name: Quality

on:
  push:
  pull_request:

jobs:
  test-and-benchmark:
    runs-on: ubuntu-latest
    timeout-minutes: 20
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
      - uses: actions/setup-python@v5
        with:
          python-version: "3.11"
          cache: pip
      - name: Check changed code whitespace
        env:
          BASE_SHA: ${{ github.event.pull_request.base.sha || github.event.before }}
          HEAD_SHA: ${{ github.sha }}
        run: |
          if [ "$BASE_SHA" = "0000000000000000000000000000000000000000" ]; then
            git diff-tree --no-commit-id --check --root -r "$HEAD_SHA"
          else
            git diff --check "$BASE_SHA...$HEAD_SHA"
          fi
      - run: pip install -r requirements.txt
      - run: pytest -q
      - name: Run synthetic golden quality gate
        env:
          INFERENCE_PROVIDER: fallback
        run: >-
          python -m simpliscribe.benchmark
          --cases data/golden_cases.v1.json
          --output data/benchmark_runs/ci.json
          --min-f1 0.85
          --max-hallucination-rate 0.10