diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 36e5837..169eb23 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -33,6 +33,24 @@ jobs: "pytest>=8.2" "pytest-asyncio>=0.23" \ ruff + - name: Install document recognition engine (pinned tag, needs ENGINES_TOKEN) + # The engine repo is private; ENGINES_TOKEN is a PAT with read access + # to GoodOneFather/TeamHUB_Engines. --no-deps keeps CI light: docling + # and ocrmypdf are imported lazily inside engine functions, only + # pdfminer.six is needed at import time. Pin must match + # app/ingestion/ENGINE_SOURCE.md. + env: + ENGINES_TOKEN: ${{ secrets.ENGINES_TOKEN }} + run: | + if [ -z "$ENGINES_TOKEN" ]; then + echo "::error::ENGINES_TOKEN secret is not available to Actions (repo Settings -> Secrets and variables -> Actions)" + exit 1 + fi + git config --global url."https://x-access-token:${ENGINES_TOKEN}@github.com/".insteadOf "https://github.com/" + pip install "pdfminer.six>=20240706" + pip install --no-deps \ + "teamhub-document-recognition-engine @ git+https://github.com/GoodOneFather/TeamHUB_Engines.git@teamhub-document-recognition-engine-v0.1.0#subdirectory=engines/teamhub-document-recognition-engine" + - name: Ruff run: ruff check app scripts tests @@ -43,12 +61,12 @@ jobs: run: python scripts/report_module_sizes.py - name: Pytest (unit only — heavy deps excluded) - # test_chunker.py skips here: it needs the pinned document recognition - # engine (private TeamHUB_Engines repo, no PAT secret configured). - # It runs in the Docker image and in dev envs with the engine installed. + # The pinned engine is installed above, so test_chunker.py and the + # engine adapter tests run for real here; -rs surfaces any skips. run: | pytest tests/test_hashing.py tests/test_quality.py \ tests/test_chunker.py tests/test_duplicates.py \ + tests/test_document_recognition_engine_adapter.py \ tests/test_hybrid_search.py -q -rs - name: docker compose config