Testing with pytest & Next Steps

Why pytest

pytest is the de-facto Python test framework: plain functions, plain assert, brilliant failure output. Install it and create a test file — pytest auto-discovers anything named test_*.py:

pip install pytest

# cart.py
def add_item(cart, item, price):
    if price < 0:
        raise ValueError("price cannot be negative")
    cart[item] = cart.get(item, 0) + price
    return cart

def total(cart):
    return sum(cart.values())

# test_cart.py
from cart import add_item, total

def test_add_item():
    cart = add_item({}, "course", 199)
    assert cart == {"course": 199}

def test_total_empty():
    assert total({}) == 0

pytest               # run everything
pytest -v            # verbose
pytest -k total      # only tests matching "total"

When an assert fails, pytest shows both sides of the comparison — no assertEqual boilerplate needed.

Testing Exceptions

import pytest

def test_negative_price_rejected():
    with pytest.raises(ValueError, match="negative"):
        add_item({}, "course", -5)

Parametrize — One Test, Many Cases

@pytest.mark.parametrize("items,expected", [
    ({}, 0),
    ({"a": 100}, 100),
    ({"a": 100, "b": 99}, 199),
])
def test_total(items, expected):
    assert total(items) == expected

Fixtures — Reusable Setup

@pytest.fixture
def filled_cart():
    return {"django": 199, "react": 149}

def test_total_filled(filled_cart):       # injected by name
    assert total(filled_cart) == 348

Fixtures replace setup/teardown methods: tests declare what they need as parameters, pytest builds it. Fixtures can depend on other fixtures, and yield-style fixtures handle teardown.

Habits That Make Tests Worth Having

• Test behaviour through public functions, not private internals.
• Name tests as specifications: test_negative_price_rejected.
• Keep tests fast — a slow suite stops getting run.
• Run pytest in CI on every push so broken code can’t merge.

Where to Go From Here

You now have the full core of Python: types, control flow, functions, the data structures, OOP, modules, files, exceptions, generators, decorators, environments, and tests. Pick a track and build:

• Web — the Django tutorial on this site takes you from models to production deployment.
• Data — the Data Science series (NumPy, pandas, statistics, scikit-learn).
• AI — the AI & ML series, then the LLM Engineering series.

The only way to consolidate a language is to ship something real with it. Pick a project this week and build it end to end.