Testing in Python

Unit testing is the practice of writing automated tests for individual components, typically functions or methods, in isolation from the rest of the system. Each test verifies a specific behavior under specific inputs. Unit tests catch regressions early and document expected behavior.

Unittest is the built-in testing framework, modeled on JUnit. Test cases are methods in classes derived from TestCase. It provides assertion methods, test discovery, fixtures via setUp and tearDown, and a command-line runner. It is reliable but verbose compared to modern alternatives.

Pytest is a popular third-party testing framework that uses plain functions and the assert keyword instead of special assertion methods. It has a powerful fixture system, parametrized tests, plugins, and excellent error reporting. As of 2026 it is the dominant testing tool in the Python ecosystem.

Fixtures are functions decorated with pytest.fixture that provide shared setup for tests. A test that takes a fixture name as a parameter receives the fixture's return value. Fixtures can have scopes from function up to session and can yield to provide cleanup. They replace setUp and tearDown methods.

Parametrized testing runs the same test function with different inputs using the pytest.mark.parametrize decorator. Each set of parameters becomes a separate test case in the report. This avoids duplicating test code for similar scenarios.

Mocking is the practice of replacing real dependencies with controllable substitutes during a test. The unittest.mock module provides Mock and MagicMock classes and the patch context manager. Mocking lets you isolate the unit under test from databases, network services, and other slow or unreliable components.

Mock is the basic mock object that records calls and supports attribute access. MagicMock is a subclass that also provides default implementations of dunder methods, so it can be used in contexts that expect features like iteration, comparison, or context management.

Monkey patching is the practice of modifying or extending objects at runtime, typically to substitute behavior for testing. Pytest provides the monkeypatch fixture for safe, scoped patching. Direct monkey patching outside tests is generally considered fragile and should be used sparingly.

Test coverage measures the percentage of code executed during tests. Tools like coverage.py and pytest-cov generate reports showing which lines and branches are tested. High coverage is necessary but not sufficient; covered code can still have bugs. The goal is meaningful tests, not just numbers. ---