microsoft
Jupyter Notebooks documentation inconsistent
On the website, it states that working with Jupyter Notebooks in VS Code requires environments have the Jupyter package installed.
The GitHub Markdown docs, which I think is correct, states that you only need ipykernel installed. This is also shown in the link that appears on the actual application modal, when you try to execute a cell in a Python environment before a kernel is detected.
Given how much a full Jupyter installation pollutes the environment as compared to just the kernel package (in both number of packages and size), I think this small nuance should be corrected.
Can you elaborate? Inside of a new virtual environment, if you try to execute a cell without ipykernel, it will prompt to install:
If you cancel, the cell output will contain this error:
Running cells with 'test-env' requires the ipykernel package.
Run the following command to install 'ipykernel' into the Python environment.
Command: ~/.venv/test-env/bin/python -m pip install ipykernel -U --force-reinstall'"
That's what the docs say: https://code.visualstudio.com/docs/datascience/jupyter-notebooks#_debug-a-jupyter-notebook
I am not talking about debugging. To be clear, the application attempts to install ipykernel even if you do not use Run By Line or Debugging features.