z`#h}$3DiFA6mh~3L3rflM!^9)IPl5*l++nja!ue0|V(JWoX%ICk@Pyx6TEn}5CMPaQdLCt7&oD@<@Dv!e=9 zAp>V}c$eI-hf)vc5n_;IJB62hoIwG!0?C%sRIM8$ty4Geg3`b@hx;GQn*OAIg&>!6 ze$0$0owS-wa~cY*K>ixBmTc4dRh~U^QomIR4*oVBx?Y0!`|ln}==V?@uRu~3!0rGo zMm^r=)(y!CoqL3RzD5*DycL$(oWZLmHAawo5*+L&XG@aWE!wTjvxV-YKw3%NI#K|L zjh)lu6aaQjW*S$V#7IgezE_2cSjTZ97Z|&s%k`-pFI)$x6^7I3v9+`)H@Ca`BhU61 zBwglicZk_q$ z%R9D-ox;vXL;al;I2%lp`II5^x9pSj@S-pxHB)Grpu#C;U#@_#wj-OrM(Y*Podl(< zMf;ZFRn1!sb|H~(U8h^)Y+6P{Yo@!72|jiMW!WvTMMYJ(N;-JhXasba*nytn zrRH=@>+HvkoP$*6i()B)Eh_ZwcEYa407dJb|CIDrAQp5^awsr&O$=dOGu*+c{7I`$ z5Phh6cBz&3--h?Z7sLD85w=)8%44DNw)F81iZm|AA<3?qK_~_mj1DCxzE!4OF>fXnmE2-Z4PLaf z9`;TNtB(Ugnhq&E|03zMwnp+^f*Ou@VJruj`}@xsF~;NZh*Fo+vVK)MX0}u}6#Q}W zdBZcc!t|NIXx^h2_u2m5-;_84h@3i#M9mf z1YwN?ZMvUxvx^Rhyj2xhIyB`(GZ-TL_Ou+lR5KjtAt*d zL%sM@$ue9HditY0=}zKv_}BlOgk*J)#_TH$r>4l;+z!<>XH`({*&l|tO1Ds-me!D?lzliXmXD(Hg)}8Ql%{?u_9s?l}=8uot zAM8Bd3h1%Iq`mB8;1fCGj{|d-%dE>dYg@+lY9oy(?t-gjp5tWU-TV~0Q|VP5+v~wT z3&bfA8uNk2+@Dzx_AuV+A@tGBR$Kell)JtBc~K?|kwKIt>9Q|!UI0;5Ep-tTO}PLi zQz(<>eud`5fTtMrXCi-5U4Cp7P6?@AsvW8OAZ3cv(r=d0tOMt|yoN@wpJ&ispspa# z-uu_E2g?dmlE#~xsAP}URu+KmkM($TrI|_Ws)u>|?doJfPy|J?-=W_Lj z#6JD1eJ5|>)v4-w!eN<-6fkc~WgO~Y zia<^!S5r*`O|H?km;pMAh|_DQ4^J96Ef&EEe(L z&qlL#oXsWe^L3d)y)SM1+azo?tM5;tv}hW#cP?6AaXJqha!YS54*rDj^}YqW94FNw zVCp8?y5}}pAl3-4wXnViB6m*d+-p&P;XkLwaJnBBZ7hpS zaxK;D^-G^VL>|bJsaQzAaXCEKGTD8ImnlfatbAkXfmLmcQ^d{&66|=u*A)qK7RZCozrRTIa&D;f+riU1d9m@quV{3ihV$4tHP99{n zHeU(I)ah*yQ5*OLBYYsVA?RoS%iGmuTeGeZMksab87-h;s51hG63c-54XMG&Cn(@B z9`)bX*XqRd+IFh$Z2?1h$C~sP%_ExFP50}Y=-q!Jvd1?5tWy=QXj$1}E40=x@j^d8 zd|x-{&npOVByqnre-Em}*2kQ`8dC-a(hh&4Q{(l29JLMBxM~^RE&Z&r&_y-<2EV(7 zOlP+DWn=jl*IZ88-Jg@n)%~Db+a}@B69UhQoAXM4mH*b{z2G-;4RNKG5yBjVmgmWn z%k$ZfeBMGE-Hp<_CbvvWYJsBhXtz#XfC;xMbP^XjaH!KRB$RGx(DCyAeDbS@HZV># zr>RAR5doAVbiK}^YBV4*8Q~-8>v8y#fr>e+J}m4qM!3b#?=5#|=ke|5O>}fz%gMk} z`%;|gd|(A5v0L#_z_f9Tqo1U`S@O|7%~-_1PR)DmT1#bh#K+*Hr8%sZdEG9VYyE+D zy7$c4nBKSwa~}!Q;2(O898lf@UZ#$SbdJWQ30&5!9>2yPlLk3VmN#ZxMHJrK_seTLSfH%Q z*L9||d*=BH)geL9^mSyhF`N{nsaX`ZJM*5AW#PuDkAY*^sC^{qC5I7vFmEF5_Y*G; zB;E)l4Odqbe0D!B4m&Y8J;Kx>a~_Zm`Z_|GPB=rA83GNwB)0c(@{q?S+iL;D1^K+K z{y1`g)LiSQzl@cIf{s+wQRIG8ol8gRHKM;vweM*IO<-wZ%bn<*Yw^F~vO7L;Xqa$* z@WafD*7%LJNTrWUKva05Px;Y=7==T4hWymq9TlsP=6RxZvyXjzIyNdy7EBUjVKRJa zby)xuKgK6}HuC!CYpNg74{T1KEne$w0rSyBaD?3ZY>+vj#}!(q!S}A^bcb-ilBLm7PMuF;iS1j%>dV1 zKFz?cmNGCjY{&YP^a8mPA%T0;N#dc$^;P&%6kDWf{Kg4;=-~w=vp8a>lwtULm&pOo zu;AAozOIi4-guzj%-YGH)_N*9R;Ui%SRSE=LU-&ZDzX7W9G5nSYh>ae?*&Av7DI$- z4HeOIPQD*E$h=1ueLivve-LoVBrLLZQ_~Ud^sOfa0}USnZ%8$*{+tQ$lhk@8&hzhL zQBBfakh?}vAj}jV=RBw%z}Rz&P&ZM6GwdsLcXHQ$(s!gX)YCeS)EzO~`~3IAu42Cj z?{M#&Mz>TVvHQCc3xWbOCtVJF$gp^>KwK3B$M_Pb7Uf-ZkH8PjGHCM=DFg*JI5k58 zJr^nhuEv;aFSL;3S!jnkhIq2#^b%5$-13cDh`SYbulA$lVVC``M)zW4Y*|NIOvr2r zUheLymzK?|QK&A;1g5)DRW`D&tmPy_&$7^d;;7u*Pu^z!-FzuRefxO+3Ee(rMWFtB z`?l?D|DGo_5oP0<^8I!^jaoCj4Y3EF#o883{yp%7qA7H{+<9G7M0@CVqN-9Ct2t^& zaJh;twQG*q-$Q=?H63>{P?{?oz!LF$txdX=Oz$drCj0c6;>&_pHgisqd~TN~Cj_3u z^HH84`lzReJdLDvHjT%v>JxY2m7&OhL$!?4%vk6dfh!N$(`mM`gCS92_IDVTg%yzp z)62&LU;5p4$JAH_{%liu!@E_*-TkkM`=zLsxDw0~=C+c>TI5UE{tUx(x_moIj;+FS z#$cAXQkUuLT@+b=%-*m->vj*GIIl!M(gDn5o!?trh8^pZYFnRQfZhzMAVgX4H@aRl^QY;zDg0OArauud!?MTbeI>}ya6ecYb&f^g!Gq@N@%51@3jj$PK>5oAHO={rkJ z1{C1|y{PIbpM|o1eOLbeeE61@a(*AN+{I%mmhc}b-HXJ_zCcH;US?R+e0p^D;RJyL zKg$xhzvj@MB6s<@ZZiRNkcbdNIQt?MmpshujSqbTp~Zqpfh%6ax3stH2KSuweHc@r znsE0TA1AIZ27LjnH8vUJkgY};x4ykEoK8p{yov#P2WBe#m}oDyw*G&O(7nwmOvyroQk z`6$9jvlVeC00A2MDGeJ=Lnvb6s3}4m!@k3CsGgtIoboeypL6=#y4mH0cRKrZ1LWRK z<~=R4_@?c;+?oBzt?2#CpABo$?|z9`IEM3)e2u6?Y|}O)(sjC~5UV`&X4+9G)$t=k z@E$lO1Fnb`$}=eK6*s;ckVPpiP7zB6?nW0_#`O>U(YR}RvRayt^^e1%tl9g=wl6}) z;3}t;L@NceSOD^*;Qi*YfO}RpQc10@%kCcUAsNySCD2U!Kh%{iL?2f`sx`(~Eqw-T zr>5M$C-11ITsg(_gWRCQ-cMsravBU!Fc09SQ z3%^%YL4KqELcO`?m9ZAj;&fAcZXbCx%x_iy?)6=NvZaFS+Z}kMGf_cB(>mPU7)CAA z2XbPaW4nDv9`rLL??)=&eA?#_NmhUYy?LUletqw654rF4&7V};qpUKozYW+9H#%<} z_HOCQc7!co&TC1Y$NOO1c;fua|1J^+uW&uw-HX&JrDWtG?su*zC%=+p1lieq(+2)5 zNEyNaa^G;BjsQ-6!759${9sM|WvG~XLh~BK{jW!M;`F)iw<-c}++BYO0zA8DXFkhb z-d_xt(N6R6Q%(D9)_x&ZJ@LfU*uf9Bb12OxSx0n*V5?jiUnNn7Q7486=Sek6_kx1w z#X`EGv@>Z4MAs`)9bR^!;uKRO#;Ftg(zX9)7-l6kJesUkk{B_);m*hJ%e%dJ8j4}* zf`zZc{&4#c^fKyGVJ6E_CvoBwanZf~$x(9y{nfwiYPKmC&-9TXwJCP)~8w5{S;L z4Gv(kqFcqKxgu5c4w^o|jF4i4F(yVpAG-^W8bO+u#qg6I6Q8`~)>SNh-1JgnOTFC@=@bQPZ zALk*EEnqSugo9M=3>l}{sA@VnMQf--A;>EwM$(@2v~IXJCFiD>cSXvsWFufmotVN zb_`0`S%OYZ$zE3+9{?yKk+N{*3D{@gnZ)&SM}RCG<*R5kf2EX~c^{!@@SnM+o`fr@ zf{K=88;sr0YAI?VI16HVl{pzBm>O{Q@j)WC#}#}UGZ;p z@Ce^-m|h?4VTSMyf%pCV->9KoQCnu>y1|(}#(WXH>V@~Hu_o!fZA!WfhiBB;fR4F+ z4DseyI91DfhIGo7xx6iTG?XIZE8%bV8retp2@Ks?3c455@k>Dyirw%1EOgkHp67<&(C7()dTqcnWYX&0qieKlv_I=nD5>c@v z?UFcRU-R+%)mOU@z+@2;T*NY2n5zox+=wYb`l(fW+2#z7=J|^77+%Nt%fqk6@djn+ z*Zza06pJqHx~-4(IO;P25fiZ)IT+QN6xmZIgL=Bf$68d?hXqqc#XsTZ!`aoDLBIn- ziPY+YdO&M1M;n{nu|NU!&wO&WTnu@2u!QaSee&fn*=H~gQwQ%!%_~Xk&E83hB}qdm zeN&?6sNK(4_&DZ2DAA>8KW^Fz zDg>;F^_2qF2|{nt9;E4nyI+wk(g1(>{a6?}1*tJ?U@GE{avN&VDu?c7kF*v_j|bj* z5r1){E)eNAeE74gzZ3|IN8jf8C%>R`A0!Kz?b^vNRvx!Tj9UGQ_kp} zyC`7y?$2V2#e+^JRjoh;R~dA#Of&1>eIBOZ$`O_dKAiKGdl)lF59vhF-+`FZ*Wb)M z85m~#SUm?^ZUmtt3js1+U1OgrpoVs4@AT5}Ycqbr| zCQeIbL3U^8GQ;)qXhF-{jlMw!##nJkywRxMobARUFCP-BQxImq!6|J<+P|G_MW>zjEEpWS|M=e2$A_3@UMv4j&( zJ~zEtn1<#Sq_T$rZ-jDlHX!}|~ z-%&!fml1hunig@emG9&+m!@WgXS+{-dcW_Hs?dhA?%~JDEnV;jZ{J89gE5V#%I%aj z&1yAY3EUBBVza2TnIA=ZcIVqFT75Bk9l{oB~kGnLHx+V!+rT{DW`@zK3BO<%(v1wR?*R>9ROw)~F1BDnGM zOLtevI&+#AC@Zrp6Ye*%9N~Z+)4RDYspAR)u7j8U;nAbpA0XqbDlCjZ1J-V-!Sq?p6feqVs9Ob>9GG{s&gw!L<0O)4@K0zLbu&=!|0+?)#ZBaLK zK;F8N5WTP{=oQiDw^XiI6RvtB+ePjB;;7u5asM6*q%}A;f?fNiSHPji|A=+x@yH9H z_mR>bz7Te+I^Y35;JwKuBg2ZH`cADPhpnu$2CP6kUfS)c+g*h)Cp;GBV4RhLOFUQXv_op^U2}DitIb%y4<<%@BRA^ZoJ>m`*~i^Ykc!H4Z1d1L1p(d zMf$RD!(x7_y|E7PEB`WvpByi?7_se5n;=cXYeUZ-jYenD)1F`twed5is9}9DN%p-Z z8wRXzgSG&$^sF_%`J~?cD6rNQ>AC!&c*bo~y58pB>`CXOJGJG0nI9${_$RRLRguPCOi7nL_pV$G;^3zD3gou~lvj2k-jbIUL z6)T;PNXJN{zb|~wlxsQ&5eG}^?!mLq_#HhXeB@kt$BFdwpNdjv*3a+Roqsap3H9*_ ze$H+-*q-cmXAoA3I7ob4H%%dlV}j;q#B(J$w6IQC^bLee$PyA_+dHy1XaP^N-S)RL z<9p=qMcv?B^6f9xZ}EKg;NLY@*d;5!&H46B-#s9~n+|IOTC^gYH` zLMK?RooDZ(mJWuvDt7WD*}vXSdE-RCubNyhCA)qj@6QPE3o)(|6Y-vl3=9`d|IzPL zaYpDTq8p4T>_%ExH&S@PCdIKAA#Nf}ojO)9tLwZI)_d6Pg@27;#`!z;kMm-yg=3p2 znuV}638G?sz!S5hAUfY6CZlRAc{wfkF>FM-0ia(tDw#*wm3J#xJKm9G@m&lYjEV|Xr_<&Mbv7+*fTccM~B-?g?X5;%Q0kGQu_$Hs`XPrSBta zFMTz*3Yl5UeAZWX7osM9$WYL?+>m$}>;3%n(H7#VT6q7dXM3ShK6Cyu=>cPHaqjtA zP2&UsTxhw0ix?y{gakS{@2bg2RvA_dwZV#KS1>hbG`7!Y#BR2yE3@TW$#kJS zn_(5sp~{%tr|Fult=GcFL)7}_nEqMZF^FT#6V9R-m`RXPWZqs~rnXf)5A0HwpLwVg z7|i8!LBBlaY|jix@EFcT=|rhmotc&PU&}rOaqGu* zQ`~zZR_lQc-lF4@*%78-ZJF=(YHjekP0%~v2wYLv+eM>-?oX?0rmH1UVPQwY?)8P< zfJ2ZVvIFW=S^b%YU9{q|zd{%EOvEJ zW9NW~V2ZO<%z_(OMaH;uA>QX)%?8<$F|H+@GtgT09bNRgrB3`)k4Nr|^;$Zt`wPS; z=l?ga(#AN`E~wAmC6HwI4-}D!`etcJH(WKwY7`d28bDK%+wlKTkeFCZZ$r)m&pq0_OCGykZ`fX){INL~ji?BC>IP@bFN1QpTZ4@SetGJnI=0Sq#EL$1iiTZL z<(zMfnfvMNwQ<%rAl&c%cxX;V z=@CcPyQa&wpdBJZygo97py2*&+Y@2z6ZoF0I7M+TmfB@v$*5%=W*?)!gFOe)kA%dB z0yiTB$g*5t>MaeE(-hdQDOc0>*>aB(AEU~ucNbIc2Jq$1TB$>(s{#9~mfNxA;O%e( zzfIw!SUeNP8r^xdrrs9Fk;i=A{<2Rjg?id*Ou9{xIZner75SGXde;v5xaoow(u)7QD^ouq#QXAiPFR?D zuRth6Qe`>9BPCyNJM`t7Dd`neRi(tCS@`oacVK!kzm`|k){ahk)-}bwJ1^Jt9X5`>>v@omqro``K7AmR+TP8Z-}`ac75<|9G0($Ug(#(ca^if!h$pKRCQ)IyHbkT8hQiiUFM zSMZ!G0HNXO@Q8%!HKL*<@Uj6o2dqzVpH~j(z9T6GvS&xOfJCjy5{M#bIZ2e^Usm8R z&$AU-VOSWy*BczbtlR8W-RSq?+jggebdt4YM~=q-ZBQ8w8WT2-jI}tP31$+@NAR4| zNaN|T*E(o(^HjGgnnq&F&3McqK*9)VFU*!#xs}w_c#@VTsiQf2!K{krR9sMEeK%1F z_DksIP)L;Q)%GFX6~vXn`gf=ETAf#vuhcdVZ=JSjwR)Ok6tb0FB47VCYq2`2X8Pi* z3+HAX3?2A56`*(@V-iy^vgpg~QF=hEA`FzdM5H`ON=5l}Q^9+P^&%x10rL9un}Lge zaHkuX2l0F%6KWFOQV|Dw*5!+nX0`m%|3TGr-LyMmATW}FU6$l&j&qGU{=8mt zgAi9DjgCmvjdH%){C}(C1Ay^UC)Ay^^I)fUGq<7qHG5SAc>opnd3*InWGjB3&t&g# zao-Uw%ahqku>Js$e0-m&v2umu#?{}+?GR#@%(XL#J>{EVt-j5g1;~6?cel~SqNpJK z9&epZo4od;yy31f?WO$-|?h9R?E)Ywo4d|#+49SRbt1f#=+nU32VmJuYDENT8eUcKWphfNXXSk zaq6APqwe?K-{15V!ib7&IHrh#pTe@3z;{KPH)dqtMc&8O)0p6=XSdQh6+%>eEMV_p zAtyx0-QUvuoCs3gDH?LNfplBtOL4mQRNV(x0}jkee|Guu z>UuIiC{#V{_6qPEzXq1_7UOL}R7Apfn)*Tvd8AP)AQ2 zpA0r`_qD5OY$YVF8wPl^X8I26VoE)R;?uRiIlMrk2MK zHHa_tC_>?BjmtMVw~*eTQ}%yJLB?bEaSJ_hDuNY3j@x-gLy>!7 zcc78T(m;}XP4rQ7h!GJ|wb^lH}h7Wy_eXMYTi(+5#q(#AK z$X>6GQ^=)BI{G$ao`orX?FM=3YpZ-*N8K_z$Xsx8qfcm#1Abp|K zpco;cs!9wVXOw9V?XgEQ?5h@1fg!!T6+J&%V`SdYG$J_^3fiPIK?Eo}8UIN0Zv?Z? zE>c0;T!i!=^q#V@YC!UzroZ7+68c5X=lln46BGn@Ly(_Fa#?!$3sLhIK7P9{d>eOY z{@ie_@!QQ%X$Zt!cN&?UKv}Zhu&>j|>|-h`!tbUQI^@{Rk%2Vqk8#>X@PPj&2w=V6 z8K{T&DO+1_m^8SFu{(A1XQjtG|R;dE)>$U!mEBW`6H|JmcHgXX9PN*qeCNgbU-E1a< z!-`~|ece$`M?Am%s{;qmz8E*qeH-pr$^x)={mG$|1;HsIwhzYCJ*`U2&Jlf{Z;%eB z!$#Hs^Q;q_rf`^@L)mO(1E}aDzM8yT4z3&+4nr6gH2q|*_*=Gbx0${@#16K2$8NpU z(S9f37_%`!b!*3NDBSP8?XQ0{XToiipDHOn>-=&0XtmQ%^`XAf`}%?xE@11