Migrate CI pipeline from Azure Pipelines to GitHub Actions

[Copilot]

Migration Plan: Azure Pipelines to GitHub Actions

• Create .github/workflows directory structure
• Create GitHub Actions workflow file that replicates Azure Pipelines functionality
  • Configure triggers for PRs on master and feature/* branches
  • Set up Python version matrix (3.10, 3.11, 3.12, 3.13)
  • Configure dependency installation (pip, poetry, spacy models)
  • Configure pytest execution with --runslow flag
  • Add caching for Poetry dependencies to improve performance
  • Add explicit permissions block for security best practices
  • Add disk space optimization to prevent storage failures
• Validate the workflow configuration
• Run final security scan
• Fix incorrect SpacyRecognizer import in test file
• Remove old azure-pipelines.yml file
• Migration complete!

Storage Optimizations

To prevent "not enough storage on device" failures when installing large NER dependencies (Flair, Stanza, HuggingFace pipelines) across multiple Python versions, the workflow includes:

• Pre-test cleanup: Removes unused pre-installed tools (.NET SDK, Android SDK, GHC, CodeQL) to free up 10-15GB of disk space
• Post-test cleanup: Clears pip and poetry caches after each test run to prevent accumulation across matrix jobs
• Disk monitoring: Tracks available storage with df -h for debugging

These optimizations ensure reliable CI execution despite the heavyweight dependencies required for comprehensive PII detection model testing.

Original prompt

I'd like to migrate the azure pipelines CI pipeline to a Github action. It should do the same process only on GH

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[omri374]

@copilot fix the wrong import for spacy_recognizer and remove the old azure pipelines yaml

[Copilot]

@copilot fix the wrong import for spacy_recognizer and remove the old azure pipelines yaml

Fixed in commit 85938eb:

• Corrected SpacyRecognizer import from presidio_analyzer.predefined_recognizers.spacy_recognizer to presidio_analyzer.predefined_recognizers
• Removed azure-pipelines.yml

[omri374]

@copilot the action sporadically fails due to not enough storage on device. Is there a way to reduce the storage?

[Copilot]

@copilot the action sporadically fails due to not enough storage on device. Is there a way to reduce the storage?

Fixed in commit 09ed59e. Added disk space optimizations:

1. Pre-cleanup: Removes unused tools before tests (.NET, Android SDK, GHC, CodeQL) - frees ~10-15GB
2. Post-cleanup: Clears pip and poetry caches after tests to free space for subsequent matrix jobs
3. Disk monitoring: Added df -h commands to track available space

These changes should prevent the sporadic "not enough storage" failures, especially with the large NER dependencies (Flair, Stanza) across 4 Python versions.