Add Feature Tracking Cookbook to the gallery

[brian-rose]

This is a new Cookbook from the Pythia Cook-off 2026.

[github-actions]

👋 Thanks for opening this PR! The Cookbook will be automatically built with GitHub Actions. To see the status of your deployment, click below.
🔍 Git commit SHA: 7a28330
✅ Deployment Preview URL: https://ProjectPythia.github.io/cookbook-gallery/_preview/36

[brian-rose]

Don't merge this yet, we need to add appropriate gallery tags to the cookbook repo.

[brian-rose]

ProjectPythia/feature-tracking-cookbook#82 added gallery tags, but we still need a description. Tracking this in ProjectPythia/feature-tracking-cookbook#83.

Once that's all done, close and reopen this PR to see a correct preview of the gallery rendering.

[brian-rose]

I think this is ready to go. Can someone from the Pythia team give a quick review?

[r-ford]

Overall, this looks pretty good. Some comments:

• The appendix notebook is raising an error when trying to pull in data via OSDF
• For relatively complex algorithms, I understand the need for the large code blocks, but it would help to have more comments throughout the algorithm explaining what is happening
• Most notebooks don't have Summary/What's next sections, but I don't really think it is a problem, since the structure of the cookbook is self-explanatory
• Just an observation: the tracked features in most of the figures are hard to see in the rendered book because the maps are wide. While it's easy enough to open them in a new tab, is there any sort of functionality in Jupyter Book to zoom into images?

I think this is ready enough for the gallery, especially if that appendix notebook error is resolved.

[brian-rose]

• The appendix notebook is raising an error when trying to pull in data via OSDF

This is an intermittent error, unfortunately. PelicanFS gives some robustness to network glitches but these failures still occur once in a while during build.

I just re-ran a new build, and it ran properly, see the current deployment here.

I could experiment with turning the --strict build flag back on for myst in this repo, so at least we would avoid publishing the failures when they occur on nightly builds.

@hrhampapura, any particular advice on how to deal with these "500 internal server error" issues?

[brian-rose]

@r-ford I'll take your other comments and turn them into issues in the cookbook repo for future tracking.

[hrhampapura]

@brian-rose One way to avoid these 500 errors (coming from bad OSDF caches) is to directly read from the data origin using: msl = xr.open_dataset(path, engine='zarr', storage_options={"direct_reads": True})

[brian-rose]

I opened ProjectPythia/feature-tracking-cookbook#89 to use strict builds for this cookbook, which I think is appropriate anyway for a cookbook that's no longer in rapid development mode, and will avoid publishing bad results from nightly builds.

Once that is merged, I think we should merge this PR and list the cookbook on the gallery.