pytest Introduce ：
pytest Is a very mature full-featured Python The test framework ：
1. Simple and flexible , Easy to use
2. Support parameterization
3. Of test cases skip and xfail, Automatic failure retry and other processing
4. Support simple unit testing and complex functional testing , It can also be used to make selenium/appnium Wait for automated testing 、 Interface automation testing （pytest+request）
5.pytest There are many third-party plug-ins , And you can customize the extension , It's easy to use pytest- allure（ perfect html Test report generation ）、pytest-xdist（ many CPU distribution ） etc.
6. It can be very good and jenkins Integrate
install pytest：pip install -U pytest among -U It's right pytest updated pytest – version： View version number
Identification and operation of test cases ：
distinguish ：
The test file ：
1.test_.py
2.test.py
Use case identification ：
1.Test* Class contains all test* Methods （ Test class cannot have __init__ Method ）
2. be not in class All in test_* Method
pytest It can also be executed unittest Use cases and methods written in framework
function ：
pytest -v ： You can run detailed log information
pytest -s ： You can print out the test samples print The content of
pytest file name .py:: Class name ： Run the classes in a module
pytest file name .py:: Class name :: Method name ： Run a method of a class in a module
pytest -v -k" Class name and not Method name "： Skip running a use case
pytest -m[ Tagnames ]：@pytest.mark.[ Tagnames ] The test case with this tag will be run
pytest -x file name ： Stop running once an error is reported
pytest --maxfail=[num]： When the running error reaches num It stops running when it's running
pytest perform - Failed to rerun ：
scene ： Rerun the test after it fails n Time , To add a delay between reruns , interval n Seconds running
install ：pip install pytest-rerunfailures
perform ：
1.pytest – reruns 3 -v -s test_class.py（ Set to run again n Time ）
2.pytest -v – reruns 5 --reruns-delay 1（ Set the delay time ）
pytest perform - Multiple assertions run even if they fail
scene ： Write multiple assertions in a method , Usually the first one doesn't pass , I won't do it now , If we want to report an error, we'll do it
install ：pip install pytest-assume
perform ：
1.pytest.assume(1==4)
2.pytest.assume(2==4)
pytest Frame structure
import pytest Allied setup,teardown Also more flexible
1. Module level （setup_module/teardown_module） Module beginning and end , Overall （ Highest priority ）
2. Function level （setup_function/teardown_function） Only for function use cases （ Not in class ）
3. Class level （setup_class/teardown_class） Run back and forth in the class only once （ In class ）
4. Method level （setup_method/teardown_method） Start from the beginning and end of the method （ In class ）
5. In class （setup/teardown） Run before and after calling the method
pytest-fixtrue Usage of
scene ：
1. Use cases 1 You need to log in first
2. Use cases 2 You don't need to log in
3. Use cases 3 Need to log in
This scenario doesn't work setup and teardown Realization
usage ： Add... Before the method @pytest.fixture()
scene ： When the test case is executed , Some use cases need login to execute , Some use cases do not require login ,setup and teardown Unable to meet .fixture Sure , Default scope（ Range ）function
step ：
1. Import pytest
2. Add... Before the login function @pytest.fixture()
3. In the test method to be used, pass （ Login function name ）, Just log in first
4. If you don't pass in, you don't log in and directly execute the test method
Application in front-end automation -conftest
scene ： When you work with other test engineers to develop , Common modules should be in different files , Where everyone can visit
solve ：conftest.py This file is used for data sharing , And it can be placed in different positions to play the role of sharing in different ranges
perform ： The system executes to the parameter login First, look for the variable with this name from this file , After the conftest.py Find out if there is
step ： Bring the login module to @pytest.fixture() Written in conftest.py in
conftest.py Configuration needs attention ：
1.conftest The file name cannot be changed
2.conftest.py The same as the running use case package Next , And there are __init__ file
3. Unwanted import Import conftest.py file ,pytest Use cases are automatically found
4. The global configuration and preliminary work can be written here , Put it under a bag , This is where the package is shared
Application in front-end automation -yieid
scene ： You can solve the problems that the test method needs to execute or depend on , How to destroy clear data after testing methods ？ The scope is at the module level . similar setupClass
solve ： By adding the first mock exam module yieid keyword ,yield Is the result returned for the first time , The second execution of the following statement returns
step ： stay @pytest.fixture(scope=module)
Add... To the login method yieid, Then add clear steps to destroy , Be careful , This method does not return a value , If you want to return to use addfinalizer
fixture Automatic application of
scene ： I don't want to make any changes to the original test method , Or all of them automatically , No exception , You can choose to apply automatically when you don't need a return value
solve ： Use fixture In the parameter autouse=True Realization
step ：
1. Add... To the method @pytest.fixture(autouse=True)
2. Add... To the test method @pytest.mark.usefixtures(“start”)
autouse: You can use the method in every test case
fixture Passing with parameters
scene ： Testing is inseparable from data , For data flexibility , General data is transmitted through parameters
solve ：fixture By fixing parameters request Pass on
step ： stay fixture add @pytest.fixture(params=[1,2,3,‘linda’]) Write... In the method parameters request
@pytest.mark.skip(‘ Do not execute this article case’)： This use case may not be executed on the method
@pytest.mark.skipif(sys.platform==‘darwin’,reason=‘ be not in macOS perform ’)： By judging the current system Give the reason for not executing this use case
import pytest
# Method name as parameter
test_user_data = ['jerry','Tome']
@pytest.fixture(scope='module')
def login_r(request):
# This is the parameter that is accepted and passed in
user = request.param
print(f"\n Open the home page and get ready to log in Landing user {user}")
return user
@pytest.mark.skip(' Do not execute this test case ')
# indirect=True, You can execute the passed parameter as a function
@pytest.mark.parametrize("login_r",test_user_data,indirect=True)
def test_case01(login_r):
a = login_r
print(f" In the test case login The return value of ：{a}")
assert a != ''
if __name__ == '__main__':
pytest.main()
skip Use scenarios
1. You don't want to run this test case while debugging
2. Mark test functions that cannot be run on some platforms
3. In some versions , Skip in other versions
4. Skip when current external resources are not available （ If the test data is taken from the database , The function of connecting to the database. If the result is not successful, skip , Because the execution also reports errors ）
solve ：
@pytest.mark.skip()： Skip this test case , You can add conditions skipIf, Only when certain conditions are met can we hope to pass , Otherwise, skip this test
Xfail scene
1. Errors that have not been implemented or repaired by functional tests , When the test passes, though it is expected to fail （ Marked as pytest.mark.xfail）, He is a xpass Will be reported in the test summary
2. You want the test to fail for some reason
solve ：
@pytest.mark.xfail
Use custom tags mark Execute only some use cases
scene ：
1. Execute only a part of the use case that meets the requirements , You can put a web The project is divided into several modules , Then specify the module name to execute
2.App When automating , If you want to Android and IOS When sharing a set of codes , You can also use the tagging function , Indicate what is IOS The use case , Which are Android The runtime knows mark Just run by name
solve ： Add... To the test case method @pytest.mark.webtest
perform ：
1.-s Parameters ： Output the... Of the test cases used print Information
2.-m： Execute relevant use cases of custom tags （ Use the command to execute ）
pytest -s test.py -m=search
pytest -s test.py -m apptest
pytest -s test.py -m ‘not ios’
Multithreading, parallel and distributed execution
scene ： The test case 1000 strip , A test case executes for one minute , A tester needs to perform 1000 minute , Usually, the cost of labor is exchanged for the cost of time , Add a few people to execute together , Time will be shortened , If 10 Individual execution only needs 1000 minute , This is a parallel test , Distributed scenarios .
solve ：pytest Distributed execution plug-ins ：pytest-xdist, Multiple cpu Or host execution Premise ： Use cases are all independent , There is no order , It can be executed at random , Repeatable execution does not affect other use cases .
install ：pip install pytest-xdist
Multiple CPU Execute use cases in parallel , direct -n 3 It's the parallel number ：pytest test.py -n 3
Execute together under multiple terminals
pytest-html Generate test reports
install ：pip install pytest-httml
Generate test reports ：pytest -v -s --html=report.html --self-contained-html
