pytest插件的7种用法
目录
- 1.pytest-repeat 重复跑
- 第一种用法: 装饰器 @pytest.mark.repeat(次数)
- 第二种用法: 命令行参数
- 第三种用法: 结合repeat-scope运行
- 2.pytest-assume 断言后继续跑
- 3.pytest-ordering 用例顺序
- 4.pytest-dependency 用例依赖
- 5.pytest-rerunfailures 用例失败重跑
- 6.pytest-xdist 分布式执行
- 7.pytest-xfail 预期失败
1.pytest-repeat 重复跑
安装包
pip install pytest-repeat
第一种用法: 装饰器 @pytest.mark.repeat(次数)
示例代码
import pytest @pytest.mark.repeat(5) def test_001(): assert 1==2 if __name__ == '__main__': pythonpytest.main(['-sv',__file__])
第二种用法: 命令行参数
语法
pytest --count=5 test.py
示例代码
import pytest def test_001(): assert 1==2 if __name__ == '__main__': pytest.main(['-sv','--count=5',__file__])
第三种用法: 结合repeat-scope运行
如果我们要对多个测试函数进行重复运行,要么加多个装饰器,要么用命令行参数
但是上述两种方法都是A重复,B重复这样,无法做到AB-AB-AB的模式
如果要实现组合重复运行,那就要用到–repeat-scope
–repeat-scope类似于pytest fifixture的scope参数,–repeat-scope也可以设置参数:session, module , class 或者 function (默认值)
function (默认)范围针对每个用例重复执行,再执行下一个用例
class 以class为用例集合单位,重复执行class里面的用例,再执行下一个
module 以模块为单位,重复执行模块里面的用例,再执行下一个
session 重复整个测试会话,即所有收集的测试执行一次,然后所有这些测试再次执行等等
示例代码1:A运行2次,B运行2次
import pytest def test_002(): assert 1==2 def test_001(): assert 1==2 if __name__ == '__main__': pytest.main(['-sv','--count=2',__file__]) #运行结果 FAILED test_demo1.py::test_002[1-2] - assert 1 == 2 FAILED test_demo1.py::test_002[2-2] - assert 1 == 2 FAILED test_demo1.py::test_001[1-2] - assert 1 == 2 FAILED test_demo1.py::test_001[2-2] - assert 1 == 2
示例代码:A-B运行2次
import pytest def test_002(): assert 1==2 def test_001(): assert 1==2 if __name__ == '__main__': pytest.main(['-sv','--count=2','--repeat-scope=session',__file__]) #AB运行1次后再运行1次AB FAILED test_demo1.py::test_002[1-2] - assert 1 == 2 FAILED test_demo1.py::test_001[1-2] - assert 1 == 2 FAILED test_demo1.py::test_002[2-2] - assert 1 == 2 FAILED test_demo1.py::test_001[2-2] - assert 1 == 2
2.pytest-assume 断言后继续跑
开发者_JAVA开发安装
pip install pytest-assume
实际测试的过程中,有可能遇到一种情况,就是你某个断言执行失败也想要做下去(比如登录的测试,测试失败后,还是要返回主页继续下一轮的测试)。而默认情况下,如果断言失败,assert后面的语句是不会执行的了。
可以应用在多重断言的场景!(可以同时做多个断言)
没有assume的示例
import pytest def test_001(): assert 1==2 #如果改为1==1,下面是不会执行的 print('\n对了会做,错了不会做') if __name__ == '__main__': pytest.main(['-s','test_order_001.py'])
有assume的示例
import pytest def test_001(): pytest.assume(1==2) print('\n对了会做,错了也会做') if __name__ == '__main__': pytest.main(['-sv','test_order_001.py'])
assume的另外一种写法:上下文管理器
def test_assume2(): with pytest.assume: assert 1==2 assert 1==3
下面图1直接用pytest.assume不会显式testid的内容,而图2中assert就能显示变量的值,要解决这个问题就可以用with的写法
3.pytest-ordering 用例顺序
安装
pip install pytest-ordering
pytest默认按字母顺序去执行的(小写英文—>大写英文—>0-9数字)
用例之间的顺序是文件之间按照ASCLL码排序,文件内的用例按照从上往下执行。
改变测试用例的执行顺序,用法是加上装饰器
@pytest.mark.run(order=[number])
示例
import pytest @pytest.mark.run(order=2) # def test_001(): assert 1==1 @pytest.mark.run(order=1) #如果没有这两句话,那么运行顺序就是001再002 def test_002(): assert 2==2 if __name__ == '__main__': pytest.main(['-v','test_order_001.py'])
其他的运行方式
方式一
第一个执行:@ pytest.mark.first第二个执行:@ pytest.mark.second倒数第二个执行:@ pytest.mark.second_to_last最后一个执行:@pytest.mark.last方式二第一个执行:@ pytest.mark.run('first')第二个执行:@ pytest.mark.run('second')倒数第二个执行:@ pytest.mark.run('second_to_last')最后一个执行:@ pytest.mark.run('last')4.pytest-dependency 用例依赖
主要解决用例之间的依赖关系。如果依赖的上下文失败后续的用例会被标识为跳过执行,相当于执
行了 pytest.mark.skip
安装
pip install pytest-dependency
函数示例:
import pytest @pytest.mark.dependency() #打上标记 def test_001(): assert 1==2 @pytest.mark.dependency(depends=['test_001']) #依赖于test_001,test_001断言 成功了才会继续这个。 def test_002(): assert 1==1 if __name__ == '__main__': pytest.main(['-sv','test_order_001.py'])
name示例:
import pytest @pytest.mark.dependency(name='a') def test_001(): assert 1==2 @pytest.mark.dependency(depends=['a']) def test_002(): assert 1==1 if __name__ == '__main__': pytest.main(['-sv','test_order_001.py'])
类示例
import pytest class Test_001(): @pytest.mark.dependency() def test_001(self): assert 1==2 @pytest.mark.dependency(depends=['Test_001::test_001']) def test_002(): assert 1==1 if __name__ == '__main__': pytest.main(['-sv','test_order_001.py'])
5.pytest-rerunfailures 用例失败重跑
安装
pip i编程客栈nstall pytest-rerunfailures
使用方法一: 装饰器
import pytest import random from arrow import now @pytest.mark.flaky(reruns=50,reruns_delay=2) #重跑50次,每次间隔2s def test_001(): print(now().format('YYYY-MM-DD HH:mm:ss')) assert 1==random.randint(1,5) #只要在多次RERUN中遇到一次成功,即可停止,并最终结果 为PASSED if __name__ == '__main__': pytest.main(['-sv',__file__])
使用方法二: 命令行
import pytest def test_001(): assert 1==2 if __name__ == '__main__': pytest.main(['-sv','--reruns=2','--reruns-delay=2',__file__]) #参数前千万不要有空格,会报错
6.pytest-xdist 分布式执行
pytest-xdist,让自动化测试用例可以分布式执行,从而大大节省测试时间。pytest-xdist 是属于进程级别的并发。
分布式测试用例的设计原则:
(1)独立运行:用例之间是独立的,并且没有依赖关系,还可以完全独立运行。
(2)随机执行:用例执www.devze.com行不强制按顺序执行,支持顺序执行或随机执行。
(3)不影响其他用例:每个用例都能重复运行,运行结果不会影响其他用例
pytest-xdist 通过一些独特的测试执行模式扩展了 pytest:
(1)测试运行并行化:如果有多个CPU或主机,则可以将它们用于组合的测试运行。这样可以加快开发速度或使用远程计算机js的特殊资源。
(2)–looponfail:在子进程中重复运行测试。每次运行之后,pytest 都会等到项目中的文件更改后再
运行之前失败的测试。重复此过程,直到所有测试通过,然后再次执行完整运行。
(3)跨平台覆盖:可以指定不同的 python 解释器或不同的平台,并在所有这些平台上并行运行测试。
用法:
其实就是参数
-n numprocesses #如 -n 2 就是用2个
-n auto #自动检测物理CPU个数-n logical #检测逻辑CPU个数逻辑CPU个数=物理cpu数量x cpu核数 x 1(不支持ht超线程技术,如果开启就是2)超线程:一个CPU核就是一个物理线程,由英特尔开发超线程技术可以把一个物理线程模拟出两个线程来使用,使得单个核心用起来像两个核一样,以充分发挥CPU的性能.参数
distributed and subprocess testing:
-n numprocesses, --numprocesses=numprocessesShortcut for '--dist=load --tx=NUM*popen'. With'auto',attempt to detect physical CPU count. With'logical',detect logical CPU count. If physical CPU countcannotbe found, falls back to logical count. This will be0when used with --pdb.--maxprocesses=maxprocesseslimit the maximum number of workers to process thetestswhen using --numprocesses=auto--max-worker-restart=MAXWORKERRESTARTmaximum number of workers that can be restarted whencrashed (set to zero to disable this feature)--dist=distmode set mode for distributing tests to execenvironments.each: send each test to all available environments.load: load balance by sending any pending test toanyavailable environment.loadscope: load balance by sending pending groups oftests in the same scope to any availableenvironment.loadfile: load balance by sending test grouped byfileto any available environment.(default) no: run tests inprocess, don't distribute.--tx=xspec add a test execution environment. some examples: --txpopen//python=python2.5 --txsocket=192.168.1.102:8888--tx ssh=user@androidcodespeak.net//chdir=testcache-d load-balance tests. shortcut for '--dist=load'--rsyncdir=DIR add directory for rsyncing to remote tx nodes.--rsyncignore=GLOB add expression for ignores when rsyncing to remotetxnodes.--boxed backward compatibility alias for pytest-forked --forked--testrunuid=TESTRUNUIDprovide an identifier shared amongst all workers asthevalue of the 'testrun_uid' fixture,,if not provided, 'testrun_uid' is filled with a newunique string on every test run.-f, --looponfail run tests in subprocess, wait for modified files andre•run failing test set until all pass.
7.pytest-xfail 预期失败
第一种用法:
import pytest @pytest.mark.xfail(True,reason='预期失败,结果成功') def test_xfail1(): assert True @pytest.mark.xfail(True,reason='预期失败,结果失败') def test_xfail2(): assert False @pytest.mark.xfail(False,reason='预期成功,结果失败') def test_xfail3(): assert False @pytest.mark.xfail(False,reason='预期成功,结果成功') def test_xfail4(): assert True if __name__ == '__main__': pytest.main(['-sv',__file__])
输出:
test_xfail.py::test_xfail1 XPASS (预期失败,结果成功)
test_xfail.py::test_xfail2 XFAIL (预期失败,结果失败)test_xfail.py::test_xfail3 FAILEDtest_xfail.py::test_xfail4 PASSED
第二种用法:
import pytest @pytest.mark.xfail(raises=AssertionError) def test_xfail2(): assert 1==2 if __name__ == '__main__': pytest.main(['-sv',__file__])
输出:
test_xfail2.py::test_xfail2 XFAIL
如果这个时候带上–runxfail参数,就会忽略所有的xfail
import pytest @pytest.mark.xfail(raises=AssertionError) def test_xfail2(): assert 1==2 if __name__ == '__main__': pytest.main(['-sv','--runxfail',__file__]) ###输出就相当于没有那个装饰器 test_xfail2.py::test_xfail2 FAILED ================================== FAILURES =================================== _________________________________ test_xfail2 _________________________________ @pytest.mark.xfail(raises=AssertionError) def test_xfail2(): > assert 1==2 E assert 1 == 2 E +1 E -2 test_xfail2.py:6: AssertionError =========================== short test summary info =========================== FAILED test_xfail2.py::test_xfail2 - assert 1 == 2 ============================== 1 failed in 0.06s ==============================
第三种用法:
def test_xfail3(): pytest.xfail() assert 1==2 if __name__ == '__main__': pytest.main(['-sv',__file__])
同上输出
到此这篇关于pytest插件的7种用法的文章就介绍到这了,更多相关pytest 插件内容请搜索我们以前的文章或继续浏览下面的相关文章希望大家以后多多支持我们!
精彩评论