Skip to content

datalake-mcp-server-client.md

Latest commit

 

History

History
38 lines (26 loc) · 1.78 KB

datalake-mcp-server-client.md

File metadata and controls

38 lines (26 loc) · 1.78 KB
Raw

Datalake MCP Server Client
GitHub Repo datalake-mcp-server-client
Python 3.13 (auto-generated client)
Package Manager uv

The datalake-mcp-server-client is an auto-generated Python client library for interacting with the datalake-mcp-server APIs.

Overview

Unlike manually crafted SDKs, this library is generated directly from the OpenAPI specification of the datalake-mcp-server. This ensures that the client remains strictly in sync with the upstream API definitions and data models with zero manual maintenance overhead.

This client is used by other internal BERDL services (such as spark_notebook_base and Jupyter tools) to interact programmatically with the Datalake MCP backend services securely.

Generation Process

The client is generated using the openapi-python-client utility. A convenience script (run.sh) is provided in the repository to automate generation:

  1. The script extracts the OpenAPI JSON spec directly from the local datalake-mcp-server codebase using an intermediate Python generator (generate_spec_from_local.py).
  2. The openapi-python-client consumes this specification and outputs static Python type hints and API client methods.
  3. The resulting code is placed cleanly into src/datalake_mcp_server_client.

Running Generation Locally

# In the datalake-mcp-server-client repo
./run.sh

(Note: Requires uv and Python 3.13 for the build environment).

Usage

Other packages can depend on this client by installing it directly from its git repository or path:

pip install "git+https://github.com/KBaseDataLakehouse/datalake-mcp-server-client.git"