JupyterClearCheck

A Python CLI to check that a Jupyter Notebook's Outputs' have been cleared.

Why does this exist?

During the development of a Jupyter Notebook, it is common to run cells multiple times. Every time a cell is run, its output is stored in the .ipynb file format. Because the outputs are saved, this can cause merge conflicts in version control systems like Git when multiple people are working on the same notebook.

nbconvert can be used to clear the outputs of a notebook, but to our knowledge, there was no package to check if the outputs have been cleared. This package aims to fill that gap.

This package can be used in a CI/CD pipeline to ensure that the outputs of a notebook have been cleared before merging a pull request reducing merge conflicts on the outputs of a notebook.

Installation

This package is available on PyPI and can be installed using pip.

pip install JupyterClearCheck

Usage

You can run the command via jupyterclearcheck or python -m jupyterclearcheck.

$ jupyterclearcheck --help

usage: jupyterclearcheck [-h] --notebook NOTEBOOK

Check a Jupyter Notebook for output cells.

options:
  -h, --help           show this help message and exit
  --notebook NOTEBOOK  Path to the Jupyter Notebook file

The notebook parameter is required and should be the path to the Jupyter Notebook file.

Contributing

We welcome contributions! Thank you in advance for your help. Feel free to submit a pull request or open an issue.

Getting Started for Development

We use Python 3.9 and PDM for dependency management. To get started, install Python 3.9 and then install PDM using pip.

pip install pdm

Once this step is done, PDM can be used to run commands related to the development of this package.

• Install dependencies: pdm install
• Run linting and formatting: pdm run lint_and_format
• Checking linting, formatting, and type annotations: pdm run check
• Run tests: pdm run test
• Build package: pdm build

Running CI pipeline locally

The CI pipeline can be run locally using the following scripts:

• Windows: ./build_and_test.sh
• Linux/MacOS: ./build_and_test.cmd.

Future Work

• Add shellcheck to lint bash scripts currently in use
• Add pre-commit hook
• Add markdown lint
• Publish on conda-forge