Upgrade notebook Gaussian Process to V4

[symeneses]

It solves #79
The number of particles n_particles was changed to 100 to get the notebook executed within an acceptable time.
Thanks to @cluhmann @fonnesbeck for your help 🔥

Co-authored-by: @AnaRita93

#DataUmbrellaPyMCSprint

Gaussian Process notebook

• Notebook follows style guide https://docs.pymc.io/en/latest/contributing/jupyter_style.html
• PR description contains a link to the relevant issue: a tracker one for existing notebooks or a proposal one for new notebooks
• Check the notebook is not excluded from any pre-commit check: https://github.com/pymc-devs/pymc-examples/blob/main/.pre-commit-config.yaml

Helpful links

• https://github.com/pymc-devs/pymc-examples/blob/main/CONTRIBUTING.md

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[symeneses]

We have 2 observations:

1. In the cell:

y_pred = [

    np.random.multivariate_normal(m, S)
    for m, S in zip(gp_trace.posterior["mu_post"].T, gp_trace.posterior["sigma_post"].T)
]

We use Transpose to get the right dimensions, but not sure if this was the best way to do it.

2. It seems that the notebooks name is misleading, as the name indicates that the notebook is about GP in general, but that's not the case. The other notebooks in the folder are named GP-XXX. Should we change the name?

[OriolAbril]

It seems that the notebooks name is misleading, as the name indicates that the notebook is about GP in general, but that's not the case. The other notebooks in the folder are named GP-XXX. Should we change the name?

I am not very opinionated on this right now, but I will oppose changing the name if it comes up in the future. Currently most of the old links are broken (in the sense that they point to v3 docs, but using the version switcher ends up in a 404 page), so changing the name right now has very little effect. Once the updated www.pymc.io/projects/examples/ links are extended, changing the name will break many references to the docs, so I don't think it would be a good idea to do anymore.

[cluhmann]

I am not understanding the Read the Docs error message. It seems to be complaining about another notebooks. @OriolAbril ?

[OriolAbril]

I am not understanding the Read the Docs error message. It seems to be complaining about another notebooks. @OriolAbril ?

Already commented above, but I'll add some extra info here for full context. The line right above the error message makes reference to a different notebook yes, but that is a sphinx warning, not an error. It is something that we should be aware of and try to fix it (it might become an actual error in a future version for example), but it is not breaking the doc build. The error references:

File "/home/docs/checkouts/readthedocs.org/user_builds/pymc-examples/envs/404/lib/python3.8/site-packages/ablog/post.py", line 40, in _split
    return {s.strip() for s in a.split(",")}
AttributeError: 'NoneType' object has no attribute 'split'

which means that what is breaking the doc build is something related to the post directive, it doesn't give specific information about what exactly in the post directive is breaking everything. That I am inferring/deducing from the bad format of the date in the post directive plus the split(",") piece; the expected format has a comma, the year only that is currently present does not. My guess is instead of splitting into two substrings, no splitting can happen and a None appears somewhere the ablog extension expects a string and we get the error.

[OriolAbril]

Looks like I did not guess the cause of the readthedocs issue 😕 . Can you try adding ablog<0.10.27 to https://github.com/pymc-devs/pymc-examples/blob/main/requirements-docs.txt#L7? It looks like they made a new release in a date that could explain the errors (I'll try to look into this later on once I have time and see if the release introduced a bug or if we are doing something wrong)

[symeneses]

Looks like I did not guess the cause of the readthedocs issue 😕 . Can you try adding ablog<0.10.27 to https://github.com/pymc-devs/pymc-examples/blob/main/requirements-docs.txt#L7? It looks like they made a new release in a date that could explain the errors (I'll try to look into this later on once I have time and see if the release introduced a bug or if we are doing something wrong)

I tested locally with ablog 0.10.26 and it worked, so let's see. I added the related issue in the comment.

[OriolAbril]

It looks like some of the updated versions (no idea what though) is breaking the thumbnail gallery: https://pymcio--404.org.readthedocs.build/projects/examples/en/404/gallery.html.

This one seems to be the last working build for pymc-examples. For now, can you check the versions used there for sphinx and sphinx gallery and pin them? Then open an issue as a reminder to look into this and fix it.

[OriolAbril]

There should be a watermark title right above this cell: https://docs.pymc.io/en/latest/contributing/jupyter_style.html#watermark

[symeneses]

The error in the gallery is still there 😭 . I updated the issue with some more details.

[OriolAbril]

I checked the preview and the gallery now and it seems to be working for me. Does it still happen for you after refreshing multiple times or trying a different browser?

[symeneses]

I checked the preview and the gallery now and it seems to be working for me. Does it still happen for you after refreshing multiple times or trying a different browser?

Niiice, it's working fine now. No need to cry 😉