Testing in Python

1. Overview#

Tests are automated checks that verify your code behaves correctly — catching regressions before they reach production. Python's standard library ships unittest, but the ecosystem has largely standardized on pytest for its simpler syntax and powerful fixture system.

---

2. Getting Started with pytest#

2.1 Installation & Basic Test#

bash
Ask AICopy
pip install pytest

Python
CopyAsk AI
# file: calculator.py
def add(a, b):
    return a + b

    def divide(a, b):
        if b == 0:
            raise ValueError("Cannot divide by zero")
        return a / b

Python
CopyAsk AI
# file: test_calculator.py
from calculator import add, divide


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

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

        def test_divide_by_zero_raises():
            import pytest

            with pytest.raises(ValueError):
                divide(10, 0)

bash
Ask AICopy
pytest # run all tests in the current directory
pytest test_calculator.py # run a specific file
pytest -v # verbose output
pytest -k "add" # run only tests matching "add"

---

3. Assertions#

pytest uses plain assert statements and rewrites them to give detailed failure output automatically.

Python
CopyAsk AI
def test_examples():
    assert 2 + 2 == 4
    assert "hello".upper() == "HELLO"
    assert [1, 2, 3] == [1, 2, 3]
    assert 5 in [1, 2, 3, 4, 5]
    assert isinstance(5, int)

When a test fails, pytest shows exactly what was compared:

code
Ask AICopy
 def test_add():
> assert add(2, 3) == 6
E assert 5 == 6

---

4. Fixtures — Reusable Test Setup#

A fixture provides data or a resource that tests need, with automatic setup and teardown.

Python
CopyAsk AI
import pytest


@pytest.fixture
def sample_users():
    return [
        {"name": "Asha", "age": 29},
        {"name": "Ravi", "age": 34},
    ]


def test_user_count(sample_users):
    assert len(sample_users) == 2


def test_first_user_name(sample_users):
    assert sample_users[0]["name"] == "Asha"

4.1 Fixtures with Teardown (Setup/Cleanup)#

Python
CopyAsk AI
import pytest


@pytest.fixture
def database_connection():
    conn = connect_to_test_db()  # setup
    yield conn  # test runs here, using `conn`
    conn.close()  # teardown — runs after the test finishes


def test_query(database_connection):
    result = database_connection.execute("SELECT 1")
    assert result is not None

4.2 Fixture Scope#

Python
CopyAsk AI
@pytest.fixture(scope="function")  # default — runs fresh for every test
def per_test_fixture(): ...


@pytest.fixture(scope="module")  # runs once per test file
def per_module_fixture(): ...


@pytest.fixture(scope="session")  # runs once for the entire test run
def per_session_fixture(): ...

---

5. Parametrized Tests#

Run the same test logic against multiple inputs without duplicating code.

Python
CopyAsk AI
import pytest
from calculator import add


@pytest.mark.parametrize(
    "a, b, expected",
    [
        (2, 3, 5),
        (-1, 1, 0),
        (0, 0, 0),
        (100, 200, 300),
    ],
)
def test_add_parametrized(a, b, expected):
    assert add(a, b) == expected

---

6. Mocking — Isolating Code from External Dependencies#

Mocking replaces a real dependency (API call, database, file system) with a fake, controllable stand-in — so tests stay fast and don't depend on external systems.

Python
CopyAsk AI
from unittest.mock import Mock, patch


def get_weather(api_client, city):
    response = api_client.fetch(city)
    return response["temperature"]


def test_get_weather():
    mock_client = Mock()
    mock_client.fetch.return_value = {"temperature": 28}

    result = get_weather(mock_client, "Bengaluru")

    assert result == 28
    mock_client.fetch.assert_called_once_with("Bengaluru")

6.1 patch — Replacing a Real Function/Module Temporarily#

Python
CopyAsk AI
# file: weather_service.py
import requests


def get_temperature(city):
    response = requests.get(f"https://api.weather.com/{city}")
    return response.json()["temperature"]

Python
CopyAsk AI
# file: test_weather_service.py
from unittest.mock import patch
from weather_service import get_temperature


@patch("weather_service.requests.get")  # patches requests.get only inside this module
def test_get_temperature(mock_get):
    mock_get.return_value.json.return_value = {"temperature": 30}

    result = get_temperature("Bengaluru")

    assert result == 30

---

7. Testing Exceptions#

Python
CopyAsk AI
import pytest
from calculator import divide


def test_divide_by_zero():
    with pytest.raises(ValueError, match="Cannot divide by zero"):
        divide(10, 0)

---

8. Organizing a Test Suite#

Architecture & Data Flow
Copy
my_project/
├── src/
│ └── calculator.py
├── tests/
│ ├── conftest.py # shared fixtures, auto-discovered by pytest
│ ├── test_calculator.py
│ └── test_weather.py
└── pytest.ini # pytest configuration

Python
CopyAsk AI
# conftest.py — fixtures here are available to ALL test files automatically
import pytest


@pytest.fixture
def sample_data():
    return {"key": "value"}

ini
Ask AICopy
# pytest.ini
[pytest]
testpaths = tests
addopts = -v --strict-markers

---

9. Measuring Coverage#

bash
Ask AICopy
pip install pytest-cov
pytest --cov=src --cov-report=term-missing

This reports what percentage of your source code is actually exercised by tests, and flags exactly which lines were missed.

---

10. Common Pitfalls#

INCORRECT: Tests That Depend on Each Other's Order#

Python
CopyAsk AI
counter = 0


def test_increment():
    global counter
    counter += 1
    assert counter == 1  # breaks if this test doesn't run first

CORRECT: Fix — Each Test Should Be Independent#

Use fixtures to reset state before every test instead of relying on shared globals or execution order.

INCORRECT: Testing Implementation Details Instead of Behavior#

Python
CopyAsk AI
def test_internal_list_used():
    calc = Calculator()
    calc.add(2, 3)
    assert calc._history == [("add", 2, 3)]  # brittle — breaks on refactor

CORRECT: Fix — Test the Public Behavior#

Python
CopyAsk AI
def test_add_returns_correct_result():
    calc = Calculator()
    assert calc.add(2, 3) == 5

---

11. Summary & Best Practices Checklist#

• Name test files test_*.py and test functions test_* so pytest auto-discovers them.
• Keep each test independent — no shared mutable state between tests.
• Use fixtures for setup/teardown instead of repeating code in every test.
• Use @pytest.mark.parametrize instead of copy-pasting near-identical tests.
• Mock external dependencies (APIs, databases, file systems) — tests should be fast and deterministic.
• Patch names where they're used, not where they're defined.
• Test behavior/outputs, not internal implementation details.
• Track coverage, but don't chase 100% blindly — focus on meaningful paths and edge cases.