easily run python unit tests

commandline, python, python-unittest
pip install pyt==1.0.5



Pyt is a lightweight wrapper around Python's unittest module that adds some nice features and enhancements over the stock unittest module.


Pyt overrides unittest's built-in pathfinding to be smarter and less verbose, so you can match tests using prefix matching which makes running a test like:

$ python -m unittest tests.foo_test.BarTestCase.test_che

as simple as:

$ pyt foo.Bar.che

But it's even less verbose if you want it to be, pyt can reach into the modules and classes to do its matching, so you don't even need to specify the module and class if you don't want to:

$ pyt che

More examples

To run all the Happy tests:

$ pyt Happy

To run all the newmodule tests:

$ pyt newmodule

To run more than one test:

$ pyt test1 test2 ...

To run every test pyt can find:

$ pyt

And the way I like to run all tests in the current directory:

$ pyt -vb

Which can also be written:

$ pyt --verbose --buffer


To see everything pyt can do

$ pyt --help


This will convert warnings into errors.

$ pyt --warnings


If your last testrun had failing tests this will rerun only the tests that failed.

$pyt --rerun

Things to be aware of

pyt uses Python's PEP 8 style conventions

pyt uses Python's code styling conventions to decide what is the module and class, so, given input like this:

$ pyt foo.bar.Baz.che

pyt will consider foo.bar to be the module, Baz to be a class (because it starts with a capital letter), and che to be a method (since it comes after the class).

Likewise, pyt uses unittest conventions, so a test module should end with _test (eg, foo.bar_test) or start with test (eg, test_foo.py) and a TestCase class should extend unittest.TestCase, and test methods should start with test_ (eg, test_che).

Vague input can cause pyt to run more tests than you expect

So if you have something like this:


And you want to run tests for foo.user and you run:

$ pyt user

it will run both tests/user_test and tests.foo.user_test, the solution is to just be more verbose when you have to be:

$ pyt foo.user

Environment Variables

If you are running the tests within pyt, you might notice there is an environment variable PYT_TEST_COUNT that contains the count of how many tests pyt found to run.


Use pip:

$ pip install pyt

You can also get it directly from the repo:

$ pip install --upgrade git+https://github.com/Jaymon/pyt#egg=pyt


Testing in 3.5 on MacOS:

$ python3.10 -m unittest pyt_test

Or, if you're really brave, you can use pyt to test itself:

$ python -m pyt tests -df