Pytest Framework

Write cleaner, more powerful tests with pytest — plain assert, parametrize for many cases, and fixtures for shared setup.

"pytest's magic: just write functions starting with test_, use plain assert, and run pytest. No classes, no self, no special assert methods. Less code, clearer intent."

— Shurai

pytest vs unittest — Side by Side

The same test written both ways. See how much cleaner pytest is:

unittest

import unittest
from math_utils import add

class TestAdd(unittest.TestCase):
    def test_add(self):
        self.assertEqual(
            add(2, 3), 5
        )

if __name__ == "__main__":
    unittest.main()

pytest

from math_utils import add

def test_add():
    assert add(2, 3) == 5

8 lines → 2 lines. Same result.

terminal

pip install pytest

Writing pytest Tests

python — test_math.py

import pytest
from math_utils import add, divide, is_even

# Plain functions — no class needed
def test_add_positive():
    assert add(2, 3) == 5

def test_add_negative():
    assert add(-1, -1) == -2

def test_divide_normal():
    assert divide(10, 2) == 5.0

def test_divide_by_zero():
    with pytest.raises(ValueError, match="zero"):
        divide(5, 0)
    # match="zero" checks the error message contains "zero"

def test_is_even():
    assert     is_even(4)
    assert not is_even(7)

terminal — running tests

pytest                          # discover + run all test_*.py files
pytest -v                       # verbose — shows each test name
pytest test_math.py             # one specific file
pytest test_math.py::test_add   # one specific test
pytest -x                       # stop at the first failure

pytest's Superpower — Better Error Messages

When a test fails, pytest shows you the exact values, not just "AssertionError":

output when a test fails

FAILED test_math.py::test_add_positive
_____ test_add_positive ______

    def test_add_positive():
>       assert add(2, 3) == 6      # expected 6 but got 5

E       AssertionError: assert 5 == 6
E        +  where 5 = add(2, 3)

1 failed in 0.05s

Parametrize — One Test, Many Cases

Instead of writing a separate test function for every input, use @pytest.mark.parametrize to run the same test with many values automatically:

python

import pytest
from math_utils import add

@pytest.mark.parametrize("a, b, expected", [
    (1,   2,   3),     # positive numbers
    (0,   0,   0),     # zeros
    (-5,  5,   0),     # negative + positive
    (100, 200, 300),   # large numbers
])
def test_add_many(a, b, expected):
    assert add(a, b) == expected

# pytest runs this function 4 times, once per row
# test_add_many[1-2-3]     PASSED
# test_add_many[0-0-0]     PASSED
# test_add_many[-5-5-0]    PASSED
# test_add_many[100-200-300] PASSED

Fixtures — Reusable Setup

A fixture is setup code that multiple tests share. Decorated with @pytest.fixture, it runs before each test that requests it:

python

import pytest

@pytest.fixture
def sample_scores():
    """Returns a fresh list of scores for each test."""
    return [85, 92, 78, 96, 65]

# Each test receives sample_scores by listing it as a parameter
def test_average(sample_scores):
    assert sum(sample_scores) / len(sample_scores) == 83.2

def test_max_score(sample_scores):
    assert max(sample_scores) == 96

def test_passing(sample_scores):
    # how many scored 75 or above?
    assert sum(1 for s in sample_scores if s >= 75) == 4

pytest vs unittest — when to use which

Use pytest when:

Writing new projects
You want parametrize
Better failure messages matter
Most professional codebases

Use unittest when:

No pip install allowed
Existing unittest codebase
Standard library only
Embedded/restricted environments

"pytest can run unittest tests too — so you never have to choose. Start with pytest on new projects. If you inherit unittest tests, just run pytest against them and they'll work."

— Shurai

🧠 Quiz — Q1

What is the minimum requirement for pytest to discover and run a test function?

🧠 Quiz — Q2

What is the big advantage of pytest's error messages over unittest?

🧠 Quiz — Q3

What does @pytest.mark.parametrize("a, b, result", [(1,2,3),(0,0,0)]) do?

🧠 Quiz — Q4

What is a pytest fixture?