Python Peeps: For Unit Testing when do you prefer to use Unittest, Doctest, Pytest ?

Examples of Unittest , Doctest and Pytest are listed bellow-respectively- if you aren't familiar with them.

# Testing Framework:
# Unittest supports test automation by writing classes.

import unittest


def fib(n):
    """Return the n-th number in the fibonacci series."""
    if n <= 0:
        return 0
    elif n == 1:
        return 1
    return fib(n - 1) + fib(n - 2)


class TestFib(unittest.TestCase):

    def test_fib_series(self):
        self.assertEqual(fib(0), 0)
        self.assertEqual(fib(1), 1)
        self.assertEqual(fib(2), 1)
        self.assertEqual(fib(3), 2)
        self.assertEqual(fib(4), 3)
        self.assertEqual(fib(5), 5)


if __name__ == "__main__":
    # normally this is unittest.main()...this is for jupyter
    unittest.main(argv=['first-arg-is-ignored'], exit=False, verbosity=2)

# Testing Framework:
# Doctest allows you to write tests in the docstrings

def fib(n):
    """Return the n-th number in the fibonacci series.

    >>> fib(0)
    0
    >>> fib(1)
    1
    >>> fib(29)
    514229
    """
    if n <= 0:
        return 0
    elif n == 1:
        return 1
    return fib(n - 1) + fib(n - 2)


if __name__ == "__main__":
    import doctest

    # doctest.testmod()
    doctest.testmod(verbose=True)

#Testing framework pytest
# Test discovery: for doctests, unittests and pytests.

import pytest


%%run_pytest[clean] -vvv

def fib(n):
    """Return the n-th number in the fibonacci series."""
    if n <= 0:
        return 0
    elif n == 1:
        return 1
    return fib(n-1) + fib(n-2)

def test_fib_series():
    assert fib(0) == 0
    assert fib(1) == 1
    assert fib(2) == 1
    assert fib(3) == 2
    assert fib(4) == 3
    assert fib(5) == 5

Comments

[rhymes]

My favorite is pytest with fixtures, they make it really easy to share code and data between tests, using dependency injection instead of encapsulating everything in unittest.TestCase: docs.pytest.org/en/latest/fixture....

[Pierre Bouillon]

I'm fine with unittest.
Putting my tests in different classes allows me to organize a bit more my code by correctly splitting what needs to (package > classes > methods) whether it is about the core or the tests.

I think documentation needs to... Well, document my methods/functions and that tests doesn't belong to it (single responsibility for my docstring?). By writting methods with unittest I can event document them in a more precise way without having 40 lines of text per method :)

And finally, for pytest, I just think unittest allow me more wrapping around the structure of my tests

Just my opinion tho, hope it helps!