Back to Details

unit-testing

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Unit Testing

单元测试

Test isolated business logic with fast, deterministic tests.
使用快速、可预测的测试来测试独立的业务逻辑。

AAA Pattern (Arrange-Act-Assert)

AAA模式（准备-执行-断言）

typescript
describe('calculateDiscount', () => {
  test('applies 10% discount for orders over $100', () => {
    // Arrange
    const order = { items: [{ price: 150 }] };

    // Act
    const result = calculateDiscount(order);

    // Assert
    expect(result).toBe(15);
  });
});
typescript
describe('calculateDiscount', () => {
  test('applies 10% discount for orders over $100', () => {
    // Arrange
    const order = { items: [{ price: 150 }] };

    // Act
    const result = calculateDiscount(order);

    // Assert
    expect(result).toBe(15);
  });
});

Test Isolation

测试隔离

typescript
describe('UserService', () => {
  let service: UserService;
  let mockRepo: MockRepository;

  beforeEach(() => {
    // Fresh instances per test
    mockRepo = createMockRepository();
    service = new UserService(mockRepo);
  });

  afterEach(() => {
    // Clean up
    vi.clearAllMocks();
  });
});
typescript
describe('UserService', () => {
  let service: UserService;
  let mockRepo: MockRepository;

  beforeEach(() => {
    // Fresh instances per test
    mockRepo = createMockRepository();
    service = new UserService(mockRepo);
  });

  afterEach(() => {
    // Clean up
    vi.clearAllMocks();
  });
});

Coverage Targets

覆盖率目标

AreaTarget
Business logic90%+
Critical paths100%
New features100%
Utilities80%+
领域目标
业务逻辑90%+
关键路径100%
新功能100%
工具类80%+

Parameterized Tests

参数化测试

typescript
describe('isValidEmail', () => {
  test.each([
    ['test@example.com', true],
    ['invalid', false],
    ['@missing.com', false],
    ['user@domain.co.uk', true],
  ])('isValidEmail(%s) returns %s', (email, expected) => {
    expect(isValidEmail(email)).toBe(expected);
  });
});
typescript
describe('isValidEmail', () => {
  test.each([
    ['test@example.com', true],
    ['invalid', false],
    ['@missing.com', false],
    ['user@domain.co.uk', true],
  ])('isValidEmail(%s) returns %s', (email, expected) => {
    expect(isValidEmail(email)).toBe(expected);
  });
});

Python Example

Python示例

python
import pytest

class TestCalculateDiscount:
    def test_applies_discount_over_threshold(self):
        # Arrange
        order = Order(total=150)

        # Act
        discount = calculate_discount(order)

        # Assert
        assert discount == 15

    @pytest.mark.parametrize("total,expected", [
        (100, 0),
        (101, 10.1),
        (200, 20),
    ])
    def test_discount_thresholds(self, total, expected):
        order = Order(total=total)
        assert calculate_discount(order) == expected
python
import pytest

class TestCalculateDiscount:
    def test_applies_discount_over_threshold(self):
        # Arrange
        order = Order(total=150)

        # Act
        discount = calculate_discount(order)

        # Assert
        assert discount == 15

    @pytest.mark.parametrize("total,expected", [
        (100, 0),
        (101, 10.1),
        (200, 20),
    ])
    def test_discount_thresholds(self, total, expected):
        order = Order(total=total)
        assert calculate_discount(order) == expected

Fixture Scoping (2026 Best Practice)

Fixture作用域（2026最佳实践）

python
import pytest
python
import pytest

Function scope (default): Fresh instance per test - ISOLATED

Function scope (default): Fresh instance per test - ISOLATED

@pytest.fixture(scope="function") def db_session(): """Each test gets clean database state.""" session = create_session() yield session session.rollback() # Cleanup
@pytest.fixture(scope="function") def db_session(): """Each test gets clean database state.""" session = create_session() yield session session.rollback() # Cleanup

Module scope: Shared across all tests in file - EFFICIENT

Module scope: Shared across all tests in file - EFFICIENT

@pytest.fixture(scope="module") def expensive_model(): """Load once per test file (expensive setup).""" return load_large_ml_model() # 5 seconds to load
@pytest.fixture(scope="module") def expensive_model(): """Load once per test file (expensive setup).""" return load_large_ml_model() # 5 seconds to load

Session scope: Shared across ALL tests - MOST EFFICIENT

Session scope: Shared across ALL tests - MOST EFFICIENT

@pytest.fixture(scope="session") def db_engine(): """Single connection pool for entire test run.""" engine = create_engine(TEST_DB_URL) Base.metadata.create_all(engine) yield engine Base.metadata.drop_all(engine)

**When to use each scope:**
| Scope | Use Case | Example |
|-------|----------|---------|
| function | Isolated tests, mutable state | db_session, mock objects |
| module | Expensive setup, read-only | ML model, compiled regex |
| session | Very expensive, immutable | DB engine, external service |
@pytest.fixture(scope="session") def db_engine(): """Single connection pool for entire test run.""" engine = create_engine(TEST_DB_URL) Base.metadata.create_all(engine) yield engine Base.metadata.drop_all(engine)

**何时使用各作用域：**
| 作用域 | 使用场景 | 示例 |
|-------|----------|---------|
| function | 独立测试、可变状态 | db_session、模拟对象 |
| module | 高成本初始化、只读 | ML模型、编译后的正则表达式 |
| session | 极高成本初始化、不可变 | 数据库引擎、外部服务 |

Indirect Parametrization

间接参数化

python
undefined
python
undefined

Defer expensive setup from collection to runtime

Defer expensive setup from collection to runtime

@pytest.fixture def user(request): """Create user with different roles based on parameter.""" role = request.param # Receives value from parametrize return UserFactory(role=role)
@pytest.mark.parametrize("user", ["admin", "moderator", "viewer"], indirect=True) def test_permissions(user): """Test runs 3 times with different user roles.""" # user fixture is called with each role assert user.can_access("/dashboard") == (user.role in ["admin", "moderator"])
@pytest.fixture def user(request): """Create user with different roles based on parameter.""" role = request.param # Receives value from parametrize return UserFactory(role=role)
@pytest.mark.parametrize("user", ["admin", "moderator", "viewer"], indirect=True) def test_permissions(user): """Test runs 3 times with different user roles.""" # user fixture is called with each role assert user.can_access("/dashboard") == (user.role in ["admin", "moderator"])

Combinatorial testing with stacked decorators

Combinatorial testing with stacked decorators

@pytest.mark.parametrize("role", ["admin", "user"]) @pytest.mark.parametrize("status", ["active", "suspended"]) def test_access_matrix(role, status): """Runs 4 tests: admin/active, admin/suspended, user/active, user/suspended""" user = User(role=role, status=status) expected = (role == "admin" and status == "active") assert user.can_modify() == expected
undefined
@pytest.mark.parametrize("role", ["admin", "user"]) @pytest.mark.parametrize("status", ["active", "suspended"]) def test_access_matrix(role, status): """Runs 4 tests: admin/active, admin/suspended, user/active, user/suspended""" user = User(role=role, status=status) expected = (role == "admin" and status == "active") assert user.can_modify() == expected
undefined

Key Decisions

关键决策

DecisionRecommendation
FrameworkVitest (modern), Jest (mature), pytest
Execution< 100ms per test
DependenciesNone (mock everything external)
Coverage toolc8, nyc, pytest-cov
决策推荐方案
框架Vitest（现代）、Jest（成熟）、pytest
执行时间每个测试<100ms
依赖无（模拟所有外部依赖）
覆盖率工具c8、nyc、pytest-cov

Common Mistakes

常见错误

  • Testing implementation, not behavior
  • Slow tests (external calls)
  • Shared state between tests
  • Over-mocking (testing mocks not code)
  • 测试实现细节而非行为
  • 测试速度慢（包含外部调用）
  • 测试间共享状态
  • 过度模拟（测试模拟对象而非实际代码）

Related Skills

相关技能

  • integration-testing
    - Testing interactions
  • msw-mocking
    - Network mocking
  • test-data-management
    - Fixtures and factories
  • integration-testing
    - 测试交互逻辑
  • msw-mocking
    - 网络模拟
  • test-data-management
    - 测试数据与工厂类

Capability Details

能力细节

pytest-patterns

pytest-patterns

Keywords: pytest, python, fixture, parametrize Solves:
  • Write pytest unit tests
  • Use fixtures effectively
  • Parametrize test cases
关键词： pytest、python、fixture、parametrize 解决问题：
  • 编写pytest单元测试
  • 高效使用fixture
  • 参数化测试用例

vitest-patterns

vitest-patterns

Keywords: vitest, jest, typescript, mock Solves:
  • Write Vitest unit tests
  • Mock dependencies
  • Test React components
关键词： vitest、jest、typescript、mock 解决问题：
  • 编写Vitest单元测试
  • 模拟依赖项
  • 测试React组件

orchestkit-strategy

orchestkit-strategy

Keywords: orchestkit, strategy, coverage, pyramid Solves:
  • OrchestKit test strategy example
  • Test coverage targets
  • Testing pyramid ratios
关键词： orchestkit、strategy、coverage、pyramid 解决问题：
  • OrchestKit测试策略示例
  • 测试覆盖率目标
  • 测试金字塔比例

test-case-template

test-case-template

Keywords: template, test, structure, arrange Solves:
  • Test case template
  • Arrange-Act-Assert structure
  • Copy-paste test starter
关键词： template、test、structure、arrange 解决问题：
  • 测试用例模板
  • 准备-执行-断言结构
  • 可复制的测试起始模板