Skip to content

Latest commit

 

History

History
102 lines (74 loc) · 3.28 KB

File metadata and controls

102 lines (74 loc) · 3.28 KB

OpenAPI MCP Server Tests

← Back to main README

This directory contains tests for the OpenAPI MCP Server project.

Test Structure

The tests are organized by module:

  • tests/api/: Tests for API-related modules

    • test_config.py: Tests for API configuration handling
    • test_discovery.py: Tests for the API discovery module
  • tests/prompts/: Tests for prompt-related modules

    • test_instructions.py: Tests for dynamic instruction generation
    • test_operation_instructions.py: Tests for operation-specific prompts
    • test_enhanced_instructions.py: Tests for enhanced API documentation
  • tests/utils/: Tests for utility modules

    • test_cache_provider.py: Tests for the cache provider module
    • test_http_client.py: Tests for the HTTP client utilities
    • test_metrics_provider.py: Tests for the metrics provider module
    • test_metrics_provider_prometheus.py: Tests for the Prometheus metrics provider (skipped if prometheus_client not installed)
    • test_metrics_provider_decorators.py: Tests for the metrics provider decorators
    • test_openapi_validator.py: Tests for the OpenAPI validation utilities
  • tests/test_init.py: Tests for module initialization

  • tests/test_main.py: Tests for the main entry point

  • tests/test_server.py: Tests for the server creation and configuration

Running Tests

To run the tests, use pytest:

# Install test dependencies
pip install "awslabs.openapi-mcp-server[test]"

# Run all tests
pytest

# Run tests with coverage
pytest --cov=awslabs

# Run specific test file
pytest tests/utils/test_cache_provider.py

# Run tests with verbose output
pytest -v

Test Coverage

The tests aim to cover:

  1. Unit Tests: Testing individual components in isolation

    • Configuration handling
    • OpenAPI spec loading and validation
    • Caching mechanisms
    • Metrics collection
    • HTTP client functionality
    • API discovery tools
    • Prompt generation utilities
  2. Integration Tests: Testing components working together

    • Server creation and configuration
    • API mounting and tool registration
    • Authentication handling
    • Dynamic prompt generation
    • Operation-specific prompts

Environment Variables for Testing

Some tests can be influenced by environment variables:

  • ENABLE_CACHETOOLS=true: Test with cachetools integration
  • ENABLE_PROMETHEUS=true: Test with Prometheus metrics (requires prometheus_client package)
  • ENABLE_TENACITY=true: Test with tenacity retry logic
  • ENABLE_OPENAPI_CORE=true: Test with openapi-core validation
  • ENABLE_OPERATION_PROMPTS=true: Test with operation-specific prompts

Mock Strategy

The tests use mocking to isolate components:

  • External HTTP requests are mocked using httpx mocks
  • File operations are mocked using mock_open
  • Environment variables are temporarily set and restored
  • Async functions are tested using pytest.mark.asyncio and AsyncMock
  • MCP server functionality is mocked using MagicMock

Adding New Tests

When adding new tests:

  1. Follow the existing module structure
  2. Use appropriate mocking to avoid external dependencies
  3. Test both success and failure paths
  4. Include tests for edge cases
  5. Ensure tests are isolated and don't depend on external state
  6. For prompt tests, verify both content generation and registration