name: Validate Training Data

on:
  pull_request:
    paths:
      - 'training/data/**/*.jsonl'
      - 'training/data/**/schema.json'
      - 'training/data/**/validate.py'

jobs:
  validate:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'

      - name: Validate scene description JSONL
        run: |
          cd training/data/scene-descriptions
          python3 validate.py *.jsonl
      - name: Validate training data provenance
        run: |
          cd training
          python3 -c "
          from training_pair_provenance import validate_provenance
          import json, sys, glob
          issues = 0
          for f in glob.glob('data/*.jsonl'):
              report = validate_provenance(f)
              print(f'{f}: {report["coverage"]:.0f}% coverage ({report["with_provenance"]}/{report["total"]})')
              if report['missing_provenance'] > 0:
                  print(f'  WARNING: {report["missing_provenance"]} pairs missing provenance')
          sys.exit(0)
          "