如何使用pytest-mock检查单元测试中是否调用了函数？

Question

我有一个单元测试，我想在其中检查函数是否被调用。如何使用pytest和pytest-mock库做到这一点？

例如，下面是一个单元测试test_hello.py。在这个测试中，我调用了函数my_function，并希望验证它是否使用给定的参数调用了hello。

def hello(name):
    return f'Hello {name}'

def my_function():
    hello('Sam')

def test_hello(mocker):
    mocker.patch('hello')
    my_function()
    hello.assert_called_once_with('Sam')

上面的代码返回以下错误：

target = 'hello'

    def _get_target(target):
        try:
>           target, attribute = target.rsplit('.', 1)
E           ValueError: not enough values to unpack (expected 2, got 1)

/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/unittest/mock.py:1393: ValueError

During handling of the above exception, another exception occurred:

mocker = <pytest_mock.MockFixture object at 0x109c5e978>

    def test_hello(mocker):
>       mocker.patch('hello')

test_hello.py:8: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/pytest_mock.py:156: in __call__
    return self._start_patch(self.mock_module.patch, *args, **kwargs)
/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/pytest_mock.py:134: in _start_patch
    p = mock_func(*args, **kwargs)
/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/unittest/mock.py:1544: in patch
    getter, attribute = _get_target(target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

target = 'hello'

    def _get_target(target):
        try:
            target, attribute = target.rsplit('.', 1)
        except (TypeError, ValueError):
            raise TypeError("Need a valid target to patch. You supplied: %r" %
>                           (target,))
E           TypeError: Need a valid target to patch. You supplied: 'hello'

/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/unittest/mock.py:1396: TypeError

Answer 1

通过以下方式解决错误

为mocked.patch分配mocked_hello

将side_effect分配给模拟函数

def bonjour(name):
    return 'bonjour {}'.format(name)


def hello(name):
    return 'Hello {}'.format(name)


def my_function():
   return hello('Sam')


def test_hellow_differnt_from_module(mocker):
    # mocked func with `test_hello.py` as module name
    mocked_hello = mocker.patch('test_hello.hello')
    # assign side_effect to mocked func
    mocked_hello.side_effect = bonjour
    # the mocked func return_value changed by side_effect
    assert mocked_hello('Sam') == 'bonjour Sam'
    # the mocked func called with Sam, but with different return value
    mocked_hello.assert_called_with('Sam')

调用一个实际函数my_function()并验证它是否调用了hello

def test_my_function(mocker):
    mocker.patch('test_hello.hello', side_effect=bonjour)
    mf = my_function()
    hello.assert_called_with('Sam')
    assert mf == 'bonjour Sam'

Answer 2

我认为只需对OP的代码进行最小程度的更改，就可以完全限定被模拟函数的名称。例如，如果代码在一个名为test.py的文件中

import pytest

def hello(name):
    return f'Hello {name}'

def my_function():
    hello('Sam')

def test_hello(mocker):
    mocker.patch('test.hello')
    my_function()
    hello.assert_called_once_with('Sam')

(我假设test.py位于sys.path的某个地方。)

但是，我认为最好将测试函数更改为：

def test_hello(mocker):
    mocked_func = mocker.patch('test.hello')
    my_function()
    mocked_func.assert_called_once_with('Sam')

这种形式对我来说似乎更清晰，不太可能在更复杂的代码中引发AttributeError。