✔️ Linting Passed

All linting checks passed. Your pull request is in excellent shape! ☀️

_{Generated for commit: 7737ae5. Link to the linter CI: here}

Looks like OpenML has issues. Its status page is all green but I am getting a 404 when I tried to download a dataset e.g. clicking the download button from this page

For further reference the fetch_openml with id=41082 is trying to download from https://api.openml.org/data/v1/download/19329737/USPS.arff

This may have been noticed by others, see openml/OpenML#1232 and openml/openml-python#1433 that seemed similar and recent.

@lesteve Can we resolve this rather quickly? Either merge this PR first or #31905. I need #31905 (Cython) to go on.

Unfortunately this seems like an OpenML issue that is a bit out of our hands ... hopefully it will be fixed in the upcoming days.

Would that be too inconvenient for you to work on your PRs that need the latest Cython with the lock-files updated, wait for reviews, and then once your PRs are reviewed hopefully the OpenML issue would have been fixed?

Let's say I can live with it.

The OpenML issue should be fixed now :).

Thanks a lot @joaquinvanschoren for the fix!

I relaunched the doc build again and it looks like we still have one issue that may be unrelated to the ARFF download issue that you fixed.

It seems like we are not able to download a parquet dataset that one of our example uses:
https://data.openml.org/datasets/0004/44063/dataset_44063.pq. Locally I am not able to download it in less than one minute (I did not try waiting for longer but that seems slow):

curl https://data.openml.org/datasets/0004/44063/dataset_44063.pq --max-time 60

Edit: I actually tried to wait for 5 minutes with --max-time 300 and same result.

I saw in openml/openml-python#1433 that there may have been some issues about slow downloads recently as well.

I saw in openml/openml-python#1433 that there may have been some issues about slow downloads recently as well.

Maybe a comment to see is this one? openml/OpenML#1232 (comment)

This is also fixed since last weekend.

Thanks, I can confirm that now downloading the problematic .pq dataset is fine locally:

curl https://data.openml.org/datasets/0004/44063/dataset_44063.pq --max-time 60 --output dataset.pq

The weird thing is that one of the doc CI build failed this morning (Monday August 18, ~5am UTC) with Too long with no output:

generating gallery for auto_examples/applications... [ 31%] plot_time_series_lagged_features.py
Too long with no output (exceeded 10m0s): context deadline exceeded

I relaunched the build to see if the problem goes away.

Note: plot_time_series_lagged_features.py is downloading https://data.openml.org/datasets/0004/44063/dataset_44063.pq.

So it seems like the doc build succeeded this time 🎉!