So let's say I have this bit of code:
import coolObject
def doSomething():
x = coolObject()
x.coolOperation()
Now it's a simple enough method, and as you can see we are using an external library(coolObject).
In unit tests, I have to create a mock of this object that roughly replicates it. Let's call this mock object coolMock.
My question is how would I tell the code when to use coolMock or coolObject? I've looked it up online, and a few people have suggested dependency injection, but I'm not sure I understand it correctly.
Thanks in advance!
def doSomething(cool_object=None):
cool_object = cool_object or coolObject()
...
In you test:
def test_do_something(self):
cool_mock = mock.create_autospec(coolObject, ...)
cool_mock.coolOperation.side_effect = ...
doSomthing(cool_object=cool_mock)
...
self.assertEqual(cool_mock.coolOperation.call_count, ...)
As Dan's answer says, one option is to use dependency injection: have the function accept an optional argument, if it's not passed in use the default class, so that a test can pass in a moc.
Another option is to use the mock library (here or here) to replace your coolObject.
Let's say you have a foo.py that looks like
from somewhere.else import coolObject
def doSomething():
x = coolObject()
x.coolOperation()
In your test_foo.py you can do:
import mock
def test_thing():
path = 'foo.coolObject' # The fully-qualified path to the module, class, function, whatever you want to mock.
with mock.patch('foo.coolObject') as m:
doSomething()
# Whatever you want to assert here.
assert m.called
The path you use can include properties on objects, e.g. module1.module2.MyClass.my_class_method. A big gotcha is that you need to mock the object in the module being tested, not where it is defined. In the example above, that means using a path of foo.coolObject and not somwhere.else.coolObject.
Related
I have one function like this one:
def function(df, path_write):
df['A'] = df['col1'] * df['col2']
write(df, path)
The function is not that simple but the question is, how can i make a unit test if the function do not return any value??
If the function returns the new df it's simple, just make:
assert_frame_equal from the library from pandas.testing import assert_frame_equal and mock the write method.
But without that return, how can i test the df line??
In general, I can think of only two kinds of functions: Those that return a value and those that produce side-effects.
With the second kind, you typically don't want the side-effects to actually happen during testing. For example, if your function writes something to disk or sends some data to some API on the internet, you usually don't want it to actually do that during the test, you only want to ensure that it attempts to do the right thing.
To roll with the example of disk I/O as a side-effect: You would usually have some function that does the actual writing to the filesystem that the function under testing calls. Let's say it is named write. The typical apporach would be to mock that write function in your test. Then you would need to verify that that mocked write was called with the arguments you expected.
Say you have the following code.py for example:
def write(thing: object, path: str) -> None:
print("Some side effect like disk I/O...")
def function(thing: object, file_name: str) -> None:
...
directory = "/tmp/"
write(thing, path=directory + file_name)
To test function, I would suggest the following test.py:
from unittest import TestCase
from unittest.mock import MagicMock, patch
from . import code
class MyTestCase(TestCase):
#patch.object(code, "write")
def test_function(self, mock_write: MagicMock) -> None:
test_thing = object()
test_file_name = "test.txt"
self.assertIsNone(code.function(test_thing, test_file_name))
mock_write.assert_called_once_with(
test_thing,
path="/tmp/" + test_file_name,
)
Check out unittest.mock for more details on mocking with the standard library. I would strongly advise to use the tools there and not do custom monkey-patching. The latter is certainly possible, but always carries the risk that you forget to revert the patched objects back to their original state after every test. That can break the entire rest of your test cases and depending on how you monkey-patched, the source of the resulting errors may become very hard to track down.
Hope this helps.
In the mock for write() you can add assert statements to ensure the form of df is as you would expect. For example:
def _mock_write(df, path):
assert path == '<expected path value>'
assert_frame_equal(df, <expected dataframe>)
So the full test case would be:
def test_function(self, monkeypatch):
# Define mock function
def _mock_write(df, path):
assert path == '<expected path value>'
assert_frame_equal(df, <expected dataframe>)
# Mock write function
monkepyatch.setattr(<MyClass>, 'write', _mock_write)
# Run function to enter mocked write function
function(test_df, test_path_write)
N.B. This is assuming you are using pytest as your test runner which supports the set up and tear down of monkeypatch. Other answers show the usage for the standard unittest framework.
Let's say I have the following modules:
# src/myapp/utils.py
import thirdparty
from myapp.secrets import get_super_secret_stuff
def get_thirdparty_client() -> return thirdparty.Client:
thirdparty.Client(**get_super_secret_stuff())
# src/myapp/things.py
from myapp.utils import get_thirdparty_client
from myapp.transformations import apply_fairydust, make_it_sparkle
def do_something(x):
thirdparty_client = get_thirdparty_client()
y = thidparty_client.query(apply_fairydust(x))
return make_it_sparkle(y)
Assume that myapp is lightly-tested legacy code, and refactoring is out of the question. Also assume (annoyingly) that thirdparty.Client does non-deterministic network I/O in its __init__ method. Therefore I intend to mock the thirdparty.Client class itself so as to make this do_something function testable.
Assume also that I must use unittest and cannot use another test framework like Pytest.
It seems like the patch function from unittest.mock is the right tool for the job. However, I'm unsure of how to apply the usual admonition to "patch where it is used."
Ideally I want to write a test that looks something like this:
# tests/test_myapp/test_things.py
from unittest import TestCase
from unittest.mock import patch
from myapp.things import do_something
def gen_test_pairs():
# Generate pairs of expected inputs and outputs
...
class ThingTest(unittest.TestCase):
#patch('????')
def test_do_something(self, mock_thirdparty_client):
for x, y in gen_test_pairs:
with self.subTest(params={'x': x, 'y': y}):
mock_thirdparty_client.query.return_value = y
self.assertEqual(do_something(x), y)
My problem is that I don't know what to write in place of the ????, because I never actually import thirdparty.Client in src/myapp/things.py.
Options I considered:
Apply the patch at myapp.utils.thirdparty.Client, which makes my test fragile and dependent on implementation details.
"Break the rules" and apply the patch at thirdparty.Client.
Import get_thirdparty_client in the test, use patch.object on it, and set its return_value to another MagicMock that I create separately, and this second mock would stand in for thirdparty.Client. This makes for more verbose testing code that can't easily be applied as a single decorator.
None of these options sounds particularly appealing, but I don't know which is considered the least bad.
Or is there another option available to me that I am not seeing?
The correct answer is 2: apply the patch to thirdparty.Client.
This is correct because it does NOT in fact break the "patch where it is used" rule.
The rule is intended to be descriptive, not literal. In this case, thirdparty.Client is considered to be "used" in the thirdparty module.
This concept is described in more detail in the 2018 PyCon talk "Demystifying the Patch Function" by Lisa Roach. The full recording is available on YouTube here: https://youtu.be/ww1UsGZV8fQ?t=537. The explanation of this particular case starts at approximately 9 minutes into the video.
I haven't been able to found a good explanation of this in the net, I'm guessing that i'm missing something trivial but I haven't been able to find it, so I came here to ask the experts :)
I have a test were I need to patch a constructor call, reading the docs as I understand, something like this should work, but:
import unittest.mock as mocker
import some_module
mock_class1 = mocker.patch('some_module.some_class')
print(mock_class1 is some_module.some_class) # Returns False
print(mock_class1) # <unittest.mock._patch>
mock_instance1 = mock_class1.return_value # _patch object has no attr return_value
Instead I get a different output if I do this
with mocker.patch('some_module.some_class') as mock_class2:
print(mock_class2 is some_module.some_class) # Returns True
print(mock_class2) # <MagicMock name=...>
mock_instance2 = mock_class2.return_value # No problem
print(mock_instance2) # <NonCallableMagicMock name=...>
Now, for the test itself, i'm using pytest-mock module which gives a mocker fixture that behaves like the first code block.
I would like to know:
why the behavior differs depending on the way that one call the mock framework
is there a clean way to trigger the behavior of the second code block without the with clause?
1) pytest mocker plugin is being developer to avoid use of context managers; and probably not everybody fond of the way the standard mock plays with functions parameter 2) not really. It is intended to be used either as content manager or function decorator.
I think it is possible use mocker package without pytest
References
https://github.com/pytest-dev/pytest-mock
https://www.packtpub.com/mapt/book/application_development/9781847198846/5/ch05lvl1sec45/integrating-with-python-mocker
What about installing pytest-mock and creating a test like this
import itertools
def test1(mocker):
mock_class1 = mocker.patch('itertools.count')
print(mock_class1 is itertools.count)
print(mock_class1)
mock_instance1 = mock_class1.return_value # Magic staff...
or may be using monkeypatching? Just do not use the standard unittest.mock with pytest
I want to test a function in python, but it relies on a module-level "private" function, that I don't want called, but I'm having trouble overriding/mocking it. Scenario:
module.py
_cmd(command, args):
# do something nasty
function_to_be_tested():
# do cool things
_cmd('rm', '-rf /')
return 1
test_module.py
import module
test_function():
assert module.function_to_be_tested() == 1
Ideally, in this test I dont want to call _cmd. I've looked at some other threads, and I've tried the following with no luck:
test_function():
def _cmd(command, args):
# do nothing
pass
module._cmd = _cmd
although checking module._cmd against _cmd doesn't give the correct reference. Using mock:
from mock import patch
def _cmd_mock(command, args):
# do nothing
pass
#patch('module._cmd', _cmd_mock)
test_function():
...
gives the correct reference when checking module._cmd, although `function_to_be_tested' still uses the original _cmd (as evidenced by it doing nasty things).
This is tricky because _cmd is a module-level function, and I dont want to move it into a module
[Disclaimer]
The synthetic example posted in this question works and the described issue become from specific implementation in production code. Maybe this question should be closed as off topic because the issue is not reproducible.
[Note] For impatient people Solution is at the end of the answer.
Anyway that question given to me a good point to thought: how we can patch a method reference when we cannot access to the variable where the reference is?
Lot of times I found some issue like this. There are lot of ways to meet that case and the commons are
Decorators: the instance we would like replace is passed as decorator argument or used in decorator static implementation
What we would like to patch is a default argument of a method
In both cases maybe refactor the code is the best way to play with that but what about if we are playing with some legacy code or the decorator is a third part decorator?
Ok, we have the back on the wall but we are using python and in python nothing is impossible. What we need is just the reference of the function/method to patch and instead of patching its reference we can patch the __code__: yes I'm speaking about patching the bytecode instead the function.
Get a real example. I'm using default parameter case that is simple, but it works either in decorator case.
def cmd(a):
print("ORIG {}".format(a))
def cmd_fake(a):
print("NEW {}".format(a))
def do_work(a, c=cmd):
c(a)
do_work("a")
cmd=cmd_fake
do_work("b")
Output:
ORIG a
ORIG b
Ok In this case we can test do_work by passing cmd_fake but there some cases where is impossible do it: for instance what about if we need to call something like that:
def what_the_hell():
list(map(lambda a:do_work(a), ["c","d"]))
what we can do is patch cmd.__code__ instead of _cmd by
cmd.__code__ = cmd_fake.__code__
So follow code
do_work("a")
what_the_hell()
cmd.__code__ = cmd_fake.__code__
do_work("b")
what_the_hell()
Give follow output:
ORIG a
ORIG c
ORIG d
NEW b
NEW c
NEW d
Moreover if we want to use a mock we can do it by add follow lines:
from unittest.mock import Mock, call
cmd_mock = Mock()
def cmd_mocker(a):
cmd_mock(a)
cmd.__code__=cmd_mocker.__code__
what_the_hell()
cmd_mock.assert_has_calls([call("c"),call("d")])
print("WORKS")
That print out
WORKS
Maybe I'm done... but OP still wait for a solution of his issue
from mock import patch, Mock
cmd_mock = Mock()
#A closure for grabbing the right function code
def cmd_mocker(a):
cmd_mock(a)
#patch.object(module._cmd,'__code__', new=cmd_mocker.__code__)
test_function():
...
Now I should say never use this trick unless you are with the back on the wall. Test should be simple to understand and to debug ... try to debug something like this and you will become mad!
Say I want to test this module:
import osutils
def check_ip6(xml):
ib_output = osutils.call('iconfig ib0')
# process and validate ib_output (to be unit tested)
...
This method is dependent on the environment, because it makes a System call (which expects a specific network interface), so its not callable on a testmachine.
I want to write a Unit test for that method which checks if the processing of ib_output works as expected. Therefore I want to mock osutils.call and let it just return testdata. What is the preferred way to do that? Do I have to do mocking or (monkey) patching?
Example test:
def test_ib6_check():
from migration import check_ib6
# how to mock os_utils.call used by the check_ib6-method?
assert check_ib6(test_xml) == True
One solution would be to do from osutils import call and then when patching things replace yourmodule.call with something else before calling test_ib6_check.
Ok, I found out that this has nothing to do with mocks, afaik I just need a monkey patch: I need to import and change the osutils.call-method and then import the method under test (and NOT the whole module, since it would then import the original call-method) afterwards. So this method will then use my changed call-method:
def test_ib6_check():
def call_mock(cmd):
return "testdata"
osutils.call = call_mock
from migration import check_ib6
# the check_ib6 now uses the mocked method
assert check_ib6(test_xml) == True