Welcome to MCP_QnA ! This repository integrates an Large Language Model (Google Gemini) with a MCP server to answer context specific questions.

Currently I have added two modules: one based on a vector database for context search and another using fuzzy search for quick matching. Enjoy exploring the project! 😃

• LLM+VectorDB
  Contains code to query a vector database and get context from research papers.

  • client.py: Implements a client using Server-Sent Events (SSE) to interact with the MCP server.
  • embed_documents.py: Embeds documents from research papers (or any other document of you choice).
  • fetch_chunks.py: Splits or fetches document chunks for processing.
  • server.py: Implements the server side for handling vector database requests.

• LLM+FuzzySearch
  Supports fuzzy searching within the available context data.

  • client.py: Implements a client for fuzzy searching.
  • context.json: Holds context data used for fuzzy search.
  • server.py: Provides the server logic for fuzzy searches.

• Client-Server Communication: Uses Server-Sent Events (SSE) for real-time communication (Easily upgradable to streamable-http based server).
• Tool Integration: The client in client.py demonstrates how to list tools and make context-based requests.

1. Clone the Repository

   git clone https://github.com/Atharva-Jayappa/MCP_QnA.git
   cd MCP_QnA
   

2. Install Dependencies Make sure Python is installed and then run:

   pip install -r requirements.txt

3. Set Up Environment Variables Store your API keys securely. For example:

   set GENAI_API_KEY=your_api_key_here

4. Run the Application For the clien script in the vector database module, run:

   cd LLM+VectorDB
   python client.py

   For the server script in the vector database module, run:

    cd LLM+VectorDB
    python server.py

   Same for the fuzzy search module:

• Vector Database Query: The client in LLM+VectorDB/client.py connects to the MCP server via SSE, lists available tools, and sends user queries to fetch embedded document context in the form of chunks.
• Fuzzy Search Query: The client in LLM+FuzzySearch/client.py allows users to perform fuzzy searches on the context data stored in context.json.
• Server-Sent Events (SSE): The server in server.py handles SSE connections, allowing real-time updates and responses to client requests.