Python在类定义之外覆盖__str__

Question

Say I have a module that I've imported that works fine but doesn't provide a good __str__ method and I want to add one.

In some hybrid language:

from foo import Foo

Foo.prototype._str__ = function myNIcePrintFunction(){

}

How do I do this please in Python please?

Answer 1

You can first define a function, and then monkey-patch it.

Like:

def foostr(self):
    return 'A foo str'

Foo.__str__ = foostr

Or with a lambda expression :

Foo.__str__ = lambda self: 'Another foo str'

There is however a small difference in the fact that this method will have a different Foo.__str__.__name__ , etc. But the str(..) builtin function will work as expected. As @chepner says, we can use the wrap function from functools to preserve most meta-data:

from functools import wraps

def foostr(self):
    return 'A foo str'

Foo.__str__ = wraps(Foo.__str__)(foostr)

But then Foo.__str__.__qualname__ will for example result in 'superfoo.__str__' (with superfoo the first class in the method resolution order (MRO) of Foo that implemented __str__ ).

For example:

>>> class Foo:
...   pass
... 
>>> str(Foo())
'<__main__.Foo object at 0x7f807046e668>'
>>> Foo.__str__ = lambda self: 'another Foo str'
>>> str(Foo())
'another Foo str'

If you have access to the implementation of the Foo class, it is however better to just define a __str__ function in the class, and not monkey patch it later in the process.

Answer 2

You can override any class function like this, while maintaining the original functionalities:

class a():
    pass

def monkey(fn):
    def newfn(self,*args, **kw):
        print('hijacked!')
        return fn(self,*args, **kw)
    return newfn

a.__str__ = monkey(a.__str__)


b = a()

print(str(b))

>>> hijacked!                                                                                                                                                                            
>>> <__main__.a object at 0x7f61f19faa58>