Pytest05-Fixture

5.Fixture ? ? 在测试过程中,fixture都可以派上用场。fixture是在测试函数运行前后,则pytest执行的外壳函数。fixture中的代码可以定制,满足多变的测试需求,包含定义传入测试中的数据集、配置测试前系统的初始状态、为批量测试提供数据源等等。来看以下简单示例,返回一个简单的fixture

import pytest@pytest.fixture() def getData(): return 28def test_getFixture(getData): assert getData=https://www.it610.com/article/=28

  • @pytest.fixture()装饰器用于声明函数是一个fixture。如果测试函数的参数列表中包含fixture名称,则pytest会检测到,并在测试函数运行之前执行该fixture。fixture可以完成任务,也可以返回数据给测试函数。
  • test_getFixture()的参数列表中包含一个fixture,名为getData,pytest会以该名称搜索fixture。pytest会优先搜索该测试所在模块,然后搜索conftest.py
后面所提到的fixture均是由@pytest.fixture()装饰器定义的函数,fixture是pytest用于将测试前后进行预备、清理工作的代码分离出核心逻辑的一种机制。
5.1 通过conftest.py共享fixture
? ? fixture的特点如下所示:
  • 1.fixture可以放在单独的测试文件中。如果希望多个测试文件共享fixture,可以在某个公共目录新建一个conftest.py文件,将fixture放在其中。
  • 2.如果希望fixture的作用范围仅限于某个测试文件,则可以将fixture写在该测试文件中
  • 3.尽管conftest.py是Python模块,但却不能被测试文件导入。因此是不允许出现import conftest的
5.2 使用fixture执行配置和销毁逻辑
? ? 在测试前准备和测试结束后清理环境,在数据库中连接使用比较多。测试前需要连接数据库,测试完成后,需要关闭数据库等,这时就可以使用fixture进行配置和清理环境了,如下所示:
1.DBOperate.py
import sqlite3 import osclass DBOperate:def __init__(self,dbPath=os.path.join(os.getcwd(),"db")): self.dbPath=dbPath self.connect=sqlite3.connect(self.dbPath)def Query(self,sql:str)->list: """传统查询语句""" queryResult = self.connect.cursor().execute(sql).fetchall() return queryResult if queryResult else []def QueryAsDict(self,sql:str)->dict: """调用该函数返回结果为字典形式""" self.connect.row_factory=self.dictFactory cur=self.connect.cursor() queryResult=cur.execute(sql).fetchall() return queryResult if queryResult else {}def Insert(self,sql:str)->bool: insertRows=self.connect.cursor().execute(sql) self.connect.commit() return True if insertRows.rowcount else Falsedef Update(self,sql:str)->bool: updateRows=self.connect.cursor().execute(sql) self.connect.commit() returnTrue if updateRows.rowcount else Falsedef Delete(self,sql:str)->bool: delRows=self.connect.cursor().execute(sql) self.connect.commit() return True if delRows.rowcount else Falsedef CloseDB(self): self.connect.cursor().close() self.connect.close()def dictFactory(self,cursor,row): """将sql查询结果整理成字典形式""" d={} for index,col in enumerate(cursor.description): d[col[0]]=row[index] return d

2.conftest.py
import pytest from DBOperate import DBOperate@pytest.fixture() def dbOperate(): # setup:connect db db=DBOperate() # 数据库操作 sql="""SELECT * FROM user_info""" res=db.QueryAsDict(sql) # tearDown:close db db.CloseDB() return res

3.test_02.py
import pytest from DBOperate import DBOperatedef test_dbOperate(dbOperate): db=DBOperate() sql = """SELECT * FROM user_info""" expect=db.QueryAsDict(sql) res=dbOperate assert expect==res

? ? 在fixture中,在执行查询语句前,db=DBOperate()相当于建立数据库连接,可视为配置过程(setup),而db.CloseDB()则相当于清理过程(teardown)过程,无论测试过程发生了什么,清理过程均会被执行。
5.3 使用--setup-show回溯fixture执行过程
? ? 如果直接运行前面的测试,则看不到其fixture的执行过程,如下所示:
>>> pytest -v .\test_02.py =================== test session starts ================================= platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 1 itemtest_02.py::test_dbOperate PASSED[100%]===================== 1 passed in 0.07s ==================================

? ? 如果希望看到其详细的执行过程及执行的先后顺序,可以使用参数--setup-show,如下所示:
>>> pytest--setup-show -v .\test_02.py ====================== test session starts ================================== platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 1 itemtest_02.py::test_dbOperate SETUPF dbOperate test_02.py::test_dbOperate (fixtures used: dbOperate)PASSED TEARDOWN F dbOperate============================ 1 passed in 0.03s =================================

? ? 从上面的运行的输出结果中可以看到,真正的测试函数被夹在中间,pytest会将每一个fixture的执行分成setup和teardown两部分。
fixture名称前面F代表其作用范围,F:表示函数级别,S:表示会话级别
5.4 使用fixture传递测试数据
? ? fixture非常适合存放测试数据,且可以返回任何数据,示例如下所示:
import pytest@pytest.fixture() def sampleList(): return [1,23,"a",{"a":1}]def test_sampleList(sampleList): assert sampleList[1]==32

运行结果如下所示:
>>> pytest -v .\test_fixture.py =========================== test session starts ============================== platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 1 itemtest_fixture.py::test_sampleList FAILED[100%]================================= FAILURES =========================================== _________________________test_sampleList ____________________________________________sampleList = [1, 23, 'a', {'a': 1}]def test_sampleList(sampleList): >assert sampleList[1]==32 Eassert 23 == 32 E+23 E-32test_fixture.py:8: AssertionError ===========================short test summary info ================================= FAILED test_fixture.py::test_sampleList - assert 23 == 32 =========================== 1 failed in 0.20s ======================================

? ? 除了指明详细的错误信息之外,pytest还给出了引起assert异常的函数参数值。fixture作为测试函数的参数,也会被纳入测试报告中。
? ? 上面的示例演示的是异常发生在测试函数中,那如果异常发生在fixture中,会怎么样?
import pytest@pytest.fixture() def sampleList(): x=23 assert x==32 return xdef test_sampleList(sampleList): assert sampleList==32

运行结果如下所示:
>>> pytest -v .\test_fixture.py ======================= test session starts ============================= platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 1 itemtest_fixture.py::test_sampleList ERROR[100%]========================== ERRORS ========================================== _________________ ERROR at setup of test_sampleList ________________________@pytest.fixture() def sampleList(): x=23 >assert x==32 Eassert 23 == 32 E+23 E-32test_fixture.py:6: AssertionError ==================== short test summary info ================================ ERROR test_fixture.py::test_sampleList - assert 23 == 32 =========================== 1 error in 0.27s =================================

? ? 在运行的输出结果中,正确定位到了fixture函数中发生assert异常的位置,其次test_sampleList并没有被标记为FAIL,而是被标记为ERROR,这个区分很清楚,如果被标记为FAIL,用户就知道失败发生在核心函数中,而不是发生在测试依赖的fixture中。
5.5 使用多个fixture
? ? 示例代码如下所示:
1.DBOperate
import sqlite3 import osclass DBOperate:def __init__(self,dbPath=os.path.join(os.getcwd(),"db")): self.dbPath=dbPath self.connect=sqlite3.connect(self.dbPath)def Query(self,sql:str)->list: """传统查询语句""" queryResult = self.connect.cursor().execute(sql).fetchall() return queryResult if queryResult else []def QueryAsDict(self,sql:str)->dict: """调用该函数返回结果为字典形式""" self.connect.row_factory=self.dictFactory cur=self.connect.cursor() queryResult=cur.execute(sql).fetchall() return queryResult if queryResult else {}def Insert(self,sql:str)->bool: insertRows=self.connect.cursor().execute(sql) self.connect.commit() return True if insertRows.rowcount else Falsedef Update(self,sql:str)->bool: updateRows=self.connect.cursor().execute(sql) self.connect.commit() returnTrue if updateRows.rowcount else Falsedef Delete(self,sql:str)->bool: delRows=self.connect.cursor().execute(sql) self.connect.commit() return True if delRows.rowcount else Falsedef CloseDB(self): self.connect.cursor().close() self.connect.close()def dictFactory(self,cursor,row): """将sql查询结果整理成字典形式""" d={} for index,col in enumerate(cursor.description): d[col[0]]=row[index] return d

2.conftest.py
import pytest from DBOperate import DBOperate@pytest.fixture() def dbOperate(): # setup:connect db db=DBOperate() yield # tearDown:close db db.CloseDB()@pytest.fixture() def mulQuerySqlA(): return ( "SELECT * FROM user_info", "SELECT * FROM case_info", "SELECT * FROM config_paras" )@pytest.fixture() def mulQuerySqlB(): return ( "SELECT * FROM user_info WHERE account in('admin')", "SELECT * FROM case_info WHERE ID in('TEST-1')", "SELECT * FROM config_paras WHERE accountMinChar==2", "SELECT * FROM report_info WHERE ID in('TEST-1')" )@pytest.fixture() def mulFixtureA(dbOperate,mulQuerySqlA): db = DBOperate() tmpList=[] for item in mulQuerySqlA: tmpList.append(db.QueryAsDict(item)) return tmpList@pytest.fixture() def mulFixtureB(dbOperate,mulQuerySqlB): db = DBOperate() tmpList = [] for item in mulQuerySqlB: tmpList.append(db.QueryAsDict(item)) return tmpList

3.test_03.py
import pytestdef test_count(mulQuerySqlA): assert len(mulQuerySqlA)==3

运行结果如下所示:
>>> pytest -v --setup-show .\test_03.py ========================= test session starts ================================ platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 1 itemtest_03.py::test_count SETUPF mulQuerySqlA test_03.py::test_count (fixtures used: mulQuerySqlA)PASSED TEARDOWN F mulQuerySqlA========================== 1 passed in 0.05s ==================================

? ? 使用fixture的优势在于:用户编写的测试函数可以只考虑核心的测试逻辑,而不需要考虑测试前的准备工作。
5.6 指定fixture作用范围
? ? fixture有一个叫scope的可选参数,称为作用范围,常用于控制fixture何时执行配置和销毁逻辑。@pytest.fixture()通常有4个可选值,分别为function、class、module和session,默认为function。各个scope的描述信息如下所示:
  • 1.scope="function"
? ? 函数级别的fixture每个测试函数仅运行一次,配置代码在测试函数运行之前运行,清理代码则在测试函数运行之后运行。
  • 2.scope="class"
? ? 类级别的fixture每个测试类仅运行一次,无论类中有多少个测试方法,都可以共享这个fixture.
  • 3.scope="moudle"
? ? 模块级别的fixture每个模块仅运行一次,无论模块中有多少个测试函数、测试方法或其他fixture都可以共享这个fixture
  • 4.scope="session"
【Pytest05-Fixture】? ? 会话级别的fixture每个会话仅运行一次,一次会话中,所有测试方法和测试函数都可以共享这个fixture。
? ? 各个作用范围的scope示例如下所示:
import pytest@pytest.fixture(scope="function") def funcScope(): pass@pytest.fixture(scope="module") def moduleScope(): pass@pytest.fixture(scope="session") def sessionScope(): pass@pytest.fixture(scope="class") def classScope(): passdef test_A(sessionScope,moduleScope,funcScope): pass@pytest.mark.usefixtures("classScope") class TestSomething: def test_B(self): pass def test_C(self): pass

运行结果如下所示:
>>> pytest --setup-show -v .\test_scope.py =================== test session starts =================================== platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 3 itemstest_scope.py::test_A SETUPS sessionScope SETUPM moduleScope SETUPF funcScope test_scope.py::test_A (fixtures used: funcScope, moduleScope, sessionScope)PASSED TEARDOWN F funcScope test_scope.py::TestSomething::test_B SETUPC classScope test_scope.py::TestSomething::test_B (fixtures used: classScope)PASSED test_scope.py::TestSomething::test_C test_scope.py::TestSomething::test_C (fixtures used: classScope)PASSED TEARDOWN C classScope TEARDOWN M moduleScope TEARDOWN S sessionScope=========================3 passed in 0.04s ===============================

? ? 以上各字母代表了不同的scope级别,C(class)、M(module)、F(function)、S(Session)。
fixture只能使用同级别或比自己更高级别的fixture。例如函数级别的fixture可以使用同级别的fixture,也可以使用类级别、模块级别、会话级别的fixture,反之则不行。
5.7 使用usefixture指定fixture
? ? 除在测试函数列表中指定fixture之外,也可以用@pytest.mark.usefixtures("fixture1","fixture2")标记测试函数或类。这种标记方法对测试类非常适用。如下所示:
@pytest.fixture(scope="class") def classScope(): pass@pytest.mark.usefixtures("classScope") class TestSomething: def test_B(self): pass def test_C(self): pass

使用usefixtures和在测试方法中添加fixture参数,两者并无太大差别,唯一区别在于后者可以使用fixture的返回值。
5.8 给fixture添加autouse选项
? ? 之前所用到的fixture都是根据测试本身来命名或针对示例中的测试类使用usefixtures,也可以通过指定autouse=True选项,使作用范围内的测试函数都运行该fixture,这种方式非常适合需要多次运行,但不依赖任何系统状态或外部数据的代码。示例代码如下所示:
import pytest import time@pytest.fixture(autouse=True,scope="session") def endSessionTimeScope(): yield print(f"\nfinished {time.strftime('%Y-%m-%d %H:%M:%S')}")@pytest.fixture(autouse=True) def timeDeltaScope(): startTime=time.time() yield endTime=time.time() print(f"\ntest duration:{round(endTime-startTime,3)}")def test_A(): time.sleep(2)def test_B(): time.sleep(5)

运行结果如下所示:
>>> pytest -v -s .\test_autouse.py ==================== test session starts ======================================= platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 2 itemstest_autouse.py::test_A PASSED test duration:2.002test_autouse.py::test_B PASSED test duration:5.006finished 2020-05-26 12:35:57=============================2 passed in 7.13s ====================================

5.9 给fixture重命名
? ? fixture的名字通常显示在使用它的测试或其他fixture函数的参数列表上,一般会和fixture函数名保持一致。pytest也允许使用@pytest.fixture(name="fixtureName")对fixture重命名。示例如下所示:
import pytest@pytest.fixture(name="Surpass") def getData(): return [1,2,3]def test_getData(Surpass): assert Surpass==[1,2,3]

? ? 在前面的示例中,使用fixture名字时,是用的函数名,而使用@pytest.fixture(name="Surpass")后,就相当于给fixture取了一别名。在调用fixture时,则可以使用别名了。运行结果如下所示:
>>> pytest --setup-show .\test_renamefixture.py =======================test session starts =============================== platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 1 itemtest_renamefixture.py SETUPF Surpass test_renamefixture.py::test_getData (fixtures used: Surpass). TEARDOWN F Surpass=========================== 1 passed in 0.05s ===============================

? ? 如果想找出重命名后的fixture定义,可以使用pytest的选项--fixtures,并提供所在测试文件名。pytest可提供所有测试使用的fixture,包含重命名的,如下所示:
>>> pytest --fixtures .\test_renamefixture.py ========================test session starts ================================================= platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 1 item --------------------------- fixtures defined from conftest ---------------------------------- mulQuerySqlA conftest.py:14: no docstring availablemulQuerySqlB conftest.py:22: no docstring availablemulFixtureA conftest.py:32: no docstring availablemulFixtureB conftest.py:40: no docstring availabledbOperate conftest.py:5: no docstring available------------------------- fixtures defined from test_renamefixture ------------------------- Surpass test_renamefixture.py:4: no docstring available

5.10 fixture参数化
? ? 在4.7中已经介绍过测试的参数化,也可以对fixture做参数化处理。下面来演示fixture参数化另一个功能,如下所示:
import pytestparas=((1,2),(3,5),(7,8),(10,-98)) parasIds=[f"{x},{y}" for x,y in paras]def add(x:int,y:int)->int: return x+y @pytest.fixture(params=paras,ids=parasIds) def getParas(request): return request.paramdef test_add(getParas): res=add(getParas[0],getParas[1]) expect=getParas[0]+getParas[1] assert res==expect

? ? fixture参数列表中的request也是pytest内建的fixture之一,代表fixture的调用状态。getParas逻辑非常简单,仅以request.param做为返回值供测试用,paras里面有4个元素,因此需要被调用4次,运行结果如下所示:
>>> pytest -v .\test_fixtrueparamize.py ============================ test session starts ================================ platform win32 -- Python 3.7.6, pytest-5.4.2, py-1.8.1, pluggy-0.13.1 -- d:\program files\python\python.exe cachedir: .pytest_cache rootdir: C:\Users\Surpass\Documents\PycharmProjects\PytestStudy\Lesson03 collected 4 itemstest_fixtrueparamize.py::test_add[1,2] PASSED[ 25%] test_fixtrueparamize.py::test_add[3,5] PASSED[ 50%] test_fixtrueparamize.py::test_add[7,8] PASSED[ 75%] test_fixtrueparamize.py::test_add[10,-98] PASSED[100%]================================ 4 passed in 0.10s =====================================

    推荐阅读