yet-sparql

This MCP server implements a SPARQL-enabled MCP endpoint. It connects to a configured SPARQL endpoint, executes queries in one of several formats (json, simplified, tabular), and exposes cache management and formatting capabilities through the MCP interface. Clients can interact with it using the FastMCP client libraries over stdio or an HTTP transport when deployed with a proxy. The server supports running in foreground (stdio) or as an HTTP server for nginx integration, and it can be daemonized or managed via systemd when installed with the provided scripts. Typical usage involves starting the server with a specified SPARQL endpoint, then issuing MCP tool calls such as query and cache actions to retrieve and format results from the SPARQL endpoint.

Prerequisites:

• Python 3.8 or newer
• pip (Python package installer)
• Git

Installation steps:

1. Clone the repository and install dependencies

git clone https://github.com/yet-market/yet-sparql-mcp-server.git
cd yet-sparql-mcp-server

# Install Python dependencies
pip install -r requirements.txt

# Install the package in editable mode (for development)
pip install -e .

2. Optional: install the provided systemd integration (if you plan to run as a service)

# From repository root, run the installation script as root
sudo ./install.sh

3. Run the server in stdio transport (default) or configure HTTP transport in environment variables or via command-line arguments as described in the README.

# Basic run (stdio transport)
python server.py --endpoint https://dbpedia.org/sparql

4. If you prefer HTTP mode (nginx) or daemonized operation, follow the README examples for --transport http and the systemd service setup. After installation, you can control and monitor the service with the provided scripts (status.sh, start.sh, stop.sh, logs.sh).

Tips and considerations:

• Ensure SPARQL endpoint is reachable from the host where the MCP server runs.
• The server supports multiple cache strategies (LRU, LFU, FIFO) and configurable TTL. Tune these in the configuration as needed for your workload.
• When deploying behind a reverse proxy, use HTTP transport to enable standard web access and integrate with nginx.
• If running as a systemd service, make sure the environment file (env) contains the correct SPARQL_ENDPOINT and any required MCP-related settings.
• For development, use the provided test suite (./test.sh) to validate service health, MCP protocol compliance, and SPARQL endpoint reachability.
• The environment variable MCP_PORT should be set to the port you want the HTTP or stdio transport to listen on when using HTTP mode.