is the install.py script still relevant given data_files support in `setup.py`?

[yifanwu]

Hi there,

Thank you for creating this reference custom kernel!

Through some investigations, I thought that the recommended path is to use data_files support in setup.py, as opposed to manually moving files into the share/jupyter/kernels/ directory. I also noticed that the kernel.json pattern (mentioned in https://jupyter-client.readthedocs.io/en/stable/kernels.html) is not used here.

Is this current repo not up to date with the most recent ipykernel developments, or did I mis-read some instructions?

Thanks so much for the help!

[ceteri]

I agree @yifanwu as far as I can tell this no longer installs correctly. Am I missing something here, or are the Jupyter docs out of sync with the committed code and both out of sync with the demo repos?

I've tried with v1.1, using Py 3.10.11 on macOS:

IPython          : 8.22.1
ipykernel        : 6.29.3
ipywidgets       : not installed
jupyter_client   : 8.6.0
jupyter_core     : 5.7.1
jupyter_server   : 2.12.5
jupyterlab       : 4.1.2
nbclient         : 0.9.0
nbconvert        : 7.16.1
nbformat         : 5.9.2
notebook         : not installed
qtconsole        : not installed
traitlets        : 5.14.1

[blink1073]

The echo kernel is using hatch to install the kernel, which works on PyPI and conda-forge. install.py is still there but could be removed, simplified, and absorbed into jupyter/echo_kernel@main/hatch_build.py.

[blink1073]

#17 is part of what we need, the other is https://github.com/jupyter/jupyter_client/issues/1012

[blink1073]

https://github.com/jupyter/jupyter_client/pull/1013

[ceteri]

thank you for the help @blink1073