Testing: Why use pytest instead of unittest?

[Joe-Heffer-Shef]

My question: what are the advantages of using pytest rather than the inbuilt unittest library?

Does it help implement testing best practices and methodologies?

Why would I install pytest when unittest is built in to the core libraries?

What can I do with pytest that I can't do with other libraries?

Pytest claims:

The pytest framework makes it easy to write small, readable tests, and can scale to support complex functional testing for applications and libraries.

unittest claims:

The unittest unit testing framework was originally inspired by JUnit and has a similar flavor as major unit testing frameworks in other languages. It supports test automation, sharing of setup and shutdown code for tests, aggregation of tests into collections, and independence of the tests from the reporting framework.

What do these concepts mean?

• test fixture
• test case
• test suite
• test runner

[ns-rse]

I use pytest but only because its the framework used when I learnt about writing tests. I've never used unittest to be able to give a comparison.

Features I like about pytest...

• Plugins, there are lots, for coverage pytest-cov, Matplotlib pytest-mpl, regression testing pytest-regtest, parallel testing pytest-xdist and more.
• Test fixtures which provide consistent context/input for tests, whether these are instantiated classes with certain properties which are reused across tests or parameterisation.
• Simple parameterisation using @pytest.mark.parametrize() fixture where you define various inputs and expected outputs which are passed into the test.

I'd be curious to find out more about unittest but having learnt and being familiar with pytest would be reluctant to switch (acquired knowledge and learning curve both being barriers to doing so).

My understanding is that a "test suite" is simply an existing set of tests for a given set of code.

Not sure about "test runner" the pytest-runner extension/plugin is deprecated but appears to be a package that runs pytest automatically on a package. No idea of its meaning in other frameworks.

[ns-rse]

Pytest Plugins for Flask testing...

• pytest-flask
• jeancochrane/pytest-flask-sqlalchemy: A pytest plugin for preserving test isolation in Flask-SQLAlchemy using database transactions. (nee pytest-flask-sqlalchemy-transactions)

I've never used Flask so can't comment on how useful they are for testing but still love using pytest for my test development and use various plugins (of which there are lots) such as pytest-regtest (although they've an issue concerning dependency on py which I suggested a solution for over a year ago but hasn't been accepted or closed), pytest-mpl and pytest-github-actions-annotate-failures.

[ns-rse]

Conditional coverage if you only want to calculate these statistics under certain conditions.

[ns-rse]

Just in time...

Pytest vs. Unittest: Which Is Better? | The PyCharm Blog

[Joe-Heffer-Shef]

Pytest Parameterisation by Neil https://blog.nshephard.dev/posts/pytest-param/

[Joe-Heffer-Shef]

Parallel test execution with pytest-xdist https://pytest-xdist.readthedocs.io/

[ns-rse]

Lots of useful Pytest extensions including...

• pytest-regtest for regression testing.
• pytest-dependency managing dependencies of tests.
• pytest-order for defining the order of tests.

[ns-rse]

Additional ways of testing.

• Hypothesis
• doctest write examples in your docstrings and have them checked.

[tdjames1]

Fixtures and mocking

• Fixture only needs to be instantiated once then can use it elsewhere
• Mocking allows simulation of server type responses
• A lot of web frameworks have testing setup
• End to end testing using selenium

https://docs.python.org/3/library/unittest.html#class-and-module-fixtures

Parameterisation

• Subtests in unittest https://docs.python.org/3/library/unittest.html#distinguishing-test-iterations-using-subtests
• Parameterisation in unittest not as neat as in pytest

Keeping tests small

• Run locally first

Order of evaluation

• https://pytest-ordering.readthedocs.io/en/develop/
• https://pypi.org/project/pytest-order/
• https://pypi.org/project/pytest-dependency/

Regression testing

Regression test plugin https://gitlab.com/uweschmitt/pytest-regtest

https://github.com/CITCOM-project/CausalTestingFramework

"Regression testing is a common technique to implement basic testing before refactoring legacy code that lacks a test suite."

Other testing frameworks are available

https://hypothesis.readthedocs.io/en/latest/ https://docs.python.org/3/library/doctest.html

[twinkarma]

Covered