Testing Strategy¶

RAG Modulo employs a comprehensive testing strategy with 947+ automated tests organized by speed and scope.

Test Organization¶

Test Categories¶

Tests are organized into four main categories:

Atomic Tests (Fastest - ~5 seconds)
Schema and data structure validation
Pydantic model tests
No database required
Unit Tests (Fast - ~30 seconds)
Individual function/class testing
Mocked dependencies
No external services
Integration Tests (Medium - ~2 minutes)
Service interaction tests
Real database operations
Vector database integration
End-to-End Tests (Slower - ~5 minutes)
Full system tests
API to database workflows
Complete feature validation

Test Markers¶

Tests use pytest markers for categorization:

@pytest.mark.atomic - Atomic schema tests
@pytest.mark.unit - Unit tests
@pytest.mark.integration - Integration tests
@pytest.mark.e2e - End-to-end tests
@pytest.mark.api - API endpoint tests
@pytest.mark.performance - Performance benchmarks

Running Tests¶

See Running Tests for detailed commands and usage.

Test Coverage¶

Minimum Coverage: 60%
Current Tests: 947+ automated tests
Coverage Report: make coverage (generates HTML report)

Testing Best Practices¶

Unit Testing¶

Mock external dependencies
Test one component at a time
Use fixtures for common setup
Keep tests fast and isolated

Integration Testing¶

Use real services (Postgres, Milvus)
Test service interactions
Validate data persistence
Clean up test data

End-to-End Testing¶

Test complete workflows
Validate API contracts
Test error handling
Verify business logic

Test Fixtures¶

Common fixtures are defined in tests/conftest.py:

db_session - Database session
test_client - FastAPI test client
sample_user - Mock user for testing
sample_collection - Mock collection data

Continuous Integration¶

All tests run in CI/CD pipeline:

On Every PR: Atomic + Unit tests (~2 min)
On Push to Main: All tests including integration (~5 min)

See CI/CD Documentation for details.

Test Organization Structure¶

tests/
├── unit/           # Unit tests with mocks
│   ├── services/   # Service layer tests
│   ├── repository/ # Repository layer tests
│   └── schemas/    # Schema validation tests
├── integration/    # Integration tests with real services
├── api/            # API endpoint tests
└── performance/    # Performance benchmarks

Writing New Tests¶

Determine Test Type
Atomic: Schema/model validation
Unit: Single component with mocks
Integration: Multiple components with real services
E2E: Full workflow testing

Add Appropriate Markers

@pytest.mark.unit
def test_search_service_query():
    # Test implementation

Use Fixtures

def test_create_collection(db_session, sample_user):
    # Test using fixtures

Assert Clearly
Use descriptive assertions
Test both success and failure cases
Validate error messages