I have a superclass with several subclasses. Each subclass overrides the "solve" method with some a different algorithm for estimating the solution to a particular problem. I'd like to efficiently create an instance of each subclass, store those instances in a list, and then take the sum (or average) of the estimates that the subclasses give:

class Technique():

    def __init__(self):
        self.weight = 1

    def solve(self, problem):

class Technique1(Technique):

    def __init__(self):

    def solve(self,problem):
        return 1

class Technique2(Technique):

    def __init__(self):

    def solve(self,problem):
        return 6

 class Technique3(Technique):

    def __init__(self):

    def solve(self,problem):
        return -13

testCase = 3
myTechniques = []
t1 = Technique1()
t2 = Technique1()
t3 = Technique1()

print(sum([t.solve(testCase) for t in myTechniques])

The code at the end feels very clunky, particularly as the number of subclasses increases. Is there a way to more efficiently create all of the instances?


You might try this:

myTechniques = [cls() for cls in Technique.__subclasses__()]

Note that it does not return indirect subclasses, only immediate ones. You might also consider what happens if some subclass of Technique is itself abstract. Speaking of which, you should be using ABCMeta and abstractmethod.

  • \$\begingroup\$ That's exactly what I was looking for. Thanks! \$\endgroup\$
    – Max Rosett
    Jul 13 '15 at 23:55
  • \$\begingroup\$ If this question were asked again today, I'd suggest __init_subclass__ as the cleanest answer. \$\endgroup\$
    – o11c
    Dec 30 '17 at 4:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.