Codecov Report

All modified and coverable lines are covered by tests ✅

Flags with carried forward coverage won't be shown. Click here to find out more.

@yuejiaointel @david-cortes-intel FYI. Not sure if python version matters for doc release work.

@yuejiaointel @david-cortes-intel FYI. Not sure if python version matters for doc release work.

It does matter because 3.13 has breaking changes in terms of how docstrings are post-processed by python, but the CI job is green so guess all our workarounds worked.

Yes I tested new python when I was working on doc release and intelex doc build will have errors, example on my fork: https://github.com/yuejiaointel/scikit-learn-intelex/actions/runs/15126053446/job/42518204573

Up to you on what makes sense - whether its closing this one and continuing to use 3.11 or leaving it open and creating a fix/ticket to address the issue with 3.13

Yes I tested new python when I was working on doc release and intelex doc build will have errors, example on my fork: https://github.com/yuejiaointel/scikit-learn-intelex/actions/runs/15126053446/job/42518204573

But the CI here shows in green, and the generated artifact looks fine. Or is there something that the CI here is not catching?

Yes I tested new python when I was working on doc release and intelex doc build will have errors, example on my fork: https://github.com/yuejiaointel/scikit-learn-intelex/actions/runs/15126053446/job/42518204573

But the CI here shows in green, and the generated artifact looks fine. Or is there something that the CI here is not catching?

I think it is a potential bug, I noticed in the --gh-pages we did not pass -W to sphinx-build command, makefile has $(o) so we don't need it in make html, but need to explicitly pass it in sphinx-build , will make a quick PR to fix this

Yes I tested new python when I was working on doc release and intelex doc build will have errors, example on my fork: https://github.com/yuejiaointel/scikit-learn-intelex/actions/runs/15126053446/job/42518204573

But the CI here shows in green, and the generated artifact looks fine. Or is there something that the CI here is not catching?

I think it is a potential bug, I noticed in the --gh-pages we did not pass -W to sphinx-build command, makefile has $(o) so we don't need it in make html, but need to explicitly pass it in sphinx-build , will make a quick PR to fix this

But the CI job doesn't use --gh-pages.

Yes I tested new python when I was working on doc release and intelex doc build will have errors, example on my fork: https://github.com/yuejiaointel/scikit-learn-intelex/actions/runs/15126053446/job/42518204573

But the CI here shows in green, and the generated artifact looks fine. Or is there something that the CI here is not catching?

I think it is a potential bug, I noticed in the --gh-pages we did not pass -W to sphinx-build command, makefile has $(o) so we don't need it in make html, but need to explicitly pass it in sphinx-build , will make a quick PR to fix this

But the CI job doesn't use --gh-pages.

Tho the doc CI does not use gh-pages, the doc release pipeline is using it

Yes I tested new python when I was working on doc release and intelex doc build will have errors, example on my fork: https://github.com/yuejiaointel/scikit-learn-intelex/actions/runs/15126053446/job/42518204573

But the CI here shows in green, and the generated artifact looks fine. Or is there something that the CI here is not catching?

I think it is a potential bug, I noticed in the --gh-pages we did not pass -W to sphinx-build command, makefile has $(o) so we don't need it in make html, but need to explicitly pass it in sphinx-build , will make a quick PR to fix this

But the CI job doesn't use --gh-pages.

Tho the doc CI does not use gh-pages, the doc release pipeline is using it

scikit-learn-intelex/.github/workflows/docs-release.yml

Line 98 in be4777e

./build-doc.sh --gh-pages

@yuejiaointel But then: there doesn't seem to be anything wrong with the updates in this PR. The CI job here gets passed the -W option currently, and doesn't result in the same import errors that were popping up in a pipeline from your fork. The generated artifact also looks Ok here as far as I can tell.

Yes I tested new python when I was working on doc release and intelex doc build will have errors, example on my fork: https://github.com/yuejiaointel/scikit-learn-intelex/actions/runs/15126053446/job/42518204573

But the CI here shows in green, and the generated artifact looks fine. Or is there something that the CI here is not catching?

I think it is a potential bug, I noticed in the --gh-pages we did not pass -W to sphinx-build command, makefile has $(o) so we don't need it in make html, but need to explicitly pass it in sphinx-build , will make a quick PR to fix this

But the CI job doesn't use --gh-pages.

Tho the doc CI does not use gh-pages, the doc release pipeline is using it

scikit-learn-intelex/.github/workflows/docs-release.yml

Line 98 in be4777e

./build-doc.sh --gh-pages

@yuejiaointel But then: there doesn't seem to be anything wrong with the updates in this PR. The CI job here gets passed the -W option currently, and doesn't result in the same import errors that were popping up in a pipeline from your fork. The generated artifact also looks Ok here as far as I can tell.

I think because docs.yml in .ci is using python 3.11, but this PR is changing python version for doc-release.yml from 3.11 to 3.13, so the CI job pass without any warnings. But if we make this change, doc-release.yml when triggered by tag/manually will fail

Got it. So then a first step before merging this PR would be to upgrade the regular docs CI to python3.13 and ensure that it works correctly (CC @ethanglaser ).