Question

有没有办法为pytest.mark.parametrize()指定一个可调用的pytest参数，以便只有在选择运行测试时动态生成参数？

为了生成参数，我要执行一些昂贵的操作，如果选择运行测试，我只想执行它们。

如，

import pytest

def my_callable():
    # do expensive operations here
    return [(1, 2), (3, 6)]


# I want my_callable to be called only if test_something
# has been selected to be run
@pytest.mark.parametrize("my_parm_1,my_parm_2", my_callable)
def test_something(my_parm_1, my_parm_2):
    assert my_parm_1 * 2 == my_parm_2

Answer 1

我认为这样做你想要的 - 昂贵的计算是在夹具内部，只有在调用测试并且昂贵的计算只进行一次时才会调用：

class TestSomething:

    _result = None

    @pytest.fixture()
    def my_callable(self):
        if TestSomething._result is None:
            # do expensive operations here
            TestSomething._result = [(1, 2), (3, 6)]

        def _my_callable(run_number):
            return TestSomething._result[run_number]
        return _my_callable

    @pytest.mark.parametrize("run_number", [0, 1])
    def test_something(self, run_number, my_callable):
        my_param_1, my_param_2 = my_callable(run_number)
        assert my_param_1 * 2 == my_param_2

Answer 2

您可以创建代理，而不是直接在mycallable中使用@pytest.mark.parametrize：

def my_callable():
    # do expensive operations here
    return [(1, 2), (3, 6)]

expensive_params = [paramset for paramset in my_callable()]

@pytest.mark.parametrize("my_parm_1,my_parm_2", expensive_params)
...

如何使用callable作为pytest parametrize（）参数？

2 个答案: