Moving transformer in TreatmentExpansionMixin

It seems that all of the classes the inherit from the TreatmentExpansionMixin so as to perform discrete treatment expansion, are doing the exact same transformer:

self._label_encoder = LabelEncoder()
self._one_hot_encoder = OneHotEncoder(categories='auto', sparse=False)
self.transformer = FunctionTransformer(
            func=(lambda T:
                  self._one_hot_encoder.transform(
                      self._label_encoder.transform(T).reshape(-1, 1))[:, 1:]),
            validate=False)

Shouldn't this piece of code move there? Or at least we can have a subclass called DiscreteTreatmentExpansionMixin, which is specific to this type of transformer and move this code there. It used to be that the OrthoForest was doing sth different, but this was a bug. Now it is doing exactly the same.