DataPizza MCP Server 🍕

A Model Context Protocol (MCP) server that provides intelligent access to datapizza-ai documentation through vector similarity search and retrieval-augmented generation.

Overview

This MCP server enables AI assistants and applications to query the comprehensive datapizza-ai documentation using natural language queries. It indexes documentation from the datapizza-ai repository and provides contextual, relevant responses through a RAG (Retrieval-Augmented Generation) pipeline.

Features

• Intelligent Documentation Search: Natural language queries across datapizza-ai documentation
• Vector-Based Retrieval: Uses OpenAI embeddings and Qdrant vector database for semantic search
• MCP Protocol Compliance: Standard Model Context Protocol implementation for broad compatibility
• Automatic Indexing: Downloads and indexes documentation from GitHub automatically
• Cloud-Ready: Supports Qdrant Cloud for scalable vector storage
• Configurable: Environment-based configuration for flexible deployment

Architecture

The server consists of four main components:

• MCP Server: FastMCP-based server exposing the query_datapizza tool
• Indexer: Downloads and processes datapizza-ai documentation into searchable chunks
• Retriever: RAG engine for semantic search and response generation
• Configuration: Environment-based settings management with validation

Prerequisites

• Python 3.10 or higher
• OpenAI API key
• Qdrant Cloud account and API key
• Internet connection for documentation indexing

Installation

1. Clone the repository:

git clone https://github.com/datapizza-labs/mcp_server_datapizza.git
cd datapizza-mcp-server

2. Navigate to the package directory:

cd datapizza-mcp-server

3. Install the package with development dependencies:

pip install -e ".[dev]"

Configuration

Create a .env file in the datapizza-mcp-server directory with the following variables:

# Required Configuration
OPENAI_API_KEY=your_openai_api_key_here
QDRANT_URL=your_qdrant_cloud_url
QDRANT_API_KEY=your_qdrant_api_key

# Optional Configuration
EMBEDDING_MODEL=text-embedding-3-small
EMBEDDING_DIMENSIONS=1536
COLLECTION_NAME=datapizza_docs
MAX_RESULTS=5
CHUNK_SIZE=1024
CHUNK_OVERLAP=200
LOG_LEVEL=INFO

Required Environment Variables

Variable	Description
OPENAI_API_KEY	OpenAI API key for generating embeddings
QDRANT_URL	Qdrant Cloud instance URL
QDRANT_API_KEY	Qdrant Cloud API key

Optional Environment Variables

Variable	Default	Description
EMBEDDING_MODEL	text-embedding-3-small	OpenAI embedding model
EMBEDDING_DIMENSIONS	1536	Embedding vector dimensions
COLLECTION_NAME	datapizza_docs	Qdrant collection name
MAX_RESULTS	5	Maximum search results returned
CHUNK_SIZE	1024	Document chunk size for indexing
CHUNK_OVERLAP	200	Overlap between document chunks
LOG_LEVEL	INFO	Logging level (DEBUG, INFO, WARNING, ERROR)

Usage

1. Index Documentation

Before using the server, index the datapizza-ai documentation:

python -m datapizza_mcp.indexer

To force re-indexing (clears existing data):

python -m datapizza_mcp.indexer --force

2. Start the MCP Server

python -m datapizza_mcp.server

Or use the provided Windows batch script:

../run_datapizza.bat

3. Query the Documentation

The server exposes a query_datapizza tool that can be called by MCP clients:

# Example query
result = await client.call_tool("query_datapizza", {
    "query": "come creare un agente con OpenAI",
    "max_results": 5
})

MCP Tools and Resources

Tools

• query_datapizza: Search datapizza-ai documentation
  • query (string): Natural language search query
  • max_results (int, optional): Maximum number of results (default: 5)

Resources

• datapizza://status: System status and configuration information

Development

Code Quality Tools

# Format code
black src/

# Lint code
ruff check src/
ruff check src/ --fix  # Auto-fix issues

# Type checking
mypy src/

# Run tests
pytest

Project Structure

datapizza-mcp-server/
├── src/datapizza_mcp/
│   ├── __init__.py          # Package exports
│   ├── config.py            # Configuration management
│   ├── server.py            # MCP server implementation
│   ├── indexer.py           # Documentation indexing
│   └── retriever.py         # RAG retrieval engine
├── pyproject.toml           # Package configuration
├── .env                     # Environment variables
└── README.md               # This file

Dependencies

Core Dependencies

• mcp: Model Context Protocol framework
• datapizza-ai-core: Core datapizza-ai functionality
• datapizza-ai-embedders-openai: OpenAI embedding integration
• datapizza-ai-vectorstores-qdrant: Qdrant vector store integration
• openai: OpenAI API client
• qdrant-client: Qdrant database client
• requests: HTTP client for GitHub API
• python-dotenv: Environment variable management

Development Dependencies

• pytest: Testing framework
• black: Code formatter
• ruff: Linter and code style checker
• mypy: Static type checker

Troubleshooting

Common Issues

1. Authentication Errors

   • Verify OPENAI_API_KEY is set correctly
   • Check Qdrant Cloud credentials (QDRANT_URL and QDRANT_API_KEY)

2. Empty Search Results

   • Ensure documentation is indexed: python -m datapizza_mcp.indexer
   • Check system status: query the datapizza://status resource

3. Connection Issues

   • Verify internet connectivity for GitHub and Qdrant Cloud access
   • Check firewall settings for outbound HTTPS connections

Debugging

Enable debug logging by setting LOG_LEVEL=DEBUG in your .env file.

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes following the code style guidelines
4. Run the full test suite and code quality checks
5. Submit a pull request

License

This project is licensed under the MIT License. See the LICENSE file for details.

Support

For issues and questions:

• GitHub Issues: datapizza-mcp-server/issues
• DataPizza AI Documentation: datapizza-ai