在 Python 中模拟类方法并更改某些对象属性 [英] Mocking a class method and changing some object attributes in Python

问题描述

我是 Python 中的新手.我想知道如何在用另一个类方法进行测试时替换(模拟)一个类方法，知道原来的只是改变了 self 的一些属性而不返回任何值.例如:

I am new to mock in Python. I want to know how to replace (mock) a class method while testing with another one, knowing that the original just changes some attributes of self without returning any value. For example:

def some_method(self):   
    self.x = 4   
    self.y = 6   

所以在这里我不能只更改模拟的 return_value.我试图定义一个新函数(应该替换原来的函数)并将其作为 side_effect 提供给模拟.但是我怎样才能让模拟函数改变类中对象的属性.这是我的代码:

So here I can't just change the return_value of the mock. I tried to define a new function (that should replace the original) and give it as side_effect to the mock. But how can I make the mocking function change attributes of the object in the class. Here is my code:

@patch('path.myClass.some_method')
def test_this(self,someMethod):

    def replacer(self):
        self.x = 5
        self.y = 16

some_method.side_effect = replacer

那么 Python 现在是如何理解替换器的 self 参数的呢?是测试类的self，还是被测类的对象self?

So how does Python now understands the self argument of replacer? Is that the self of the test class, or the self as the object of the class under test?

推荐答案

如果我不明白您要做什么，请提前道歉，但我认为这可能有效:

Apologies in advance if I don't understand what you are trying to do, but I think this might work:

import unittest
from unittest.mock import patch

class MyClass:

    def __init__(self):
        self.x = 0
        self.y = 0

    def some_method(self):   
        self.x = 4   
        self.y = 6    

class OtherClass:

    def other_method(self):
        self.x = 5
        self.y = 16

class MyTestClass(unittest.TestCase):

    @patch('__main__.MyClass.some_method', new=OtherClass.other_method)
    def test_patched(self):
        a = MyClass()
        a.some_method()
        self.assertEqual(a.x, 5)
        self.assertEqual(a.y, 16)

    def test_not_patched(self):
        a = MyClass()
        a.some_method()
        self.assertEqual(a.x, 4)
        self.assertEqual(a.y, 6)

if __name__ == "__main__":
    unittest.main()

这在打补丁的时候用 other_method() 替换了 some_method()，它为属性 x、y 设置了不同的值，并且在运行测试时，它给出了结果:

This replaces some_method() with other_method() when patched, which sets different values for attributes x, y, and when the test is run, it gives the results:

..
----------------------------------------------------------------------
Ran 2 tests in 0.020s

OK

回答有关如何在不模拟类的情况下在测试函数中执行的问题...

to answer question about how to do inside the test function without mocking a class...

def test_inside_patch(self):
    def othermethod(self):
        self.x = 5
        self.y = 16
    patcher = patch('__main__.MyClass.some_method', new=othermethod)
    patcher.start()
    a = MyClass()
    a.some_method()
    self.assertEqual(a.x, 5)
    self.assertEqual(a.y, 16) 
    patcher.stop()

确保在补丁程序上调用 start() 和 stop() ，否则您可能会遇到补丁处于活动状态而您不希望它处于活动状态的情况.请注意，在测试代码函数中定义模拟函数，我没有使用补丁作为装饰器，因为模拟函数必须在补丁中使用'new'关键字之前定义.如果你想使用补丁作为装饰器，你必须在补丁之前的某个地方定义模拟函数，在 MyTestClass 中定义它也可以，但似乎你真的希望在测试函数代码中定义模拟函数.

Make sure you call start() and stop() on the patcher otherwise you can get into a situation where the patch is active and you don't want it to be. Note that to define the mock function inside the test code function, I didn't use patch as a decorator, because the mock function has to be defined before using the 'new' keyword in patch. If you want to use patch as a decorator you have to define the mock function someplace before the patch, defining it inside of MyTestClass also works, but it seems you really want to have the mock function defined inside your test function code.

添加了我看到的 4 种方法的摘要...

added summary of 4 ways I see to do this...

# first way uses a class outside MyTest class
class OtherClass:
    def other_method(self):
        ...

class MyTest(unittest.TestCase):

    @patch('path_to_MyClass.some_method', new=OtherClass.other_method)
    def test_1(self)
        ...

    # 2nd way uses class defined inside test class    
    class MyOtherClass:
        def other_method(self):
            ...
    @patch('path_to_MyClass.some_method', new=MyOtherClass.other_method)    
    def test_2(self):
        ...

    # 3rd way uses function defined inside test class but before patch decorator 
    def another_method(self):
        ...
    @patch('path_to_MyClass.some_method', new=another_method)    
    def test_3(self):
        ...

    # 4th way uses function defined inside test function but without a decorator
    def test_4(self):
        def yet_another_method(self):
            ...
        patcher = patch('path_to_MyClass.some_method', new=yet_another_method)
        patcher.start()
        ...
        patcher.stop()

这些都没有使用副作用，但它们都解决了模拟类方法和更改某些属性的问题.您选择哪一种取决于应用程序.

None of these uses a side_effect, but they all solve the problem of mocking a class method and changing some attributes. Which one you choose depends on the application.

这篇关于在 Python 中模拟类方法并更改某些对象属性的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！