MCP Server for Splunk

# Test Suite for MCP Server for Splunk A comprehensive test suite covering all aspects of the MCP Server for Splunk implementation. ## Test Structure ### Core Test Files 1. **`test_splunk_tools.py`** - Tests for all 12 MCP Splunk tools (19 tests) - Health checks (`get_splunk_health`) - Index operations (`list_indexes`) - Metadata tools (`list_sourcetypes`, `list_sources`) - Search operations (`run_oneshot_search`, `run_splunk_search`) - App/user management (`list_apps`, `list_users`) - KV Store operations (`list_kvstore_collections`, `get_kvstore_data`, `create_kvstore_collection`) - Configuration management (`get_configurations`) 2. **`test_transport.py`** - Transport layer testing (27 tests) - stdio transport configuration and execution - streamable-http transport with various configurations - Environment variable handling - Error conditions and security scenarios - Docker environment compatibility 3. **`test_splunk_client.py`** - Splunk connection testing (6 tests) - Connection establishment and failure scenarios - Credential validation - Port handling and conversion 4. **`test_mcp_server.py`** - Integration tests (17 tests - currently excluded) - Requires additional FastMCP client dependencies - End-to-end testing with real MCP clients - Currently excluded from default test runs ### Test Configuration - **`conftest.py`** - Comprehensive pytest fixtures and mocks - **`pyproject.toml`** - Pytest configuration and coverage settings - **`scripts/run_tests.sh`** - Enhanced test runner with multiple options ## Running Tests ### Quick Start ```bash # Run all core tests (excluding integration tests) uv run scripts/run_tests.sh # Or directly with pytest uv run pytest --ignore=tests/test_mcp_server.py ``` ### Test Categories ```bash # Transport layer tests only uv run scripts/run_tests.sh -k transport # Splunk tools tests only uv run scripts/run_tests.sh -k splunk_tools # Splunk client tests only uv run scripts/run_tests.sh -k splunk_client # Integration tests (requires additional setup) uv run pytest tests/test_mcp_server.py ``` ### Test Options ```bash # Quick tests without coverage uv run scripts/run_tests.sh --no-coverage -x # Verbose output uv run scripts/run_tests.sh -v # Specific test pattern uv run scripts/run_tests.sh -k health_check # Fail on first error uv run scripts/run_tests.sh -x ``` ## Coverage Current test coverage: **63% overall** - `src/server.py`: 59% (272 statements) - `src/splunk_client.py`: 85% (41 statements) - `src/__init__.py`: 100% (2 statements) Coverage reports are generated in: - Terminal output - HTML report: `htmlcov/index.html` ## Test Results ✅ **52 core tests passing** - 19 Splunk tools tests - 27 transport tests - 6 Splunk client tests ❌ **17 integration tests excluded** (require FastMCP client dependencies) ## Mock Architecture The test suite uses extensive mocking to simulate Splunk services: - **Mock Splunk Service**: Simulates complete Splunk SDK client - **Mock Context**: FastMCP context with lifespan management - **Mock Search Results**: Realistic search response data - **Mock KV Store**: Collection and document structures - **Mock Configurations**: Splunk configuration stanzas ## Dependencies Testing dependencies managed via `uv`: - `pytest` - Test framework - `pytest-asyncio` - Async test support - `pytest-mock` - Enhanced mocking - `pytest-cov` - Coverage reporting ## Future Enhancements 1. **Integration Test Completion**: Add FastMCP client fixtures for full end-to-end testing 2. **Performance Tests**: Add benchmarking for large search results 3. **Error Scenario Expansion**: More comprehensive failure mode testing 4. **Parallel Execution**: Add pytest-xdist for faster test runs 5. **Property-Based Testing**: Use hypothesis for edge case discovery ## Contributing When adding new tests: 1. Follow the existing class-based organization 2. Use descriptive test names explaining the scenario 3. Include comprehensive assertions 4. Update this README for new test categories 5. Maintain >60% coverage threshold ## Notes - Integration tests (`test_mcp_server.py`) are excluded by default due to missing FastMCP client fixtures - All core functionality is thoroughly tested with unit tests - The test suite is compatible with both uv and traditional pip environments - Mock objects closely mirror real Splunk SDK behavior for accuracy