testing-python

name	testing-python
description	Python testing with pytest: fixtures, parametrize, mocking with unittest.mock, and coverage

Testing — Python

Framework

• pytest for all projects. Avoid unittest for new code.
• Run: pytest / pytest --cov=src --cov-report=term-missing

File Conventions

• Test files: tests/test_<module>.py or co-located src/utils/test_format.py.
• Conftest: shared fixtures in tests/conftest.py.

Test Structure

def test_format_date_returns_iso_string_for_valid_date():
    # Arrange
    date = datetime(2026, 1, 15)

    # Act
    result = format_date(date)

    # Assert
    assert result == "2026-01-15"

Fixtures

# conftest.py
@pytest.fixture
def db_session():
    session = create_test_session()
    yield session
    session.rollback()
    session.close()

# test file
def test_user_created(db_session):
    user = create_user(db_session, name="Alice")
    assert user.id is not None

Parametrize

@pytest.mark.parametrize("value,expected", [
    ("", True),
    ("  ", True),
    ("hello", False),
])
def test_is_blank(value, expected):
    assert is_blank(value) == expected

Mocking

from unittest.mock import patch, MagicMock

def test_sends_email_on_signup(mock_smtp):
    with patch("myapp.mailer.smtplib.SMTP") as mock_smtp:
        send_welcome_email("[e-mail]")
        mock_smtp.return_value.sendmail.assert_called_once()

Coverage Targets

• Business logic: 85%+ branch coverage.
• Run pytest --cov in CI; fail below threshold with --cov-fail-under=80.