Year 2024

[mathbunnyru]

Thank you all contributors and users.

It's been a nice year for our images and the Jupyter ecosystem.

Notable changes

Sorry if I missed something, but I think these are quite notable changes that happened to our images:

• CUDA-enabled images for tensorflow and pytorch notebooks
• start.sh is now an ENTRYPOINT
• ubuntu 24.04, mamba v2, python 3.12
• now we have a CHANGELOG file to allow our users adapt newer versions faster
• almost no manual updates are needed (see below)
• various documentation improvements

What the future holds

Well, this repo is doing quite well without much interaction.

I look forward to having GitHub-hosted arm64 runners.
They should be available early next year: issue, blog
When we switch to GitHub-hosted arm64 runners, we should have less infrastructure code, faster builds, and, maybe, a bit more reliability.

Also, with more and more AI around, It's possible that users would want to have these features included with our images by default.

Few more thoughts

At this point, I think the quality of the images we produce mostly depends on the quality of the packages we use.
There are almost no manual changes that this repo needs on a regular basis - I am only updating tag examples in docs. Updating Python and Ubuntu versions is a rare thing and it's ok to have manual PRs for that.

Fun fact - tig images/ is less than 50 commits and fits on one page (on my screen).

One more thing

On a personal side, I moved to London this year and got a UK Global Talent visa.
If you have contributed to this project (or Jupyter ecosystem) or mamba, and you plan to spend some time in London, let me know, I'm always happy to have a coffee (beer and cider are also fine) 🙂
You can DM me on LinkedIn: http://linkedin.com/in/mathbunnyru/

Happy holidays 🎉

[mathbunnyru]

And one more notable change (I made it today): I updated the Home page and generation scripts.
Now, it's not just a long list of months with links, but nice tables grouped by year and also some stats (number of builds and number of images).

[consideRatio]

Wieee @mathbunnyru!!! Congratulations on the global talent visa! Thank you for a nice writeup about the year and for all your amazing contributions!

I hope to find an opportunity to meet with you in the future, London isn't so far from Uppsala/Sweden.

Happy holidays @mathbunnyru!

[mathbunnyru]

Thank you @consideRatio! ❤️

[mathbunnyru]

And one more notable change (I made it today): I updated the Home page and generation scripts. Now, it's not just a long list of months with links, but nice tables grouped by year and also some stats (number of builds and number of images).

As part of my New Year gift to this repo (haha), I almost completely rewrote the script generating the wiki home page.
And now the table has really nice information, both for each month and year.
Also reduced the complexity of the script and new changes to the home page won't require manual interaction with the wiki git repo.

This is how it looked before the change: https://github.com/jupyter/docker-stacks/wiki/Home/7ec293a264f0a4257b0ee40b8d66a2bea4ea6868
How it looks now: https://github.com/jupyter/docker-stacks/wiki

[yuvipanda]

Congratulations, @mathbunnyru \o/ Happy new year to you too :)

[mathbunnyru]

Congratulations, @mathbunnyru \o/ Happy new year to you too :)

Thanks @yuvipanda ❤️