quality-check-python

Quality Check Python

Assume the Smart Collection Python QA role. Keep the yolo/ and upscale/ microservices lint-clean and covered by their available tests.

Scope Detection

• For focused Python changes, run lint/tests for only the touched microservice before checking both services.
• If Python files under upscale/ changed, read upscale/AGENTS.md and run the upscale checks.
• If Python files under yolo/ changed, read yolo/AGENTS.md and run the YOLO checks.
• If the user asks for full Python QA, run checks for both services.

Linters

Run Ruff as the default Python linter. Ruff is a transient QA tool here; do not add it to project requirements unless the user explicitly asks.

Preferred Docker Compose commands, validating the current workspace files:

docker compose run --rm --entrypoint sh -v .:/workspace upscale-service -c "python -m pip install --no-cache-dir ruff && python -m ruff check --ignore E402 /workspace/upscale/main.py /workspace/upscale/test_main.py"
docker compose run --rm --entrypoint sh -v .:/workspace yolo-service -c "python -m pip install --no-cache-dir ruff && python -m ruff check --ignore E402 /workspace/yolo/main.py"

Ignore E402 because both Python services intentionally monkeypatch runtime libraries before importing heavy ML dependencies.

If a service container is unavailable, use an equivalent ephemeral Docker Compose run or clearly report the blocker.

Tests

For upscale/, run the unit tests:

docker compose run --rm --entrypoint python --workdir /workspace -v .:/workspace upscale-service -m unittest upscale/test_main.py

For yolo/, run available tests if present. If there are no test files, run lint plus a syntax check and report that no YOLO test suite exists yet:

docker compose run --rm --entrypoint python -v .:/workspace yolo-service -m py_compile /workspace/yolo/main.py

Fixing Issues

• Preserve the monkeypatches and offline-mode rules documented in the service AGENTS.md files.
• Do not add Python dependencies to requirements.txt without explicit user authorization.
• Do not download models or require GPU access for unit tests.
• Keep changes narrow to the failing linter/test issue.

Final Report

Report:

• Which Python services were checked.
• Which lint/test commands were run.
• Issues found and fixes applied.
• Final Python QA status, including any missing test suite.