dperdomoplain/mcp_example

Stack Overflow MCP Server

This is a Model Context Protocol (MCP) server that provides tools to query Stack Overflow data from a SQL Server database. The project uses uv as a Python package and virtual environment manager.

Project Setup

The following commands explain how to set up the project:

1. Initialize the project

uv init --name mcp_example --python 3.13 . 

This command initializes a new Python project in the current directory (.) with the name mcp_example and specifies that Python 3.13 should be used.

2. Create virtual environment

2.1 Create virtual environment specifying path

uv venv --python /opt/homebrew/bin/python3

2.2 Create virtual environment on windows, linux and mac

uv venv --python 3.13

Creates a virtual environment using uv. The virtual environment is stored in the .venv directory.

3. Activate virtual environment

source .venv/bin/activate

Activates the virtual environment so that dependencies are installed in the project's isolated environment.

4. Install dependencies

uv add "mcp[cli]"
uv add pyodbc
uv add pydantic
uv add pydantic-settings

Installs the required packages:

• mcp with CLI extras for Model Context Protocol
• pyodbc for SQL Server connectivity
• pydantic for data validation
• pydantic-settings for configuration management

5. Debugging Inspector

npx @modelcontextprotocol/inspector

Features

This MCP server includes the following tools:

• get_most_viewed_posts: Retrieves the most viewed posts from the Stack Overflow database
• get_last_post_with_activity: Gets the most recent posts with activity from the Stack Overflow database

Architecture

The project uses a clean, modern architecture with Pydantic for data validation:

• sql_server.py: Contains the SQLServerConnection class and Pydantic models
• main.py: Contains the MCP server and tool definitions
• Pydantic Models: Strong typing and validation for all data structures
• Environment Variables: Support for loading configuration from environment
• Context Management: Automatic connection cleanup using Python context managers
• Error Handling: Comprehensive error handling with detailed error messages

Configuration

Environment Variables

Create a .env file with your SQL Server configuration:

SERVER=localhost
DATABASE=TestDB
USERNAME=your_username
PASSWORD=your_password
PORT=1433
TRUSTED_CONNECTION=false

Pydantic Models

The project uses Pydantic models for data validation:

• SQLServerConfig: Configuration with validation
• QueryRequest: SQL query requests with validation
• QueryResponse: Structured query responses

SQL Server Setup

Prerequisites

You need to install the ODBC Driver 17 for SQL Server:

On macOS:

Download and install from the Microsoft SQL Server ODBC Driver download page

On Windows:

Download and install from the Microsoft SQL Server ODBC Driver download page

On Linux:

Download and install from the Microsoft SQL Server ODBC Driver download page

Database Schema

This MCP server expects a SQL Server database with Stack Overflow data. The database should contain at least a Posts table with the following relevant columns:

• ViewCount: Number of times the post has been viewed
• LastActivityDate: The date of the last activity on the post
• Other standard Stack Overflow post fields

You can obtain Stack Overflow data dumps from the Stack Exchange Data Dump and import them into your SQL Server database.

Running the Server

uv run main.py

The server will start and listen for MCP client connections.