Rasdaman

You can interact with Rasdaman through a dedicated MCP server that exposes Rasdaman functionality as tools your client or LLM can call. This setup lets you query datacubes, inspect coverages, or construct WCPS queries in a natural, tool-based workflow while the MCP server handles the underlying WCS/WCPS interactions with Rasdaman.

How to use

Use an MCP client to connect to the Rasdaman MCP server in either stdio (local) mode or HTTP (standalone server) mode. You will call exposed tools to list datacubes, describe a specific coverage, validate WCPS syntax, and execute WCPS queries. The server translates tool invocations into Rasdaman-compatible requests and returns results in a straightforward format.

How to install

Prerequisites you need before installing include Python 3 and a compatible package manager. Ensure you have network access to install Python packages.

Install the MCP server package from PyPI.

pip install rasdaman-mcp

Choose your preferred mode after installation: stdio for direct client integration, or HTTP to run a standalone web server.

If you plan to run in stdio mode (default), start the MCP tool with your Rasdaman credentials.

rasdaman-mcp --username rasguest --password rasguest

If you prefer an HTTP server, start the MCP server with HTTP transport and specify Rasdaman connection details.

rasdaman-mcp --transport http --host 127.0.0.1 --port 8000 --rasdaman-url "http://localhost:8080/rasdaman/ows"

Additional setup and notes

In HTTP mode, once the server is running you connect your MCP-aware client to http://127.0.0.1:8000/mcp. In stdio mode, you integrate directly with your client workflow by invoking the rasdaman-mcp command in your startup sequence.

Additional sections

Both usage modes provide tools that you can call to work with Rasdaman data. Use the list_coverages tool to see all available datacubes, describe_coverage to fetch metadata for a specific datacube, and execute_wcps_query to run a WCPS query and obtain results. The WCPS crash course and a validator help you craft and verify valid queries before execution.

Configuration and environment

The server supports configuring the Rasdaman connection either via command-line arguments or environment variables. This enables flexible deployment in containers or different runtime environments.

Available tools

list_coverages

Lists all available datacubes in Rasdaman that you can query.

describe_coverage

Retrieves metadata for a specific datacube, identified by coverage_id.

wcps_query_crash_course

Provides a quick reference on WCPS syntax to help you craft queries.

validate_wcps_query

Validates the syntax of a WCPS query without executing it.

execute_wcps_query

Executes a WCPS query and returns results, either as a string or as a file path.