@umituz/react-native-ai-gemini-provider 3.0.20 → 3.0.22

package/dist/index.d.ts

@@ -5,7 +5,7 @@
 export type { GeminiConfig, GeminiGenerationConfig, GeminiHarmCategory, GeminiHarmBlockThreshold, GeminiContent, GeminiPart, GeminiInlineDataPart, GeminiMessagePart, GeminiSafetySetting, GeminiModelOptions, GeminiChatConfig, GeminiResponse, GeminiCandidate, GeminiFinishReason, GeminiSafetyRating, GeminiUsageMetadata, GeminiErrorInfo, GeminiApiError, } from "./domain/entities";
 export { GeminiErrorType, GeminiError, GEMINI_MODELS, DEFAULT_MODELS } from "./domain/entities";
 export { geminiClient } from "./infrastructure/services/GeminiClient";
-export { createChatSession } from "./infrastructure/services/ChatSession";
+export { createChatSession, type ChatSendResult } from "./infrastructure/services/ChatSession";
 export { textGeneration } from "./infrastructure/services/TextGeneration";
 export { structuredText } from "./infrastructure/services/StructuredText";
 export { streaming } from "./infrastructure/services/Streaming";

package/dist/infrastructure/services/ChatSession.d.ts

@@ -1,24 +1,15 @@
 import type { GeminiChatConfig, GeminiMessagePart } from "../../domain/entities";
+export interface ChatSendResult {
+    text: string;
+    finishReason?: string;
+}
 /**
  * Creates a Gemini chat session with full support for system instructions,
  * safety settings, generation config, and multi-turn conversation history.
- *
- * Usage:
- * ```ts
- * const session = createChatSession({
- *   model: "gemini-2.5-flash",
- *   systemInstruction: "You are a helpful assistant.",
- *   generationConfig: { temperature: 0.7, maxOutputTokens: 512 },
- *   history: geminiHistory,
- * });
- *
- * const text = await session.send([{ text: "Hello" }]);
- * ```
  */
 export declare function createChatSession(config?: GeminiChatConfig): {
     /**
-     * Send a message and return the full response text.
-     * Accepts text parts and/or inline data parts (images, audio).
+     * Send a message and return the response text + finish reason.
      */
-    send(parts: GeminiMessagePart[]): Promise<string>;
+    send(parts: GeminiMessagePart[]): Promise<ChatSendResult>;
 };

package/dist/infrastructure/services/index.d.ts

@@ -3,4 +3,4 @@ export { textGeneration } from "./TextGeneration";
 export { structuredText } from "./StructuredText";
 export { streaming } from "./Streaming";
 export { geminiProvider, GeminiProvider } from "./GeminiProvider";
-export { createChatSession } from "./ChatSession";
+export { createChatSession, type ChatSendResult } from "./ChatSession";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@umituz/react-native-ai-gemini-provider",
-  "version": "3.0.20",
+  "version": "3.0.22",
   "description": "Google Gemini AI text generation provider for React Native applications",
   "main": "./src/index.ts",
   "types": "./dist/index.d.ts",

package/src/index.ts

@@ -34,7 +34,7 @@ export {
 
 // Services
 export { geminiClient } from "./infrastructure/services/GeminiClient";
-export { createChatSession } from "./infrastructure/services/ChatSession";
+export { createChatSession, type ChatSendResult } from "./infrastructure/services/ChatSession";
 export { textGeneration } from "./infrastructure/services/TextGeneration";
 export { structuredText } from "./infrastructure/services/StructuredText";
 export { streaming } from "./infrastructure/services/Streaming";

package/src/infrastructure/services/ChatSession.ts

@@ -3,25 +3,17 @@ import { geminiClient } from "./GeminiClient";
 import { DEFAULT_MODELS } from "../../domain/entities";
 import type {
   GeminiChatConfig,
-  GeminiContent,
   GeminiMessagePart,
 } from "../../domain/entities";
 
+export interface ChatSendResult {
+  text: string;
+  finishReason?: string;
+}
+
 /**
  * Creates a Gemini chat session with full support for system instructions,
  * safety settings, generation config, and multi-turn conversation history.
- *
- * Usage:
- * ```ts
- * const session = createChatSession({
- *   model: "gemini-2.5-flash",
- *   systemInstruction: "You are a helpful assistant.",
- *   generationConfig: { temperature: 0.7, maxOutputTokens: 512 },
- *   history: geminiHistory,
- * });
- *
- * const text = await session.send([{ text: "Hello" }]);
- * ```
  */
 export function createChatSession(config: GeminiChatConfig = {}) {
   const model = geminiClient.getModel({
@@ -42,13 +34,16 @@ export function createChatSession(config: GeminiChatConfig = {}) {
 
   return {
     /**
-     * Send a message and return the full response text.
-     * Accepts text parts and/or inline data parts (images, audio).
+     * Send a message and return the response text + finish reason.
      */
-    async send(parts: GeminiMessagePart[]): Promise<string> {
+    async send(parts: GeminiMessagePart[]): Promise<ChatSendResult> {
       const result = await chat.sendMessage(parts as Part[]);
       if (!result.response) throw new Error("No response from Gemini SDK");
-      return result.response.text();
+      const candidate = result.response.candidates?.[0];
+      return {
+        text: result.response.text(),
+        finishReason: candidate?.finishReason ?? undefined,
+      };
     },
   };
 }

package/src/infrastructure/services/index.ts

@@ -3,4 +3,4 @@ export { textGeneration } from "./TextGeneration";
 export { structuredText } from "./StructuredText";
 export { streaming } from "./Streaming";
 export { geminiProvider, GeminiProvider } from "./GeminiProvider";
-export { createChatSession } from "./ChatSession";
+export { createChatSession, type ChatSendResult } from "./ChatSession";

package/README.md

@@ -1,191 +0,0 @@
-# @umituz/react-native-ai-gemini-provider
-
-Production-ready Google Gemini AI provider for React Native applications with built-in error handling, rate limiting, and caching.
-
-## Features
-
-- ✅ **Production Ready** - Real API integration with proper error handling
-- ✅ **Type-Safe** - Full TypeScript support with structured output
-- ✅ **Cost Optimized** - Uses `gemini-2.5-flash-lite` by default (1000 free requests/day)
-- ✅ **Rate Limiting** - Built-in rate limiter to prevent API throttling
-- ✅ **Caching** - LRU cache with TTL for performance optimization
-- ✅ **Retry Logic** - Exponential backoff with jitter for resilient API calls
-- ✅ **No Mock Data** - Production-grade implementation
-
-## Installation
-
-```bash
-npm install @umituz/react-native-ai-gemini-provider
-npm install @google/generative-ai
-```
-
-## Configuration
-
-### 1. Get API Key
-
-Visit [Google AI Studio](https://aistudio.google.com/apikey) and create an API key.
-
-### 2. Environment Setup
-
-Create `.env` file:
-
-```env
-EXPO_PUBLIC_GEMINI_API_KEY=your_api_key_here
-```
-
-### 3. Initialize Provider
-
-```typescript
-import { geminiClientCoreService } from '@umituz/react-native-ai-gemini-provider';
-
-geminiClientCoreService.initialize({
-  apiKey: process.env.EXPO_PUBLIC_GEMINI_API_KEY,
-  textModel: 'gemini-2.5-flash-lite',
-  maxRetries: 3,
-  baseDelay: 1000,
-  maxDelay: 10000,
-});
-```
-
-## Usage
-
-### Simple Text Generation
-
-```typescript
-import { geminiTextService } from '@umituz/react-native-ai-gemini-provider';
-
-const response = await geminiTextService.generateText(
-  'gemini-2.5-flash-lite',
-  'Write a short greeting in Turkish'
-);
-console.log(response); // "Merhaba! Bugün size nasıl yardımcı olabilirim?"
-```
-
-### Structured Output (JSON)
-
-```typescript
-import { geminiStructuredTextService } from '@umituz/react-native-ai-gemini-provider';
-
-interface AnalysisResult {
-  score: number;
-  feedback: string;
-  suggestions: string[];
-}
-
-const schema = {
-  type: 'object',
-  properties: {
-    score: { type: 'number' },
-    feedback: { type: 'string' },
-    suggestions: { type: 'array', items: { type: 'string' } },
-  },
-  required: ['score', 'feedback', 'suggestions'],
-};
-
-const result = await geminiStructuredTextService.generateStructuredText<AnalysisResult>(
-  'gemini-2.5-flash-lite',
-  'Analyze this text',
-  schema
-);
-
-console.log(result.score); // 85
-console.log(result.feedback); // "Good quality overall..."
-```
-
-### With Rate Limiting & Caching
-
-```typescript
-import { 
-  geminiTextService, 
-  rateLimiter,
-  SimpleCache 
-} from '@umituz/react-native-ai-gemini-provider';
-
-const cache = new SimpleCache<string, string>({ 
-  maxSize: 50, 
-  ttl: 60000 // 1 minute 
-});
-
-const cacheKey = `text_${prompt}`;
-const cached = cache.get(cacheKey);
-
-if (cached) {
-  return cached;
-}
-
-const response = await rateLimiter.execute(() =>
-  geminiTextService.generateText('gemini-2.5-flash-lite', prompt)
-);
-
-cache.set(cacheKey, response);
-return response;
-```
-
-## Production Example
-
-See `examples/prod-ai-service.ts` for a complete production-ready implementation with:
-
-- ✅ Content analysis
-- ✅ Motivation letter analysis
-- ✅ Chat assistant
-- ✅ Text summarization
-- ✅ Batch processing
-- ✅ Error handling with retry
-
-## Models
-
-| Model | Input | Output | Free/Day | Best For |
-|-------|-------|--------|-----------|----------|
-| `gemini-2.5-flash-lite` | $0.10 | $0.40 | 1000 | High volume, simple tasks |
-| `gemini-2.5-flash` | $0.15 | $0.60 | 20 | Balanced speed/quality |
-| `gemini-2.5-pro` | $1.25 | $10.00 | 25 | Complex reasoning |
-
-**Recommendation:** Use `gemini-2.5-flash-lite` for most use cases.
-
-## API Reference
-
-### Services
-
-- `geminiClientCoreService` - Client initialization and configuration
-- `geminiTextService` - Text generation (alias: `geminiTextGenerationService`)
-- `geminiStructuredTextService` - JSON structured output
-- `geminiRetryService` - Retry logic with exponential backoff
-
-### Utilities
-
-- `rateLimiter` - Request rate limiting
-- `SimpleCache` - LRU cache with TTL
-- `measureAsync` / `measureSync` - Performance measurement
-
-## Error Handling
-
-```typescript
-try {
-  const result = await geminiTextService.generateText('gemini-2.5-flash-lite', prompt);
-} catch (error) {
-  if (error instanceof Error) {
-    if (error.message.includes('API key')) {
-      // Handle missing/invalid API key
-    } else if (error.message.includes('429') || error.message.includes('quota')) {
-      // Handle rate limit - wait and retry
-    }
-  }
-}
-```
-
-## Best Practices
-
-1. **Always initialize** the client before making requests
-2. **Use rate limiting** to prevent API throttling
-3. **Implement caching** for frequently used prompts
-4. **Handle errors** gracefully with retry logic
-5. **Use structured output** for consistent JSON responses
-6. **Choose the right model** based on complexity and cost
-
-## License
-
-MIT
-
-## Support
-
-For detailed integration guide, see: [factory/templates/docs/ai/GEMINI_AI_PROVIDER.md](https://github.com/umituz/react-native-app-factory/blob/main/factory/templates/docs/ai/GEMINI_AI_PROVIDER.md)