我在Python中使用mock
相当困难:
def method_under_test():
r = requests.post("http://localhost/post")
print r.ok # prints "<MagicMock name='post().ok' id='11111111'>"
if r.ok:
return StartResult()
else:
raise Exception()
class MethodUnderTestTest(TestCase):
def test_method_under_test(self):
with patch('requests.post') as patched_post:
patched_post.return_value.ok = True
result = method_under_test()
self.assertEqual(type(result), StartResult,
"Failed to return a StartResult.")
测试实际上返回正确的值,但是r.ok
是一个Mock对象,而不是True
。如何在Python的mock
库中模拟属性?
您需要使用return_value
和PropertyMock
:
with patch('requests.post') as patched_post:
type(patched_post.return_value).ok = PropertyMock(return_value=True)
这意味着:当调用requests.post
时,在该调用的返回值上,为属性ok
设置PropertyMock
以返回值True
。
一种紧凑而简单的方法是使用new_callable
patch
的属性强制patch
使用PropertyMock
而不是MagicMock
来创建模拟对象。传递给patch
的其他参数将用于创建PropertyMock
对象。
with patch('requests.post.ok', new_callable=PropertyMock, return_value=True) as mock_post:
"""Your test"""
对于模拟版本'1.0.1',问题中提到的更简单的语法得到了支持,并且可以正常工作!
示例代码更新(py。使用Test代替unittest):
import mock
import requests
def method_under_test():
r = requests.post("http://localhost/post")
print r.ok
if r.ok:
return r.ok
else:
raise Exception()
def test_method_under_test():
with mock.patch('requests.post') as patched_post:
patched_post.return_value.ok = True
result = method_under_test()
assert result is True, "mock ok failed"
运行下面的代码:(确保你安装了pytest)
$ py.test -s -v mock_attributes.py
======= test session starts =======================
platform linux2 -- Python 2.7.10 -- py-1.4.30 -- pytest-2.7.2 -- /home/developer/miniconda/bin/python
rootdir: /home/developer/projects/learn/scripts/misc, inifile:
plugins: httpbin, cov
collected 1 items
mock_attributes.py::test_method_under_test True
PASSED
======= 1 passed in 0.03 seconds =================