Test Coverage¶

Carbon Connect maintains comprehensive test coverage across the backend, frontend, and pipeline layers.

Coverage Targets¶

The quality gates workflow enforces a minimum coverage threshold of 69% (--cov-fail-under=69) to allow for incremental improvement toward the 85% target.

Backend Tests: 533 Tests¶

Test Breakdown by Category¶

Running Coverage¶

Backend Coverage¶

# Run with coverage report
poetry run pytest --cov=backend --cov-report=html

# Run with terminal output
poetry run pytest --cov=backend --cov-report=term-missing

# Run with minimum threshold enforcement
poetry run pytest --cov=backend --cov-fail-under=69

# Run specific category
poetry run pytest tests/unit/services/test_matching_engine.py --cov=backend.app.services.matching_engine

Coverage Report¶

After running with --cov-report=html, open htmlcov/index.html in a browser to view the detailed coverage report.

Frontend Coverage¶

cd frontend
npm run test -- --coverage

CI Coverage Enforcement¶

Quality Gates Workflow¶

The quality-gates.yml workflow runs pytest with coverage checks:

- name: Run Tests with Coverage
  run: |
    poetry run pytest tests/ \
      --maxfail=3 \
      --cov=backend \
      --cov-report=xml:coverage.xml \
      --cov-report=term-missing \
      --cov-fail-under=69

Codecov Integration¶

Coverage reports are automatically uploaded to Codecov on every push to main:

- name: Upload coverage
  uses: codecov/codecov-action@v4
  with:
    files: coverage.xml
    token: ${{ secrets.CODECOV_TOKEN }}

Improving Coverage¶

Priority Areas¶

When improving coverage, focus on these areas in order:

1. Business logic -- Matching engine, carbon scoring, grant normalization
2. API endpoints -- Request validation, authorization, error responses
3. Error handling -- Network failures, invalid data, edge cases
4. Integration points -- Database queries, cache operations

Writing High-Value Tests¶

Tests that provide the most coverage value:

# Test the happy path AND error cases
@pytest.mark.asyncio
async def test_grant_search_with_all_filters(client):
    """Test search with every filter applied simultaneously."""
    response = await client.get("/api/v1/grants", params={
        "query": "renewable energy",
        "country": "DE",
        "nace_code": "D35.1",
        "is_carbon_focused": True,
        "status": "active",
        "min_funding": 50000,
        "max_funding": 500000,
        "page": 1,
        "per_page": 10,
        "sort_by": "deadline",
    })
    assert response.status_code == 200

Tests to Avoid¶

• Tests that only verify mock behavior (not real functionality)
• Tests that duplicate coverage already provided by type checking
• Overly specific tests that break with minor refactors
• Tests that take longer than 5 seconds without good reason