Added Python module isolation-forest-onnx which can convert an isolation forest model to ONNX format.

[jverbus]

Added Python module isolation-forest-onnx which can convert an isolation forest model to ONNX format.

Description

This PR introduces the isolation-forest-onnx module to the repository. The module provides a converter for the LinkedIn Isolation Forest model to the ONNX format, enabling seamless integration and deployment across various platforms supporting ONNX.

Key Changes:

• Updated the gradle build configuration to support Python in addition to the existing Scala from the isolation-forest module.
• Added the isolation-forest-onnx module, including:
  • The core IsolationForestConverter class for converting Isolation Forest models to ONNX format.
  • Full testing coverage, including end-to-end correctness testing.

Testing:

• Unit Tests: Added coverage for core functionality with pytest and coverage. These tests include end-to-end correctness testing comparing the output of both the Spark/Scala and ONNX formatted models. All new tests are passing.
• Manual Verification: Tested the module by converting a sample Isolation Forest model to ONNX and verifying the ONNX output.

Next steps

After this PR is committed, there will be future PRs to do the following:

• Publish build artifacts to PyPI: Add jobs in GitHub Actions to publish artifacts to the PyPI repository, so they can be easily used via pip install isolation-forest-onnx.
• Update README.md: Update the README.md to provide instructions on how to use the ONNX converter.

Contributors

Special thanks to the following technical contributors to this converter (alphabetical order): Angie Zhu, Deepesh Nathani, Devendra Jaisinghani, James Verbus, Siva Popuri, Ying Su

[wangzhongqing]

这是来自QQ邮箱的假期自动回复邮件。王中情：您好，已收到您的邮件，我会尽快给您回复。谢谢您的邮件！祝您生活愉快！

[angie-z]

LGTM!

Thank you so much for adding this module!