Move dependency tests from PreCommit Coverage to PostCommit

[Abacn]

Please add a meaningful description for your change here

PreCommit Python Coverage costs ~2h30min to run outstandingly. Splitting the dependency compatibility unit tests from this test suite and put into a new PostCommit

---

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

• Mention the appropriate issue in your description (for example: addresses #123), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, comment fixes #<ISSUE NUMBER> instead.
• Update CHANGES.md with noteworthy changes.
• If this contribution is large, please file an Apache Individual Contributor License Agreement.

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI or the workflows README to see a list of phrases to trigger workflows.

[Abacn]

btw code coverage report seems broken for a while:

[2024-03-21T20:24:10.281Z] ['error'] There was an error running the uploader: Error uploading to https://codecov.io: Error: There was an error fetching the storage URL during POST: 400 - Could not determine repo and owner

[github-actions]

Assigning reviewers. If you would like to opt out of this review, comment assign to next reviewer:

R: @liferoad for label python.
R: @damccorm for label build.

Available commands:

• stop reviewer notifications - opt out of the automated review tooling
• remind me after tests pass - tag the comment author after tests pass
• waiting on author - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)

The PR bot will only process comments in the main thread (not review comments).

[Abacn]

checked that these marked tests also run on Python PreCommit Coverage and the specific Precommits (e.g. Dataframe)

beam/sdks/python/apache_beam/dataframe/io_test.py

Line 131 in e8e272b

@pytest.mark.uses_pyarrow

[damccorm]

Each of these categories are not strictly dependency tests, they also test some other functionality which is not necessarily tested elsewhere. I agree we don't need the full matrix of all versions in our precommits, but I think it would be good to keep at least 1 version per type that also runs in the precommit. So for example, we could change this line to 2 lines:

preCommitPyCoverage.dependsOn "testPy38pyarrow-8"
postCommitPyDep.dependsOn "testPy38pyarrow-8"

but leave the rest of the pyarrow versions for the dependency task

[Abacn]

I thought these tests are also running on usual precommit, with dependency version unrestricted (see comment above). Let me open a draft PR to double check

[Abacn]

From test result of #30226 Only pyarrow and pandas tests were executed in other PreCommit (PreCommit dataframe), so these two tests can be fully moved to postcommit; other tests leave at least one version in PreCommit coverage

[Abacn]

Still takes 2h 0min. The very slow part is to install certain dependency:

Fri, 22 Mar 2024 17:01:15 GMT
py38-embeddings: install_deps> target/.tox-py38-embeddings/py38-embeddings/bin/python /runner/_work/beam/beam/sdks/python/test-suites/tox/py38/build/srcs/sdks/python/target/.tox-py38-embeddings/py38-embeddings/bin/pip install --retries 10 --pre sentence-transformers==2.2.2
Fri, 22 Mar 2024 17:31:16 GMT
.pkg_external: pip==24.0,setuptools==69.1.0,wheel==0.42.0
Fri, 22 Mar 2024 17:32:47 GMT
py38-pytorch-200: install_deps> target/.tox-py38-pytorch-200/py38-pytorch-200/bin/python /runner/_work/beam/beam/sdks/python/test-suites/tox/py38/build/srcs/sdks/python/target/.tox-py38-pytorch-200/py38-pytorch-200/bin/pip install --retries 10 --pre mpmath==1.3.0 'torch<2.1.0,>=2.0.0'
Fri, 22 Mar 2024 17:55:20 GMT
.pkg_external: freeze> python -m pip freeze --all
Fri, 22 Mar 2024 18:12:39 GMT
py38-transformers-430: install_deps> target/.tox-py38-transformers-430/py38-transformers-430/bin/python /runner/_work/beam/beam/sdks/python/test-suites/tox/py38/build/srcs/sdks/python/target/.tox-py38-transformers-430/py38-transformers-430/bin/pip install --retries 10 --pre tensorflow==2.12.0 'torch<1.14.0,>=1.9.0' 'transformers<4.31.0,>=4.30.0'
Fri, 22 Mar 2024 18:28:15 GMT
.pkg_external: freeze> python -m pip freeze --all

took 30 min to install sentence-transformers; 20 min to install torch and mpmath; another 10 min to install tensorflow and torch

[Abacn]

sentence-transformers has several very large dependencies:

• torch is 700 MB+
• nvidia-cudnn-cu12 are both 700MB+
• other nvidia dependencies are also several hundreds MBs

[tvalentyn]

Thanks for making this change. Do we still need to explicitly mention new suites anywhere under release/src/main/scripts or this is no longer necessary?

[Abacn]

It's not necessary. job_phrase for postcommit currently is not effective (#28909). I just copy-paste from existing yml which had this structure.