Testing Patterns

Apply these patterns when writing or reviewing tests.

Structure: Arrange-Act-Assert (AAA)

Every test should follow the AAA pattern with clear visual separation:

def test_user_creation_with_valid_data():
    # Arrange
    user_data = {"email": "test@example.com", "name": "Test User"}

    # Act
    user = UserService.create(user_data)

    # Assert
    assert user.email == "test@example.com"
    assert user.is_active is True

it("creates a user with valid data", () => {
  // Arrange
  const userData = { email: "test@example.com", name: "Test User" };

  // Act
  const user = UserService.create(userData);

  // Assert
  expect(user.email).toBe("test@example.com");
  expect(user.isActive).toBe(true);
});

Naming Convention

Test names should describe the scenario and expected outcome:

• Python: test_<unit>_<scenario>_<expected_result>
• TypeScript: "<unit> <scenario> <expected result>"

Good: test_login_with_invalid_password_returns_401 Bad: test_login, test_case_1

Test Data

Use Factories (not fixtures for mutable data)

• Python: factory_boy or custom factory functions
• TypeScript: builder pattern or factory functions
• Never hardcode IDs or timestamps -- use factories to generate them
• Share setup via factory defaults, override per-test as needed

Isolation

• Each test must be independent -- no shared mutable state
• Use fresh database transactions per test (Django: TestCase with automatic rollback)
• Mock external services at the boundary (HTTP, email, queues)

What to Test

Must Test

• Happy path for every public function/endpoint
• Error/edge cases: null inputs, empty collections, boundary values
• Authorization: ensure protected routes reject unauthorized access
• Validation: verify invalid input is rejected with correct errors

Don't Test

• Private/internal methods directly (test through public interface)
• Framework code (Django ORM, React rendering engine)
• Trivial getters/setters with no logic

Coverage Expectations

• New code: aim for 80%+ line coverage
• Critical paths (auth, payments, data mutations): 95%+
• Don't chase 100% -- focus on meaningful assertions over line coverage

Mocking Guidelines

• Mock at boundaries: HTTP clients, databases, file systems, clocks
• Prefer dependency injection over patching
• Verify mock interactions only when the side effect IS the behavior
• Don't mock the thing you're testing