What problem does it solve? Building a Python backend on Databricks Apps requires correct OAuth authentication, resource wiring via app.yaml, framework-specific deployment commands, and connectivity to SQL warehouses, Lakebase, and model-serving endpoints — mistakes in any of these cause apps to crash or silently run stale code. ## Core Features & Use Cases - Framework Selection & Patterns: Databricks-specific guidance for FastAPI (default), Flask, Dash, Streamlit, Gradio, and Reflex, including correct app.yaml commands and production servers (uvicorn, Gunicorn). - Auth & Resources: Service principal auth via SDK Config(), on-behalf-of user tokens via x-forwarded-access-token, and valueFrom resource references for SQL warehouses, Lakebase, secrets, and serving endpoints. - Deployment & LLM Integration: CLI and Asset Bundles deployment workflows, plus working examples for foundation-model chat, parallel LLM calls, and structured JSON outputs. - Use Case: A data team needs an internal FastAPI service that queries a SQL warehouse with per-user Unity Catalog filters and deploys it as a Databricks App — this Skill provides the exact auth, app.yaml, and deployment steps. ## Quick Start Use the databricks-apps-python skill to scaffold a FastAPI backend for a Databricks App that queries my SQL warehouse and deploy it with the Databricks CLI.
