Using reflection to peep into a module and its classes

Question

I've got a problem which is a bit difficult to explain. I have a module which consists of more than one class: someModule.py

#imports over here class Default(Base): def __init__(self): a = Rectangle() #all these guys derive from Shape b = Circle() c = Sphere() class Foo: #members over here #other classes/functions/whatever we can define here, except the boiler plate code to check __main__

What i want to do is create an object of the class which derives from a specific base class (eg. Base) at run time and manipulate those data members which derive from another specific base class (eg. Shape). Meaning i want to write such a script that takes module name and performs the above task. Any ideas how i can do this using inspect or something else? I've taken a look at inspect but didn't quite find the methods that should get the job done. I may be missing something.

Rik Poggi · Accepted Answer · 2012-02-05 08:37:26Z

There is no way to know what's inside __init__ before the creation of the instance.

You can only check them after, and one way to do it is with vars():

defy = Default() for name,value in vars(defy).items(): if isinstance(value, Shape): # manipulate

To do the above on all the classes in someModule.py that are also subclasses of Base:

import someModule instances = [] for cls_name,cls in vars(someModule): if issubclass(cls, Base): obj = cls() for name,value in vars(cls).items(): if isinstance(value, Shape): # manipulate instances.append(obj)

Instead, if you want to manipulate which Shape subclass is going to be instanciated, you'll have to make them class attributes, example:

class Default(Base): default_shapes = [Rectangle, Circle, Sphere] def __init__(self): self.shapes = [shape() for shape in self.__class__.default_shapes]

voithos · Accepted Answer · 2012-02-05 06:32:59Z

Here's one way: make sure that your base classes (e.g. Base, Shape) are new-style classes (they inherit from object). By doing so, you can use the built-in __subclasses__() function to get a list of all derived classes. Then you can simply create an object and go through the fields, comparing them with the classes that you're interested in.

baseDerived = Base.__subclasses__() shapeDerived = Shape.__subclasses__() # Iterate through derived classes for derivedClass in baseDerived: derivedObject = derivedClass() # Iterate through data attributes for attr, dataMember in derivedObject.__dict__.iteritems(): if dataMember.__class__ in shapeDerived: # Do stuff with the data member here

Collectives™ on Stack Overflow

Using reflection to peep into a module and its classes

2 Answers 2

Comments

Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Related