Pytest multi file execution sequence control

1. only one py file

• 1. Use pytest Do interface tests , If you test case Only exists in a single .py file , So the test case By default, it is executed from top to bottom , If used pytest-order plug-in unit

2. If there are multiple py file

• 1. Use pytest Do interface tests , If you test case Exist in multiple .py In file , The default is based on the file name ascii Code sequence execution , After entering the document , By default, each unit test interface is executed from top to bottom .

  test_user.py  #  User correlation 
  	class TestUser:
  		def test_user_create:
  		def test_user_login:
  		def test_user_delete
  		
  test_order.py  #  Order related 
  	class TestOrder:
  		def test_order_create:
  		def test_order_list:
  		def test_order_delete
  		
  test_stock.py  #  Inventory related 
  	class TestStock:
  		def test_stock_add:
  		def test_stock_list:
  		def test_stock_reduce
  

  1. By file name ascii Sort ：test_order > test_stock > test_user

  2.test_order_create > test_order_list > test_order_delete > test_stock_add > test_stock_list > …

• 2. If single .py In the test file pytest-order plug-in unit , Then... Is added to the file order Test cases will be executed first , Those not added will follow 1 Sequential execution , In this way, the sequence of unit tests will cross execute in multiple files .（ So single .py The file is in use pytest-order In the case of plug-ins , Suggest that each case Take them all order=x, And x Don't be the same ）

  test_user.py  #  User correlation 
  	class TestUser:
  		@pytest.mark.run(order=1)
  		def test_user_create:
  		def test_user_login:
  		@pytest.mark.run(order=2)
  		def test_user_delete
  		
  test_order.py  #  Order related 
  	class TestOrder:
  		def test_order_create:
  		def test_order_list:
  		def test_order_delete
  		
  test_stock.py  #  Inventory related 
  	class TestStock:
  		def test_stock_add:
  		def test_stock_list:
  		def test_stock_reduce
  

  1. because test_user In the document case Used pytest-order plug-in unit , So the priority implementation uses order Sort of case

  2.test_user_create > test_user_delete> test_order_create> … > test_stock_add > … > test_user_delete

• 3. If more than one .py The file uses pytest-order plug-in unit , If each order The specified order does not conflict , Just follow order Execute in the specified order , If there is a conflict , That will appear in multiple .py Cross execution in the file , Probably Not in line with our expectations .

  test_user.py  #  User correlation 
  	class TestUser:
  		@pytest.mark.run(order=1)
  		def test_user_create:
  		def test_user_login:
  		@pytest.mark.run(order=2)
  		def test_user_delete
  		
  test_order.py  #  Order related 
  	class TestOrder:
  		def test_order_create:
  		def test_order_list:
  		def test_order_delete
  		
  test_stock.py  #  Inventory related 
  	class TestStock:
  		@pytest.mark.run(order=1)
  		def test_stock_add:
  		@pytest.mark.run(order=2)
  		def test_stock_list:
  		def test_stock_reduce
  

  1.test_stock and test_user There is order Conflict , So according to the file name ascii Order

  2.test_stock_add > test_user_create > test_stock_list > test_user_delete > order relevant > test_stock_reduce > test_user_login

4. Multiple py The file is modified according to the file name ascii Code sorting method

• demand ： Don't go back and forth in multiple files case, Ensure that the sequence of test cases is ： User module --> Order module --> Inventory module

• Mode one ： By changing the file name , Make the file name ascii Sort code , Test with us case The execution sequence is consistent , Make sure case There is no pytest-order plug-in unit

  test_1_user.py  #  User correlation 
  	class TestUser:
  		def test_user_create:
  		def test_user_login:
  		def test_user_delete
  		
  test_2_order.py  #  Order related 
  	class TestOrder:
  		def test_order_create:
  		def test_order_list:
  		def test_order_delete
  		
  test_3_stock.py  #  Inventory related 
  	class TestStock:
  		def test_stock_add:
  		def test_stock_list:
  		def test_stock_reduce
  

  But usually , We .py Files are named according to modules , So we can achieve the expected execution order by modifying the file name , Not very friendly

• Mode two ： If you use pytest-order Plug in to control , It must be ensured that each document is order Value cannot be repeated , After a .py file order The minimum value must be greater than the previous .py Maximum file size , This ensures that the file execution order

  In this way, after adding test cases , You may need to modify a lot order The order

  test_user.py  #  User correlation 
  	class TestUser:
  		@pytest.mark.run(order=1)
  		def test_user_create:
  		@pytest.mark.run(order=3)
  		def test_user_login:
  		@pytest.mark.run(order=2)
  		def test_user_delete
  		
  test_order.py  #  Order related 
  	class TestOrder:
  		@pytest.mark.run(order=4)
  		def test_order_create:
  		@pytest.mark.run(order=5)
  		def test_order_list:
  		@pytest.mark.run(order=6)
  		def test_order_delete
  		
  test_stock.py  #  Inventory related 
  	class TestStock:
  		@pytest.mark.run(order=7)
  		def test_stock_add:
  		@pytest.mark.run(order=8)
  		def test_stock_list:
  		@pytest.mark.run(order=9)
  		def test_stock_reduce
  

• Mode three ： adopt pytest The hook method provided pytest_collection_modifyitems, Yes case The execution sequence is modified

  # conftest.py
  
  def pytest_collection_modifyitems(config, items)
      #  Expect the use case sequence to follow .py File execution 
      appoint_classes = {
        "TestUser": [], "TestOrder": [], "TestStock": []}
  
      for item in items:
          for cls_name in appoint_classes:
              if item.parent.name == cls_name:
                  appoint_classes[cls_name].append(item)
      items.clear()
      for cases in appoint_classes.values():
          items.extend(cases)
  

  Users only need to add their new test modules class Add to... In the expected order appoint_classes Then you can , Simple and flexible