Create unit test for PredefinedModels.php utility

I was wondering if we should create an issue to update the existing test implementation to use attributes.

The MR has a lot of validation and test problems.

Hey

I would love to help here! Could I assign myself this and freshly work on the Unit Test?

@hrishikesh-dalal that sounds great! Go ahead and assign yourself and let us know if you have any questions.

Sure thanks! I will get it completed by today itself

@hrishikesh-dalal

I checked the MR you created: https://git.drupalcode.org/project/ai/-/merge_requests/1238

Firstly, thank you for the work so far, it’s really appreciated.

I haven’t done a deep code review yet, but I have some initial thoughts comparing the two merge requests:

- MR339 has broader test coverage and validates against the real YAML files shipped with the module. This is useful because it ensures the actual predefined model definitions in the module remain correct and don’t break over time.

- MR1238 takes a more isolated unit-testing approach by using a virtual filesystem. This makes the test more portable, deterministic, and aligned with unit-testing best practices, since it doesn’t depend on the module’s physical file structure.

My initial feeling is that MR 1238 follows better unit-testing practices, while MR 339 provides stronger real-world validation and broader scenario coverage. Ideally, a combination of both approaches would give the best overall test reliability and coverage.

See if it’s possible to achieve that balance.

Regarding the broken PHPStan checks, this is a known issue, and other MRs targeting the 2.x branch also have it. A good way to confirm is to look at other MRs and issues targeting the same branch.

For example, the Needs review queue for 2.x:
https://www.drupal.org/project/issues/search/ai?status%5B%5D=8&version%5...

You’ll notice that several 2.x MRs are currently affected. I haven’t investigated those errors in detail, but there are likely existing issues tracking them.

Appreciate the feedback, @abhisekmazumdar. I see the point regarding the coverage in MR 339. I'll bridge the gap by bringing those scenarios into my isolated setup in MR 1238.

Thanks for clarifying the PHPStan issue as well—saved me some unnecessary debugging. I'll update the MR shortly.

Thank you!

I reviewed MR !1238 as the active candidate for this issue and did not find any blocking issues.

Summary

- Scope is tight and matches the issue (single new unit test file).
- The test coverage is solid for the targeted utility behavior.
- Latest MR pipeline is successful overall.

Blockers

- None.

Optional recommendation

- Add explicit negative-path wrapper tests for:
- getPredefinedModelsAsOptions(NULL|invalid)
- getPredefinedModel(invalid, ...)

marking as rtbc