Jupyter Notebook Splitter

This tool splits a Jupyter Notebook into Sub-Notebooks depending on cell metadata. It converts a Master Notebook into a Teacher Notebook and a Student Notebook; or into a Slides Notebook, a Tasks Notebook, and a Solutions Notebook.

Installation

Although the Notebook Splitter is only a single file it can be installed via pip

pip install notebook-splitter

Usage

Overview

TL;DR: See notebook-splitter --help.

1. Add cell metadata to your Jupyter Notebook: Add an exercise key (default, can be changed) to the metadata (JSON); give it values (tags) on which to create Sub-Notebooks

   {
       "exercise": "task"
   }
   // another cell
   {
       "exercise": "solution"
   }

2. Use --keep and --remove flags of the Notebook Splitter to keep and remove cells with according tags; export it to the respective Notebook:

   notebook-splitter input.ipynb --keep task --remove solution   -o tasks.ipynb
   notebook-splitter input.ipynb --keep solution --remove task   -o solutions.ipynb
   notebook-splitter input.ipynb --remove task --remove solution -o slides.ipynb

Examples in Action

See the examples directory in this repository.

Options

• Repeated Parameters: --keep and --remove parameters on the command line of the script can be given multiple times: --keep task --keep onlytask --remove solution
• Remove All: As a special parameter value, --remove all will remove all cells except those for which a --keep value is specified (--keep all is the default)
• Stdin/Stdout: If no output file is given with -o/--output, the resulting Notebook will be printed to stdout; if no input file as a parameter is given, the input Notebook will be read from stdin (good for Linux-like daisy-chaining of tools)
• Change Basekey: In the above example, the cell meta data key of discrimination is exercise which is the default. With --basekey, this can be changed.

Limitations

The values to the --keep and --remove parameters create sets of values to keep and remove. One could implement this tool probably quite cleverly with set operations (with the added complication of the --remove all ). If you can, feel free to file a merge request!