Testing Components Together (Sociable Testing)

What this covers: Testing real component (class) interactions while controlling external dependencies
Time to read: ~15 minutes
Prerequisites: Unit Testing Fundamentals, Solitary Unit Tests, Test Doubles
Best for: Verifying components (classes) work together correctly, catching integration bugs while maintaining test speed

Sociable unit tests verify how components interact with their real dependencies while keeping external I/O operations under control. This approach identifies bugs in how business logic components work together - bugs that solitary tests miss.

Sociable Tests Are Unit Tests

Sociable tests are unit tests, not integration tests. They keep external I/O (databases, APIs, file systems) mocked to stay fast and side-effect-free. The integration bugs they catch are issues in how business logic components work with each other - not issues with real external systems.
For the distinction, see Unit Testing Fundamentals: Quick Reference.

Overview

This tutorial covers:

1. Setting up a sociable test with real dependencies
2. Handling external dependencies and I/O operations
3. Managing multiple dependencies using .expose()
4. Choosing when to configure mocks (before vs after compilation)
5. Using a decision framework to choose what to mock

Step 1: Set Up the First Sociable Test

This example tests a UserService that depends on an EmailValidator.

1.1 Create the Services

src/services/email-validator.ts

@Injectable()
export class EmailValidator {
  isValid(email: string): boolean {
    return email.includes('@') && email.includes('.');
  }
}

src/services/user.service.ts

@Injectable()
export class UserService {
  constructor(private validator: EmailValidator) {}

  createUser(email: string) {
    if (!this.validator.isValid(email)) {
      throw new Error('Invalid email');
    }
    return { email };
  }
}

1.2 Write the Sociable Test

The .expose() method tells Suites to use the real implementation instead of a mock.

src/services/user.service.spec.ts

import { TestBed } from '@suites/unit';
import { UserService, EmailValidator } from './services';

describe('UserService', () => {
  let userService: UserService;

  beforeAll(async () => {
    const { unit } = await TestBed.sociable(UserService)
      .expose(EmailValidator)  // Use real EmailValidator
      .compile();

    userService = unit;
  });

  it('validates email using real logic', () => {
    const result = userService.createUser('test@example.com');
    expect(result.email).toBe('test@example.com');
  });

  it('rejects invalid email', () => {
    expect(() => userService.createUser('invalid'))
      .toThrow('Invalid email');
  });
});

The real EmailValidator runs its actual validation logic. If the validator has a bug, this test will detect it.

Step 2: Handle External Dependencies

Most services interact with external systems like databases. This example extends the previous one.

2.1 Add Database Dependency

src/services/user.service.ts

@Injectable()
export class UserService {
  constructor(
    private validator: EmailValidator,
    @Inject('DATABASE') private db: DatabaseClient  // Token injection
  ) {}

  async createUser(email: string) {
    if (!this.validator.isValid(email)) {
      throw new Error('Invalid email');
    }
    return this.db.users.save({ email });
  }
}

2.2 Test with Mocked I/O

Dependencies injected using tokens are automatically mocked.

src/services/user.service.spec.ts

import { TestBed, Mocked } from '@suites/unit';

describe('UserService', () => {
  let userService: UserService;
  let database: Mocked<DatabaseClient>;

  beforeAll(async () => {
    const { unit, unitRef } = await TestBed.sociable(UserService)
      .expose(EmailValidator)
      .compile();

    userService = unit;
    database = unitRef.get<DatabaseClient>('DATABASE');
  });

  it('saves valid user', async () => {
    database.users.save.mockResolvedValue({ id: 1, email: 'test@example.com' });

    const result = await userService.createUser('test@example.com');

    expect(result.id).toBe(1);
    expect(database.users.save).toHaveBeenCalledWith({ email: 'test@example.com' });
  });
});

Third-party packages use @Inject('TOKEN') because they're not @Injectable() classes. Dependencies injected this way are always mocked automatically.

Step 3: Manage Multiple Dependencies

As services grow, you need to manage more dependencies. Here's how to handle them:

3.1 Service with Many Dependencies

src/services/order.service.ts

@Injectable()
export class OrderService {
  constructor(
    private pricingService: PricingService,
    private taxCalculator: TaxCalculator,
    private inventoryChecker: InventoryChecker,
    private discountEngine: DiscountEngine,
    @Inject('DATABASE') private db: DatabaseClient,
    @Inject('EMAIL_SERVICE') private email: EmailClient
  ) {}

  async processOrder(items: OrderItem[], region: string) {
    const subtotal = this.pricingService.calculateTotal(items);
    const tax = this.taxCalculator.calculateTax(subtotal, region);
    const total = subtotal + tax;

    const order = await this.db.orders.create({ items, subtotal, tax, total });
    await this.email.send({ to: order.customerEmail, template: 'order-confirmation' });

    return order;
  }
}

3.2 Using .expose()

List each dependency that should use its real implementation.

src/services/order.service.spec.ts

describe('OrderService', () => {
  let orderService: OrderService;
  let database: Mocked<DatabaseClient>;

  beforeAll(async () => {
    const { unit, unitRef } = await TestBed.sociable(OrderService)
      .expose(PricingService)      // Use real implementation
      .expose(TaxCalculator)       // Use real implementation
      .expose(InventoryChecker)    // Use real implementation
      .expose(DiscountEngine)      // Use real implementation
      .compile();

    orderService = unit;
    database = unitRef.get<DatabaseClient>('DATABASE');
  });

  it('processes order with real calculations', async () => {
    const items = [{ price: 10, quantity: 6 }];
    database.orders.create.mockResolvedValue({ id: 123 });

    await orderService.processOrder(items, 'US');

    // All exposed services run real code
    expect(database.orders.create).toHaveBeenCalledWith({
      items,
      subtotal: 54,    // Real discount calculation
      tax: 4.32,       // Real tax calculation
      total: 58.32
    });
  });
});

Understanding the Control Boundary

Suites can only control explicit dependencies (passed through constructors). It cannot control implicit dependencies (direct imports). For more, see Test Doubles.

Step 4: Choose When to Configure Mocks

You can configure mocks before compilation (for consistent values) or after compilation (for test-specific scenarios).

Before Compilation

When to use: Set default values that all tests in the suite need.

Best for: Shared test data, reducing repetitive setup code

src/services/user.service.spec.ts

beforeAll(async () => {
  const { unit, unitRef } = await TestBed.sociable(UserService)
    .expose(EmailValidator)
    .mock('DATABASE')
    .final({
      users: {
        save: async () => ({ id: 42, email: 'test@example.com' })
      }
    })
    .compile();

  userService = unit;
  // Database always returns the same value
});

it('creates user', async () => {
  const result = await userService.createUser('test@example.com');
  expect(result.id).toBe(42);  // Always 42
});

After Compilation

When to use: Configure different behaviors for individual tests.

Best for: Test-specific scenarios, error cases, varying responses

src/services/user.service.spec.ts

beforeAll(async () => {
  const { unit, unitRef } = await TestBed.sociable(UserService)
    .expose(EmailValidator)
    .compile();

  userService = unit;
  database = unitRef.get<DatabaseClient>('DATABASE');
});

it('handles save error', async () => {
  // Configure for this specific test
  database.users.save.mockRejectedValue(new Error('Connection failed'));

  await expect(userService.createUser('test@example.com'))
    .rejects.toThrow('Connection failed');
});

it('handles successful save', async () => {
  // Different configuration for different test
  database.users.save.mockResolvedValue({ id: 99 });

  const result = await userService.createUser('test@example.com');
  expect(result.id).toBe(99);
});

Decision Framework

Use this flowchart to decide how to handle each dependency:

Rules:

• External I/O (databases, APIs, file systems) → Use token injection → Auto-mocked
• Business logic to test → Use .expose() → Runs real code
• Everything else → Leave as default → Auto-mocked

Summary

Sociable tests work alongside solitary tests to provide comprehensive coverage:

• Solitary tests: Verify individual class behavior in isolation
• Sociable tests: Verify components work together correctly

Takeaways

• Sociable tests verify how components interact using real implementations for business logic
• External systems (I/O) are always mocked using token injection (@Inject('TOKEN'))
• Use .expose() to specify which dependencies should use real implementations
• Configure mocks before compilation for consistent values, or after for test-specific scenarios
• Follow the decision tree: External I/O → token injection, business logic → expose, everything else stays mocked

Next Steps

After understanding sociable testing, explore these resources:

• Test Doubles: Core concepts of mocking and stubbing
• Suites Examples Repository: Working examples of sociable testing patterns