DataFusion Documentation

This folder contains the source content of the python api. This is published to https://arrow.apache.org/datafusion-python/ by a GitHub action when changes are merged to the main branch.

Dependencies

It's recommended to install build dependencies and build the documentation inside a Python virtualenv.

• Python
• pip3 install -r requirements.txt

Build & Preview

Run the provided script to build the HTML pages.

./build.sh

The HTML will be generated into a build directory.

Preview the site on Linux by running this command.

firefox build/html/index.html

Release Process

The documentation is served through the arrow-site repo. To release a new version of the docs, follow these steps:

1. Run ./build.sh inside docs folder to generate the docs website inside the build/html folder.
2. Clone the arrow-site repo
3. Checkout to the asf-site branch (NOT master)
4. Copy build artifacts into arrow-site repo's datafusion folder with a command such as

• cp -rT ./build/html/ ../../arrow-site/datafusion/ (doesn't work on mac)
• rsync -avzr ./build/html/ ../../arrow-site/datafusion/

5. Commit changes in arrow-site and send a PR.