Build Streamlit app for looking and interacting with results

[ngupta23]

Add a Streamlit app that can be used to compare results as well as filter to a subset of the benchmarks for viewing (filters)

Use Cases

• Comparison between Keys (for filtering)
  • os
  • python_version
  • library
  • library_version
  • dataset
  • group
  • model
  • model_engine
  • model_engine_version
  • execution_engine
  • execution_engine_version
  • execution_mode
  • execution_mode_version
  • num_cpus
  • backup_model
• Comparison over time
• If no env changes, but pycaret settings change (in setup or create_model), do results become better or worse?
• Comparison of Ray and Spark
• Comparison of Native vs. Fugue
• Compare the performance of Pycaret across different python versions
• How does pycaret/ray/spark scale with the number of CPUs
• Running performance tests (nightly/weekly) - if nothing changes, performance should not change
• Metrics when a new version of pycaret is released (compared to the previous version)

Related to #10