I created a line magic for debugging to creates cells dynamically

JupyterLab extensions
1

jupyter_ipython_load_node
https://docs.kedro.org/en/latest/notebooks_and_ipython/kedro_and_notebooks.html#debugging-a-kedro-project-within-a-notebook

I am working for a machine learning framework called kedro. The idea of the line magic is to lower the barrier for debugging. While live debugger is powerful, it is not leveraged widely by our main users group. i.e. data scientist. In addition, it is more convenient to do interactive stuff in a notebook which you can’t in a debugger (i.e. plotting a graph). The concepts of Kedro is not too important here.

By leveraging ipylab and the resource in this forum, I am able to create multiple cells interactively with Jupyter Lab or Notebook >7.0.

I want to share this experience because I find it interesting but I am also frustrated with the fragment developer experience on different notebook platforms. I think Notebook 7.0 is a great step to unify the experience between Notebook & Lab. For any Jupyter extension developer, it is a consistent challenge to develop something that work across different Notebook-like platform (SageMakers, Colab, Databricks, VSCode notebook).

Right now the magic supports mainly:

  • Notebook & Lab
  • VS Code notebook & IPython
  • Databricks (limited support)

Credits: