apparently now you can run Jupyter notebooks in the browser without a server with WebAssembly using JupyterLite?? https://jupyterlite.readthedocs.io/en/stable/

I learned about this because @paddymul made a pull request to my pandas-cookbook tutorial on github to add support for it

you can try it out here, it seems to work: https://jvns.github.io/pandas-cookbook/lab/index.html

JupyterLite — JupyterLite 0.4.3 documentation

@b0rk Just got back into working with Jupyter recently. Genuinely excited to check this out. Thanks for sharing!

@b0rk oh wow! Just fiddling with the one you linked is so killer.

I switched to dotnet interactive notebooks a few years ago (I just really like f#). Given dotnet actually has a really good webassembly story these days... I should see if I can't make that work directly in-browser.

@b0rk It works. There’s tiny differences, alas. We used it for humble data Jupyter workshops (https://github.com/HumbleData) for a while, but for this year switched back to local install at EuroPython because (as of a July 2024) we couldn’t get the “%load” magic to work which we use to deliver cell solutions. Also it made sense to help beginners set up Jupyter on their own devices.
Humble Data

Humble Data has 7 repositories available. Follow their code on GitHub.

GitHub
@b0rk It even works on an iPhone! At least print (“hello”) did. *boggled*

@b0rk Yeah, it's amazing, and I use it a lot. Two things to know:

a) The recent work with `uv` has made running jupyter locally *really* easy, and

b) jupyterlite stores your notebooks in some kind of local storage. Over the course of months or years, at least for me, they tend to all disappear every now and then

@zellyn thanks! I'm wondering if I should update the instructions for the my pandas cookbook repo to suggest using `uv`, just today I noticed that it still recommends anaconda which is a super out of date recommendation