knowledge-server CLAUDE.md

Development Tips

• Always source venv before running any command

Testing and Code Quality

Running Tests

Tests use a fast embedding model (sentence-transformers/all-MiniLM-L6-v2) for speed. Most tests that were previously slow should now complete quickly.

source venv/bin/activate
python -m pytest tests/ -v

Bedrock Tests

Tests that require AWS Bedrock (LLM calls) are marked with @pytest.mark.bedrock and are skipped by default to keep tests fast. To run them:

# Run only bedrock tests
python -m pytest tests/ -m bedrock -v

# Run all tests including bedrock
python -m pytest tests/ -m "not slow" -v

# Run everything (including bedrock and slow tests)
python -m pytest tests/ --tb=short

Test Environment Configuration

The project includes a .env.test file with optimized settings for testing:

• Fast embedding model: sentence-transformers/all-MiniLM-L6-v2 (22M parameters)
• MPS device for fast GPU acceleration on Apple Silicon
• Reduced token limits for faster processing

If you need to test with production embedding models, temporarily modify the test files or use the main .env configuration.

Code Formatting

The project uses black (line length 100), isort, and flake8 for code formatting and linting:

source venv/bin/activate
black src/ tests/
isort src/ tests/
flake8 src/ tests/

Pre-commit Cleanup

Before committing, run the complete cleanup:

source venv/bin/activate
black src/ tests/
isort src/ tests/
python -m pytest tests/