Ai Ml C

AI-powered GitHub assistant using MCP (Model Context Protocol), Haystack, and Gemini

Python Streamlit MCP Gemini

🚀 Features

• 🔍 GitHub Analysis: Fetch issues, PRs, commits, and discussions via MCP
• 🤖 AI-Powered Insights: Analyze repository health with Gemini AI
• 🔧 Auto-Fix Workflow: Detect issues and create automated PRs
• 💻 CLI Interface: Interactive command-line experience
• 🌐 Web UI: Modern Streamlit interface with buttons
• 📊 Jupyter Demo: Interactive notebook for experimentation
• 🔒 Secure: Environment-based API keys, no hardcoded secrets

📋 Prerequisites

• Python 3.8+
• GitHub Personal Access Token (with repo scope)
• Google API Key for Gemini

🛠️ Quick Setup

1. Clone and navigate:

   git clone https://github.com/saumya1317/ai_Ml_c_mcp.git
   cd ai_Ml_c_mcp
   

2. Install dependencies:

   pip install -r ai_mlc/requirements.txt
   

3. Set environment variables:

   # Windows PowerShell
   $Env:GITHUB_PERSONAL_ACCESS_TOKEN = "your_github_token"
   $Env:GITHUB_USERNAME = "your_github_username"
   $Env:GOOGLE_API_KEY = "your_gemini_api_key"
   
   # Linux/macOS
   export GITHUB_PERSONAL_ACCESS_TOKEN="your_github_token"
   export GITHUB_USERNAME="your_github_username"
   export GOOGLE_API_KEY="your_gemini_api_key"
   

🎯 Usage

CLI Interface

python ai_mlc/cli.py

Example queries:

• "Find errors in octocat/Hello-World"
• "List open PRs in owner/repo"
• "Summarize recent commits in owner/repo"

Web Interface

streamlit run ai_mlc/stream_app.py

Jupyter Demo

jupyter notebook ai_mlc/demo/agent_demo.ipynb

🏗️ Architecture

ai_mlc/
├── cli.py                      # Main CLI application
├── stream_app.py               # Streamlit web interface
├── mcp_demo_haystack.py        # MCP + Haystack integration
├── requirements.txt            # Python dependencies
├── demo/
│   └── agent_demo.ipynb        # Interactive Jupyter demo
└── github_agent/
    ├── agent.py                # LangChain agent setup
    ├── gemini_llm.py           # Gemini AI integration
    ├── github_api_tools.py     # GitHub REST API tools
    └── __init__.py             # Package initialization

🔧 Key Components

MCP Integration

• GitHub MCP Server: Real-time repository data
• Haystack Framework: Agent orchestration
• Tool Integration: Issues, PRs, commits, discussions

AI Analysis

• Gemini 1.5 Flash: Advanced reasoning and code analysis
• Error Detection: Automated bug identification
• Fix Generation: AI-suggested code improvements

Auto-PR Workflow

1. Analyze repository issues
2. Generate fix suggestions
3. Create feature branch
4. Apply changes
5. Open pull request

🛡️ Security

• ✅ Environment-based API keys
• ✅ No hardcoded secrets
• ✅ Token usage clarification
• ✅ Timeout protection
• ✅ Error handling

📚 Examples

CLI Session

Welcome to the GitHub AI Assistant (Merge Wars)
Type 'exit' to quit at any time.

Enter your GitHub question: Find errors in octocat/Hello-World

=== octocat/Hello-World ===
Agent: Found 3 open issues:
1. Documentation needs updating
2. Missing error handling in main.py
3. Performance issue in data processing

Suggested fix:
FILE: main.py
def process_data(data):
    try:
        return data.upper()
    except AttributeError:
        return None

Web UI Features

• Repository Input: Multi-repo support
• Action Buttons: Issues, PRs, Commits, Discussions
• Error Analysis: Automated fix suggestions
• PR Creation: One-click pull request generation

🔍 Troubleshooting

Common Issues

• MCP Timeouts: Check GITHUB_PERSONAL_ACCESS_TOKEN and repo scope
• Gemini Errors: Verify GOOGLE_API_KEY is set correctly
• Import Errors: Ensure all dependencies are installed

Environment Setup

# Verify environment variables
echo $GITHUB_PERSONAL_ACCESS_TOKEN
echo $GITHUB_USERNAME  
echo $GOOGLE_API_KEY

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

Demo video

Watch the demo

Direct link: https://drive.google.com/file/d/1Uk3BtSCJJ5a7CcA9nFKzcOG0iL5pvdYx/view?usp=sharing

📄 License

This project is open source and available under the MIT License.

🙏 Acknowledgments

• Model Context Protocol for GitHub integration
• Haystack for agent framework
• Google Gemini for AI capabilities
• Streamlit for web interface

---