How ahodroj/mcp-iceberg-service works
Project type and purpose
• This is a standalone Python microservice (“MCP Iceberg Catalog”) that implements the Model Context Protocol to expose a SQL interface to Apache Iceberg data lakes.
• It is invoked by Claude Desktop over stdio (via the UV runner), parses SQL commands, executes them against an Iceberg catalog and S3-backed warehouse, then returns results.
File structure
• Top level:
• Dockerfile, pyproject.toml, requirements.txt, smithery.yaml – containerization and metadata
• src/mcp_server_iceberg/server.py – entry point
• __init__.py – package marker
• uv.lock – UV CLI lockfile
• Naming indicates a single-service Python package, no frontend.
• Deployment via UV (the “uv” CLI) and Docker.
README for architecture insights
• Three main logical components:
• MCP Protocol Handler (stdio, request/response)
• Query Processor (SQL parsing via sqlparse, command handling: LIST, DESCRIBE, SELECT, INSERT)
• Iceberg Integration (PyIceberg + PyArrow for metadata & data operations)
• External dependencies:
• Iceberg REST catalog (HTTP)
• S3-compatible storage for tables
• Claude Desktop as the client orchestrator