Integrate Model Context Protocol (MCP)

[jsham042]

Summary:
Set up the Model Context Protocol (MCP) in our Python project using uv, define and expose a tool that uses a structured data type (BMI calculator), and verify interaction via MCP server.

Description:
This task involves integrating MCP into our Python project to allow structured and typed tool execution over an MCP server. We'll use a dataclass as input for one tool (BMI calculation) to demonstrate this capability.

Acceptance Criteria:

Project environment set up using uv

MCP SDK installed and working

MCP server created and running

At least one MCP tool implemented using a custom data class

Tool accepts input and returns output correctly

Code tested and documented

Steps to Implement:

1. Set Up Environment

bash
Copy
Edit
pip install uv
uv init my_mcp_project
cd my_mcp_project
uv venv
source .venv/bin/activate # On Windows use .venv\Scripts\activate
2. Install MCP SDK

bash
Copy
Edit
uv add mcp
3. Create a Data Type

python
Copy
Edit

bmi_input.py

from dataclasses import dataclass

@DataClass
class BMIInput:
weight_kg: float
height_m: float
4. Set Up MCP Server

python
Copy
Edit

main.py

from mcp.server.fastmcp import FastMCP
from bmi_input import BMIInput

mcp = FastMCP("BMIServer")

@mcp.tool()
def calculate_bmi(input_data: BMIInput) -> float:
"""Calculate BMI from weight and height."""
return input_data.weight_kg / (input_data.height_m ** 2)

if name == "main":
mcp.run()