Testing with pytest

Introduction

pytest is a simple yet powerful testing framework for Python. It encourages concise, readable test code and scales to support complex functional testing for applications and libraries. It provides expressive assertions, auto-discovery, and a rich fixture system, making it ideal for both small and large projects.

Key Features

Rich assert introspection (no need for self.assert* methods)
Auto-discovery of test files, classes, and functions
Flexible fixture model for shared test setup and teardown
Compatible with unittest and nose test suites
Thriving plugin ecosystem (e.g., pytest-cov, pytest-mock)
Seamless integration with pre-commit and CI tools

Installation

Use uv to add pytest as a development dependency:

uv add --dev pytest

Configuration

Configure pytest through pyproject.toml:

[tool.pytest.ini_options]
testpaths = ["tests"]
markers = [
    "slow: marks tests as slow",
]

This instructs pytest to discover tests in the tests/ directory and register a custom slow marker.

Usage

1. Writing Tests

import pytest
from src.temperature import celsius_to_fahrenheit, fahrenheit_to_celsius

@pytest.mark.parametrize(
    "celsius, expected_fahrenheit",
    [(0, 32.0), (100, 212.0), (-40, -40.0), (37, 98.6), (25.5, 77.9)],
)
def test_celsius_to_fahrenheit(celsius, expected_fahrenheit):
    assert celsius_to_fahrenheit(celsius) == pytest.approx(expected_fahrenheit)

@pytest.mark.parametrize(
    "fahrenheit, expected_celsius",
    [(32, 0.0), (212, 100.0), (-40, -40.0), (98.6, 37.0), (77.9, 25.5)],
)
def test_fahrenheit_to_celsius(fahrenheit, expected_celsius):
    assert fahrenheit_to_celsius(fahrenheit) == pytest.approx(expected_celsius)

2. Fixtures

@pytest.fixture
def sample_data():
    return {"name": "Sunil", "age": 30}

def test_sample_data(sample_data):
    assert sample_data["name"] == "Sunil"

3. Markers

Use markers to group or filter tests:

@pytest.mark.slow
def test_large_dataset():
    ...

Run only slow tests:

pytest -m slow

Register custom markers in pyproject.toml:

[tool.pytest.ini_options]
markers = [
    "slow: marks tests as slow",
]

4. Common CLI Options

-v: Verbose mode
-q: Quiet mode
-k <expr>: Filter by name substring
-m <marker>: Run tests with marker
–maxfail=<N>: Stop after N failures
–disable-warnings: Suppress warnings

Example:

pytest -v -k "fahrenheit" --maxfail=2 --disable-warnings

5. Naming Conventions

pytest automatically discovers tests that match:

Files named test_*.py or *_test.py
Functions prefixed with test_
Classes prefixed with Test (no __init__ method)

6. Running Tests

Run all tests using:

uv run pytest

7. Integration with pre-commit

Add the following hook to .pre-commit-config.yaml:

- repo: local
  hooks:
    - id: test
      name: Run tests
      entry: uv run pytest
      language: python
      args: ["--maxfail=1", "--disable-warnings", "-q"]
      always_run: true
      pass_filenames: false

This ensures that your test suite runs before every commit.

Additional Resources

Pytest documentation: https://docs.pytest.org/
Parametrization guide: https://docs.pytest.org/en/latest/how-to/parametrize.html

Next Step

Now that testing is set up, the next step is to integrate test coverage using the coverage tool. This helps ensure your tests are exercising all critical parts of your code.

Uninstall

To remove pytest:

uv remove --dev pytest