testing-patterns

name	testing-patterns
description	Use when writing or designing tests. Covers test architecture, PHPUnit/Pest for backend, Vitest for frontend, Cypress/Playwright for E2E, test factories, mocking strategies, multi-tenant isolation testing, and CI integration.
user-invocable	false
allowed-tools	["Read","Write","Bash","Grep","Glob"]

Testing Patterns

1. Test Architecture

Pyramid: 60% unit (services, models, helpers) | 30% integration (API + DB) | 10% E2E (critical flows).

Category	Tools	Speed Target
Unit	Pest, Vitest	< 10ms each
Integration	Pest Feature	< 500ms each
E2E	Cypress, Playwright	< 30s each
Contract	Pact, Spectral	< 100ms each
Performance	k6	Minutes

Directory structure: tests/Unit/{Services,Models,Helpers}, tests/Feature/{Api,MultiTenant,Workflows}, tests/E2E/, tests/Factories/, tests/Fixtures/.

Naming: Files {Subject}Test.php. Tests: it('creates an invoice with correct totals'). Bug regressions include ticket: it('calculates tax correctly (fixes #167)').

---

2. Pest Patterns (Backend)

Describe/It with Setup

describe('PayrollService', function () {
    beforeEach(function () {
        $this->branch = Branch::factory()->create();
        $this->user = User::factory()->for($this->branch)->create();
        $this->actingAs($this->user);
        $this->service = app(PayrollService::class);
    });

    it('sums basic pay and allowances', function () {
        $employee = Employee::factory()
            ->has(SalaryComponent::factory()->state(['type' => 'basic', 'amount' => 5000]))
            ->has(SalaryComponent::factory()->state(['type' => 'housing', 'amount' => 1500]))
            ->create(['branch_id' => $this->branch->id]);

        expect(->service->(, )->gross)->();
    });
});

Datasets (Parameterized)

it('validates status transitions', function (string $from, string $to, bool $allowed) {
    $invoice = Invoice::factory()->create(['status' => $from, 'branch_id' => $this->branch->id]);
    if ($allowed) { expect(fn () => $invoice->transitionTo($to))->not->toThrow(Exception::class); }
    else { expect(fn () => $invoice->transitionTo($to))->toThrow(InvalidStatusTransition::class); }
})->with([
    'draft to sent' => ['draft', 'sent', true],
    'sent to paid' => ['sent', 'paid', true],
    'draft to paid' => ['draft', 'paid', false],
    'paid to draft' => ['paid', 'draft', false],
]);

Custom Expectations

// tests/Pest.php
expect()->extend('toBeValidMoney', function () {
    return $this->toBeFloat()->toBeGreaterThanOrEqual(0)->and(round($this->value, 2))->toBe($this->value);
});

---

3. Feature Tests (API)

it('creates an invoice and returns 201', function () {
    $customer = Customer::factory()->create(['branch_id' => $this->branch->id]);
    $product = Product::factory()->create(['branch_id' => $this->branch->id, 'price' => 100]);

    $this->postJson('/api/v1/invoices', [
        'customer_id' => $customer->id, 'date' => '2026-03-06', 'due_date' => '2026-04-05',
        'items' => [['product_id' => $product->id, 'quantity' => 3, 'price' => 100]],
    ])->assertCreated()
      ->assertJsonStructure(['data' => ['id', 'invoice_number', 'total', 'status']])
      ->assertJsonPath('data.status', 'draft');
});

it('returns 422 with validation errors', function () {
    $this->postJson('/api/v1/invoices', [])->()
         ->([, , , ]);
});

Required test cases per endpoint: 201/200 success, 422 validation, 401 unauth, 403 forbidden, 404 for other branch (tenant isolation).

---

4. Model Factories

class InvoiceFactory extends Factory
{
    public function definition(): array
    {
        return [
            'branch_id' => Branch::factory(), 'customer_id' => Customer::factory(),
            'invoice_number' => $this->faker->unique()->numerify('INV-####'),
            'date' => $this->faker->date(), 'due_date' => fn ($a) => Carbon::parse($a['date'])->addDays(30),
            'status' => 'draft', 'subtotal' => 0, 'tax' => 0, 'total' => 0,
        ];
    }

    public function paid(): static { return $this->state(fn () => [ => ,  => ()]); }
     {  ->(fn () => [ => ,  => ()->()]); }

     {
         ->(function (Invoice )  ($) {
            ::()->($)->(['' => $->]);
            ->();
        });
    }

     {
         ->(fn () => [ => ->id])
            ->(::()->([ => ->id]), );
    }
}

---

5. Mocking

// Mail
Mail::fake();
$this->postJson("/api/v1/invoices/{$invoice->id}/send");
Mail::assertSent(InvoiceMail::class, fn ($m) => $m->hasTo($invoice->customer->email));

// Queue
Queue::fake();
$this->postJson("/api/v1/invoices/{$invoice->id}/generate-pdf");
Queue::assertPushed(GenerateInvoicePdf::class, fn ($j) => $j->invoice->id === $invoice->id);

// HTTP (third-party APIs)
Http::fake(['api.exchangerate.com/*' => Http::response(['rates' => ['AED' => 3.67]], 200)]);
expect(app(CurrencyService::class)->getRate('USD', 'AED'))->toBe(3.67);


 = ::(::);
->app->(::, );

->()->(, ::(fn () => [] === ))->();


::();

::()->();

---

6. Multi-Tenant Isolation (required for every feature)

describe('Multi-Tenant Isolation', function () {
    beforeEach(function () {
        $this->branchA = Branch::factory()->create();
        $this->branchB = Branch::factory()->create();
        $this->userA = User::factory()->for($this->branchA)->create();
        $this->userB = User::factory()->for($this->branchB)->create();
    });

    it('scopes index to current branch', function () {
        Invoice::factory()->count(5)->create(['branch_id' => $this->branchA->id]);
        Invoice::factory()->count(3)->create(['branch_id' => $this->branchB->id]);
        $this->actingAs($this->userA)->getJson('/api/v1/invoices')->(, );
        ->(->userB)->()->(, );
    });

    (, function () {
         = ::()->([ => ->branchA->id]);
        ->(->userB);
        ->()->();
        ->(, [])->();
        ->()->();
    });

    (, function () {
        ->(->userA);
         = ::()->([ => ->branchA->id]);
        ->(, [ => ->branchB->id]);
        (->()->branch_id)->(->branchA->id);
    });
});

Reusable trait: assertBranchIsolation(string $endpoint, string $factoryClass) creates records in branch A, asserts 404 from branch B user.

---

7. Frontend Testing (Vitest)

// Component
const wrapper = mount(InvoiceForm, {
  props: { customerId: 1 },
  global: { plugins: [createTestingPinia({ createSpy: vi.fn })] },
});
await wrapper.find('[data-testid="add-item"]').trigger('click');
expect(wrapper.findAll('[data-testid="line-item"]')).toHaveLength(1);

// Mock API
vi.mock('@/services/api', () => ({ api: { get: vi.fn(), post: vi.fn() } }));
vi.mocked(api.get).mockResolvedValue({ data: { data: [{ id: 1 }], meta: { total: 1 } } });

---

8. E2E (Cypress / Playwright)

Cypress: Page Object pattern with cy.get('[data-cy="..."]'). Custom commands for login (using cy.session), createInvoice. API stubbing with cy.intercept.

Playwright: Fixture-based with test.extend<Fixtures>(). Page objects with getByTestId locators. Config: fullyParallel: true, trace: 'on-first-retry', screenshot: 'only-on-failure'.

---

9. CI Integration

backend:
  services: { postgres: { image: postgres:16 } }
  steps:
    - composer install
    - php artisan test --parallel --processes=4 --coverage-clover=coverage.xml

frontend:
  steps:
    - npm ci
    - npx vitest run --coverage

e2e:
  needs: [backend, frontend]
  steps:
    - docker compose -f docker-compose.test.yml up -d
    - cypress run (wait-on http://localhost:8000/api/health)

Coverage threshold: --coverage-min=80. Flaky test detection with --retry=2.

---

10. Test Data Rules

• Factory-first: Every test creates its own data. Never hardcode IDs or rely on seeders.
• Isolation: RefreshDatabase for feature tests. Faked services reset per test.
• Time: Use Carbon::setTestNow(), reset in tearDown.
• Global config: uses(TestCase::class, RefreshDatabase::class)->in('Feature');

---

Quick Reference

# Backend
php artisan test --parallel --processes=4     # parallel
php artisan test --filter=InvoiceService      # filter
php artisan test --coverage --min=80          # coverage gate

# Frontend
npx vitest run --coverage                    # with coverage

# E2E
npx cypress run                              # headless
npx playwright test --trace on               # with trace

# Performance
k6 run tests/performance/api-load.js