Dev/master

.github/workflows/ci.yml

@@ -17,8 +17,12 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12"]
+        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12", "3.13"]
         os: [ubuntu-latest, macos-latest, windows-latest]
+        # Exclude Python 3.13 on Windows due to segmentation fault issues
+        exclude:
+          - os: windows-latest
+            python-version: "3.13"
     steps:
       - uses: actions/checkout@v4
 
@@ -27,19 +31,77 @@ jobs:
         with:
           python-version: ${{ matrix.python-version }}
 
+      - name: Install system dependencies (Ubuntu)
+        if: matrix.os == 'ubuntu-latest'
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y pkg-config gfortran g++
+          # Install OpenBLAS for scipy builds (needed for Python 3.13)
+          if [[ "${{ matrix.python-version }}" == "3.13" ]]; then
+            sudo apt-get install -y libopenblas-dev liblapack-dev
+          fi
+
+      - name: Install system dependencies (macOS)
+        if: matrix.os == 'macos-latest'
+        run: |
+          brew update
+          brew install pkg-config
+          # Install gcc, gfortran, and OpenBLAS for Python versions that need compilation
+          if [[ "${{ matrix.python-version }}" == "3.13" ]]; then
+            brew install gcc gfortran openblas
+            # Add Homebrew's bin directory to PATH to ensure gcc and gfortran are found
+            echo "$(brew --prefix)/bin" >> $GITHUB_PATH
+          fi
+
       - name: Install dependencies for tox
         run: |
           python -m pip install --upgrade pip
-          pip install tox
+          pip install tox tox-gh-actions
           pip install poetry
-          pip install pytest
-          poetry install --no-root --without dev
+
+      - name: Install project dependencies with retry
+        # Use poetry install with retry logic for network issues
+        shell: bash
+        run: |
+          # Try with system dependencies first, fall back to source if needed
+          poetry config virtualenvs.create true
+          poetry config virtualenvs.in-project false
+
+          # For Python 3.13, try to install with --no-build-isolation first
+          if [[ "${{ matrix.python-version }}" == "3.13" ]]; then
+            # Use environment variables to help with compilation
+            if [[ "${{ matrix.os }}" == "macos-latest" ]]; then
+              export LDFLAGS="-L$(brew --prefix zlib)/lib"
+              export CPPFLAGS="-I$(brew --prefix zlib)/include"
+            fi
+
+            # Try to install with system site packages enabled for system libraries
+            poetry install --no-root --without dev --no-interaction || \
+            echo "First attempt failed, retrying with different options..."
+
+            # Fallback: try using pip directly with pre-compiled wheels where possible
+            python -m pip install --upgrade setuptools wheel
+
+            # For problematic packages, try to get pre-compiled wheels
+            if [[ "${{ matrix.os }}" == "macos-latest" ]]; then
+              python -m pip install --only-binary :all: numpy scipy matplotlib || \
+              python -m pip install numpy scipy matplotlib --no-build-isolation
+            fi
+          else
+            poetry install --no-root --without dev --no-interaction
+          fi
 
       - name: Run unit tests
-        run: tox -e py
+        shell: bash
+        env:
+          # For Python 3.13 compilation
+          NPY_DISTUTILS_APPEND_FLAGS: 1
+        run: |
+          # Use tox-gh-actions to select the right tox environment
+          tox
 
   linters:
-    runs-on: ubuntu-latest  # Линтеры и доки только на одной версии
+    runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
 
@@ -51,7 +113,7 @@ jobs:
       - name: Install dependencies for linters
         run: |
           python -m pip install --upgrade pip
-          pip install tox
+          pip install tox tox-gh-actions
           pip install poetry
           poetry install --no-root
 

README.md

@@ -56,7 +56,7 @@ Diff-in-Diff) and CUPED methods, to rigorously test hypotheses and validate expe
 - **Data Tests**: Incorporates SMD, KS, PSI, and Repeats tests to affirm the robustness of effect estimations.
 - **Feature Selection**: Employs LGBM and Catboost feature selection to pinpoint the most impactful features for causal
   analysis.
-- **AB Testing Suite**: Features a suite of AB testing tools for comprehensive hypothesis evaluation.
+- **AB Testing Suite**: Features a suite of AB testing tools for comprehensive hypothesis evaluation, including CUPED and CUPAC variance reduction methods with detailed reports.
 - **Stratification support**: Stratify groups for nuanced analysis
 - **Weights support**:  Empower your analysis by assigning custom weights to features, enhancing the matching precision
   to suit your specific research needs
@@ -150,9 +150,12 @@ data = Dataset(
 test = ABTest()  # Classic A/B test
 test = ABTest(multitest_method="bonferroni")  # A/Bn test with Bonferroni corrections
 test = ABTest(additional_tests=['t-test', 'u-test', 'chi2-test'])  # Use can choose tests
+test = ABTest(cuped_features={'post_spends': 'pre_spends'})  # CUPED variance reduction
+test = ABTest(cupac_features={'post_spends': ['pre_spends', 'feature1']})  # CUPAC variance reduction
 
 result = test.execute(data)
 result.resume  # Resume of results
+result.variance_reduction_report  # Variance reduction report for CUPED/CUPAC
 ```
 More about A/B test [here](https://github.com/sb-ai-lab/HypEx/tree/master/examples/tutorials/ABTestTutorial.ipynb)
 

docs/conf.py

@@ -76,19 +76,19 @@
 highlight_language = "python"
 
 html_theme_options = {
-    'logo_only': False,
-    'prev_next_buttons_location': 'bottom',
-    'style_external_links': True,
-    'vcs_pageview_mode': 'blob',
-    'style_nav_header_background': '#2980B9',
+    "logo_only": False,
+    "prev_next_buttons_location": "bottom",
+    "style_external_links": True,
+    "vcs_pageview_mode": "blob",
+    "style_nav_header_background": "#2980B9",
     # Toc options
-    'collapse_navigation': True,
-    'sticky_navigation': True,
-    'navigation_depth': 4,
-    'includehidden': True,
-    'titles_only': False,
-    'globaltoc_collapse': True,
-    'globaltoc_maxdepth': 3,
+    "collapse_navigation": True,
+    "sticky_navigation": True,
+    "navigation_depth": 4,
+    "includehidden": True,
+    "titles_only": False,
+    "globaltoc_collapse": True,
+    "globaltoc_maxdepth": 3,
 }
 
 # Add any paths that contain custom static files (such as style sheets) here,
@@ -97,16 +97,16 @@
 html_static_path = ["_static"]
 
 html_css_files = [
-    'custom.css',
+    "custom.css",
 ]
 
 html_show_sourcelink = False
 html_sidebars = {
-    '**': [
-        'globaltoc.html',
-        'relations.html',
-        'sourcelink.html',
-        'searchbox.html',
+    "**": [
+        "globaltoc.html",
+        "relations.html",
+        "sourcelink.html",
+        "searchbox.html",
     ]
 }
 
@@ -130,7 +130,7 @@
     "ignore-module-all": True,
     "show-inheritance": True,
     "exclude-members": EXCLUDED_MEMBERS,
-    'inherited-members': False,
+    "inherited-members": False,
 }
 
 # order of members in docs, usefully for methods in class

examples/experiments/performance_test/performance_test.py

@@ -49,12 +49,12 @@ def __init__(self, fixed_data_params: dict | None = None):
 
     @staticmethod
     def _generate_synthetic_data(
-            n_columns: int,
-            n_rows: int,
-            n2c_ratio: float,
-            rs: int | None,
-            num_range: tuple,
-            n_categories: int,
+        n_columns: int,
+        n_rows: int,
+        n2c_ratio: float,
+        rs: int | None,
+        num_range: tuple,
+        n_categories: int,
     ) -> pd.DataFrame:
         if rs is not None:
             np.random.seed(rs)
@@ -72,11 +72,11 @@ def _generate_synthetic_data(
         return pd.DataFrame(
             np.hstack((numerical_data, categorical_data)),
             columns=[f"num_col_{i}" for i in range(n_numerical)]
-                    + [f"cat_col_{i}" for i in range(n_categorical)],
+            + [f"cat_col_{i}" for i in range(n_categorical)],
         )
 
     def create_dataset(
-            self, params: dict
+        self, params: dict
     ) -> tuple[Dataset, dict[str, int | tuple[int, int] | float]]:
         all_params = self.fixed_data_params.copy()
         all_params.update(params)
@@ -92,9 +92,9 @@ class ExperimentProfiler:
     default_experiment_params: ClassVar[dict] = {"n_iterations": 10}
 
     def __init__(
-            self,
-            fixed_experiment_params: dict | None = None,
-            experiment: type = AATest,
+        self,
+        fixed_experiment_params: dict | None = None,
+        experiment: type = AATest,
     ):
         fixed_experiment_params = fixed_experiment_params or {}
         self.fixed_experiment_params = self.default_experiment_params.copy()
@@ -116,12 +116,12 @@ class PerformanceTester:
     resume: ClassVar[defaultdict] = defaultdict(dict)
 
     def __init__(
-            self,
-            dataProfiler: DataProfiler,
-            experimentProfiler: ExperimentProfiler,
-            iterable_params: list | None = None,
-            use_memory: bool = True,
-            rewrite: bool = True,
+        self,
+        dataProfiler: DataProfiler,
+        experimentProfiler: ExperimentProfiler,
+        iterable_params: list | None = None,
+        use_memory: bool = True,
+        rewrite: bool = True,
     ):
         self.dataProfiler = dataProfiler
         self.experimentProfiler = experimentProfiler
@@ -153,14 +153,16 @@ def execute(self, file_name, analysis="onefactor"):
                     "analysis",
                     *list(self.experimentProfiler.fixed_experiment_params.keys()),
                     *list(self.dataProfiler.fixed_data_params.keys()),
-                    "time", "M1", "M2"
+                    "time",
+                    "M1",
+                    "M2",
                 ]
                 writer.writerow(row_items)
         with alive_bar(
-                self.get_number_params(),
-                bar="squares",
-                spinner="dots_waves2",
-                title=f"Analysis : {analysis}",
+            self.get_number_params(),
+            bar="squares",
+            spinner="dots_waves2",
+            title=f"Analysis : {analysis}",
         ) as bar:
             for params, data, experiment in tqdm(self.get_params()):
                 combined_params = {**data[1], **experiment[1]}
@@ -184,7 +186,7 @@ def execute(self, file_name, analysis="onefactor"):
                 process.join()
                 monitor.join()
 
-                max_memory_mb = return_dict2["max_memory"] / 1024 ** 2
+                max_memory_mb = return_dict2["max_memory"] / 1024**2
 
                 with open(file_name, "a", newline="") as file:
                     writer = csv.writer(file)
@@ -234,14 +236,14 @@ def function_performance(self, func, param_dict, return_dict):
 
         return_dict["results"] = [
             exec_time,
-            memory_usage / 10 ** 6 if self.use_memory else None,
+            memory_usage / 10**6 if self.use_memory else None,
         ]
 
 
 def performance_test_plot(
-        params: dict,
-        output_path: str,
-        title="The results of the one-factor performance test of the AA Test",
+    params: dict,
+    output_path: str,
+    title="The results of the one-factor performance test of the AA Test",
 ):
     df = pd.read_csv(output_path)
     df = df[df.analysis == "onefactor"]