Python Frameworks¶

Work In Progress

Databricks Apps is not opinionated on which python framework to use. All of them should work. Some like Streamlit or a bit special but work out-of-the-box. Full-stack ones like Django or Superset need to the configured correctly to delegate auth and use SQLLite.

Plotly Dash¶

Great framework. High initial learning curve.

Gradio¶

Great to create AI demos. Can be limiting as the app gets more complex.

Preset Environment Variables¶

Variable	Description
GRADIO_SERVER_NAME	Server name for Gradio. Set to `0.0.0.0`.
GRADIO_SERVER_PORT	Port value used by Gradio. Set to `$DATABRICKS_APP_PORT`

How to get request headers¶

TODO: link to app examples

Streamlit¶

Preset Environment Variables¶

Variable	Description
STREAMLIT_SERVER_ADDRESS	Server address used by Streamlit. Set to `0.0.0.0`
STREAMLIT_SERVER_PORT	Port used by Streamlit. Set to `$DATABRICKS_APP_PORT`
STREAMLIT_SERVER_ENABLE_XSRF_PROTECTION	Set to `false`. See Proxy
STREAMLIT_SERVER_ENABLE_CORS	Set to `false`. See Proxy
STREAMLIT_SERVER_HEADLESS	true
STREAMLIT_BROWSER_GATHER_USAGE_STATS	false

How to get request headers¶

TODO: link to app examples

Flask¶

Django¶

Solara¶

TODO

FastAPI¶

Variable	Description
UVICORN_HOST	Uvicorn host. Set to `0.0.0.0`.
UVICORN_PORT	Uvicorn port. Set to `$DATABRICKS_APP_PORT`

React¶

Totally possible. Just need to build the static assets first and serve them from your python app.