Ewald
094cb552c2
- Require the same Comfy auth on all Majoor GET routes as on writes, closing a read-confidentiality gap for remote/multi-user deployments (no-op when Comfy auth is disabled, the default single-user setup). - Gate nightly and stable releases on a quality job (pytest, mypy, frontend typecheck/build/tests, dist/ drift check) before publishing, and refresh the committed dist/ bundle to match a clean build. - Pin softprops/action-gh-release and Comfy-Org/publish-node-action to commit SHAs. - Resolve ComfyUI's CacheProvider through the public comfy_api.latest.Caching alias instead of the underscore-private _caching module. - Make asset-delete DB cleanup atomic: a failed cleanup statement now aborts the transaction instead of committing a partial delete.
Tests - Majoor Assets Manager
Test suite for ComfyUI-Majoor-AssetsManager.
Test Structure
tests/
├── conftest.py # Pytest fixtures & shared setup
├── README.md # Test documentation
├── run_tests_all.bat # Windows full-suite runner
├── __init__.py
├── assets/ # Asset fixtures and media samples
├── compat/ # Backward-compatibility coverage
├── config/ # Configuration tests and runner
├── core/ # Core route and path helpers
├── database/ # Schema and SQLite behavior
├── features/ # Main feature, service, and route coverage
├── metadata/ # Metadata extraction coverage
├── parser/ # Parser-focused regression samples
├── quality/ # Quality-gate tests
├── rating_tags/ # Ratings, tags, and sync behavior
├── regressions/ # Targeted regression coverage
├── routes/ # Route-level scenarios
└── security/ # Security-specific tests
Running Tests
All tests (Windows)
Double-click run_tests.bat in the project root, or:
cd ComfyUI/custom_nodes/ComfyUI-Majoor-AssetsManager
python -m pytest tests/ -q
Single test file
python -m pytest tests/core/test_routes.py -v
Single folder
python -m pytest tests/metadata/ -v
Single test function
python -m pytest tests/core/test_routes.py::test_routes -v
With coverage
pytest tests/ --cov=mjr_am_backend --cov=mjr_am_shared --cov-report=html
Test Categories
core/ - Core functionality
- Result pattern (Ok/Err)
- IndexService scan/search
- FTS5 full-text search
- API route handlers
- End-to-end integration
metadata/ - Metadata extraction
- ExifTool/FFprobe extraction
- PNG/WEBP/video generation info
- Workflow JSON parsing
- Quality score calculations
rating_tags/ - Ratings & tags
- Rating/tag extraction from files
- OS file sync
- Import operations
assets/ - Assets & search
- Batch operations
- Delete operations
- FTS5 search
- Staging deduplication
database/ - Database & schema
- Schema migration/healing
- SQLite query patterns
- Geninfo self-healing
config/ - Config & security
- CSRF protection
- Origin validation
- Rate limiting
- Path traversal prevention
- Settings persistence
features/ - Feature tests
- Collections CRUD
- Multi-file ZIP export
- Date histogram
- Viewer grid masks
regressions/ - Regression tests
- Audit issue fixes
- Scan batching regressions
- Logging system
Writing Tests
Test isolation
Each test should be independent using in-memory databases:
@pytest.mark.asyncio
async def test_example():
db = Sqlite(":memory:")
await migrate_schema(db)
# ... test logic
await db.aclose()
Result pattern
Always verify the Result pattern:
def test_service_result():
result = service.do_something()
assert isinstance(result, Result)
if result.ok:
assert result.data is not None
else:
assert result.error is not None
Mock external tools
Don't depend on ExifTool/FFprobe availability:
class MockExifTool:
def is_available(self):
return False
Fixtures
Common fixtures are defined in conftest.py:
test_db- In-memory SQLite with schematemp_dir- Temporary directory for test filesmock_exiftool- Mocked ExifTool adapter
Resources
Last updated: April 5, 2026