MCP Server

MCP Server

A Microsoft Copilot Plugin (MCP) server that provides various tools for querying information, including legal documents, Microsoft product licensing, weather data, and Azure pricing.

Overview

This server provides a set of API tools that can be used by Microsoft Copilot to access and query various information sources:

• Legal Information: Query the Chinese Criminal Law articles and provisions
• Microsoft Product Licensing: Get information about licensing for various Microsoft products (Visual Studio, SQL Server, Windows Server, Microsoft 365, Dynamics 365)
• Privacy Regulations: Query information about GDPR and China's Personal Information Protection Law
• Weather Data: Get weather alerts and forecasts for locations in the United States
• Azure Pricing: Query Azure pricing information using OData filter expressions
• Utility Tools: Count Chinese characters in a text string

The server uses LlamaIndex for vector search and retrieval, Azure OpenAI for embeddings and language model capabilities, and MongoDB/Azure Cosmos DB for vector storage.

Prerequisites

• Python 3.12 or higher
• MongoDB or Azure Cosmos DB with MongoDB API
• Azure OpenAI API access
• Docker (optional, for containerized deployment)

Installation

Local Development

1. Clone the repository:

   git clone <repository-url>
   cd mcp-server
   

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Create a .env file with the following variables:

   MONGODB_CONNECTION_STRING="your_mongodb_connection_string"
   AZURE_OPENAI_API_KEY="your_azure_openai_api_key"
   AZURE_OPENAI_ENDPOINT="your_azure_openai_endpoint"
   AZURE_OPENAI_CHAT_MODEL="gpt-4"
   AZURE_OPENAI_EMBEDDING_MODEL="text-embedding-ada-002"
   AZURE_OPENAI_TEMPERATURE="0.1"
   

Docker Deployment

1. Build the Docker image:

   docker build -t mcp-server .
   

2. Run the container:

   docker run -p 8080:8080 --env-file .env mcp-server
   

Usage

Starting the Server

Start the server with:

python mcp-server.py --host 0.0.0.0 --port 8080

The server will be available at http://localhost:8080.

Available Endpoints

• /sse - Server-Sent Events endpoint for MCP communication
• /messages/ - MCP message handling endpoint

Available Tools

The server provides the following tools:

Licensing Information Tools

• query_visual_studio_license - Get information about Visual Studio 2022 licensing
• query_sql_server_license - Get information about SQL Server 2022 licensing
• query_windows_server_license - Get information about Windows Server 2022 licensing
• query_m365_license - Get information about Microsoft 365 Enterprise licensing
• query_dynamics365_license - Get information about Dynamics 365 licensing

Privacy Regulation Tools

• query_gdpr - Get information about GDPR regulations
• query_china_pipl - Get information about China's Personal Information Protection Law

Legal Information Tools

• get_article_by_code - Get information by article code from Chinese Criminal Law
• search_by_content - Search for legal articles by content
• get_by_article_name - Get legal information by article name/cause of action
• get_specific_article - Get a specific paragraph from an article
• get_all_law_contents - Get all contents of the Chinese Criminal Law

Weather and Utility Tools

• get_alerts - Get weather alerts for a US state
• get_forecast - Get weather forecast for a location
• get_azure_price - Get Azure price for a service using OData filter expressions
• count_chinese_characters - Count the number of Chinese characters in a string

Development

Adding New Tools

To add a new tool to the server:

1. Define a new async function with the @mcp.tool() decorator in mcp-server.py
2. Implement the functionality within the function
3. Add appropriate error handling
4. Restart the server to make the new tool available

Troubleshooting

MongoDB Connection Issues

• Ensure the MongoDB connection string is correctly formatted
• For Azure Cosmos DB, make sure the cluster supports vector search or use the fallback option

Azure OpenAI API Issues

• Verify that your API key and endpoint are correct
• Check if the model deployment names match those specified in your environment variables

License

[Your License Information]

Contact

[Your Contact Information]