akhilsahu/mmt_flight_mcp

Multi-Agent Flight Search System

This project implements a multi-agent system using LangGraph for orchestrating specialized agents to perform tasks like searching for flights, providing weather forecasts, and performing calculations.

Features

• Flight Search: Search for flights on MakeMyTrip.
• Weather Forecast: Get weather forecasts for a given location.
• Math Calculations: Perform basic mathematical calculations.
• Multi-Agent Orchestration: Uses a supervisor agent to delegate tasks to the appropriate specialized agent.

Architecture

The system is built using the following components:

• LangGraph: A library for building stateful, multi-agent applications with LLMs.
• Supervisor Agent: A central agent that manages the workflow and delegates tasks to other agents based on the user's query.
• Specialized Agents:
  • FlightSearchAgent: Interacts with the flight search tool.
  • WeatherAgent: Interacts with the weather tool.
  • MathAgent: Interacts with the math tool.
• MCP Tool Servers: Each specialized agent communicates with a corresponding MCP (Model-Client-Protocol) server that exposes the tools for that agent's domain. These servers run as separate processes.

How to Run

Prerequisites

• Python 3.10+
• An OpenAI API key

1. Setup Environment

1. Clone the repository:

   git clone <repository-url>
   cd <repository-directory>
   

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

   Note: The mcp library is used in this project. If you have issues with this dependency, you may need to install it from a custom source.

4. Create a .env file in the root directory and add your OpenAI API key:

   OPENAI_API_KEY="your-openai-api-key"
   

2. Run the Tool Servers

Open three separate terminal windows and run the following commands, one in each terminal:

• Math Server:

  python mcp_tool/math_server.py
  

• Weather Server:

  python mcp_tool/weather_server.py
  

• Flight Search Server:

  python mcp_tool/flight_search_server.py
  

3. Run the Main Application

Once the tool servers are running, open a new terminal and run the main application:

python main.py

The application will then process the example queries defined in main.py.

Example Usage

You can modify the queries list in main.py to ask the system different questions:

if __name__ == "__main__":
    import asyncio
    
    async def main():
        # Example queries
        queries = [
            "What is 25 * 47 + 123 /4?",
            "What's the weather in New York?",
            # "Search for flights from LKO to IXL on 10/10/2024" # This requires the flight server
        ]
        
        for query in queries:
            await run_multi_agent_system(query)
            print("\n" + "="*60 + "\n")
    
    asyncio.run(main())