mkdocs-jupyterlite 0.2.0

Embed interactive JupyterLite notebooks in your MkDocs site.

mkdocs-jupyterlite

A MkDocs plugin that enables embedding interactive jupyterlite notebooks in your docs.

Say you have a notebook example.ipynb in your awesome project, and you want users to be able to play around with it. In the past, you could use a tool like Binder to achieve this. But, that requires a full docker environment and a remote server. By using JupyterLite, you can run Jupyter notebooks directly in the browser without any server-side dependencies.

However, to use jupyterlite, you have to manually install jupyterlite and then run a build step to package your notebooks, other files, and python dependencies into a single static site.

This plugin automates that process for you.

Installation

Step 1: Install the plugin from PyPI:

python -m pip install mkdocs-jupyterlite

Step 2: Configure your mkdocs.yml file

See the mkdocs.yml that configures this project's site.

site_name: mkdocs-jupyterlite
site_url: https://nickcrews.github.io/mkdocs-jupyterlite/
repo_url: https://github.com/nickcrews/mkdocs-jupyterlite/

nav:
  - Home: index.md
  - Notebook 1: notebook.ipynb

plugins:
  - jupyterlite:
      enabled: true
      notebook_patterns:
        # include all
        - "**/*.ipynb",
        # exclude drafts
        - "!**/draft_*.ipynb",  
        # re-include a specific draft
        - "project/drafts/draft_keep.ipynb",
        # exclude an anchored notebook
        - "!/top_secret.ipynb",

Here are the details on the configuration options:

enabled

bool, whether or not the plugin is enabled. Defaults to true.

notebook_patterns

A list of patterns that uses gitignore semantics to include and exclude files. The last matching pattern will be used to determine if a file is a notebook.

For all files that match, the content of the page will be an iframe that embeds the JupyterLite Notebook html.

Contributing

I want this to be usable for other people, so file an issue if you want to use this in your site, but run into any problems.

Possible improvements:

• Include custom python wheels into the JupyterLite environment.
• Passing an entire jupyter-lite.json config file.
• Instead of using an iframe, actually inline the contents of the generated HTML?
• Fix the TOC so clicking headers actually scrolls in the iframe.