wilsbj

I will concede that the test could be clearer, but I would argue that the test is working as intended. The test checks that all sktime base classes appear in...

Perhaps a simpler test would be to traverse down the `__bases__` of the class and assert that the last element in this list at each level is a subclass of...

We are testing that the mixin classes appear higher in the MRO. We do this by removing the super classes of a test base class from the MRO and checking...

> the condition wir are testing is: for all mixins, there is at least one base class at a higher (larger mro index) position from which all classes inheriting from...

The implemented test: For all objects (including estimators), we check that any sktime base classes that appear in the MRO appear last, except for the super classes of these base...

The simplest sktime object and it's MRO would look like this: ```python from sktime.base import BaseObject class MyObject(BaseObject): pass MyObject.mro() # [__main__.MyObject, # sub class of BaseObject # sktime.base._base.BaseObject, #...

Thanks for working on this @Ankush1oo8. In addition to comments by @fkiraly, I would like the test to set another variable in `__init__` that is derived from `a`. E.g. `self.b_...

Thanks @Ankush1oo8. Does the test still pass if you remove the hardcoded definition for `reset`? The expectation is that the definition of `reset` on `BaseObject` will already perform this update...

A reasonable question. Having a way to set config on nested objects seems to be needed. I like `composite.set_config(comp2__foo="bar")` as an interface. A related question, should set_config have a recursive...

I initially though `deep` would be appropriate here, but `deep` in `set_params` only intends to set a given param in one place, while this idea is to set the same...