Master Python Mocking: From Basics to Advanced Unit Test Techniques

Why Mocking Is Essential

When writing Python unit tests you often encounter problems such as calling real payment gateways, missing databases in CI, unstable external APIs, or nondeterministic time functions. The root cause is a lack of isolation, and using mocks and stubs resolves these issues.

Mock vs Stub

Stub provides predefined responses, returning fixed values to replace external services. Mock records calls, arguments, and allows assertions about interactions. In unittest.mock, a Mock object can act as both a stub and a mock.

Core unittest.mock Objects

Mock – a generic stand‑in; MagicMock – supports magic methods; patch – dynamically replaces objects (the most commonly used tool).

Basic Usage Examples

Examples show creating a Mock, setting return values, asserting calls, using MagicMock for iterables, and patching functions with @patch.

Five Practical Scenarios

Mock third‑party HTTP requests to avoid real network calls: @patch('payment.requests.post').

Stub the datetime function to make time controllable in reports.

Mock SQLite connections to test database logic without a real DB.

Simulate network errors and verify exception handling using side_effect.

Mock class instantiation to replace __init__ and verify method calls.

Advanced Techniques

Use spec to restrict Mock attributes and catch misspellings.

Use patch as a context manager for temporary, local mocks.

Mock chained calls (common in ORMs) by setting return values on successive mock objects.

Common Pitfalls

Typical mistakes include using the wrong import path in patch, over‑mocking internal logic, ignoring side_effect, not asserting call behavior, and polluting global state by replacing module attributes directly.

Conclusion

Mocking focuses tests on the code under test, while stubbing answers “if X is returned, does my code work?”. Mastering unittest.mock enables fast, reliable, parallelizable tests that run in CI without external services, covering success and failure branches.

Actionable Advice

Open a test file, locate real external calls, replace them with @patch, and add assert_called* checks to verify interactions, turning your tests into isolated experiments.