MCP Demo 2 - Model Context Protocol Integration

A complete demonstration of the Model Context Protocol (MCP) with LangChain integration, featuring:

• MCP Weather Server: FastMCP server exposing weather tools via SSE transport
• MCP Weather Client: Client for connecting to MCP servers and calling tools
• LangChain Integration: Tool wrapper for using MCP tools with LangChain agents
• Ollama Integration: Local LLM integration for natural language interactions
• Automated Tests: Comprehensive test suite for validation

Setup

1. Create and activate virtual environment:

uv env
.venv\Scripts\activate  # Windows
# or
source .venv/bin/activate  # Linux/Mac

2. Install dependencies:

uv pip install -r requirements.txt

Running the MCP Weather Server

1. Start the MCP server:

python mcp_weather_server.py

The server will start on http://localhost:8000 with:

• MCP SSE endpoint: /sse - For MCP client connections
• Available tools: get_weather_tool, set_weather_tool, list_cities_tool

Testing the System

Automated Tests

1. Start the MCP server (in one terminal):

python mcp_weather_server.py

2. Run the test suite (in another terminal):

python run_tests.py

Test Coverage:

• ✅ Server connectivity and health checks
• ✅ Tool discovery and listing
• ✅ Weather operations (get, set, list cities)
• ✅ Error handling for invalid inputs
• ✅ Multiple operation sequences
• ✅ Data persistence verification

Running the Host

1. Start the ollama server and run a model:

ollama serve
ollama run llama3.2:3b

2. In another terminal (with the same virtual environment), run the host:

python mcp_host.py

API Endpoints

Server Endpoints

• GET /sse - Server-Sent Events endpoint for MCP communication

MCP Tools

• get_weather(city) - Get current weather for a city
• list_cities() - List all available cities
• update_weather(city, temp, humidity, condition) - Update weather data