Testing Python Code with pytest: Fixtures and Parametrize

pytest is the standard testing tool for Python. Tests are plain functions using plain assert, and pytest's output shows exactly what differed.

First tests

# test_pricing.py
import pytest
from pricing import apply_discount

def test_ten_percent_discount():
    assert apply_discount(1000, percent=10) == 900

def test_discount_cannot_be_negative():
    with pytest.raises(ValueError, match="percent"):
        apply_discount(1000, percent=-5)

uv run pytest -q
uv run pytest -k discount -x      # only matching tests, stop at first failure

Parametrize: many cases, one test

@pytest.mark.parametrize(
    ("price", "percent", "expected"),
    [
        (1000, 0, 1000),
        (1000, 25, 750),
        (999, 10, 899),     # rounding down
        (0, 50, 0),
    ],
)
def test_apply_discount(price, percent, expected):
    assert apply_discount(price, percent=percent) == expected

Fixtures: reusable setup

# conftest.py - fixtures here are available to every test file
import pytest

@pytest.fixture
def db():
    conn = create_test_database()
    yield conn                 # the test runs here
    conn.close()               # teardown

@pytest.fixture
def user(db):
    return db.users.create(email="test@example.com")

# test_orders.py
def test_new_user_has_no_orders(db, user):
    assert db.orders.for_user(user.id) == []

Built-in fixtures

def test_writes_report(tmp_path):
    out = tmp_path / "report.csv"
    write_report(out)
    assert out.read_text().startswith("id,total")

def test_uses_api_key(monkeypatch):
    monkeypatch.setenv("API_KEY", "test-key")
    assert load_config().api_key == "test-key"

Coverage

uv add --dev pytest-cov
uv run pytest --cov=src --cov-report=term-missing

Aim coverage at the code that matters (business rules, edge cases), rather than chasing 100%.