Setup repo to build .whl for pypi and conda-forge

[MridulS]

Hey @quigleyj97! Currently the only way to install jupyterlab-spreadsheet requires users to use npm, I have added the required setup files to create a .whl distribution of the package so users can use pypi (and conda-forge) to install this on their machines. This follows the "recommended" setup config from https://github.com/jupyterlab/extension-cookiecutter-ts/.

To test this locally you can try building it with the following set of command:

$ pip install build
$ python -m build -s     # builds the source package
$ python -m build          # builds the .whl

This will build jupyterlab-spreadsheet/dist/jupyterlab_spreadsheet-0.4.1-py3-none-any.whl and this can be installed with pip.

Happy to help with the pypi/conda-forge-feedstock setup to automate the builds and releases for this :)

Let me know!

[MridulS]

Just wanted to bump this up in the notifications :)

[quigleyj97]

Thank you for doing this! I’ll play around with this tonight and come back with any Qs

[quigleyj97]

Hi again! Sorry for the delay (life has been very busy lately), but coming back to this with the impending 4.0 release of JupyterLab. I think having a python package could help in other ways too (like enabling server-side behaviors, such as 'live' XLSX sheets with the help of xlwings). I'll see if I can get this working this weekend.

[MridulS]

No worries! Let me know if I can help :)

[quigleyj97]

Ok, closing this in favor of #60, which builds off your PR to simplify the build and adds some additional Python files to get JL to recognize the Python package distribution as a valid Python package. I still need to fix some visual bugs but we should be ready to release in time for JupyterLab 4.0 with support for installing via pip. Thank you so much for your contributions!