脚本之家,脚本语言编程技术及教程分享平台!
分类导航

Python|VBS|Ruby|Lua|perl|VBA|Golang|PowerShell|Erlang|autoit|Dos|bat|shell|

服务器之家 - 脚本之家 - Python - pytest fixtures函数及测试函数的参数化解读

pytest fixtures函数及测试函数的参数化解读

2023-02-23 12:00把苹果咬哭的测试笔记 Python

这篇文章主要介绍了pytest fixtures函数及测试函数的参数化解读,有需要的朋友可以借鉴参考下,希望能够有所帮助,祝大家多多进步,早日升职加薪

pytest fixtures测试函数参数化

Pytest会在以下几个级别启用测试参数化:

  • pytest.fixture(),可以对fixture函数进行参数化。
  • @pytest.mark.parametrize,可以在测试函数或类中定义多组参数和fixture。
  • pytest_generate_tests,可以自定义参数化方案或扩展。

一、@pytest.mark.parametrize:参数化测试函数

1. 常规用法

对测试函数的参数进行参数化,直接使用内置的装饰器pytest.mark.parameterized即可。

?
1
2
3
4
import pytest
@pytest.mark.parametrize("test_input,expected", [("3+5", 8), ("2+4", 6), ("6*9", 42)])
def test_eval(test_input, expected):
    assert eval(test_input) == expected

从代码里可以看出,在装饰器里定义了三个不同的元组。我们把("test_input,expected", [("3+5", 8), ("2+4", 6), ("6*9", 42)])
拆开看:

  • "test_input,expected":这个字符串里定义了2个参数,test_input和expected。
  • ("3+5", 8), ("2+4", 6), ("6*9", 42):这里3个元组,没个元组里有2个元素,依次序分别对应test_input和expected。
  • 3个元组外层的[]:列表里就是参数化具体的传参了,因为里面传了3个不同的元组,所以测试函数test_eval会分别执行3次。
?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
============================= test session starts =============================
platform win32 -- Python 3.9.4, pytest-6.2.3, py-1.10.0, pluggy-0.13.1
rootdir: D:\PycharmProjects\wms-api\interface, configfile: pytest.inicollected 3 items
test_module1.py ..F
demo\test_module1.py:3 (test_eval[6*9-42])
54 != 42
Expected :42
Actual   :54
 <Click to see difference>
test_input = '6*9', expected = 42
    @pytest.mark.parametrize("test_input,expected", [("3+5", 8), ("2+4", 6), ("6*9", 42)])
    def test_eval(test_input, expected):
>       assert eval(test_input) == expected
E       AssertionError: assert 54 == 42
E        +  where 54 = eval('6*9')
test_module1.py:6: AssertionError

运行结果可以看到最后一次失败了,因为第三次运行测试函数取的参数是 ("6*9", 42),54不等于42,所以断言失败。

2. 在参数化中标记单个测试实例

在参数化中标记单个测试实例,比如之前提到过的mark.xfail,这个可以标记测试函数为失败。那么在参数化中,如果想让其中的某个参数运行
的时候测试失败,就可以这样用:

?
1
2
3
4
5
6
7
import pytest
@pytest.mark.parametrize(
    "test_input,expected",
    [("3+5", 8), ("2+4", 6), pytest.param("6*9", 42, marks=pytest.mark.xfail)],
)
def test_eval(test_input, expected):
    assert eval(test_input) == expected

运行一下:

?
1
2
test_module1.py                                                       [100%]
======================== 2 passed, 1 xfailed in 0.05s =========================..x

3. 多个参数化组合,笛卡尔积

如果在测试函数上加了多个参数化装饰器,那么得到的参数组合是一个笛卡尔积:

?
1
2
3
4
5
6
import pytest
@pytest.mark.parametrize("x", [0, 1])
@pytest.mark.parametrize("y", [2, 3])
def test_foo(x, y):
    print("\nx:", x)
    print("y:", y)

应该会组合成4组数据x=0/y=2, x=1/y=2, x=0/y=3, 和x=1/y=3,测试函数执行4次:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
test_module1.py .
x: 0
y: 2
.
x: 1
y: 2
.
x: 0
y: 3
.
x: 1
y: 3
                                                     [100%]
============================== 4 passed in 0.01s ==============================

二、用钩子函数pytest_generate_tests example拓展

如果有些场景需要动态的确定参数或者fixture的使用范围,那么可以使用pytest_generate_tests这个钩子函数,该函数会在收集测试函数时候被调用。

通过传入的metafunc对象,可以检查请求测试函数的上下文,还可以进一步的调用metafunc.parameterize()来实现参数化。

举例,有个测试函数需要接受输入的字符串作为参数,而且通过pytest命令行获取到,那么就要编写一个获取参数的fixture函数来给测试函数调用。

?
1
2
3
# content of test_strings.py
def test_valid_string(stringinput):
    assert stringinput.isalpha()

新建conftest.py文件,fixture函数写在这里:

?
1
2
3
4
5
6
7
8
9
10
11
# content of conftest.py
def pytest_addoption(parser):
    parser.addoption(
        "--stringinput",
        action="append",
        default=[],
        help="list of stringinputs to pass to test functions",
    )
def pytest_generate_tests(metafunc):
    if "stringinput" in metafunc.fixturenames:
        metafunc.parametrize("stringinput", metafunc.config.getoption("stringinput"))

现在用命令行方式来运行这个测试函数:

pytest -q --stringinput="hello" --stringinput="world" test_strings.py

会运行2次。

?
1
2
3
D:\PycharmProjects\wms-api\interface\demo>pytest -q --stringinput="hello" --stringinput="world" test_strings.py
..                                                                                                                                                                     [100%]
2 passed in 0.01s

再换个输入参数,让测试函数失败:

pytest -q --stringinput="!" test_strings.py

?
1
FAILED test_strings.py::test_valid_string[!] - AssertionError: assert False1 failed in 0.04s

如果没有字符串输入,那么测试函数它将被跳过。因为metafunc.parameterize()被调用时&#xff0c;传过去的是一个列表:

pytest -q -rs test_strings.py

?
1
2
SKIPPED [1] test_strings.py: got empty parameter set ['stringinput'], function test_valid_string at $REGENDOC_TMPDIR/test_strings.py:2
1 skipped in 0.12s

注意,在调用metafunc时, 如果使用不同的参数集进行多次参数化,这些参数集上的所有参数名称都不能重复,否则将会报错。

总结

文中讲到的3种用法,实际应用中第一种最常见,第二种次之,至于第三种,可以作为一些用法启发,更多关于pytest fixtures测试函数参数化的资料请关注服务器之家其它相关文章!

原文链接:https://blog.csdn.net/wessonlan/article/details/124813092

延伸 · 阅读

精彩推荐