Automated testing - unittest framework

Catalog

Framework analysis  

test fixture（ Test firmware ）

test case（ The test case ）

test suite（ test suite ）

test runner（ The test execution ）

Batch script execution

Build test suite

addTest（）

makeSuit()

TestLoader()

discover（）

The execution order of use cases

Ignore use case execution

unittest Assertion  

HTML Report generation

Exception capture and error screenshot

Data driven

---

Framework analysis  

unittest yes python The unit test framework of , It mainly has the following functions ：

• Provides use case organization and execution
• Provide rich comparison methods
• Provide rich logs

unittest There are four important concepts ,test fixture,test case,test suite,test runner.

test fixture（ Test firmware ）

Build and destroy a test case environment , It's just one. fixture, By covering setUp() and tearDown() Method to implement .

• setUp（）： Preparation of test environment and data
• tearDown（）： Clean up after the test case execution

test case（ The test case ）

A test case is a complete test process , Including the construction of preparation environment before testing （setUp）、 The code to implement the test process , And the restoration of the environment after testing （tearDown）. A use case is a method , The name to test_ start .

test suite（ test suite ）

Verification of a function often requires multiple test cases , Multiple test cases can be assembled and executed together , This produces the test suite TestSuite The concept of . Must be Inherit （ object-oriented ） How to use .

test runner（ The test execution ）

Test execution is also a very important concept , stay unittest In the frame , adopt TextTestRunner Class provides the run() Method to execute test suite/test case.

Batch script execution

Build test suite

Suppose we have written test1.py And test2.py Two documents , So how can we execute these two files at the same time ？

addTest（）

TestSuite Class addTest() Methods can assemble test methods from different test classes into test suites , however addTest() Only one test method in one class can be assembled into the test suite at a time .

import unittest
from src0716 import test1
from src0716 import test2
def createsuite():
    #addTest
    suite = unittest.TestSuite()
    suite.addTest(test1.Baidu1("test_hao"))   #addTest( Script name . Class name (" Method name "))
    suite.addTest(test1.Baidu1("test_hbaidu"))
    suite.addTest(test2.Baidu2("test_hao"))
    suite.addTest(test2.Baidu2("test_baidusearch"))
    return suite
if __name__=="__main__":
    suite = createsuite()
    runner = unittest.TextTestRunner(verbosity=2)  #verbosity 0,1,2 The larger the number, the more detailed the background print 
    runner.run(suite)

But there are two inconveniences in the above practice ：

1. You need to import all related py file , such as import test1, Each new script needs to be imported .
2. addTest Only one test method can be added at a time , If one py In file 10 Test methods , If they are to be assembled into the test suite , It needs to be increased 10 Time .

makeSuit()

stay unittest The framework provides makeSuite() Methods ,makeSuite It can realize all the tests in the test case class case Test suite composed of TestSuite ,unittest call makeSuite When , Just pass in the test class name .

suite.addTest(unittest.makeSuite(test1.Baidu1))
suite.addTest(unittest.makeSuite(test2.Baidu2))

TestLoader()

TestLoader Test suite for creating classes and modules , In general , use TestLoader().loadTestsFromTestCase(TestClass) To load the test class .

suite1 = unittest.TestLoader().loadTestsFromTestCase(test1.Baidu1)
suite2 = unittest.TestLoader().loadTestsFromTestCase(test2.Baidu2)
suite = unittest.TestSuite([suite1, suite2])

however makeTest（） and TestLoader（） You still need to import a related file every time you add a script .

discover（）

discover Is recursively to its subdirectory, starting from the specified directory , Find all test modules and return an object containing them TestSuite , Then load the only test file matching the pattern , That is to execute all test cases of test scripts in a folder ,discover The parameters are

unittest.defaultTestLoder.discover("../src",pattern="test*.py",top_level_dir=None)
#                                    Folder          The file where the test case is located 

The execution order of use cases

unittest By default, the order in which the framework loads test cases is based on ASCII The order of the codes , The order of numbers and letters is ：0~9,A~Z,a~z .addTest（） It is executed in the order of addition .

Ignore use case execution

Tag test cases that you don't want to run ：@unittest.skip("skiping")

@unittest.skip("skipping")
def test_baidusearch(self):
    driver = self.driver
    driver.get(self.base_url + "/")
    driver.find_element_by_id("kw").click()
    driver.find_element_by_id("kw").clear()
    driver.find_element_by_id("kw").send_keys(u" test ")
    driver.find_element_by_id("su").click()
    driver.find_element_by_id("su").click()

unittest Assertion  

Assertion ： Judge whether the actual results are consistent with the expected results

Common assertions ：

self.assertEqual(arg1,arg2,msg=" XXX")	Prediction expression arg1 And arg2 equal , Otherwise output XXX
self.assertNotEqual(arg1,arg2,msg="XXX")	Prediction expression arg1 And arg2 It's not equal , Otherwise output XXX
self.assertTrue(arg1,msg="XXX")	Prediction expression arg1 It's true , Otherwise output XXX
self.assertFalse(arg2,msg="XXX")	Prediction expression arg2 For false , Otherwise output XXX

HTML Report generation

Run a test suite , There are hundreds of test cases , How to view the execution results of test cases centrally and clearly ？ After the script is executed, it can generate HTML The report

HTMLTestRunner.py file , And put in python38/Lib Under the table of contents , Download address ：http://tungwaiyip.info/software/HTMLTestRunner.html

Generate HTML Reporting steps ：

1. Create a repository HTML Report folder ：

curpath=sys.path[0]

if not os.path.exists(curpath+'/resultreport'):

     os.makedirs(curpath+'/resultreport')

2. Solve the problem of duplicate naming （ Name it by time ）：

now=time.strftime("%Y-%m-%d %M %S",time.localtime(time.time()))

filename=curpath+'/resultreport/'+now+'resultreport.html'

3. The output of the report ：

Exception capture and error screenshot

Error screenshot API：get_screenshot_as_file()

Purpose ： Keep the test site

 

Data driven

python Of unittest No built-in data-driven function . So if you use unittest, At the same time, I want to use data-driven , Then you can use it ddt To complete .ddt Installation

ddt Use

Guide pack ：from ddt import ddt,unpack,data,file_date ; At the same time, use labels on classes @ddt

Data driven way ：

@data(value) Pass one parameter at a time

@data(value1,value2....) Multiple parameters at a time , Need to use @unpack mapping

@file_data("json file "）

@data（* Method of parsing data （txt/csv file ））