SimoneFassio/DOREMUS_MCP

DOREMUS Music Knowledge Graph - MCP Server

A Model Context Protocol (MCP) server for accessing the DOREMUS Knowledge Graph, providing comprehensive access to classical music metadata including composers, works, performances, recordings, and instrumentation.

Overview

This MCP server enables LLMs to query the DOREMUS Knowledge Graph (https://data.doremus.org) using natural language, with optimized tools for:

Quick Start

Prerequisites

• Docker and Docker Compose (for containerized deployment)
• Or Python 3.11+ and Poetry (for local development)

Using Docker (Recommended for Production)

1. Build and run the container:

docker-compose up --build

2. Access the MCP server at:

http://localhost:8000/mcp

Local Development with Poetry

1. Install Poetry (if not already installed):

curl -sSL https://install.python-poetry.org | python3 -

2. Install server dependencies:

# Install core + server dependencies only
poetry install

3. Run the server:

poetry run python -m src.server.main

Setting Up Evaluators/Client

For running evaluations and the LangChain client:

# Install with evaluation dependencies
poetry install --with eval

# Run evaluations
poetry run python evaluators/test_query.py

Dependency Groups

This project uses Poetry with dependency groups:

• main: Core dependencies (langchain, requests, etc.)
• server: Server-only dependencies (fastmcp, uvicorn, starlette)
• eval: Evaluation/client dependencies (langgraph-cli, pandas, jupyter)

Install combinations as needed:

# Server only
poetry install

# Everything for development
poetry install --with eval,dev

# Production (no optional groups)
poetry install --only main

Architecture

DOREMUS_MCP/
├── src/
│   ├── server/              # MCP Server implementation
│   │   ├── server.py        # Main FastMCP server with tools
│   │   ├── query_builder.py # Parametric SPARQL query builder
│   │   ├── find_paths.py    # Graph path finding utilities
│   │   └── __init__.py
│   └── rdf_assistant/       # LangChain assistant for evaluations
│       ├── doremus_assistant.py
│       ├── extended_mcp_client.py
│       ├── prompts.py
│       └── eval/
├── evaluators/              # Evaluation scripts
│   ├── client_geminiCLI.py
│   └── test_query.py
├── tests/                   # Unit tests
│   └── test_server.py
├── data/                    # Data files
│   ├── graph.csv
│   └── cq.json             # Competency questions
├── docs/                    # Documentation
│   ├── ENDPOINT_GUIDE.md
│   └── EXAMPLES.md
├── pyproject.toml          # Poetry dependencies and config
├── Dockerfile              # Container configuration
├── docker-compose.yml      # Docker Compose setup
└── README.md               # This file

Resources

• DOREMUS Project: http://www.doremus.org/
• SPARQL Endpoint: https://data.doremus.org/sparql/

License

This MCP server implementation is provided as-is for accessing the publicly available DOREMUS Knowledge Graph.

Acknowledgments

Built on top of the DOREMUS (DOing REusable MUSical data) project, funded by the French National Research Agency (ANR-14-CE24-0020).

Special thanks to:

• DOREMUS consortium members
• Bibliothèque nationale de France
• Philharmonie de Paris
• Radio France

Support

For issues related to:

• This MCP server: Open an issue on the repository
• DOREMUS data or ontology: Contact the DOREMUS project
• SPARQL endpoint: Check https://data.doremus.org/

Last Updated: December 2025