R Econometrics MCP Server

# CLAUDE.md This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository. ## Project Overview RMCP (R Model Context Protocol) is a statistical analysis server that bridges AI assistants with R statistical computing capabilities through the Model Context Protocol. ## Development Commands ### Setup & Running ```bash poetry install # Install dependencies poetry run rmcp start # Start server in stdio mode poetry run rmcp http # Start HTTP server with SSE ``` ### Testing ```bash pytest tests/unit/ # Run unit tests (no R required) pytest tests/integration/ # Run integration tests (requires R) pytest tests/e2e/ # Run end-to-end tests pytest -xvs tests/test_file.py::test_name # Run single test ``` ### Code Quality ```bash black rmcp tests streamlit # Format code isort rmcp tests streamlit # Sort imports flake8 rmcp tests streamlit # Lint code mypy rmcp # Type checking ``` ## Architecture ### Core Components - **Transport Layer** (`rmcp/transport/`): Handles stdio and HTTP+SSE communication - **Core Server** (`rmcp/core/`): MCPServer, Context management, JSON-RPC protocol - **Registries** (`rmcp/registries/`): Dynamic registration for tools, resources, prompts - **Tools** (`rmcp/tools/`): 44 statistical analysis tools across 11 categories - **R Integration** (`rmcp/r_integration.py`): Python-R bridge via subprocess + JSON ### Adding New Statistical Tools 1. Create tool function in appropriate file under `rmcp/tools/` 2. Use `@tool` decorator with input/output JSON schemas 3. Implement corresponding R script in `r_assets/scripts/` 4. Write tests in `tests/unit/tools/` and `tests/integration/` ### Key Design Patterns - **Registry Pattern**: All tools/resources/prompts use centralized registries - **Context Pattern**: Request-scoped state management with lifecycle - **Transport Abstraction**: Common interface for stdio/HTTP transports - **Virtual Filesystem**: Security sandbox for file operations (`rmcp/security/vfs.py`) ### Testing Strategy - Unit tests mock R integration for speed - Integration tests verify actual R execution - E2E tests validate full MCP protocol flows - Fixtures in `conftest.py` provide server/context factories ## Important Notes - Python 3.10+ required - R must be installed for integration/E2E tests - All R communication uses JSON via subprocess - HTTP transport includes session management and SSE for streaming - VFS security restricts file access to configured paths only