DEVELOPING.md

Developing

Setup

Requires uv - see the project README for installation instructions.

make setup

# <authentication is the same as in production>

Run using the MCP inspector

uv run fastmcp dev src/mcp_server_datahub/__main__.py:create_app --with-editable .

In the inspector UI, add environment variables for DATAHUB_GMS_URL and DATAHUB_GMS_TOKEN, then click Connect.

Note: Use fastmcp dev (not mcp dev), since this project uses the standalone FastMCP package.

Run using an MCP client

Use this configuration in your MCP client e.g. Claude Desktop, Cursor, etc.

{
  "mcpServers": {
    "datahub": {
      "command": "<full-path-to-uv>",  // e.g. /Users/hsheth/.local/bin/uv
      "args": [
        "--directory",
        "path/to/mcp-server-datahub",  // update this with an absolute path
        "run",
        "mcp-server-datahub"
      ],
      "env": {  // required if ~/.datahubenv does not exist
        "DATAHUB_GMS_URL": "<your-datahub-url>",
        "DATAHUB_GMS_TOKEN": "<your-datahub-token>"
      }
    }
  }
}

Run linting

# Check linting
make lint-check

# Fix linting
make lint

Run tests

The test suite is currently very simplistic, and requires a live DataHub instance.

make test

Publishing

We use setuptools-scm to manage the version number.

CI will automatically publish a new release to PyPI when a GitHub release is created.