Mocking Demo

Demo code for a study group session with Boston Python on the topic of mocking

This repository includes an example Python package and associated code for running tests with pytest. In the main branch, the test code is incomplete to allow people to experiment with completing it using unittest.mock. This code is intended for educational purposes only.

References

• Python documentation for unittest.mock: https://docs.python.org/3/library/unittest.mock.html
• Open-Meteo weather api: https://open-meteo.com/

Example Python Package

The example Python package outfit-picker uses openmeteo-requests to request a weather forecast with the maximum temperature for the day. Then some basic logic is used to pick clothing items based on the predicted temperature.

There are two version of the code, one written using a class (classes.py) and one written using functions (functions.py) to enable practice using unittest.mock with both styles of code.

Tests

Setup

To run the tests, it is recommended to first clone this repository, then install the Python package and dev requirements in a virtual environment.

With your desired virtual environment activated, install the Python package from source and install the dev requirements.

pip install .
pip install -r dev-requirements.txt

After the dependencies are installed, the test code can by run using pytest. Note: The tests on the main branch arte not expected to pass.

pytest tests/

Test Code

The test code in the main branch is incomplete so that you can experiment with completing it.

• test_exception_handling has one test intended to verify that the code goes down the expected path when the request to openmeteo fails with the error requests.exceptions.ConnectionError.
• test_outfit_logic_classes has several tests that verify that the output changes depending on the max temperature, using the version of the outfit-picker code that was written using a class. This file has many tests that do essentially the same thing in different ways.
• test_outfit_logic_functions is the same as the above except using the version of the code written using functions.
• test_parameters has tests to verify that a function is gettting called with the expected arguments.

Although there is no single "right" answer for completing the code, there is a version in the completed branch of the git repository where all tests should pass. To get this version of the code, run git checkout completed.

To supplement the test code, here are some questions to consider.

1. Could the outfit-picker code have been written differently so that the code to test the outfit logic (in test_outfit_logic_classes.py and/or test_outfit_logic_functions.py) did not need to use mocking at all? How?
2. In test_exception_handling, unittest.mock.patch is used on sys.stdout. What does that do?