pytest-randomly 5.0.0

Features

All of these features are on by default but can be disabled with flags.

• Randomly shuffles the order of test items. This is done first at the level of modules, then at the level of test classes (if you have them), then at the order of functions. This also works with things like doctests.

• Generates a base random seed or accepts one for reproduction with --randomly-seed. The base random seed is printed at the start of the test run, and can be passed in to repeat a failure caused by test ordering or random data.

• At the start of the test run, and before each test setup, run, and teardown, it resets Python’s global random seed to a fixed value, using random.seed(). The fixed value is derived from the base random seed, the pytest test ID, and an offset for setup or teardown. This ensures each test gets a different but repeatable random seed.

• pytest-randomly also resets several libraries’ random states at the start of every test, if they are installed:

  • factory boy

  • Faker

    The faker pytest fixture is also affected, as pytest-randomly defines the faker_seed fixture.

  • Model Bakery

  • NumPy

    Only its legacy random state is affected.

  • Polyfactory

    Only the default random generator is affected. Factories that have called seed_random() with their own seed are unaffected.

• If additional random generators are used, they can be registered under the pytest_randomly.random_seeder entry point and will have their seed reset at the start of every test. Register a function that takes the current seed value.

• Works with pytest-xdist.

About

Randomness in testing can be quite powerful to discover hidden flaws in the tests themselves, as well as giving a little more coverage to your system.

By randomly ordering the tests, the risk of surprising inter-test dependencies is reduced - a technique used in many places, for example Google’s C++ test runner googletest. Research suggests that “dependent tests do exist in practice” and a random order of test executions can effectively detect such dependencies [1]. Alternatively, a reverse order of test executions, as provided by pytest-reverse, may find less dependent tests but can achieve a better benefit/cost ratio.

By resetting the random seed to a repeatable number for each test, tests can create data based on random numbers and yet remain repeatable, for example factory boy’s fuzzy values. This is good for ensuring that tests specify the data they need and that the tested system is not affected by any data that is filled in randomly due to not being specified.

I have written a blog post covering the history of pytest-randomly, including how it started life as the nose plugin nose-randomly.

Additionally, I appeared on the Test and Code podcast to talk about pytest-randomly.

Installation

Install with:

python -m pip install pytest-randomly

Python 3.10 to 3.15 supported.

Usage

Pytest will automatically find the plugin and use it when you run pytest. The output will start with an extra line that tells you the random seed that is being used:

$ pytest
...
platform darwin -- Python ...
Using --randomly-seed=1553614239
...

If the tests fail due to ordering or randomly created data, you can restart them with that seed using the flag as suggested:

pytest --randomly-seed=1234

Or more conveniently, use the special value last:

pytest --randomly-seed=last

(This only works if pytest’s cacheprovider plugin has not been disabled.)

Since the ordering is by module, then by class, you can debug inter-test pollution failures by narrowing down which tests are being run to find the bad interaction by rerunning just the module/class:

pytest --randomly-seed=1234 tests/module_that_failed/

You can disable behaviours you don’t like with the following flags:

• --randomly-dont-reset-seed - turn off the reset of random.seed() at the start of every test

• --randomly-dont-reorganize - turn off the shuffling of the order of tests

The plugin appears to Pytest with the name ‘randomly’. To disable it altogether, you can use the -p argument, for example:

pytest -p no:randomly

Entry Point

If you’re using a different randomness generator in your third party package, you can register an entrypoint to be called every time pytest-randomly reseeds. Implement the entrypoint pytest_randomly.random_seeder, referring to a function/callable that takes one argument, the new seed (int).

For example in your setup.cfg:

[options.entry_points]
pytest_randomly.random_seeder =
    mypackage = mypackage.reseed

Then implement reseed(new_seed).

References