Sui MCP Server

🚀 MCP Server Integrated with FAISS for RAG Project Documentation

This project combines the MCP server with FAISS (an efficient similarity search and machine learning library) to implement RAG (Retrieval Augmented Generation). It offers a complete workflow for document retrieval and generation, supporting the extraction of Move files from GitHub and integrating large language models (LLMs) for enhanced answer generation.

🚀 Quick Start

The MCP server project integrated with FAISS provides a comprehensive solution for RAG. It enables efficient document retrieval, supports Move file extraction from GitHub, and integrates LLMs for enhanced response generation.

✨ Features

• Document Retrieval: Conducts efficient similarity searches via FAISS.
• Move File Extraction: Supports downloading and processing relevant Move files from GitHub.
• RAG Pipeline: Implements a complete RAG workflow, including document retrieval, context formation, and LLM response generation.
• API Integration: Offers RESTful API interfaces to support remote document indexing and retrieval operations.

📦 Installation

Environment Dependencies

• Python 3.8+
• FAISS library
• requests library
• transformers library
• torch library

Configuration Steps

1. Clone the project repository:

git clone https://github.com/your-repository/mcp_server.git
cd mcp_server

2. Install dependencies:

pip install -r requirements.txt

3. Configure environment variables: Create a .env file and add the following content:

GITHUB_TOKEN=your_github_token_here
OPENAI_API_KEY=your_openai_api_key_here

📚 Documentation

Project Structure

mcp_server/
├── __init__.py             # Package initialization file
├── main.py                # Main server file
├── mcp_api.py             # MCP API implementation
├── index_move_files.py    # File indexing tool
├── local_query.py         # Local query tool
├── download_move_files.py # GitHub Move file extractor
├── rag_integration.py     # LLM integration module
├── pyproject.toml         # Package configuration file
├── requirements.txt       # Dependency management
├── .env.example           # Environment variable example
├── README.md              # Project documentation
├── data/                  # Data storage directory
├── docs/                  # Sample documents
│   └── move_files/        # Downloaded Move files
├── models/                # Model implementation
│   └── vector_store.py    # FAISS vector storage implementation
└── utils/
    ├── document_processor.py  # Document processing tool
    └── github_extractor.py    # GitHub file extraction tool

Core Components

MCP API Implementation (mcp_api.py)

• Provides the /mcp/action endpoint, supporting the following operations:
  • retrieve_documents: Retrieve relevant documents based on the query.
  • index_documents: Index documents from the specified directory.

RAG Integration (rag_integration.py)

• Implements the complete RAG workflow:
  1. Search query
  2. Document retrieval
  3. Context formation
  4. LLM response generation

GitHub Move File Extraction (download_move_files.py)

• Supports two file extraction methods:
  • GitHub API (recommended): Requires a valid GitHub Token to be configured.
  • Command line: Supports downloading files by specifying query parameters via the command line.

💻 Usage Examples

Initialize the Server

python main.py --port 8000

Call the API

Send a POST request to http://localhost:8000/mcp/action with the following request body:

{
  "operation": "retrieve_documents",
  "params": {
    "query": "your_query_content",
    "top_k": 5
  }
}

Download Move Files

Download relevant Move files from GitHub using the following command:

python download_move_files.py --query "your-query" --output-dir ./docs/move_files

🔧 Technical Details

• Vector Storage: Implements efficient document vectorization and similarity search using FAISS.
• Model Integration: Supports multiple LLM models, such as GPT, BERT, etc.
• File Processing: Provides flexible document parsing and preprocessing capabilities.

📄 License

This project is licensed under the MIT License. For details, please refer to the LICENSE file.