coverage-analysis

Coverage Analysis Guide

Purpose

Provide Claude with a comprehensive reference for running code coverage analysis in Your Project. This skill covers the coverage.sh script usage, interpreting results, and best practices.

Activation Cues

• Running coverage analysis
• Checking test coverage before PRs
• Investigating which code paths need more tests
• Generating coverage reports

Quick Reference

# Run coverage with HTML report (default)
./coverage.sh

# Run coverage with text report only (faster)
./coverage.sh --no-html

# Include integration tests
./coverage.sh --integration

# Include integration tests, text only
./coverage.sh --integration --no-html

Output Locations

Output	Location
Text report	/tmp/$PROJECT_NAME/coverage/coverage_report.txt
HTML report	/tmp/$PROJECT_NAME/coverage/index.html
Quick link	file:///tmp/coverage.html (symlink)

Understanding Results

Test Summary

Test Summary:
  Total tests: 232
  Passed: 204
  Failed: 28

Coverage Summary

Overall Coverage: 69%

Key Files Coverage:
main.py                    44%
llm_service.py            74%
game_state.py             57%
firestore_service.py      53%

Coverage Thresholds

CI Workflow Thresholds

Coverage	CI Result	Badge Color
≥80%	Pass	Green (MINIMUM_GREEN: 80)
60-79%	Pass (warning)	Orange (MINIMUM_ORANGE: 60)
<60%	Fail	Red

Quality Guidelines

Coverage	Status	Action	CI Threshold
80%+	Excellent	Maintain	Green (MINIMUM_GREEN: 80)
60-79%	Good	Minor improvements	Orange (MINIMUM_ORANGE: 60)
40-59%	Fair	Add targeted tests	Warning
<40%	Poor	Prioritize testing	Error

Key Files to Monitor

Priority coverage targets for Your Project:

1. narrative_response_schema.py - LLM response parsing
2. llm_service.py - AI integration
3. world_logic.py - Core game mechanics
4. firestore_service.py - Data persistence
5. game_state.py - State management
6. main.py - API endpoints

Running Single Test Files

For focused development, run individual test files:

# Using venv with proper PYTHONPATH
source venv/bin/activate
PYTHONPATH="$PWD:$PYTHONPATH" TESTING=true python $PROJECT_ROOT/tests/test_specific.py

Common Issues

ModuleNotFoundError: No module named 'mvp_site'

Cause: PYTHONPATH doesn't include project root.

Fix: The coverage.sh script sets PYTHONPATH="$PROJECT_ROOT:$PYTHONPATH" automatically. For manual runs:

export PYTHONPATH="$PWD:$PYTHONPATH"

Test Fails with Missing Dependencies

Fix: Install test dependencies:

pip install numpy fastembed onnxruntime

Browser/Integration Tests Failing

Some tests require additional setup:

• Browser tests need Playwright: playwright install
• Integration tests may need running services

GitHub Actions Integration

For CI coverage reporting, use py-cov-action/python-coverage-comment-action@v3.32 (SHA-pinned in workflow):

- name: Coverage comment
  uses: py-cov-action/python-coverage-comment-action@fb02115d6115e7b3325dc3295fe1dcfb1919248a  # v3.32
  with:
    GITHUB_TOKEN: ${{ github.token }}
    MINIMUM_GREEN: 80
    MINIMUM_ORANGE: 60

See .github/workflows/coverage.yml for the full workflow with threshold checks.

Best Practices

1. Run before PRs: Check coverage impact before submitting
2. Focus on critical paths: Prioritize business logic coverage
3. Don't chase 100%: Focus on meaningful test coverage
4. Review uncovered lines: Use HTML report for line-by-line analysis
5. Mock external services: Use TESTING=true for isolated testing

Related Skills

• end2end-testing.md - E2E test patterns
• evidence-standards.md - Test evidence collection