使用外部函数更新内部python类参数

Question

I have an external function my_update_function which is not part of my class my_class (currently for design reasons). Now, I would like to update the internal parameters of my_class ie self.x and self.y using the external function my_update_function .

Please see this MWE, which describes my questions:

def my_update_function(arg1,args2):
    """
    Updates arg1 and arg2.
    """
    return arg1*arg1,arg2*arg2

class my_class(object):
    def __init__(self):
        self.x = 2
        self.y = 3
    def simulate(self):
        my_update_function(self.x,self.y)

However, I am wondering if there is some way (either by class re-design or otherwise) that I can update self.x and self.y without having to save the output of the function my_update_function as parameter values (and without having to include my_update_function as part of my class my_class ) ie I would rather not write this:

def simulate(self):
    self.x, self.y = my_update_function(self.x,self.y)

but would want it to be like this:

def simulate(self):
    my_update_function(self.x,self.y)

so that my_update_function communicates the updated parameter values to the class without having to explicitly store them as outputs.

Please let me know if it is not clear what I mean, and I shall update the question.

Answer 1

def my_update_function(obj):
    obj.x, obj.y = obj.x * obj.x, obj.y * obj.y

then in your class:

class my_class(object):
    def __init__(self):
        self.x = 2
        self.y = 3
    def simulate(self):
        my_update_function(self)

Answer 2

Since python does not have pointers, a common trick is to use lists, which are passed by reference, and hence their contents is mutable :

>>> def my_update_function(l):
...     l[0], l[1] = l[0] * l[0], l[1] * l[1]
...
>>> l = [2, 3]
>>> my_update_function(l)
>>> l
[4, 9]

I'm adding this as a second answer to not create confusion.

Using an object instead of a list works the same way (both are passed by reference), and you already have an object at hand, so you better off using the my_class object.