Pytest插件pytest-order指定用例顺序

安装

 pip install pytest-order

• 注意不是pytest-ordering
• 说起来这里有个故事

关于pytest-ordering和pytest-order

https://github.com/ftobia/pytest-ordering

https://github.com/pytest-dev/pytest-order

• 在写这个文章之前，我用的一直是pytest-ordering

• 但我在安装pytest-ordering的时候一直看到有pytest-order

• 今天特意看了下，好家伙

  • GITHUB上写了这么一句:pytest-ordering is no longer maintained, please use https://pypi.org/project/pytest-order/
  • pytest-order is a fork of pytest-ordering that provides additional features like ordering relative to other tests.
  • pytest-order works with Python 3.6 - 3.10, with pytest versions >= 5.0.0 for all versions except Python 3.10, and for pytest >= 6.2.4 for Python 3.10. pytest-order runs on Linux, macOS and Windows.

• 所以pytest-ordering在当前的pytest及python版本中可能会出现问题，而pytest-order是同步更新的，你可以放心食用

• 那么问题来了，pytest-order怎么用呢？看官方的示例跟pytest-ordering还是有区别的（如果没有区别，装了2个的话你可能会分不清哪个调用的，当然我们不推荐你用2个）

  • 虽然用起来很简单，但的确蛮细节的，网上你看到的基本都是pytest-ordering的用法，然它在3年前就停止更新了
  • 我把它的git@github.com:pytest-dev/pytest-order.git，下下来，它写了很多的example，嗯~

根据索引排序

• 其实就一个数字

• 也可以是与之对应的特定字符，如first等

  字符	数字index
  first	0
  second	1
  last	-1
  second_to_last	-2
  eighth_to_last	-8

• 示例1

   import pytest
   
   
   @pytest.mark.order(index=2)
   def test_three():
       print('three')
       assert 3 == 3
   
   
   @pytest.mark.order('second')
   def test_two():
       print('two')
       assert 2 == 2
   
   @pytest.mark.order(-1)
   def test_last():
       print('last')
       assert 'last' == 'last'
   
   @pytest.mark.order(0)
   def test_one():
       print('one')
       assert 1 == 1
   
   
   if __name__ == '__main__':
       pytest.main(['-sv', __file__])
   
  

   test_order_v1.py::test_one one
   PASSED
   test_order_v1.py::test_two two
   PASSED
   test_order_v1.py::test_three three
   PASSED
   test_order_v1.py::test_last last
   PASSED
   
   ============================== 4 passed in 0.06s ==============================
   
   进程已结束，退出代码为 0
   
  

• index是从0开始的；切记index=2其实是第三个，first其实是0.所以我们不建议混用

• -1是最后一个没有问题，索引体系跟list的类似，还是比较好理解的。建议用数字，学习成本就比较低。

• 这个装饰器可以用到类上

   import pytest
   
   
   @pytest.mark.order(2)
   class TestA:
       def test_one(self):
           assert 1 == 1
       def test_two(self):
           assert 1 == 1
   
   @pytest.mark.order(1)
   class TestB:
       def test_one(self):
           assert 1 == 1
       def test_two(self):
           assert 1 == 1
   
   
   if __name__ == '__main__':
       pytest.main(['-sv', __file__])
   
  

   test_order_v1.py::TestB::test_one PASSED
   test_order_v1.py::TestB::test_two PASSED
   test_order_v1.py::TestA::test_one PASSED
   test_order_v1.py::TestA::test_two PASSED
  

排在指定用例后面

• 直接看示例

   import pytest
   
   
   @pytest.mark.order(after='test_hallo')
   def test_hello():
       assert 1 == 1
   
   
   def test_hallo():
       assert 1 == 1
   
   
   if __name__ == '__main__':
       pytest.main(['-sv', __file__])
   
  

• 还可以这样写，指定类

   @pytest.mark.order(after="TestB::test_c")
   @pytest.mark.order(after="Test2")
   
  

• 可以指定文件夹/文件::类::测试用例(即测试函数名)，也可以用before

   @pytest.mark.order(before="test_module_c/test_submodule.py::test_2")
   
  

• 组合也可以

   @pytest.mark.order(index=0, after="test_second")
   @pytest.mark.order(after=["test_second", "other_module.py::test_other"])
   
  

• 如果是参数化，那就直接用测试函数名

   import pytest
   
   @pytest.mark.order(after=["test_second"])
   def test_first():
       assert True
   
   @pytest.parametrize(param, [1, 2, 3])
   def test_second(param):
       assert True
  

说在最后

• 关于用例的顺序相关的插件是不少的，比如

  • pytest-randomly：随机顺序
  • pytest-reverse：反转（通过一个hook亦可实现）
  • pytest-random-order ：随机顺序
  • pytest-depends：依赖
  • pytest-find-dependencies：寻找依赖

• 写完发现他有个doc，白整了~

   https://pytest-order.readthedocs.io/en/latest/