我在使用假设附加包进行 python 单元测试时遇到问题,Faker . 我想测试我网站的登录过程,我已经有了单元测试场景,但我想用 hypothesis 自动化场景. 这是我用于验证电子邮件和密码的简单代码。
import unittest
from Environtments.ParametrizedTestCase import ParametrizedTestCase
from hypothesis import given
from faker import Faker
class TestLogin(ParametrizedTestCase):
fake = Faker('en_US')
@given(email = fake.email(),password = fake.password()
def test_login(self,email,password):
assert '@' in email
assert len(password) >= 6
if __name__ == '__main__':
suite = unittest.TestSuite()
suite.addTest(ParametrizedTestCase.parametrize(TestLogin, param=""))
unittest.TextTestRunner(verbosity=2).run(suite)
我总是发现这个错误。我错过了什么吗?
test_login (__main__.TestLogin) ... You can add @seed(311936867547523412638507752560457398354) to this test to reproduce this failure.
ERROR
======================================================================
ERROR: test_login (__main__.TestLogin)
----------------------------------------------------------------------
Traceback (most recent call last):
File "tester_uf_exmp_hypo.py", line 10, in test_login
def test_login(self,email,password):
File "/Library/Python/2.7/site-packages/hypothesis/core.py", line 1001, in wrapped_test
state.run()
File "/Library/Python/2.7/site-packages/hypothesis/core.py", line 725, in run
runner.run()
File "/Library/Python/2.7/site-packages/hypothesis/internal/conjecture/engine.py", line 435, in run
self._run()
File "/Library/Python/2.7/site-packages/hypothesis/internal/conjecture/engine.py", line 771, in _run
self.reuse_existing_examples()
File "/Library/Python/2.7/site-packages/hypothesis/internal/conjecture/engine.py", line 626, in reuse_existing_examples
self.test_function(last_data)
File "/Library/Python/2.7/site-packages/hypothesis/internal/conjecture/engine.py", line 153, in test_function
self._test_function(data)
File "/Library/Python/2.7/site-packages/hypothesis/core.py", line 691, in evaluate_test_data
escalate_hypothesis_internal_error()
File "/Library/Python/2.7/site-packages/hypothesis/core.py", line 663, in evaluate_test_data
result = self.execute(data, collect=True)
File "/Library/Python/2.7/site-packages/hypothesis/core.py", line 578, in execute
result = self.test_runner(data, run)
File "/Library/Python/2.7/site-packages/hypothesis/executors.py", line 58, in default_new_style_executor
return function(data)
File "/Library/Python/2.7/site-packages/hypothesis/core.py", line 552, in run
args, kwargs = data.draw(self.search_strategy)
File "/Library/Python/2.7/site-packages/hypothesis/internal/conjecture/data.py", line 158, in draw
return self.__draw(strategy, label=label)
File "/Library/Python/2.7/site-packages/hypothesis/internal/conjecture/data.py", line 175, in __draw
return strategy.do_draw(self)
File "/Library/Python/2.7/site-packages/hypothesis/core.py", line 180, in do_draw
return self.base.do_draw(data)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/lazy.py", line 157, in do_draw
return data.draw(self.wrapped_strategy)
File "/Library/Python/2.7/site-packages/hypothesis/internal/conjecture/data.py", line 148, in draw
if strategy.is_empty:
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/strategies.py", line 144, in accept
recur(self)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/strategies.py", line 141, in recur
mapping[strat] = getattr(strat, calculation)(recur)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/collections.py", line 67, in calc_is_empty
return any(recur(e) for e in self.element_strategies)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/collections.py", line 67, in <genexpr>
return any(recur(e) for e in self.element_strategies)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/strategies.py", line 141, in recur
mapping[strat] = getattr(strat, calculation)(recur)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/strategies.py", line 541, in calc_is_empty
return recur(self.mapped_strategy)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/strategies.py", line 141, in recur
mapping[strat] = getattr(strat, calculation)(recur)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/lazy.py", line 87, in calc_is_empty
return recur(self.wrapped_strategy)
File "/Library/Python/2.7/site-packages/hypothesis/searchstrategy/lazy.py", line 110, in wrapped_strategy
*self.__args, **self.__kwargs
File "/Library/Python/2.7/site-packages/hypothesis/strategies.py", line 651, in fixed_dictionaries
check_strategy(v)
File "/Library/Python/2.7/site-packages/hypothesis/internal/validation.py", line 45, in check_strategy
check_type(SearchStrategy, arg, name)
File "/Library/Python/2.7/site-packages/hypothesis/internal/validation.py", line 39, in check_type
% (typ_string, name, arg, type(arg).__name__))
InvalidArgument: Expected SearchStrategy but got u'2)Y4+APqu@' (type=unicode)
----------------------------------------------------------------------
Ran 1 test in 2.013s
FAILED (errors=1)
我把代码改成这个。
from hypothesis.extra.fakefactory import fake_factory
@given(email = fake_factory('email'),password = fake_factory('password'))
这是警告。但是伪造者有效。这会有问题吗?
tester_uf_exmp_hypo.py:9: HypothesisDeprecationWarning: hypothesis.extra.fakefactory has been deprecated, because it does not support example discovery or shrinking. Consider using a lower-level strategy (such as st.text()) instead.
@given(email = fake_factory('email'),password = fake_factory('password'))
test_login (__main__.TestLogin) ... kaareberge@pedersen.no
!l3QCGDd8I
myslrshd@bn.com
(@s(RwUM0z
请您参考如下方法:
您的第一个案例是使用错误类型的参数调用 @given - 假设要求它们是生成示例值的“策略”,而不是值本身。您可以在回溯的底部看到这一点 - got u'2)Y4+APqu@' (type=unicode) 是 fake.password() 的结果,而不是(例如)hypothesis.strategies.text()。
第二个代码片段确实有效 - 欢呼!
但是 - 这就是我写弃用警告的原因 - 它放弃了假设的大部分优势:因为这只是假工厂的包装,无法最小化示例并且测试过程没有足够的进行覆盖率引导优化的信息。
相反,您可以考虑什么是密码 - 在本例中,可能是“一个包含 6 到 40 个字符的 unicode 字符串”。在假设中,这将是:
import hypothesis.strategies as st
@given(password=st.text(min_size=6, max_size=40))
def test(password):
...
然后你会得到缩小的、以配置文件为导向的示例选择,以及更多 :D
“电子邮件地址”也是字符串,但具有更多结构 - 幸运的是,Hypothesis 也涵盖了这一点。您可以使用 .map() 从部分组成策略,或者使用 from_regex() 策略,例如,或者只是从 hypothesis.provisional.emails 复制我们的内部实现(或者甚至只是导入它,如果您不介意可以随时删除的不稳定 API)。
TLDR:您可以混合使用 Hypothesis 和 fake-factory,但我只使用 Hypothesis。
