Fix abstract RandomRegularDepthOneFit and test_random_regular.py not running #34
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
conradhaupt
requested review from
AlbertoBaiardi,
ElePT and
eggerdj
as code owners
Collaborator
|
Thanks a lot for taking a look into the CI, as you can see, it's a bit of a pandora box... |
conradhaupt
changed the title
Contributor
Author
|
I've updated the PR with fixes for |
stestr may be failing on CI because of DeprecationWarnings raised by tarfile on python3.12. Juliacall will download Julia as a tar file on Ubuntu runners, extract the file, and install Julia. Other platform runners get different files. With python3.12, tarfile raised a DeprecationWarning about a change in the interface. It appears that stestr will fail on test discovery if a warning is raised. Therefore, only the Python3.12 tests appeared to fail on Linux devices where Julia was not already installed. Locally, an installed version of Julia meant that tests passed as juliacall didn't have to install it from a Tar file.
It appears stestr does not pass PYTHONWARNINGS to subunit calls to tests. This means that warnings are still raised by tests, which fail stestr. This is useful if we're testing for warnings; however, installing Julia appears to raise some deprecation warnings which we do not want to fail on. Here we import from *.pauli_propagation to force Julia installation and print a message if this succeeds. Failure indicates that Julia installation failed too.
ElePT
approved these changes
Nov 5, 2025
Collaborator
ElePT
left a comment
ElePT
left a comment
There was a problem hiding this comment.
Thanks a lot @conradhaupt, all changes look good, and your effort in fixing the CI is really appreciated.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Summary
Tests in
test/training/models/were not running with stestr (CI and Tox) as it is missing an__init__.pyfile (see #33). Running these tests showed thatRandomRegularDepthOneFitwas abstract, and thus tests failed. This PR also fixes this by implementingRandomRegularDepthOneFit.minimization(), raising an error as minimization is not defined for the trainer (see below for details).Details and comments
stestr requires the test folder to be importable as a module, and thus
__init__.pyfiles are needed in every folder withtest_*.pyfiles. An__init__.pyfile was missing intest/training/models/up until now and thustest_random_regular_fit.pywas never run with CI unit-tests. As an example, see the latest commit as of this PR (ab7eaea) and associated Linux tests. We would expect there to be a line printed by stestr containing the stringtest.training.models.test_random_regular_fit.TestRandomRegularModel, but there is not, showing that these tests were never run.The tests for
RandomRegularDepthOneFitwere failing because it was missingminimization()and was abstract. The fix in this PR defines the method but with a raised error, asminimization()is not defined for this type of trainer. The class maps a problem instance to the closest known optimised parameters for a regular graph, based on average degree. There is no minimization carried out in this scenario.Version updated
Not currently necessary.