Python testing using mock and pytest

Python
Testing
using Mock
& PyTest

unittest.mock
Defn: unittest.mock is a library for testing in Python. It
allows you to replace parts of your system under test with
mock objects and make assertions about how they have
been used.
• Using Mock you can replace/mock any dependency of
your code.
• Unreliable or expensive parts of code are mocked using
Mock, e.g. Networks, Intensive calculations, posting on
a website, system calls, etc.

• As a developer you want your calls to be right
rather than going all the way to final output.
• So to speed up your automated unit-tests you
need to keep out slow code from your test runs.

>>> from unittest.mock import Mock
>>> m = Mock()
>>> m
<Mock id='140457934010912'>
>>> m.some_value = 23
>>> m.some_value
23
>>> m.other_value
<Mock name='mock.other_value' id='140457789462008'>
Mock Objects - Basics

>>> m.get_value(value=42)
<Mock name='mock.get_value()' id='140457789504816'>
>>> m.get_value.assert_called_once_with(value=42)
>>> m.get_value.assert_called_once_with(value=2)
raise AssertionError(_error_message()) from cause
AssertionError: Expected call: get_value(value=2)
Actual call: get_value(value=42)

• Flexible objects that can replace any part of
code.
• Creates attributes when accessed.
• Records how objects are being accessed.
• Using this history of object access you can make
assertions about objects.
More about Mock objects

>>> from unittest.mock import Mock
>>> config = {
... 'company': 'Lenovo',
... 'model': 'Ideapad Z510',
... 'get_sticker_count.return_value': 11,
... 'get_fan_speed.side_effect': ValueError
... }
>>> m = Mock(**config)
>>> m.company
'Lenovo'
>>> m.get_sticker_count()
11
>>> m.get_fan_speed()
raise effect
ValueError
Customize mock objects

Using spec to define attr
>>> user_info = ['first_name', 'last_name', 'email']
>>> m = Mock(spec=user_info)
>>> m.first_name
<Mock name='mock.first_name' id='140032117032552'>
>>> m.address
raise AttributeError("Mock object has no attribute %
r" % name)
AttributeError: Mock object has no attribute 'address'

Automatically create all specs
>>> from unittest.mock import create_autospec
>>> import os
>>> m = create_autospec(os)
>>> m.
Display all 325 possibilities? (y or n)
m.CLD_CONTINUED m.forkpty
m.CLD_DUMPED m.fpathconf
m.CLD_EXITED m.fsdecode
[CUT]
m.fchown m.walk
m.fdatasync m.write
m.fdopen m.writev
m.fork

Using Mock through patch
• Replaces a named object with Mock object
• Also can be used as decorator/context manager
that handles patching module and class level
attributes within the scope of a test.

1 # main.py
2 import requests
3 import json
4
5 def upload(text):
6 try:
7 url = 'http://paste.fedoraproject.org/'
8 data = {
9 'paste_data': text,
10 'paste_lang': None,
11 'api_submit': True,
12 'mode': 'json'
13 }
14 reply = requests.post(url, data=data)
15 return reply.json()
16 except ValueError as e:
17 print("Error:", e)
18 return None
19 except requests.exceptions.ConnectionError as e:
20 print('Error:', e)
21 return None
22 except KeyboardInterrupt:
23 print("Try again!!")
24 return None
25
26 if __name__ == '__main__':
27 print(upload('Now in boilerplate'))

1 # tests.py
2 import unittest
3 import requests
4 from unittest.mock import patch
5 from main import upload
6
7 text = 'This is ran from a test case'
8 url = 'http://paste.fedoraproject.org/'
9 data = {
10 'paste_data': text,
11 'paste_lang': None,
12 'api_submit': True,
13 'mode': 'json'
14 }
15 class TestUpload(unittest.TestCase):
16 def test_upload_function(self):
17 with patch('main.requests') as mock_requests:
18 result = upload(text) # call our function
19 mock_requests.post.assert_called_once_with(url, data=data)
20
21 def test_upload_ValueError(self):
22 with patch('main.requests') as mock_requests:
23 mock_requests.post.side_effect = ValueError
24 result = upload(text)
25 mock_requests.post.assert_any_call(url, data=data)
26 self.assertEqual(result, None)

patching methods #1
>>> @patch('requests.Response')
... @patch('requests.Session')
... def test(session, response):
... assert session is requests.Session
... assert response is requests.Response
...
>>> test()

patching methods #2
>>> with patch.object(os, 'listdir', return_value=
['abc.txt']) as mock_method:
... a = os.listdir('/home/hummer')
...
>>> mock_method.assert_called_once_with
('/home/hummer')
>>>

Mock return_value
>>> m = Mock()
>>> m.return_value = 'some random value 4'
>>> m()
'some random value 4'
OR
>>> m = Mock(return_value=3)
>>> m.return_value
3
>>> m()
3

Mock side_effect
• This can be a Exception, Iterable or function.
• If you pass in a function it will be called with
same arguments as the mock, unless function
returns DEFAULT singleton.

#1 side_effect for Exception
>>> m = Mock()
>>> m.side_effect = ValueError('You are always gonna get this!!')
>>> m()
raise effect
ValueError: You are always gonna get this!!

>>> m = Mock()
>>> m.side_effect = [1, 2, 3, 4]
>>> m(), m(), m(), m()
(1, 2, 3, 4)
>>> m()
StopIteration
#2 side_effect for returning
sequence of values

>>> m = Mock()
>>> side_effect = lambda value: value ** 3
>>> m.side_effect = side_effect
>>> m(2)
8
#3 side_effect as function

Installation
For Python3
$ pip3 install -U pytest
For Python2
$ pip install -U pytest
or
$ easy_install -U pytest

What is pytest?
● A fully featured Python Testing tool.
● Do automated tests.

Tests with less Boilerplate
1 import unittest
2
3 def cube(number):
4 return number ** 3
5
6
7 class Testing(unittest.TestCase):
8 def test_cube(self):
9 assert cube(2) == 8
Before py.test

1 def cube(number):
2 return number ** 3
3
4 def test_cube():
5 assert cube(2) == 8
6
7 # Here no imports or no classes are needed
After py.test

Running Tests
pytest will run all files in the current directory and
its subdirectories of the form test_*.py or
*_test.py or else you can always feed one file at a
time.
$ py.test cube.py
=============================== test session starts============================================
platform linux -- Python 3.4.3, pytest-2.8.3, py-1.4.30, pluggy-0.3.1
rootdir: /home/hummer/Study/Nov2015PythonPune/pyt, inifile:
collected 1 items
cube.py .
===============================1 passed in 0.01 seconds========================================

$ py.test
Run entire test suite
$ py.test test_bar.py
Run all tests in a specific file
$ py.test -k test_foo
Run all the tests that are named test_foo
By default pytest discovers tests in
test_*.py and *_test.py

pytest fixtures
• Fixtures are implemented in modular manner, as each
fixture triggers a function call which in turn can trigger
other fixtures.
• Fixtures scales from simple unit tests to complex
functional test.
• Fixtures can be reused across class, module or test
session scope.

1 import pytest
2
3 def needs_bar_teardown():
4 print('Inside "bar_teardown()"')
5
6 @pytest.fixture(scope='module')
7 def needs_bar(request):
8 print('Inside "needs bar()"')
9 request.addfinalizer(needs_bar_teardown)
10
11 def test_foo(needs_bar):
12 print('Inside "test_foo()"')

[hummer@localhost fixtures] $ py.test -sv fix.py
========================= test session starts ======================================
platform linux -- Python 3.4.3, pytest-2.8.3, py-1.4.30, pluggy-0.3.1 --
/usr/bin/python3
cachedir: .cache
rootdir: /home/hummer/Study/Nov2015PythonPune/pyt/fixtures, inifile:
collected 1 items
fix.py::test_foo Inside "needs bar()"
Inside "test_foo()"
PASSEDInside "bar_teardown()"
========================= 1 passed in 0.00 seconds==================================
[hummer@localhost fixtures] $

References
• http://www.toptal.com/python/an-introduction-
to-mocking-in-python
• https://docs.python.org/dev/library/unittest.
mock.html
• http://pytest.org/latest/contents.html
• http://pythontesting.net/

Python testing using mock and pytest

In this document

More Related Content

What's hot

Viewers also liked

Similar to Python testing using mock and pytest

More from Suraj Deshmukh

Recently uploaded

Python testing using mock and pytest