pytest Intro — Less Boilerplate

• Install pytest and run it with one command.
• Write tests as plain functions using plain assert.
• Read pytest's rich failure output.
• See side-by-side why pytest beats unittest on ergonomics.

pip install pytest

# unittest — needs a class and self.assertEqual
import unittest
class TestAdd(unittest.TestCase):
    def test_add(self):
        self.assertEqual(add(2, 3), 5)

# pytest — just a function and a plain assert
def test_add():
    assert add(2, 3) == 5

A pytest test file

# test_mathtools.py
from mathtools import is_prime, factorial

def test_is_prime():
    assert is_prime(17)
    assert not is_prime(4)

def test_factorial():
    assert factorial(5) == 120

def test_factorial_negative():
    import pytest
    with pytest.raises(ValueError):
        factorial(-1)

Run it

$ pytest
========================= test session starts =========================
collected 3 items

test_mathtools.py ...                                            [100%]

========================== 3 passed in 0.01s ==========================

The killer feature: assert rewriting

def test_factorial():
    assert factorial(5) == 119      # wrong on purpose

    def test_factorial():
>       assert factorial(5) == 119
E       assert 120 == 119
E        +  where 120 = factorial(5)

test_mathtools.py:7: AssertionError

From a plain assert, pytest shows the actual value (120), the expected (119), AND the call that produced it — without you writing any message. unittest needs assertEqual to get half of that.

Common commands

pytest                  run all test_*.py in the tree
pytest test_bank.py     one file
pytest -v               verbose: one line per test
pytest -k "discount"    only tests matching a keyword
pytest -x               stop at first failure
pytest -q               quiet
pytest test_bank.py::test_overdraw   one specific test

It runs unittest too

pytest discovers and runs your existing unittest.TestCase classes as well — so you can adopt it gradually without rewriting everything.

Here's the Lesson 7 bank suite rewritten in pytest — notice how much shrinks:

# test_bank.py (pytest version)
import pytest
from bank import BankAccount

def test_starts_at_zero():
    assert BankAccount().balance == 0

def test_deposit_increases_balance():
    acc = BankAccount()
    acc.deposit(100)
    assert acc.balance == 100

def test_withdraw_decreases_balance():
    acc = BankAccount(100)
    acc.withdraw(40)
    assert acc.balance == 60

def test_overdraw_raises():
    with pytest.raises(ValueError):
        BankAccount(50).withdraw(100)

def test_negative_deposit_raises():
    with pytest.raises(ValueError, match="positive"):
        BankAccount().deposit(-5)

$ pytest test_bank.py -v
test_bank.py::test_starts_at_zero PASSED
test_bank.py::test_deposit_increases_balance PASSED
test_bank.py::test_withdraw_decreases_balance PASSED
test_bank.py::test_overdraw_raises PASSED
test_bank.py::test_negative_deposit_raises PASSED

5 passed in 0.01s

Read the diff

No class, no self, no unittest.main(). pytest.raises(ValueError, match="positive") checks both the exception AND its message in one line. The whole file is shorter and reads like plain Python. This is why nearly every modern Python project uses pytest.

pytest: plain functions, plain assert, no class needed. It rewrites assertions to show actual vs expected automatically, and pytest.raises(Exc, match=...) checks exceptions + messages in one line. It even runs your unittest classes, so adoption is gradual. Run with pytest [-v] [-k] [-x]. From here on, we use pytest. Next: the discovery rules.