Testing

Testing Best Practices

This skill provides guidance for writing user-centric tests using Vitest, Testing Library, and Nuxt Test Utils.

Core Principles

User-Centric Philosophy: Test what users see and do, not implementation details.

Never Test Implementation:

• ❌ Don't access wrapper.vm or internal component state
• ❌ Don't test props directly
• ❌ Don't query by CSS selectors or test IDs
• ❌ Don't shallow mount components
• ❌ Never include skipped tests with it.skip

Always Test Behavior:

• ✅ Query by accessibility (roles, labels, text)
• ✅ Interact like a user (click, type, keyboard)
• ✅ Assert on what users see
• ✅ Full component rendering with dependencies
• ✅ Ensure it works on Mobile and Desktop

Test Stack

Component Tests

Use @nuxt/test-utils/runtime for Nuxt components:

import { renderSuspended } from '@nuxt/test-utils/runtime'
import { screen } from '@testing-library/vue'
import userEvent from '@testing-library/user-event'
import { describe, it, expect, beforeEach } from 'vitest'
import MyComponent from './MyComponent.vue'

describe('MyComponent', () => {
  let user: ReturnType<typeof userEvent.setup>

  beforeEach(() => {
    user = userEvent.setup()
  })

  describe('form rendering', () => {
    it('should display all form fields', async () => {
      await renderSuspended(MyComponent)

      expect(screen.getByLabelText(/email/i)).toBeDefined()
      expect(screen.getByLabelText(/password/i)).toBeDefined()
      expect(screen.getByRole('button', { name: /submit/i })).toBeDefined()
    })
  })

  describe('form input', () => {
    it('should allow user to type in fields', async () => {
      await renderSuspended(MyComponent)

      const emailInput = screen.getByLabelText(/email/i) as HTMLInputElement
      await user.type(emailInput, 'test@example.com')

      expect(emailInput.value).toBe('test@example.com')
    })
  })
})

Key imports:

• renderSuspended - Render component in full Nuxt environment
• screen - Query rendered output
• userEvent - Simulate realistic user interactions
• Test file naming: *.nuxt.test.ts for Nuxt components

File Naming Convention:

• Component tests: *.nuxt.test.ts (e.g., Login.nuxt.test.ts)
• Page tests: *.nuxt.test.ts (e.g., login.nuxt.test.ts)
• Server/API tests: *.test.ts (e.g., index.get.test.ts)
• Utility tests: *.test.ts (e.g., format.test.ts)
• E2E tests: *.spec.ts (e.g., login.spec.ts)

API Tests

Test server routes using $fetch:

import { describe, it, expect, beforeAll } from 'vitest'

describe('/api/workouts', () => {
  let authCookie: string

  beforeAll(async () => {
    // Setup test user session
    const response = await $fetch('/api/auth/login', {
      method: 'POST',
      body: { email: 'test@example.com', password: 'password' }
    })
    authCookie = response.headers.get('set-cookie') || ''
  })

  it('should return user workouts', async () => {
    const workouts = await $fetch('/api/workouts', {
      headers: { cookie: authCookie }
    })

    expect(Array.isArray(workouts)).toBe(true)
    expect(workouts[0]).toHaveProperty('id')
    expect(workouts[0]).toHaveProperty('name')
  })

  it('should return 401 without auth', async () => {
    await expect($fetch('/api/workouts')).rejects.toThrow('401')
  })
})

Key points:

• Use $fetch within tests to call API routes
• Test authentication, validation, errors, and success cases
• Test file naming: *.test.ts for API tests

Database Tests

Test query functions directly:

import { describe, it, expect, beforeEach } from 'vitest'
import { db } from '~~/server/database'
import { queryUserWorkouts } from '~~/server/database/queries/workouts'
import { users } from '~~/server/database/schema'

describe('queryUserWorkouts', () => {
  let userId: string

  beforeEach(async () => {
    // Create test user
    const [user] = await db.insert(users).values({
      email: 'test@example.com',
      name: 'Test User'
    }).returning()
    userId = user.id
  })

  it('should return empty array when user has no workouts', async () => {
    const workouts = await queryUserWorkouts(userId)
    expect(workouts).toEqual([])
  })

  it('should enforce RLS - only return user's workouts', async () => {
    // Create workout for another user
    const [otherUser] = await db.insert(users).values({
      email: 'other@example.com',
      name: 'Other'
    }).returning()

    // Create workout for other user
    await db.insert(workouts).values({
      userId: otherUser.id,
      name: 'Other's Workout'
    })

    // Query should return empty for original user
    const userWorkouts = await queryUserWorkouts(userId)
    expect(userWorkouts).toEqual([])
  })
})

Key points:

• Test query functions in isolation
• Use transactions for test data
• Verify RLS policies work correctly
• Test file naming: *.test.ts for database tests

Query Priority

Use queries in this order (from Testing Library docs):

1. getByRole - Preferred for interactive elements

screen.getByRole('button', { name: /sign in/i })
screen.getByRole('link', { name: /forgot password/i })
screen.getByRole('heading', { name: /welcome/i })
screen.getByRole('textbox', { name: /email/i })

2. getByLabelText - Best for form inputs

screen.getByLabelText(/email/i)
screen.getByLabelText(/password/i)

3. getByPlaceholderText - When label isn't available

screen.getByPlaceholderText(/enter your email/i)

4. getByText - For non-interactive text

screen.getByText(/welcome back/i)
screen.getByText(/successfully logged in/i)

5. getByTestId - Last resort only (avoid if possible)

screen.getByTestId('submit-button') // Use getByRole instead

Query variants:

• getBy* - Throws if not found, throws if multiple
• queryBy* - Returns null if not found
• findBy* - Async, waits for element to appear

Regex matching: Always use case-insensitive regex for text matching:

// ✅ Good
screen.getByText(/sign in/i)
screen.getByRole('button', { name: /submit/i })

// ❌ Bad
screen.getByText('Sign In') // Breaks if text changes

Common Test Patterns

Pattern 1: Form Submission

Test user filling out and submitting a form:

describe('login form', () => {
  let user: ReturnType<typeof userEvent.setup>

  beforeEach(() => {
    user = userEvent.setup()
  })

  it('should submit form with valid credentials', async () => {
    const navigateTo = vi.fn()
    vi.stubGlobal('navigateTo', navigateTo)

    await renderSuspended(LoginForm)

    // Fill form
    await user.type(screen.getByLabelText(/email/i), 'test@example.com')
    await user.type(screen.getByLabelText(/password/i), 'password123')

    // Submit
    await user.click(screen.getByRole('button', { name: /sign in/i }))

    // Assert navigation
    await waitFor(() => {
      expect(navigateTo).toHaveBeenCalledWith('/dashboard')
    })
  })
})

Pattern 2: Interactive Elements

Test buttons, toggles, and other interactive elements:

describe('password visibility toggle', () => {
  it('should toggle password visibility', async () => {
    await renderSuspended(LoginForm)

    const passwordInput = screen.getByLabelText(/password/i) as HTMLInputElement
    const toggleButton = screen.getByRole('button', { name: /show password/i })

    expect(passwordInput.type).toBe('password')

    await user.click(toggleButton)
    expect(passwordInput.type).toBe('text')

    await user.click(toggleButton)
    expect(passwordInput.type).toBe('password')
  })
})

Pattern 3: Conditional Rendering

Test components that show/hide based on state:

it('should show success message after submission', async () => {
  await renderSuspended(ForgotPasswordForm)

  // Initially no success message
  expect(screen.queryByText(/check your email/i)).toBeNull()

  // Submit form
  await user.type(screen.getByLabelText(/email/i), 'test@example.com')
  await user.click(screen.getByRole('button', { name: /send/i }))

  // Success message appears
  await waitFor(() => {
    expect(screen.getByText(/check your email/i)).toBeDefined()
  })

  // Form is hidden
  expect(screen.queryByLabelText(/email/i)).toBeNull()
})

Pattern 4: Navigation

Test links and navigation behavior:

it('should navigate to signup page', async () => {
  await renderSuspended(LoginPage)

  const signupLink = screen.getByRole('link', { name: /create account/i })
  expect(signupLink.getAttribute('href')).toBe('/signup')
})

it('should redirect if already logged in', async () => {
  const navigateTo = vi.fn()
  vi.stubGlobal('navigateTo', navigateTo)

  // Mock logged in state
  mockNuxtImport('useUserSession', () => ({
    loggedIn: ref(true)
  }))

  await renderSuspended(LoginPage)

  await waitFor(() => {
    expect(navigateTo).toHaveBeenCalledWith('/')
  })
})

Pattern 5: Multi-Step Flows

Test wizards and multi-step forms:

it('should progress through email and OTP steps', async () => {
  await renderSuspended(EmailOtpPage)

  // Step 1: Email form visible
  expect(screen.getByLabelText(/email/i)).toBeDefined()
  expect(screen.queryByLabelText(/verification code/i)).toBeNull()

  // Submit email
  await user.type(screen.getByLabelText(/email/i), 'test@example.com')
  await user.click(screen.getByRole('button', { name: /send code/i }))

  // Step 2: OTP form visible
  await waitFor(() => {
    expect(screen.queryByLabelText(/email/i)).toBeNull()
    expect(screen.getByLabelText(/verification code/i)).toBeDefined()
  })
})

Async State Handling

Always test all four states: loading, error, empty, and success.

Loading State

it('should show loading state during submission', async () => {
  await renderSuspended(LoginForm)

  const submitButton = screen.getByRole('button', { name: /sign in/i })

  // Initially not disabled
  expect(submitButton.disabled).toBe(false)

  // Fill and submit
  await user.type(screen.getByLabelText(/email/i), 'test@example.com')
  await user.type(screen.getByLabelText(/password/i), 'password')
  await user.click(submitButton)

  // Button disabled during loading
  expect(submitButton.disabled).toBe(true)
  expect(screen.getByText(/signing in/i)).toBeDefined()
})

Error State

it('should show error message on API failure', async () => {
  // Mock API error
  vi.mocked($fetch).mockRejectedValueOnce(new Error('Invalid credentials'))

  await renderSuspended(LoginForm)

  await user.type(screen.getByLabelText(/email/i), 'test@example.com')
  await user.type(screen.getByLabelText(/password/i), 'wrong')
  await user.click(screen.getByRole('button', { name: /sign in/i }))

  await waitFor(() => {
    expect(screen.getByText(/invalid credentials/i)).toBeDefined()
  })
})

Empty State

it('should show empty state when no data', async () => {
  // Mock empty response
  vi.mocked($fetch).mockResolvedValueOnce([])

  await renderSuspended(WorkoutList)

  await waitFor(() => {
    expect(screen.getByText(/no workouts yet/i)).toBeDefined()
    expect(screen.getByRole('button', { name: /create workout/i })).toBeDefined()
  })
})

Success State

it('should display data when loaded', async () => {
  // Mock successful response
  vi.mocked($fetch).mockResolvedValueOnce([
    { id: '1', name: 'Morning Workout' },
    { id: '2', name: 'Evening Workout' }
  ])

  await renderSuspended(WorkoutList)

  await waitFor(() => {
    expect(screen.getByText(/morning workout/i)).toBeDefined()
    expect(screen.getByText(/evening workout/i)).toBeDefined()
  })
})

Mocking Strategies

Mock Nuxt Composables

import { mockNuxtImport } from '@nuxt/test-utils/runtime'

mockNuxtImport('useUserSession', () => ({
  user: ref({ id: '1', email: 'test@example.com' }),
  loggedIn: ref(true),
  clear: vi.fn()
}))

mockNuxtImport('useToast', () => ({
  message: vi.fn()
}))

Mock $fetch

import { vi } from 'vitest'

vi.mock('#app', () => ({
  $fetch: vi.fn()
}))

// In test
vi.mocked($fetch).mockResolvedValueOnce({ success: true, data: {} })

Mock navigateTo

const navigateTo = vi.fn()
vi.stubGlobal('navigateTo', navigateTo)

// Assert
expect(navigateTo).toHaveBeenCalledWith('/dashboard')

Test Organization

Group tests with describe blocks using lowercase naming:

describe('component name', () => {
  let user: ReturnType<typeof userEvent.setup>

  beforeEach(() => {
    user = userEvent.setup()
  })

  describe('form rendering', () => {
    it('should display heading', async () => { /* ... */ })
    it('should display all form fields', async () => { /* ... */ })
    it('should display submit button', async () => { /* ... */ })
  })

  describe('form input', () => {
    it('should allow user to type email', async () => { /* ... */ })
    it('should allow user to type password', async () => { /* ... */ })
  })

  describe('form validation', () => {
    it('should show error for invalid email', async () => { /* ... */ })
    it('should show error for short password', async () => { /* ... */ })
  })

  describe('form submission', () => {
    it('should submit with valid data', async () => { /* ... */ })
    it('should show error on API failure', async () => { /* ... */ })
    it('should redirect on success', async () => { /* ... */ })
  })
})

IMPORTANT: Naming Convention

• Main describe() block: Use lowercase with spaces (e.g., describe('component name', ...) or describe('page name', ...))
• Nested describe() blocks: Use lowercase with spaces (e.g., describe('form rendering', ...))
• Test descriptions: Always start with "should" and describe behavior

Common group names (lowercase):

• form rendering
• form input
• form validation
• form submission
• interactive elements
• navigation
• loading states
• error handling
• conditional rendering
• password visibility toggle
• alternative login/signup methods
• initial values

Common Pitfalls

❌ Pitfall 1: Testing Props

// Wrong
it('accepts email prop', () => {
  const wrapper = mount(Component, { props: { email: 'test@example.com' } })
  expect(wrapper.props('email')).toBe('test@example.com')
})

// Right
it('should display email in message', async () => {
  await renderSuspended(Component, { props: { email: 'test@example.com' } })
  expect(screen.getByText(/test@example\.com/i)).toBeDefined()
})

❌ Pitfall 2: Accessing Internal State

// Wrong
it('has loading state', () => {
  const wrapper = mount(Component)
  expect(wrapper.vm.isLoading).toBe(false)
})

// Right
it('should show loading spinner', async () => {
  await renderSuspended(Component)
  // Trigger loading state
  await user.click(screen.getByRole('button', { name: /submit/i }))
  expect(screen.getByRole('status', { name: /loading/i })).toBeDefined()
})

❌ Pitfall 3: Empty Smoke Tests

// Wrong
it('renders', () => {
  const wrapper = mount(Component)
  expect(wrapper.exists()).toBe(true)
})

// Right
it('should display form heading and inputs', async () => {
  await renderSuspended(Component)
  expect(screen.getByRole('heading', { name: /sign in/i })).toBeDefined()
  expect(screen.getByLabelText(/email/i)).toBeDefined()
  expect(screen.getByLabelText(/password/i)).toBeDefined()
})

❌ Pitfall 4: Using CSS Selectors

// Wrong
const button = wrapper.find('.submit-btn')
const input = wrapper.find('#email-input')

// Right
const button = screen.getByRole('button', { name: /submit/i })
const input = screen.getByLabelText(/email/i)

❌ Pitfall 5: Not Handling Async

// Wrong
it('should show success message', async () => {
  await renderSuspended(Component)
  await user.click(screen.getByRole('button'))
  expect(screen.getByText(/success/i)).toBeDefined() // May fail
})

// Right
it('should show success message', async () => {
  await renderSuspended(Component)
  await user.click(screen.getByRole('button'))
  await waitFor(() => {
    expect(screen.getByText(/success/i)).toBeDefined()
  })
})

Running Tests

Vitest (Unit & Component Tests)

# Run all tests
pnpm test

# Run tests in watch mode
pnpm test:watch

# Run tests with UI
pnpm test:ui

# Run tests with coverage
pnpm test:coverage

# Run specific test file
pnpm test path/to/file.test.ts

# Run specific project (unit or nuxt)
pnpm test --project=unit
pnpm test --project=nuxt

# Run tests matching pattern
pnpm test --grep "should submit form"

Playwright (E2E Tests)

# Run all E2E tests
pnpm test:e2e

# Run E2E tests in headed mode (browser visible)
HEADED=true pnpm test:e2e

# Run specific E2E test file
pnpm test:e2e tests/e2e/auth/login.spec.ts

# View E2E test report
pnpm exec playwright show-report

Test Configuration

Vitest Setup

• Config file: vitest.config.ts
• Setup file: vitest.setup.ts
• Two test projects:
  • unit: Node environment for server/shared utils
  • nuxt: Nuxt environment for components/pages
• Coverage: V8 provider with multiple reporters (text, json, html, clover)

Playwright Setup

• Config file: playwright.config.ts
• Test directory: ./tests/e2e
• Base URL: http://localhost:3000
• Browser: Chromium (Desktop Chrome)
• Workers: 6 locally, 1 in CI
• Retries: 0 locally, 2 in CI

Coverage Thresholds

Current thresholds (as of latest config):

• Statements: 56.08%
• Branches: 37.5%
• Functions: 53.7%
• Lines: 56.18%

Check coverage:

pnpm test:coverage

E2E Testing Patterns

Test Helpers

Use shared helper functions for common operations:

Auth Helpers (tests/helpers/auth.ts):

import { loginAsTestUser, logout, registerNewUser } from './helpers/auth'

test('should login successfully', async ({ page }) => {
  await loginAsTestUser(page, 'josephanson@hotmail.co.uk', 'Testtest1')
  await expect(page).toHaveURL(/\/feed/)
})

test('should logout successfully', async ({ page }) => {
  await loginAsTestUser(page)
  await logout(page)
  await expect(page).toHaveURL('/')
})

Test Fixtures (tests/fixtures/users.ts):

import { testUser, adminUser, newUser } from '../../fixtures/users'

test('should login with test user', async ({ page }) => {
  await loginAsTestUser(page, testUser.email, testUser.password)
})

E2E Test Structure

import { expect, test } from '@playwright/test'
import { testUser } from '../../fixtures/users'
import { loginAsTestUser } from '../../helpers/auth'

test.describe('Feature Name', () => {
  test('should perform user action', async ({ page }) => {
    // Setup
    await loginAsTestUser(page, testUser.email, testUser.password)

    // Action
    await page.getByRole('button', { name: /click me/i }).click()

    // Assert
    await expect(page).toHaveURL(/\/expected-url/)
  })
})

E2E Best Practices

1. Use Test Helpers: Reuse loginAsTestUser, logout, and other helpers
2. Use Fixtures: Import user data from tests/fixtures/users.ts
3. Wait for States: Use waitForLoadState('networkidle') and waitForURL()
4. Accessible Queries: Use getByRole, getByPlaceholder, getByText when possible
5. Timeouts: Set appropriate timeouts for async operations
6. Clean State: E2E tests should handle authentication state properly

Mocking in Component Tests

Mock $fetchResult

For components that use $fetchResult (custom fetch wrapper):

const mockFetchResult = vi.fn()
vi.stubGlobal('$fetchResult', mockFetchResult)

beforeEach(() => {
  vi.clearAllMocks()
  mockFetchResult.mockResolvedValue({ success: true })
})

// In test
await waitFor(() => {
  expect($fetchResult).toHaveBeenCalledWith(
    '/api/endpoint',
    expect.objectContaining({
      method: 'PUT',
      body: expect.objectContaining({
        field: 'value',
      }),
    }),
  )
})

Working with Portals and UISelect

When testing Shadcn UISelect components that render options in portals:

it('should show options when clicked', async () => {
  await renderSuspended(Component)

  const selects = screen.getAllByRole('combobox')
  await user.click(selects[0])

  // Wait for portal rendering - options render in body
  await waitFor(() => {
    const options = screen.queryAllByRole('option')
    expect(options.length).toBeGreaterThan(0)
  }, { timeout: 1000 })

  expect(screen.getByRole('option', { name: /option text/i })).toBeDefined()
})

Additional Resources

For detailed migration examples, see:

• .claude/specs/05-test-coverage-improvement/MIGRATION-PATTERNS.md - Before/after examples
• .claude/specs/05-test-coverage-improvement/design.md - Test patterns and architecture
• Testing Library Docs
• Nuxt Test Utils
• Playwright Docs