@crewdle/ai-firebase-service

A TypeScript AI Firebase service wrapper that provides workflow execution and file processing capabilities for Crewdle applications. Built on top of @crewdle/firebase-service.

Features

• 🤖 AI Workflows: Execute AI workflows with message history and file attachments
• 📦 File Processing: Handle file uploads with base64 encoding and caching
• 🔄 Streaming Support: Stream AI responses in real-time
• 💾 Image Caching: Automatic caching of fetched images for performance
• 🎯 TypeScript: Full type safety and IntelliSense support
• 🔧 Flexible: Works with any Firebase project configuration
• 📱 Angular Integration: Injectable service with module support

Installation

npm install @crewdle/ai-firebase-service @crewdle/firebase-service

Note: Firebase is a peer dependency, so you need to install it separately:

npm install firebase

Quick Start

Angular Usage (Recommended)

1. Install the package:

npm install @crewdle/ai-firebase-service @crewdle/firebase-service firebase

2. Configure in your app module:

import { NgModule } from '@angular/core';
import { AiFirebaseServiceModule } from '@crewdle/ai-firebase-service';
import { environment } from '../environments/environment';

@NgModule({
  imports: [
    AiFirebaseServiceModule.forRoot({
      firebaseConfig: environment.firebaseConfig,
      functionsRegion: 'us-central1' // optional
    })
  ]
})
export class AppModule { }

Alternative: Using providers directly (for standalone apps):

import { bootstrapApplication } from '@angular/platform-browser';
import { AiFirebaseService, AI_FIREBASE_SERVICE_CONFIG } from '@crewdle/ai-firebase-service';
import { environment } from './environments/environment';

bootstrapApplication(AppComponent, {
  providers: [
    {
      provide: AI_FIREBASE_SERVICE_CONFIG,
      useValue: {
        firebaseConfig: environment.firebaseConfig,
        functionsRegion: 'us-central1'
      }
    },
    AiFirebaseService
  ]
});

3. Inject and use in your components/services:

import { Component, inject } from '@angular/core';
import { AiFirebaseService, WorkflowMessage } from '@crewdle/ai-firebase-service';

@Component({
  selector: 'app-example',
  template: '...'
})
export class ExampleComponent {
  private readonly aiService = inject(AiFirebaseService);

  async generateContent() {
    const messages: WorkflowMessage[] = [
      { role: 'user', content: 'Hello, can you help me?' }
    ];
    
    const response = await this.aiService.generateContent('workflow-id', messages);
    console.log(response);
  }
}

Non-Angular Usage

import { AiFirebaseService } from '@crewdle/ai-firebase-service';

const aiService = AiFirebaseService.create({
  firebaseConfig: {
    apiKey: "your-api-key",
    authDomain: "your-project.firebaseapp.com",
    projectId: "your-project-id",
    // ... other config
  },
  functionsRegion: "us-central1" // optional, defaults to us-central1
});

API Reference

Workflow Execution

// Generate content using a workflow
const messages: WorkflowMessage[] = [
  { role: 'user', content: 'Analyze this data' }
];

const response = await aiService.generateContent('workflow-id', messages);
console.log('Generated content:', response);

// Generate content with file attachments
const files = [new File(['content'], 'data.txt', { type: 'text/plain' })];
const responseWithFiles = await aiService.generateContent('workflow-id', messages, files);

Streaming Responses

// Stream AI responses in real-time
const messages: WorkflowMessage[] = [
  { role: 'user', content: 'Write a story about AI' }
];

for await (const chunk of aiService.streamContent('workflow-id', messages)) {
  console.log('Chunk:', chunk);
  // Update UI with streaming content
}

Image Processing

// Fetch and cache images from URLs
const imageFile = await aiService.fetchUrl('https://example.com/image.jpg');
console.log('Image file:', imageFile.name, imageFile.size);

// Use fetched image in workflow
const messages: WorkflowMessage[] = [
  { role: 'user', content: 'Describe this image' }
];
const description = await aiService.generateContent('vision-workflow', messages, [imageFile]);

Cache Management

// Check cached images count
const cacheCount = aiService.getCachedImageCount();
console.log('Cached images:', cacheCount);

// Clear image cache
aiService.clearImageCache();

Access Underlying Services

// Get underlying Firebase service if needed
const firebaseService = aiService.getFirebaseService();
const firestore = firebaseService.getFirestore();

Advanced Usage

Custom File Processing

// Process multiple files with different types
const files = [
  new File(['{"data": "value"}'], 'data.json', { type: 'application/json' }),
  new File(['Hello world'], 'text.txt', { type: 'text/plain' })
];

const response = await aiService.generateContent('data-analysis-workflow', [
  { role: 'user', content: 'Analyze these files' }
], files);

Error Handling

try {
  const response = await aiService.generateContent('workflow-id', messages);
} catch (error) {
  console.error('Failed to generate content:', error);
}

Streaming with Error Handling

try {
  for await (const chunk of aiService.streamContent('workflow-id', messages)) {
    // Process chunk
    console.log(chunk);
  }
} catch (error) {
  console.error('Streaming failed:', error);
}

TypeScript Support

The package includes full TypeScript definitions:

import { 
  AiFirebaseService, 
  WorkflowMessage,
  WorkflowResponse,
  WorkflowFile,
  StreamChunk,
  FetchUrlResponse,
  FirebaseServiceConfig
} from '@crewdle/ai-firebase-service';

Requirements

• Node.js >= 16
• TypeScript >= 4.7
• Angular >= 15 (for Angular usage)
• @crewdle/firebase-service >= 1.0.0
• firebase >= 9.0.0

License

MIT © Crewdle