The target audience for the Calkit extension is scientists and other researchers who often create collections of notebooks for different aspects of their projects but don’t automate their execution, so they need to be run one-by-one, which is a source of inefficiency and potential mistakes.
Here users can define and attach an environment to each notebook, for which all information is stored locally in the project folder, and the extension ensures it matches its spec before launching the kernel and attaching it to the notebook.
DVC is leveraged to keep track of notebook staleness according to declared inputs and outputs.
It can be installed with:
uv tool install calkit-python
or
pip install calkit-python
Or you can try it out without installing:
uvx calk9 jupyter lab
Tutorial video: https://youtu.be/8q-nFxqfP-k
Source code: GitHub - calkit/calkit: Single-button reproducible research project management.
Docs: JupyterLab - Calkit