Jupyter notebooks

We generally recommend packaging reusable code components into Python modules where possible. However occasionally we support researchers who prefer a notebook environment or projects which want to provide examples and tutorials in notebook format. Notebooks can also be a valid alternative to Python scripts for running and recording the results of numerical experiments for example.

Live executable environments

Name	Short description	🚦
binder	Turns a Git repository into a collection of interactive notebooks	🟢
Google Colab	Colab is a hosted Jupyter Notebook service that requires no setup to use and provides free access to computing resources, including GPUs and TPUs.	🟢
GitHub codespaces with JupyterLab	A codespace is a development environment that is hosted in the cloud and can run Python and Jupyter notebooks	🟠

Linting

Many of our recommended linters don’t work out-of-the box on Jupyter notebooks, however nbQA has been found to work well. This is a translation tool that can be used to run our recommended linters over notebooks.

There is one exception: we’ve found that black’s own Jupyter extension works more reliably than black via nbQA.

Name	Short description	🚦
black[jupyter]	(also black-jupyter) black with the optional Jupyter extension.	🟢
nbQA	Recommended for all other linters (ruff, isort) and mypy.	🟢
pre-commit	Has cell output cleanup hooks (if desired). Also found to work well with nbQA.	🟢
nbqa black	black via nbQA.	🟠

Other

Name	Short description	🚦
jupytext	Automatically convert notebooks to/from a plain text format that is more human readable and better for version control than the standard notebook format.	🟢