Simon Blanke

> However, if we do not change pyproject.toml, I believe nothing would change from the user perspective? Only the imports are more restricted, and no scikit-learn imports happen on module...

added a test 'test_check_objects_attr_name_explicit'. The asserts are not that useful, but it at least covers some additional code if `attr_name` is not `None`

added another test 'test_check_objects_attr_name_custom_tag'. It sets a custom tag via `set_tags`. I just realized, that I used a private method `_check_objects` for the tests. Normally, that should be a no-go,...

@fkiraly I improved the test quality by triggering an error in each test and catching the error type.

> I think the coupling to `self` should happen at least one layer higher, i.e., not in the checker function, but where it is being called. > > Otherwise we...

> are there any other tags that get read in there? No, I only see `attr_name = self.get_tag("named_object_parameters")` here. For the decoupling I would like to separate the code into...

okay there is a lot going on here (or maybe this is just in my head). First I added [this note](https://github.com/sktime/skbase/pull/466/files#diff-e6197ccf8ea00f7080233f04b26c12e30a6e790958541927d724179387010da5R390), because `_check_objects` does not really make sense to me...

There is a linter error in a file I did not change: "skbase/tests/test_base.py:389:5: B043 Do not call delattr with a constant attribute value, it is not any safer than normal...

> There is a linter error in a file I did not change: "skbase/tests/test_base.py:389:5: B043 Do not call delattr with a constant attribute value, it is not any safer than...

The idea behind the sleep parameter is indeed for testing/benchmarking. Some mathematical test-functions have interesting shapes, but are evaluated very quickly. Adding a sleep parameter enables to imitate the computational...