Mock Generator
A tool to generate the basic mocks and asserts for faster unit testing.
🎉
New: you can now use pytest-mock-generator, for more fluid pytest development. See documentation later in this doc.
Introduction
A typical unit test looks like this (AAA pattern):
- Arrange – Setup and prepare the various objects and prerequisites.
- Act – Invoke the tested code by calling the tested function.
- Assert – Verify the outcome. This can be the return value of the tested function and/or some side effects.
When using mocks, much time is wasted on the wiring. The Arrange and Assert sections are notorious in that regard. Only a fraction of the time is spent on the actual logic of the test.
This tool is meant to save you time, by generating the Arrange and Assert sections for you. The generated code can then be used as is, or altered according to your needs.
Usage
Note: All examples assume usage of the pytest-mock which is a fixture for pytest.
Getting Started
Let's assume you have a module named tested_module.py which holds a function to process a string sent to it and then add it to a zip file:
import zipfile
def process_and_zip(zip_path, file_name, contents):
processed_contents = "processed " + contents # some complex logic
with zipfile.ZipFile(zip_path, 'w') as zip_container:
zip_container.writestr(file_name, processed_contents)
This is the unit under test, or UUT.
Although this is a very short function, writing the test code takes a lot of time. It's the fact that the function uses a context manager makes the testing more complex than it should be. If you don't believe me, try to manually write mocks and asserts which verify that zip_container.writestr was called with the right parameters.
In any case, you start with a test skeleton:
import mock_autogen
from tests.sample.code.tested_module import process_and_zip
def test_process_and_zip(mocker):
# Arrange: todo
# Act: invoking the tested code
process_and_zip('/path/to.zip', 'in_zip.txt', 'foo bar')
# Assert: todo
Now it's time to use Mock Generator instead of manually writing the 'Arrange' and 'Assert' sections.
Generating the 'Arrange' section
To generate the 'Arrange' section, simply put this code at the beginning of your test function skeleton and run it:
mock_autogen.generate_uut_mocks(process_and_zip)
This will generate the 'Arrange' section for you:
# mocked dependencies
mock_ZipFile = mocker.MagicMock(name='ZipFile')
mocker.patch('tests.sample.code.tested_module.zipfile.ZipFile', new=mock_ZipFile)
The generated code is returned, printed to the console and also copied to the clipboard for your convenience. Just paste it (as simple as ctrl+V) at the start of your test function:
import mock_autogen
from tests.sample.code.tested_module import process_and_zip
def test_process_and_zip(mocker):
# mocked dependencies
mock_ZipFile = mocker.MagicMock(name='ZipFile')
mocker.patch('tests.sample.code.tested_module.zipfile.ZipFile', new=mock_ZipFile)
# Act: invoking the tested code
process_and_zip('/path/to.zip', 'in_zip.txt', 'foo bar')
# Assert: todo
Excellent! Arrange section is ready.
Generating the Assert section
Now it's time to add the asserts. Add the following code at the 'Assert' step:
mock_autogen.generate_asserts(mock_ZipFile)
The mock_ZipFile is the mock object you generated earlier. Now execute the test function to get the assert section:
assert 1 == mock_ZipFile.call_count
mock_ZipFile.assert_called_once_with('/path/to.zip', 'w')
mock_ZipFile.return_value.__enter__.assert_called_once_with()
mock_ZipFile.return_value.__enter__.return_value.writestr.assert_called_once_with('in_zip.txt', 'processed foo bar')
mock_ZipFile.return_value.__exit__.assert_called_once_with(None, None, None)
Wow, that's a handful of asserts! Some are very useful:
- Checking that we opened the zip file with the right parameters.
- Checking that we wrote the correct data to the proper file.
- Finally, ensuring that
__enter__and__exit__are called, so there are no open file handles which could cause problems.
You can remove any generated line which you find unnecessary.
Paste that code right after the act phase, and you're done!
The complete test function:
from tests.sample.code.tested_module import process_and_zip
def test_process_and_zip(mocker):
# mocked dependencies
mock_ZipFile = mocker.MagicMock(name='ZipFile')
mocker.patch('tests.sample.code.tested_module.zipfile.ZipFile', new=mock_ZipFile)
# Act: invoking the tested code
process_and_zip('/path/to.zip', 'in_zip.txt', 'foo bar')
assert 1 == mock_ZipFile.call_count
mock_ZipFile.assert_called_once_with('/path/to.zip', 'w')
mock_ZipFile.return_value.__enter__.assert_called_once_with()
mock_ZipFile.return_value.__enter__.return_value.writestr.assert_called_once_with('in_zip.txt', 'processed foo bar')
mock_ZipFile.return_value.__exit__.assert_called_once_with(None, None, None)
Can you imagine the time it would have taken you to code this on your own?
What's Next
Using pytest-mock-generator
You can now use pytest-mock-generator to generate mocks:
pip install pytest-mock-generator
Then use mg fixture in your tests, instead of having to import mock_autogen, like so:
from tests.sample.code.tested_module import process_and_zip
def test_process_and_zip(mocker, mg):
# Arrange: todo
# replace with the output of the below command
mg.generate_uut_mocks(process_and_zip)
# Act: invoking the tested code
process_and_zip('/path/to.zip', 'in_zip.txt', 'foo bar')
# Assert: todo
Asserting Existing Mocks
At times, you may be editing a test code already containing mocks, or you choose to write the mocks yourself, to gain some extra control.
Mock Generator can generate the assert section for standard Python mocks, even if they were not created using the Mock Generator.
Put this after the 'Act' (replace mock_obj with your mock object name):
import mock_autogen
mock_autogen.generate_asserts(mock_obj)
Take the generated code and paste it at the 'Assert' section.
Generating the 'Arrange' and 'Assert' sections in one call
You can make the generate_uut_mocks_with_asserts call create the generate_asserts code for you (instead of having to call generate_uut_mocks):
import mock_autogen
mock_autogen.generate_uut_mocks_with_asserts(function_under_test)
Mocking Everything
So far you have seen examples of mocking the dependencies of a single function. Mock Generator can generate mocks for objects, classes and entire modules!
A great way to learn about those capabilities is to see them in action:
from mock_autogen import PytestMocker
PytestMocker(a_module_or_a_class_you_test_and_want_to_mock_its_dependencies).mock_everything().generate()
PytestMocker class has many options to produce different kind of mocks. See its documentation for further details.
Wrapping up
I hope that you were convinced that this tool can save you a lot of time.
If you have improvement suggestions, bug reports, or would like to contribute pull requests, let me know.
1 Jan 03, 2022
2 Aug 12, 2021
10 Aug 02, 2022
736 Dec 16, 2022
19 Jan 04, 2023
2.2k Jan 04, 2023
2k Jan 06, 2023
40 Dec 16, 2022
1.9k Dec 30, 2022
756 Dec 22, 2022
2 Jan 02, 2022
34 Dec 27, 2022
502 Dec 30, 2022
1.6k Dec 30, 2022
3 Mar 04, 2022
548 Dec 28, 2022
9 Feb 10, 2022
15 Dec 15, 2022
1 Nov 22, 2021
2 Jun 17, 2022