mcpydoc 0.2.0

Model Context Protocol (MCP) server providing documentation and code analysis for Python packages

MCPyDoc - Python Package Documentation MCP Server

MCPyDoc is a Model Context Protocol (MCP) server that provides comprehensive documentation and code analysis capabilities for Python packages. It enables AI agents like Cline and GitHub Copilot to understand and work with Python codebases more effectively.

✨ Features

• 📚 Package Documentation: Get comprehensive docs for any Python package
• 🔍 Symbol Search: Find classes, functions, and modules by pattern
• 💻 Source Code Access: Retrieve actual implementation code
• 🏗️ Structure Analysis: Analyze complete package architecture
• 🔧 Type Hints: Extract and analyze type annotations
• 📖 Docstring Parsing: Support for Google, NumPy, and Sphinx formats
• 🏃 High Performance: Efficient caching and optimized operations
• 🛡️ Error Handling: Robust error management and validation

🚀 Quick Start

Installation

# Install from PyPI (Recommended)
pip install mcpydoc

# Or install from source
git clone https://github.com/amit608/MCPyDoc.git
cd MCPyDoc
pip install .

Once installed and configured with your AI agent, the server will automatically start when needed.

🔌 Integration with you favorite coding AI agent

Configuration

Add MCPyDoc to your Cline/Claude Code/Cursor/Github Copilot MCP configuration:

{
  "mcpServers": {
    "mcpydoc": {
      "command": "python",
      "args": ["-m", "mcpydoc"],
      "env": {},
      "description": "Python package documentation and code analysis server"
    }
  }
}

Recommended Rules

To help the coding AI agent use MCPyDoc effectively, add these rules to your rules file (Cline: .clinerules, Cursor: .cursor/rules/, Github Copilot:copilot-instructions.md, Claude Code: .claude/settings.json):

# Python Package Analysis Guidelines

## When to Use MCPyDoc Tools

### Always use MCPyDoc when:
- User asks about Python package functionality or capabilities
- Need to understand how to use a specific Python library
- Looking for specific functions or classes in a package
- Want to see source code implementation details
- Need to understand package structure or available modules
- Troubleshooting Python package usage issues

### Recommended workflow:
1. **Start with `analyze_structure`** to understand package organization
2. **Use `search_symbols`** to find relevant functions/classes
3. **Use `get_package_docs`** for comprehensive documentation of packages/modules
4. **Use `get_package_docs` with `module_path`** for specific method/class documentation
5. **Use `get_source_code`** when you need to see implementation details

### For specific method documentation:
- **Option 1**: `get_package_docs` with `module_path` parameter (e.g., `package="requests", module_path="Session.get"`)
- **Option 2**: `search_symbols` to find the method, then `get_package_docs` for details
- **Option 3**: `get_source_code` for the method implementation and its docstring

### Examples of when to use MCPyDoc:
- "I need to understand how to use the SQLAlchemy Session.query() method - show me its documentation and parameters"
- "How do I use requests to make HTTP calls?"
- "What classes are available in the pathlib module?"
- "Show me the source code for json.loads"
- "What functions does the os module provide for file operations?"
- "How do I use pandas for data analysis?"

### Don't use MCPyDoc for:
- Questions about Python language syntax
- General programming concepts
- Non-Python packages or libraries
- Questions already answered in previous context

Place this in your .clinerules file to help Cline automatically use MCPyDoc when working with Python packages.

📊 Supported Package Types

• ✅ Standard Library - Built-in modules (json, os, sys, etc.)
• ✅ Third-Party Packages - pip-installed packages
• ✅ Local Packages - Development packages in current environment
• ✅ Virtual Environments - Proper path resolution

🛠️ API Reference

Core Methods

get_module_documentation(package, module_path=None, version=None)

Get comprehensive documentation for a package or specific module.

Parameters:

• package (str): Package name
• module_path (str, optional): Dot-separated path to specific module
• version (str, optional): Specific version to use

Returns: ModuleDocumentationResult

search_package_symbols(package, pattern=None, version=None)

Search for symbols in a package.

Parameters:

• package (str): Package name
• pattern (str, optional): Search pattern (case-insensitive)
• version (str, optional): Specific version to use

Returns: List[SymbolSearchResult]

get_source_code(package, symbol_name, version=None)

Get source code for a specific symbol.

Parameters:

• package (str): Package name
• symbol_name (str): Dot-separated path to symbol
• version (str, optional): Specific version to use

Returns: SourceCodeResult

analyze_package_structure(package, version=None)

Analyze complete package structure.

Parameters:

• package (str): Package name
• version (str, optional): Specific version to use

Returns: PackageStructure

🏗️ Architecture

MCPyDoc uses a clean, modular architecture:

mcpydoc/
├── __init__.py          # Package interface
├── __main__.py          # CLI entry point
├── server.py            # Core MCPyDoc class
├── mcp_server.py        # MCP JSON-RPC server
├── analyzer.py          # Package analysis engine
├── documentation.py     # Docstring parsing
├── models.py            # Pydantic data models
├── exceptions.py        # Custom exceptions
└── utils.py             # Utility functions

Key Components

• Analyzer: Package introspection and symbol discovery
• Documentation Parser: Multi-format docstring parsing
• MCP Server: JSON-RPC protocol implementation
• Models: Type-safe data structures with Pydantic
• Exception Handling: Comprehensive error management

📝 License

MIT License - see LICENSE file for details.

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Run the test suite
6. Submit a pull request

---

Made with ❤️ for the Python community