MajoorWaldi-ComfyUI-Majoor-AssetsManager/docs/TESTING.md

Testing

Version: 2.3.3
Last Updated: February 28, 2026

This project uses pytest (backend) and Vitest (frontend) for comprehensive test coverage. On Windows, batch runners are provided for convenience and generate both:

• JUnit XML (.xml)
• Styled HTML report (.html)

Reports are written to:

• tests/__reports__/

Open the index:

• tests/__reports__/index.html

Test Coverage

Backend (Python)

• Core functionality (index, search, routes)
• Metadata extraction (ExifTool, FFprobe, geninfo)
• Database operations (schema, migrations)
• Security (CSRF, auth, path validation)
• Feature tests (collections, batch ZIP, viewer)
• Regression tests

Frontend (JavaScript)

• Vue.js components
• API client
• UI utilities
• Event handling
• Drag & drop

---

Quick Commands

All Tests (Cross-Platform)

# From repo root
python -m pytest -q

# With verbose output
python -m pytest -v

# With coverage
pytest tests/ --cov=mjr_am_backend --cov-report=html

Single Test File

python -m pytest tests/core/test_index.py -v

Single Test Folder

python -m pytest tests/metadata/ -v

Single Test Function

python -m pytest tests/core/test_index.py::test_scan_recursive -v

Frontend Tests

# Run JavaScript tests
npm run test:js

# Watch mode
npm run test:js:watch

Canonical Quality Gate

# Full repo gate
python scripts/run_quality_gate.py

# Fast Python-only gate (useful before pushing)
python scripts/run_quality_gate.py --python-only --skip-tests

# Tox wrapper for the fast Python-only gate
tox -e quality

The canonical gate runs encoding/BOM checks, ruff, mypy, bandit, pip-audit, xenon/radon complexity checks, backend tests, frontend tests, and npm audit. The pip-audit step audits requirements.txt directly so the local package itself does not need to be published on PyPI for the gate to pass. CI uses the same script for the Python quality job so local and CI behavior stay aligned.

The Python coverage gate currently enforces a minimum combined backend/shared coverage threshold of 60%, which gives the CI pipeline a regression floor without making legacy cleanup block unrelated work.

During the migration to stricter quality thresholds, ruff is enforced on changed Python files while repository-wide hygiene, security, and complexity checks continue to run across the repo. This keeps the gate hard for new work without turning legacy cleanup into a single big-bang change.

Batch runners (Windows)

From the repo root:

• Full suite: run_tests.bat (delegates to tests/run_tests_all.bat)
• Quick suite (skips test_comfy_output): run_tests_quick.bat
• Metadata / parser suite: run_tests_parser.bat

Category runners:

• tests/config/run_tests_config.bat
• tests/core/run_tests_core.bat
• tests/database/run_tests_database.bat
• tests/features/run_tests_features.bat
• tests/metadata/run_tests_metadata.bat
• tests/rating_tags/run_tests_rating_tags.bat
• tests/regressions/run_tests_regressions.bat

All batch runners support /nopause for non-interactive runs:

run_tests_quick.bat /nopause

Test artifacts (DB/WAL/SHM)

Some tests create SQLite files. Pytest temp files are stored under:

• tests/__pytest_tmp__/

This includes *.db, *.db-wal, *.db-shm, and other runtime artifacts.

Parser samples (metadata extraction)

tests/metadata/test_parser_folder_scan.py scans every file under:

• tests/parser/ (recursive)

If you don't want to commit large samples, you can point the test to an external folder:

$env:MJR_TEST_PARSER_DIR = "C:\path\to\parser"
python -m pytest tests/metadata/test_parser_folder_scan.py -q