GitHub Repository Analysis Tool

A powerful tool for analyzing GitHub repositories using the Model Context Protocol (MCP) and AI-powered insights from Large Language Models (LLMs).

Features

• 🔍 Repository Information Analysis: Fetch detailed information about GitHub repositories.
• 🔄 Commit History Analysis: Analyze commit histories to identify patterns and insights.
• 🤖 AI-Powered Insights: Utilize Ollama to generate intelligent responses and analyses based on repository data.
• 🎨 Beautiful Command-Line Interface: Enjoy a user-friendly CLI for seamless interaction.
• 🔌 MCP-Based Client-Server Architecture: Leverage a modular architecture for efficient data fetching and analysis.

Architecture

The application employs a client-server architecture based on the Model Context Protocol (MCP):

• Server: Provides tools for fetching GitHub data through a standardized MCP interface.

  • get_repo_info: Fetches repository information.
  • get_commit_history: Retrieves commit history for analysis.

• Client: Connects to the server and provides analysis features

  • Communicates with the server using MCP's stdio transport
  • Uses Ollama for AI-powered analysis
  • Provides an interactive command-line interface

Prerequisites

• Python 3.10 or higher
• Ollama installed with the qwen2.5:7b model
• GitHub personal access token (for higher API rate limits)

Installation

To get started with the GitHub Repository Analysis Tool, follow these steps:

1. Clone the Repository:

   git clone https://github.com/frangkli/github-analysis.git
   cd github-analysis

2. Set Up Environment and Dependencies:

   uv sync

3. Activate Virtual Environment:

   source .venv/bin/activate

4. Install Ollama: Ensure you have Ollama installed with the qwen2.5:7b model for AI-powered insights.

5. Set Up GitHub Personal Access Token: For higher API rate limits, create a GitHub personal access token and set it as an environment variable:

   export GITHUB_TOKEN=your_token_here  # On Windows use `set GITHUB_TOKEN=your_token_here`

Usage

The tool can be run in two modes: server and client.

1. First, start the server:

github-analysis server

2. Then, in another terminal, run the client:

github-analysis client <owner> <repo>

You can also run the client in Ollama-only mode, which bypasses the MCP server and tools:

github-analysis client <owner> <repo> --disable-tools

Analysis Options

The tool provides three types of analysis:

1. Repository Analysis: Analyzes repository metadata, languages, and statistics
2. Commit Analysis: Analyzes recent commit history and patterns
3. Custom Analysis: Allows you to ask custom questions about the repository

When running in Ollama-only mode (--disable-tools), the analysis will be based on general knowledge rather than real-time repository data.

Development

Project Structure

github_analysis/
├── client/
│   └── client.py     # MCP client implementation
├── server/
│   ├── server.py     # MCP server implementation
│   └── exceptions.py # Custom exceptions
└── main.py           # CLI entry point
tests/
└── test_mcp.py       # Some sample tests to test LLM output

Architecture Details

• MCP Server:

  • Implements the Model Context Protocol
  • Provides tools for GitHub API interaction
  • Uses stdio transport for communication

• MCP Client:

  • Connects to the server using stdio transport
  • Manages tool calls and response handling
  • Integrates with Ollama for analysis

Contributing

Contributions are welcome. If you have suggestions or improvements, please open an issue or submit a pull request.

• New Server Tools:

  • Add new functions with the @mcp.tool() decorator in server.py
  • Implement the tool's functionality using GitHub's API

• New Analysis Types:

  • Add new handler methods in the client class
  • Create appropriate system prompts for Ollama
  • Update the menu options

License

This project is licensed under the MIT License - see the LICENSE file for details.