Python Medical MCP Server

This project provides a Python-based server with a dual interface:

1. A direct Model Context Protocol (MCP) interface (handled by your existing script, e.g., your_original_mcp_script.py).
2. An HTTP REST API interface (via FastAPI in api_server.py) for web clients like Bolt.new.

Both interfaces utilize a shared core medical logic module (medical_core.py).

Features

• Shared Medical Logic: Core medical assessment functions are centralized in medical_core.py.
• FastAPI HTTP Server: Exposes medical logic via RESTful endpoints.
  • POST /api/medical/trauma-assessment: For performing trauma assessments.
  • GET /api/status: To check the server's operational status.
• CORS Enabled: Configured for frontend applications.
• Pydantic Validation: Automatic request and response validation for API endpoints.
• PM2 Ready: Instructions provided for running both the direct MCP script and the API server using PM2.

Project Structure

python_mcp_server/
├── your_original_mcp_script.py  (Your existing script for direct MCP communication)
├── medical_core.py        (Shared core medical logic)
├── api_server.py          (FastAPI HTTP server)
├── requirements.txt       (Python dependencies)
├── .gitignore             (Standard Python .gitignore)
└── README.md              (This file)

Setup and Installation

1. Clone the repository (if you haven't already for your existing script): If this code is part of a larger repository that you've cloned, navigate to the python_mcp_server directory.

2. Create and Activate a Python Virtual Environment: It's highly recommended to use a virtual environment. If you don't have one for this project yet:

   cd python_mcp_server
   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install Dependencies:

   pip install -r requirements.txt
   

Running the Servers

You will typically run two Python processes:

• Your original script for direct MCP communication.
• The new api_server.py for HTTP REST API access.

It's recommended to use a process manager like PM2.

Using PM2

Ensure PM2 is installed (npm install -g pm2).

1. Start your existing Python MCP script (for direct MCP protocol): (Replace your_original_mcp_script.py with the actual name of your script. Ensure it's updated to import logic from medical_core.py.)

   # Navigate to python_mcp_server directory first if not already there
   # Ensure your virtual environment is NOT active when running PM2 with --interpreter,
   # or provide the full path to the python interpreter from the venv.
   # Best practice: Deactivate venv, then let PM2 use the venv's interpreter path.
   # Example: pm2 start your_original_mcp_script.py --name "python-mcp-direct" --interpreter /path/to/your/venv/bin/python
   
   # Simpler if running from an already activated venv (but PM2 might behave differently):
   pm2 start your_original_mcp_script.py --name "python-mcp-direct" --interpreter python3
   

2. Start the FastAPI HTTP API server (e.g., on port 8002):

   # Ensure you are in the python_mcp_server directory
   pm2 start "uvicorn api_server:app --host 0.0.0.0 --port 8002" --name "python-mcp-http-api"
   

   Note: Adjust the port 8002 if needed.

3. Check Status:

   pm2 list
   

4. Save PM2 Process List (for auto-restart on boot):

   pm2 save
   

   (You might need to run pm2 startup first to configure PM2 startup scripts for your OS.)

Manual Startup (for development/testing without PM2)

1. Terminal 1 (Direct MCP Script):

   cd python_mcp_server
   source venv/bin/activate
   python your_original_mcp_script.py
   

2. Terminal 2 (FastAPI HTTP Server):

   cd python_mcp_server
   source venv/bin/activate
   uvicorn api_server:app --reload --host 0.0.0.0 --port 8002
   

API Endpoints

• GET /api/status

  • Description: Returns the status of the HTTP API server.
  • Response: {"status": "Medical MCP API Server is running", "version": "0.1.0", "active_protocols": ["trauma_assessment"]}

• POST /api/medical/trauma-assessment

  • Description: Performs a trauma assessment.
  • Request Body (JSON):

    {
      "mechanismOfInjury": "Fall from 10ft ladder",
      "reportedSymptoms": ["severe leg pain", "dizziness"],
      "conscious": true,
      "age": 35, // optional
      "gender": "female", // optional
      "obviousBleeding": false // optional
    }
    

  • Response Body (JSON):

    {
      "severity_level": "serious",
      "immediate_actions": ["Ensure scene safety.", "Call for emergency medical services..."],
      "assessment_steps": ["Check for responsiveness...", "..."],
      "red_flags": ["High-risk mechanism of injury..."],
      "next_steps": ["Reassess ABCDEs frequently...", "..."]
    }
    

Development

• The core medical logic resides in medical_core.py.
• The FastAPI application and its endpoints are defined in api_server.py.
• Remember to update the origins list in api_server.py for CORS if your frontend URLs change.