Add initial tests for diagnostic metrics

[paul-buerkner]

I have started writing tests for the diagnostics module, specicially the diagnostic metrics. I have a few questions for @LarsKue before I continue adding more tests.

• Is the way I wrote the tests generally in line with how bayesflow tests should look like?
• When I run these tests, it runs them twice, for batch_size 2 and 3. I don't want or need to batch_size to play any role here. It's not something I use in these tests nor do I know where it's coming from. How can I deactivate that the tests are run multiple times?
• Next on my list would be the diagnostic plots. Is there something I shall test about python plots beyond the fact that they run through?

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

All modified and coverable lines are covered by tests ✅

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Files with missing lines	Coverage Δ
bayesflow/utils/dict_utils.py	60.62% <100.00%> (+39.37%)	⬆️

... and 14 files with indirect coverage changes

[LarsKue]

Looks good. The tests and fixtures are generally idiomatic and in line with the rest of bayesflow. I would avoid testing against constant shapes, unless this is explicitly what you want. For example, should the shape of ce["values"] be (3,) or identical to some other shape, e.g. y.shape[1]?

As for disabling the parametrization over the batch size, I think we can turn autouse off on that fixture. If this causes other problems, I will have to make some other changes first, so it should be fine if these tests are run twice for now.

I do not think there is much we can/should do about testing plots. You could verify that the functions return the correct object. But in general, as long as the function runs, we can expect the plotting libraries to do the heavy lifting here.

[paul-buerkner]

Great, thank you!

[paul-buerkner]

I have now added tests for all metrics and most plots. @Kucharssim will add some tests for the model comparison plots soon in another PR. I will merge this PR as soon as all checks are passing.