Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
CI Add wheel builds for Python 3.11 #24446
CI Add wheel builds for Python 3.11 #24446
Changes from 46 commits
99c99b9
99f2f36
a005d02
18a1807
f7923b1
a75bb7d
0b42e61
d073620
6add9af
b69242c
bce8672
6c1a129
a48568b
2f411a9
5345e01
7e3af0a
4c47cb6
00dbfd8
cbbf87e
b77b6a3
40c1d03
45cee88
1816933
26a41ab
18eac63
0c4e80d
c51315e
fca7ca6
65dbb59
f0938db
b4b6f5f
2d2baed
2d1cd7f
82a21b6
33fe747
9950435
fd2371f
8fe7c36
21e769d
3210043
657b563
6db7e7c
c9a59e2
ed2c6a0
a3a8100
5a43727
66ab6de
75e53eb
b12be0b
a1e9d7d
1c50c26
File filter
Filter by extension
Conversations
Jump to
There are no files selected for viewing
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Shouldn't we do that only for the 3.11 build? The other builds should work with the released version of numpy / scipy (once scipy/scipy#17224 backported as part of scipy 1.9.3).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am not completely aware of the scipy wheels publication procedure.
In my opinion if it had been possible to test scikit-learn nightly wheels against scipy nightly wheels the failure in #24612 and the one scipy/scipy#17224 is willing to fix (both related to 3.10 too), would have been detected before the release of scipy 1.9.2 , and the backporting would not have been necessary.
The question is how far do you want to go with testing.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Now that SciPy 1.9.3 has been released, I think we need to remove this step from this PR if we want to publish stable wheels once Python 3.11 is officially release.
We in the meantime create another PR for adapting scikit-learn nightly builds' dependences (incorporating this step in if needed). What do you think?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sorry @cmarmo I had missed your reply. Indeed the testing against the nightly builds of our dependencies is a different goal than publishing our own nightly built wheels for project that depend upon us. I think that for the latter we should build always build the wheels we publish against stable dependencies not have our wheel publishing infrastructure blocked by a bug/regression in the dev branch of one of our dependencies.
But I agree with your suggestions and @jjerphan's that we might want to upgrade our nightly
[scipy-dev]
tests (currently configured on Azure pipelines and only for linux) to also run the tests on all the platforms, probably by reusing the config done on github actions for both goals (testing against dev dependencies and publishing our own dev wheels).But let's do that in a dedicated follow-up PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As discussed in https://github.com/scikit-learn/scikit-learn/pull/24446/files#r1000353644, let's focus this PR on building and publishing 3.11 wheels against stable dependencies and remove the use of dev dependencies from this PR.
We can tackle nightly testing against dev dependencies (without uploading) in a dedicated PR.
To do that, we could introduce a second github actions scheduled event with a specific env variable/config flag to also disable the final upload to anaconda or pypi.
We will also have to re-implement the
[scipy-dev]
commit message trigger currently implemented in our Azure Pipelines config and delete that config from Azure Pipelines.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@cmarmo I did not have time to write that comment and you addressed it concurrently :)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think this sets the variables for
pip
to pick up laterThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I understand your comment, but the script is working eg here, the variable is set: do you mind explaining to me what is the difference with the export command? Sorry for bothering and thanks for your patience.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I agree with @cmarmo: using
ARG
should be enough to set variables.