Use state sampler stub to defer metrics updates when DoFn#process is executed in subprocess.

[claudevdm]

---

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.

[tvalentyn]

It looks strange to a reader unfamiliar with this change why we are abstracting only one particular method of StateSampler. Furthermore, we are reasoning about the logic how to accumulate the metrics in the stub, which might go out-of sync.

I wonder whether we can replace the statesampler reference in the subprocess with a unittest.mock.MagicMock, and then replay the calls returned by the_mock.mock_calls() on the real reference without introducing another class. As long as the mock doesn't need to return any values, this should work. @robertwb do you think this would be too hacky? Are there some concerns if the replay happens only after the DoFn.process() call finishes?

[robertwb]

That's an interesting idea, however I'd rather avoid dependencies on test classes like MagicMock. (It may also have issues with the pickling required here). I do think if we're going to introduce an interface here, we should list all the relevant methods and have the stub explicitly raise an error for the unsupported ones to clarify that it's a partial implementation.

[tvalentyn]

@robertwb do we need to apply the interface onto the cython implementation as well in this case? Does cython support inheriting from a python ABC?

[tvalentyn]

Discussed with @robertwb offline - cdef Cython classes cannot inherit from python classes, there is a limited support for multiple inheritance but comes with a potential performance hit, and this is performance sensitive code; we can leave the cython codepath alone and add a comment about the existence of the StateSampler interfaces defined in python codebase but not use them in Cythonized codepath.

[robertwb]

That's an interesting idea, however I'd rather avoid dependencies on test classes like MagicMock. (It may also have issues with the pickling required here). I do think if we're going to introduce an interface here, we should list all the relevant methods and have the stub explicitly raise an error for the unsupported ones to clarify that it's a partial implementation.

[robertwb]

Any updates on this?

[claudevdm]

I will be able to pick this up again early next week. Is the consensus to go with the interface approach but list all the relevant method/raise an error if unimplemented?

[tvalentyn]

Is the consensus to go with the interface approach but list all the relevant method/raise an error if unimplemented?

sgtm

[claudevdm]

I've updated with a more complete interface. Can you please take a look @robertwb ?

[github-actions]

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

R: @damccorm for label python.

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).

[github-actions]

Reminder, please take a look at this pr: @damccorm

[github-actions]

Assigning new set of reviewers because Pr has gone too long without review. If you would like to opt out of this review, comment assign to next reviewer:

R: @liferoad for label python.

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)

[github-actions]

Reminder, please take a look at this pr: @liferoad

[github-actions]

Assigning new set of reviewers because Pr has gone too long without review. If you would like to opt out of this review, comment assign to next reviewer:

R: @damccorm for label python.

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)

[tvalentyn]

Overall lgtm, PTAL at lint issues. Thanks!

 class StubStateSampler(StateSamplerInterface):
ERROR: /runner/_work/beam/beam/sdks/python/test-suites/tox/pycommon/build/srcs/sdks/python/apache_beam/transforms/core.py Imports are incorrectly sorted.
--- /runner/_work/beam/beam/sdks/python/test-suites/tox/pycommon/build/srcs/sdks/python/apache_beam/transforms/core.py:before	2024-11-18 21:13:46.912794
+++ /runner/_work/beam/beam/sdks/python/test-suites/tox/pycommon/build/srcs/sdks/python/apache_beam/transforms/core.py:after	2024-11-18 21:23:41.718949
@@ -21,6 +21,7 @@