CEDAR MCP Server

# Integration Tests for CEDAR MCP Server This directory contains comprehensive integration tests for the CEDAR MCP server that test real API interactions with both CEDAR and BioPortal APIs. ## Prerequisites ### 1. Install Test Dependencies ```bash pip install -r requirements-dev.txt ``` ### 2. Set Up API Keys Create a `.env.test` file in the project root with your actual API keys: ```bash # .env.test CEDAR_API_KEY=your-cedar-api-key-here BIOPORTAL_API_KEY=your-bioportal-api-key-here ``` **Important:** Never commit `.env.test` to version control. It's already in `.gitignore`. ### 3. Get API Keys #### CEDAR API Key 1. Go to [cedar.metadatacenter.org](https://cedar.metadatacenter.org) 2. Create an account or log in 3. Click on the person icon → Profile 4. Copy your API key from the profile page #### BioPortal API Key 1. Go to [bioportal.bioontology.org](https://bioportal.bioontology.org) 2. Create an account or log in 3. Go to Account Settings → API Key 4. Copy your API key ## Running Tests ### Run All Tests ```bash pytest ``` ### Run Integration Tests Only ```bash pytest -m integration ``` ### Run Unit Tests Only ```bash pytest -m unit ``` ### Run Tests with Coverage ```bash pytest --cov=src/cedar_mcp --cov-report=html ``` ### Run Specific Test Files ```bash # Test external API functions pytest test/test_external_api.py # Test processing functions pytest test/test_processing.py # Test server functions pytest test/test_server.py ``` ### Run Specific Test Classes ```bash pytest test/test_external_api.py::TestGetChildrenFromBranch ``` ### Run Specific Test Methods ```bash pytest test/test_external_api.py::TestGetChildrenFromBranch::test_get_children_valid_branch ``` ### Skip Slow Tests ```bash pytest -m "not slow" ``` ## Test Categories ### Integration Tests (`@pytest.mark.integration`) - Test real API calls to CEDAR and BioPortal - Require actual API keys in `.env.test` - May be slower due to network requests - Test actual data transformation with real responses ### Unit Tests (`@pytest.mark.unit`) - Test logic without external API dependencies - Use mocked data and responses - Fast execution - Test edge cases and error handling ### Slow Tests (`@pytest.mark.slow`) - Tests that take longer to run (>5 seconds) - Complex API interactions - Large data processing ## Troubleshooting ### Tests Skipped Due to Missing API Keys ``` SKIPPED [1] test/conftest.py:15: CEDAR_API_KEY not found in .env.test ``` **Solution**: Create `.env.test` file with valid API keys. ### Network Timeout Errors ``` requests.exceptions.ConnectTimeout: HTTPSConnectionPool ``` **Solution**: Check internet connection and API service status. ### Authentication Errors ``` 401 Unauthorized ``` **Solution**: Verify API keys are valid and not expired. ### BioPortal Rate Limiting ``` 429 Too Many Requests ``` **Solution**: Add delays between tests or reduce test frequency. ## Contributing When adding new tests: 1. **Follow Naming Convention**: `test_*.py` files, `test_*` functions 2. **Add Appropriate Markers**: `@pytest.mark.integration` or `@pytest.mark.unit` 3. **Handle API Failures Gracefully**: Use try/except and meaningful assertions 4. **Document Test Purpose**: Clear docstrings explaining what each test validates 5. **Use Fixtures**: Reuse existing fixtures or create new ones in `conftest.py` ## Using the Test Runner The project includes a convenient test runner script: ```bash # Run all tests python run_tests.py # Run only integration tests python run_tests.py --integration # Run only unit tests (no API calls) python run_tests.py --unit # Run fast tests (excluding slow ones) python run_tests.py --fast # Run with coverage report python run_tests.py --coverage # Run specific test modules python run_tests.py --external-api python run_tests.py --processing python run_tests.py --server # Run with verbose output python run_tests.py --verbose # Run without warnings python run_tests.py --no-warnings ```