【Pytest：Python 单元测试工具学习】

• Brief Introduction to Pytest
• Installation
• A Simple Testing
• Testing Whether Throws an Exception
• Test Classes
• Some Basic API For Pytest
• Conclusion about Pytest
• References

• 由于是课程作业，所以是全英文的，词汇量不是很难吧。

• Pytest is a very mature and full-featured Python testing framework. It is simple, flexible and easy to use.
  It supports simple unit testing and complex function testing, and can also cooperate with third-party plug-ins for automatic testing.
  There are many third-party plug-ins that can be customized and extended.
  We can see more details in the official documents of pytest:
  https://www.osgeo.cn/pytest/contents.html

• In CMD, enter the following command to install pytest:
  pip install -U pytest
  We can check the installed version of pytest by entering the following command. My running results are as follows:
  

• Firstly write a simple two functions, the second functions is for testing the first function.
  
  Then, in the cmd or console (because I use Visual Studio Code, I will run in console), run pytest command to see whether the assert runs well.
  
  Cause 1 not equal to 2, so assert(False) turns the FAILURES, gives us AssertionError.
  Change the line5, and rerun the pytest, then will pass the test.
  
  

• We could use pytest.raises(Exception) to check whether this Exception is raised.
  
  As we expected, the test_1() passed and test_2() failed because it didn’t raise the ‘SystemExit’ Exception.
  

• If we have several tests, we could class them together, but we need to add prefix ‘Test’ for class name , or we will skip test the class.
  
  
  It’s beneficial for us to make several relevant tests into one class.

• According to the official document, I will introduce several important API here.

1. Fail(msg,pytrace) -> NoReturn
   
   It means if we use ‘pytest.fail(mes)’ , it will show the error message.
   If we let second argument be false, then the trace will not show the fail message.
   
   
   See, only test_1 shows the trace, but test_2 only shows ‘Special Error2’
2. Pytest.raises
   It declares that the code blocks will raise expected exception. Otherwise, it will cause fail exception.
   It could use like I mentioned above:
   
   Also, could detect the defined exception, for example, ValueError
   
3. Pytest.approx
   Some times, we want to compare two float numbers whether they are equal, we may write in the form of ‘0.1 + 0.2 == 0.3’. However, this may be wrong due to the storage way of floating numbers.
   So, we may write ‘abs((0.1 + 0.2) - 0.3) < 1e-6’. However, it’s more complicated for us to understand. And, 1e-6 seems good for number 1, but bad for number 1e18.
   So , we use approx so that it’s better to check the equality of two numbers, or even list and etc.
   
   

• Pytest has many advantages:

1. Automatically find test functions and cases.
2. Convenient declarations to match exceptions.
3. Run testing cases, modules, and classes smoothly.
4. Various pluggable units are available.
5. Compatible with unnitest or nose.

For testing part, it will show all the traces so that we could know which testing cases are not pass, as well as their exceptions.

1. Pytest 官方文档：https://www.osgeo.cn/pytest/contents.html