Stop Injecting HttpClient Into Angular Components : Here's the Architecture That Actually Scales

${this.baseUrl}/${endpoint}, payload); } }

### 2. Repository Pattern (Data Transformation & Caching)
Transforms raw payloads, implements in-memory caching, and exposes reactive state.

```typescript
import { Injectable, signal, computed, effect } from '@angular/core';
import { toSignal } from '@angular/core/rxjs-interop';
import { ApiService } from './api.service';
import { User, UserResponse } from './models';

@Injectable({ providedIn: 'root' })
export class UserRepository {
  private readonly api = inject(ApiService);
  
  private readonly _users = signal<User[]>([]);
  private readonly _loading = signal(false);
  private readonly _error = signal<string | null>(null);

  readonly users = this._users.asReadonly();
  readonly loading = this._loading.asReadonly();
  readonly error = this._error.asReadonly();
  
  readonly activeUsers = computed(() => 
    this._users().filter(u => u.status === 'active')
  );

  constructor() {
    this.load();
  }

  load(): void {
    this._loading.set(true);
    this._error.set(null);

    const usersSignal = toSignal(
      this.api.get<UserResponse[]>('users').pipe(
        // Interceptors handle auth/retry; repository handles mapping
      ),
      { initialValue: [] }
    );

    effect(() => {
      const raw = usersSignal();
      this._users.set(raw.map(this.mapUser));
      this._loading.set(false);
    });
  }

  private mapUser(response: UserResponse): User {
    return { id: response.id, name: response.full_name, status: response.is_active ? 'active' : 'inactive' };
  }
}

3. Functional Interceptors (Cross-Cutting Concerns)

Centralized auth, logging, and error transformation without touching component code.

import { HttpInterceptorFn } from '@angular/common/http';
import { inject } from '@angular/core';
import { AuthService } from '../auth/auth.service';
import { catchError, retry, throwError } from 'rxjs';

export const authInterceptor: HttpInterceptorFn = (req, next) => {
  const auth = inject(AuthService);
  const token = auth.getToken();
  
  const authReq = token 
    ? req.clone({ setHeaders: { Authorization: `Bearer ${token}` } })
    : req;

  return next(authReq).pipe(
    retry({ count: 2, delay: 500 }),
    catchError(err => {
      if (err.status === 401) auth.logout();
      return throwError(() => new Error(err.error?.message || 'Network error'));
    })
  );
};

4. Component Consumption (Zero HTTP Awareness)

Components only read signals and trigger repository methods.

import { Component, inject } from '@angular/core';
import { UserRepository } from './user.repository';
import { AsyncPipe } from '@angular/common';

@Component({
  selector: 'app-user-list',
  standalone: true,
  imports: [AsyncPipe],
  template: `
    @if (userRepo.loading()) { <div>Loading...</div> }
    @if (userRepo.error()) { <div class="error">{{ userRepo.error() }}</div> }
    <ul>
      @for (user of userRepo.activeUsers(); track user.id) {
        <li>{{ user.name }}</li>
      }
    </ul>
  `
})
export class UserListComponent {
  readonly userRepo = inject(UserRepository);
}

Pitfall Guide

1. Leaking HTTP Semantics to the UI: Exposing Observable or HttpClient responses directly to components forces UI code to manage subscriptions and async pipes, defeating the purpose of the data layer.
2. Signal Mutation Outside Repository Boundaries: Calling .set() or .update() on repository signals from components breaks encapsulation and makes state tracing impossible. Always use repository methods.
3. Interceptor Error Swallowing: Catching errors in interceptors without re-throwing or transforming them into a consistent error shape causes silent failures and breaks downstream catchError operators.
4. Unbounded Caching: Storing every fetched response in signals without TTL, cache invalidation triggers, or memory limits leads to heap growth and stale data in long-running SPAs.
5. Over-Engineering the Repository: Injecting business logic, form validation, or UI-specific transformations into repositories violates separation of concerns. Repositories should only handle data fetching, mapping, and caching.
6. Ignoring Race Conditions: Failing to use switchMap or cancellation logic in repository methods causes stale responses to overwrite fresh data when users trigger rapid requests (e.g., search inputs).
7. Testing the HTTP Client Instead of the Data Flow: Mocking HttpClient in component tests creates brittle test suites. Test repository outputs (signals/observables) and interceptor behavior in isolation instead.

Deliverables

• Architecture Blueprint: Layered diagram detailing data flow from HttpClient → Interceptors → API Service → Repository → Signal State → Component. Includes dependency injection boundaries and change detection optimization strategies.
• Implementation Checklist:
  • Register functional interceptors in app.config.ts
  • Configure provideHttpClient(withInterceptors([...]))
  • Define strict TypeScript interfaces for API responses
  • Implement repository caching strategy (TTL/invalidation)
  • Replace all component HttpClient injections with repository signals
  • Add unit tests for repository mapping & interceptor error paths
  • Verify no async pipe or subscribe() remains in UI components
• Configuration Templates:
  • app.config.ts with provideHttpClient, withInterceptors, and withFetch()
  • tsconfig.json strict mode flags (strictNullChecks, noImplicitReturns)
  • environment.ts structure for base URLs, retry policies, and cache TTLs
  • ESLint/Prettier rules enforcing signal immutability and repository encapsulation