Contributing Code to Jupyter Book

This guide covers the technical setup and architecture for code contributors.

Application Design¶

The Jupyter Book application is a Python package containing a Node.js application. It’s functionally equivalent to the MyST-MD engine.

Python Shim¶

The Python package ensures users have Node.js to run the underlying application. For users without Node.js, it uses nodeenv to download a local copy.

JavaScript Application¶

The core application is written in JavaScript, currently extending the mystmd CLI. In future, mystmd may become smaller with some markup extensions moving to Jupyter Book.

Jupyter Book 2’s CLI adds some extra functionality (like a way to upgrade from Jupyter Book 1), but is intentionally as lightweight as possible.

Building the Python Package¶

Jupyter Book uses hatch to build the Python package. Configuration is in pyproject.toml, using plugins:

• hatch-jupyter-builder - builds the Node.js application

• hatch-deps-selector - manages dependencies for conda-forge vs PyPI

• hatch-nodejs-version - provides metadata from the JS package

Build the application:

hatch build