nbtoolbelt - Tools to Work with Jupyter Notebooks

• validate: validate notebooks
• head: show head or tail of notebooks
• dump: dump notebook info and source on terminal
• stats: summarize notebooks with statistics
• view: view notebook, including all embedded images, LaTeX, and HTML in a browser
• cat: concatenate multiple notebooks
• clean: clean notebooks by removing specified elements
• run: execute notebooks, with pre/post cleaning
• split: split notebooks into MarkDown, code, and raw
• punch: punch holes into notebooks and fill them (for creating exercises)

Available as library functions and as configurable command-line scripts.

Installation

pip install nbtoolbelt

Documentation

Documentation is available on Read the Docs.

On the command line, you can use the options -h or --help.

Usage

On the command line:

nbtb [-h] [options] tool [options] nb.ipynb ...

As library: see documentation

Testing

pip install nbtoolbelt[test]

nbtoolbelt comes with a set of automatic test cases for pytest.

Developing

Some useful commands, and where to run them:

• In nbtoolbelt/docs/,

  • clean build directory: make clean
  • create html documentation tree: make html
  • create pdf documentation: make latexpdf
  • determine size of documentation: wc `find . -name '*.rst'`

• In nbtoolbelt/test/,

  • run all test cases: pytest .

• In nbtoolbelt/,

  • test package configuration: python setup.py check -r -s
  • create source distribution and wheel: python setup.py sdist bdist_wheel

• In nbtoolbelt/dist/

  • create digital signature: gpg --detach-sign -a ...
  • upload to PyPI: twine upload ...

• In nbtoolbelt/src/,

  • determine size of code: wc `find . -name '*.py'`

Steps to add a feature:

1. Add issue.

2. Design interface.

   1. Add (failing) test cases in nbtoolbelt/test/.

   2. Add documentation in nbtoolbelt/docs/.

3. Implement feature in nbtoolbelt/src/nbtoolbelt/.

4. Update nbtoolbelt/src/nbtoolbelt/_version.py.

5. Update CHANGELOG.rst.

6. Commit and push changes.

7. Close issue, indicating commit hash.

License

Copyright (c) 2017 - Eindhoven University of Technology, The Netherlands

This software is made available under the terms of the MIT License.

Resources

• Python: Python 3
• Jupyter
• Jupyter notebook format
  • Documentation
  • Includes a format validator based on JSON schemas, such as nbformat.v4.schema.json
• JSON Schema
• ./jq: a lightweight and flexible command-line JSON processor
• Python 3 Patterns, Recipes and Idioms