[Python] Unneeded Abstract Method

I'm trying to write some abstract classes that can calculate measurements. Measurements can have one or multiple submeasures. Submeasures have a denominator, numerator, and a stratification. So the code looks something like this:

class Denominator(ABC):
"""
Doc String
"""

@abstractmethod
def _set_populace(self) -> None:
"""
Doc String
"""
pass

@abstractmethod
def _remove_exclusions(self) -> None:
"""
Doc String
"""
pass

class Submeasure(Denominator,Numerator,Stratification):
"""
Doc String
"""
# call denominator methods...
# call numerator methods...
# call stratification methods...


and then (concrete) Measurements can look like

class ASC(Measurement):
"""
Doc String
"""

def __init__(self,sub1_data:list[pd.DataFrame],sub2_data:list[pd.DataFrame] = None):
super().__init__("ASC")
self.__sub1__: Submeasure = _Sub_1(self.get_name() + "_sub_1",sub1_data)
self.__sub2__: Submeasure = _Sub_2(self.get_name() + "_sub_2",sub2_data)


Most submeasure 2s are subsets of submeasure 1s, so I can something like this do avoid doing excess calculations

def __get_sub2_subset(self,sub1_results:dict[strd.DataFrame]) -> tuple[pd.DataFrame,pd.DataFrame]:
"""
Doc String
"""
# sub 2's populace is a subset of sub 1's numerator,
# so all that's needed to do is filter sub 1's results where numerator == True
populace = self.__sub1__.get_name()
stratification = self.__sub1__.get_name() + '_stratification'

populace = sub1_results.get(populace)
populace = populace[populace['numerator']].copy()
populace = populace.drop('numerator',axis=1)

stratification = sub1_results.get(stratification)
stratification = stratification[stratification['patient_id'].isin(populace['patient_id'])].copy()
return populace,stratification


and then just pass that to self.__sub2

sub1_subset = self.__get_sub2_subset(sub1_results)
self.__sub2__._set_sub1_subset(sub1_subset)


But now sub2's denomiator method (remove_exclusions) doesn't need to be implemented and is just taking up space...

@override
def _remove_exclusions(self):
"""
Doc String
"""
# not sure what to do with this b/c sub2 piggy backs off of sub1's final data
pass


How am i supposed fix this?

I could just make another abstraction called Exclusions but then I'd need to somehow have some Submeasure(Denominator,Exclusions,Numerator,Stratification) and some Submeasure(Denominator,Numerator,Stratification)... but that doesn't seem right to me because how do i make sure that sub1 is the first Submeasure and that sub2 is the second Submeasure? would the code be able to figure it out by me implementing the Exclusion abstract methods?

It's been a while since I've done this type of coding so I know that this should be a common question, I just cant remember what it's called/where to look it up.

Thanks

Continue reading...