# Pytest introduction (technical talk) --- Do we need to test our software? ---- Not properly tested software is a poor quality software. Poor quality software damage company reputation (in a very optimistic scenario). :-1::-1::-1::-1::money_with_wings: :money_with_wings: :money_with_wings::money_with_wings: ---- And now to get some intuition what you can feel about testing. Imaging you're building the rocket with your team. Now you're going to flight to the Mars in this rocket. Are you ok to sit there without ensuring that EVERYTHING was tested? Sorry for the exaggeration:) --- Pytest is a software testing framework. https://docs.pytest.org/en/stable/ ---- :hearts: Pytest is awesome: :hearts: - It is the most popular testing framework in the world In according with 2020 python developer [survey](https://www.jetbrains.com/lp/python-developers-survey-2020/):  And the adoption is very rapid. Starting from 2017 pytest is strongly dominating over other frameworks ---- except the percent of people who is not testing their software at all. Those percent was always big :rolling_on_the_floor_laughing: --- ## Technical part --- Installation: ```bash $ pip install pytest ``` ---- Project tree ``` ├── README.md ├── poetry.lock ├── pyproject.toml # pytest configuration ├── src │   └── wlb_bot │   ├── __init__.py │   ├── __main__.py │   └── app.py └── tests # place for tests ├── __init__.py ├── conftest.py # fixtures ├── handlers │   └── test_start.py # pytest automatically discovering tests in nested dirs └── test_app.py # testing module ``` ---- How does test look like? ```python= def test_basic(): assert True ``` --- ## How execute tests From top of my head the most frequently used commands are: ```bash= $ pytest # to run the pytest session $ pytest --collect-only # to list collected tests $ pytest -k 'some pattern or expression here' # to run tests matching the pattern $ pytest --lf # run last failed tests only $ pytest --pdb # fall into python debugger if test fails $ pytest /some/path # run tests in the /some/path ``` ---- Otherwise check `pytest --help` or https://docs.pytest.org --- ## How to write tests ---- Open terminal: cd $CODE/talks/ --- ## Interesting things about fixtures ---- A lot of nice examples are in the documentation (https://docs.pytest.org/en/stable/fixture.html), like: - Fixtures as factories - Parametrizing fixtures (by adding a fixture your test will be parametrized) https://github.com/oamg/convert2rhel/blob/main/convert2rhel/unit_tests/conftest.py#L64 - Test parametrization https://docs.pytest.org/en/stable/parametrize.html#pytest-mark-parametrize-parametrizing-test-functions --- ## Pytest markers ---- Pytest markers helps you categorize your tests into groups and then decide which groups to run. ```python= @pytest.mark.webtest def test_send_http(): pass ``` ```bash= $ pytest -m webtest ``` Will execute only tests marked as `webtest` ---- Usecases: - Runnig tests in different environments (some group in container, some in VM etc.) - Running long duration tests separately ---- There are builtin markers: ```pytest @pytest.mark.skip() @pytest.mark.skipif(condition) ``` --- ## Pytest plugins ---- You can see it as a pytest fixture provisioned from another package. ---- Most popular plugins for me: - pytest-cov (coverage report) - pytest-xdist (test parallelization, isolation, remote execution) - pytest-asyncio (for running coroutines in easy way) - pytest-sugar (for beautiful test session reports) - pytest-mock (pytest friendly Mock with nice reports) - pexpect More on https://github.com/augustogoulart/awesome-pytest --- ## Pytest configuration ---- pyproject.toml or pytest.ini check https://docs.pytest.org/en/stable/customize.html ---- Example: ``` [tool.pytest.ini_options] minversion = "6.0.0" addopts = "-vv -s --tb=native" testpaths = "tests/" log_cli = true log_cli_level = "DEBUG" log_cli_format = "| %(asctime)s | %(name)s | %(levelname)s | %(filename)s | %(message)s" ``` --- Talk is available at: https://github.com/zhukovgreen/talks/blob/main/pytest-generic.md --- ## :question: Questions :question: --- Thank you for your attention! --- with :heart: to Python https://zhukovgreen.pro