如何在一处设置带Mock属性的类补丁，避免重复调用助手函数？

需要测试ExampleClass类，该类初始化时会创建SubComponent实例，其example_function_double_value方法依赖SubComponent的value属性：

class ExampleClass:
    def __init__(self, specific_args):
        # Do something with specific args
        self.specific_args = specific_args
        self.sub_component = SubComponent(...)

    def example_function_double_value(self) -> float:
        return SomeFunction(self.specific_args, self.sub_component.value)

当前单元测试通过@mock.patch('SubComponent_Path.SubComponent')替换SubComponent，并通过助手函数set_patch设置其value为3，但每个测试用例都要手动调用该助手函数：

import unittest
from unittest import mock

def set_patch(patch: mock.Mock):
    patch_return_value = mock.Mock()
    patch_return_value.value = 3
    patch.return_value = patch_return_value

@mock.patch('SubComponent_Path.SubComponent')
class ExampleUnitTest(unittest.TestCase):

    def test_case1(self, patch):
        set_patch(patch)
        specific_args1 = # Generate specific arguments for this test case
        test_class = ExampleClass(specific_args1)

        # Some assertion where I expect `subcomponent.value to be 3`

    def test_case2(self, patch):
        set_patch(patch)
        specific_args2 = # Generate specific arguments for this test case
        test_class = ExampleClass(specific_args2)

        # Some other assertion where I expect `subcomponent.value to be 3`

有没有更优雅的实现方式，不用在每个测试用例开头都调用助手函数？

优雅的实现方式 #

方法1：用测试类的setUp统一处理 #

利用unittest的setUp方法，在每个测试用例执行前自动完成mock的初始化，省去重复调用助手函数的步骤：

import unittest
from unittest import mock

@mock.patch('SubComponent_Path.SubComponent')
class ExampleUnitTest(unittest.TestCase):
    def setUp(self):
        # 创建带value属性的mock对象
        mock_sub = mock.Mock()
        mock_sub.value = 3
        # 给patch设置返回值
        self.sub_patch.return_value = mock_sub

    def test_case1(self, patch):
        self.sub_patch = patch  # 把patch参数绑定到实例属性，供setUp使用
        specific_args1 = # 生成当前测试的参数
        test_class = ExampleClass(specific_args1)
        # 执行你的断言逻辑

    def test_case2(self, patch):
        self.sub_patch = patch
        specific_args2 = # 生成当前测试的参数
        test_class = ExampleClass(specific_args2)
        # 执行你的断言逻辑

方法2：提前配置patch的return_value #

这是最简洁的方案，直接在@mock.patch里预先设置好带value属性的mock对象，所有测试用例自动复用这个配置：

import unittest
from unittest import mock

# 预先创建好符合要求的mock对象
mock_sub_component = mock.Mock()
mock_sub_component.value = 3

# 直接把这个mock对象设为patch的返回值
@mock.patch('SubComponent_Path.SubComponent', return_value=mock_sub_component)
class ExampleUnitTest(unittest.TestCase):
    def test_case1(self, patch):
        specific_args1 = # 生成参数
        test_class = ExampleClass(specific_args1)
        # 断言逻辑

    def test_case2(self, patch):
        specific_args2 = # 生成参数
        test_class = ExampleClass(specific_args2)
        # 断言逻辑

如果后续有个别测试需要修改value，直接在对应测试方法里重新赋值mock_sub_component.value = 新数值即可，完全不影响其他用例。

方法3：自定义装饰器封装逻辑 #

如果需要更灵活的复用，可以把mock配置逻辑封装成自定义装饰器，替代原生的@mock.patch：

import unittest
from unittest import mock

def patch_sub_component():
    def decorator(test_cls):
        # 先应用原生patch
        patched_cls = mock.patch('SubComponent_Path.SubComponent')(test_cls)
        # 重写setUp方法，加入mock配置
        original_setUp = patched_cls.setUp
        def new_setUp(self, *args, **kwargs):
            original_setUp(self, *args, **kwargs)
            mock_sub = mock.Mock()
            mock_sub.value = 3
            self.sub_patch.return_value = mock_sub
        patched_cls.setUp = new_setUp
        # 把patch参数绑定到实例属性
        for name, method in patched_cls.__dict__.items():
            if name.startswith('test_') and callable(method):
                def wrapped(self, patch, *args, **kwargs):
                    self.sub_patch = patch
                    return method(self, patch, *args, **kwargs)
                setattr(patched_cls, name, wrapped)
        return patched_cls
    return decorator

# 使用自定义装饰器
@patch_sub_component()
class ExampleUnitTest(unittest.TestCase):
    def test_case1(self, patch):
        specific_args1 = # 生成参数
        test_class = ExampleClass(specific_args1)
        # 断言逻辑

    def test_case2(self, patch):
        specific_args2 = # 生成参数
        test_class = ExampleClass(specific_args2)
        # 断言逻辑

内容的提问来源于stack exchange，提问作者seedship