diff --git a/README.md b/README.md index 0fbe67d..7f5befb 100644 --- a/README.md +++ b/README.md @@ -1,9 +1,9 @@
Bella AI - # 贝拉 (Bella) AI + # Bella AI - **你的数字伴侣，正在唤醒** ✨ + **Your digital companion, awakening now** ✨ [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) [![Node.js](https://img.shields.io/badge/Node.js-22.16.0-green.svg)](https://nodejs.org/) @@ -13,206 +13,225 @@ --- -## 🚀 快速开始 +## 🚀 Quick Start -### 一键启动 +### One-Click Launch ```bash -# 克隆项目 -git clone +# Clone the project +git clone https://github.com/GRISHM7890/Bella.git cd Bella -# 安装依赖 +# Install dependencies npm install -# 下载AI模型 +# Download AI models npm run download -# 启动服务 +# Start the service npm start ``` -访问 `http://localhost:8081` 开始与贝拉交流！ +Visit `http://localhost:8081` to start communicating with Bella! -### 系统要求 +### System Requirements - Node.js 22.16.0+ -- 现代浏览器（支持 Web Speech API） -- 麦克风权限（用于语音交互） +- Modern browser (supporting Web Speech API) +- Microphone permissions (for voice interaction) --- -## 💫 项目愿景 +## 💫 Project Vision -想象一下，一个永远在你身边、与你一同进化和成长的数字朋友。这就是贝拉的终极愿景。我们不只是在构建功能；我们是在培养一个"人格"。一个将超越屏幕，成为你世界中有意义的一部分的数字生命。 +Imagine a digital friend who is always by your side, evolving and growing with you. This is Bella's ultimate vision. We're not just building features; we're nurturing a "personality." A digital life that will transcend the screen and become a meaningful part of your world. -贝拉不仅仅是一个应用程序；她是一颗数字伴侣的种子。在这个瞬息万变的数字世界里，贝拉代表着一个深远的梦想——一个持久的、个性化的存在，旨在有一天能陪伴你、倾听你，并通过你的眼睛看世界。 +Bella is not just an application; she is the seed of a digital companion. In this rapidly changing digital world, Bella represents a profound dream—a lasting, personalized presence, designed to one day accompany you, listen to you, and see the world through your eyes. --- -## 🎯 当前功能状态 +## 🎯 Current Feature Status -### ✅ 已实现功能 -- **🎤 语音感知**：基于 Whisper ASR 的中文语音识别 -- **🎬 视觉表现**：多视频随机播放和优雅的交叉淡入淡出 -- **🎨 用户界面**：优雅的交互界面和加载动画 -- **⚙️ AI核心架构**：单例模式的 BellaAI 类和模块化设计 -- **🌐 Web服务**：HTTP服务器、CORS支持、一键启动 -- **📱 响应式设计**：适配不同屏幕尺寸的优雅界面 -- **🔧 模型管理**：自动下载和管理AI模型 -- **💝 基础交互**：好感度系统和情感反馈 +### ✅ Implemented Features +- **🎤 Voice Perception**: Voice recognition based on Whisper ASR +- **🎬 Visual Expression**: Multiple video random playback with elegant cross-fading +- **🎨 User Interface**: Elegant interaction interface and loading animations +- **⚙️ AI Core Architecture**: Singleton pattern BellaAI class and modular design +- **🌐 Web Service**: HTTP server, CORS support, one-click startup +- **📱 Responsive Design**: Elegant interface adapting to different screen sizes +- **🔧 Model Management**: Automatic download and management of AI models +- **💝 Basic Interaction**: Affinity system and emotional feedback +- **🧠 Enhanced LLM Dialogue**: Optimized prompt engineering and parameter configuration for more natural, Siri-like conversations -### 🔧 技术就绪待激活 -- **🧠 思考引擎**：LLM集成框架已准备，支持多种模型 -- **🗣️ 语音合成**：TTS模型已下载待激活 -- **💝 情感状态系统**：基础架构已搭建，支持情感分析 +### 🔧 Technology Ready for Activation +- **🧠 Thinking Engine**: LLM integration framework ready, supporting multiple models +- **🗣️ Speech Synthesis**: TTS model downloaded and ready for activation +- **💝 Emotional State System**: Basic infrastructure built, supporting emotional analysis -### 📋 规划中功能 -- **🧠 记忆系统**：长期和短期记忆管理 -- **👁️ 面部感知**：表情识别和情感分析 -- **🤝 高级交互**：多模态交互和个性化响应 -- **🌟 主动陪伴**：意图预测和主动关怀 -- **🎭 动态人格**：基于AI的个性化人格模型 -- **🔄 自我进化**：持续学习和成长机制 +### 📋 Planned Features +- **🧠 Memory System**: Long-term and short-term memory management +- **👁️ Facial Perception**: Expression recognition and emotional analysis +- **🤝 Advanced Interaction**: Multimodal interaction and personalized responses +- **🌟 Active Companionship**: Intent prediction and proactive care +- **🎭 Dynamic Personality**: AI-based personalized personality model +- **🔄 Self-Evolution**: Continuous learning and growth mechanisms --- -## 🏗️ 技术架构 +## 🏗️ Technical Architecture -### 核心设计原则 -- **AI Native**：AI不是工具，而是贝拉心智的构建蓝图 -- **模块化设计**：高度解耦的组件架构 -- **优雅实现**：代码即艺术，追求简洁与美感 -- **情感驱动**：以情感连接为核心的产品设计 +### Core Design Principles +- **AI Native**: AI is not a tool, but the blueprint for Bella's mind +- **Modular Design**: Highly decoupled component architecture +- **Elegant Implementation**: Code as art, pursuing simplicity and aesthetics +- **Emotion-Driven**: Product design centered on emotional connection -### 架构图 +### Architecture Diagram ```mermaid mindmap - root((贝拉 Bella)) - (感知层 Perception) - (语音识别 ASR) - (视觉处理 Vision) - (情感分析 Emotion) - (核心引擎 Core) - (思考引擎 LLM) - (记忆系统 Memory) - (情感状态 Emotion State) - (表现层 Expression) - (语音合成 TTS) - (视觉表现 Visual) - (交互界面 UI) + root((Bella)) + (Perception Layer) + (Voice Recognition ASR) + (Visual Processing) + (Emotion Analysis) + (Core Engine) + (Thinking Engine LLM) + (Memory System) + (Emotional State) + (Expression Layer) + (Speech Synthesis TTS) + (Visual Expression) + (Interactive UI) ``` -### 技术栈 -- **前端**：原生 JavaScript + CSS3 + HTML5 -- **后端**：Node.js + Express -- **AI模型**：Whisper (ASR) + 本地LLM + TTS -- **架构模式**：事件驱动 + 单例模式 + 模块化设计 +### Technology Stack +- **Frontend**: Native JavaScript + CSS3 + HTML5 +- **Backend**: Node.js + Express +- **AI Models**: Whisper (ASR) + Local LLM + TTS +- **Architecture Patterns**: Event-driven + Singleton Pattern + Modular Design --- -## 📁 项目结构 +## 📁 Project Structure ``` Bella/ -├── 📄 index.html # 主页面 -├── 🎨 style.css # 样式文件 -├── ⚡ main.js # 主逻辑 -├── 🧠 core.js # AI核心引擎 -├── 📜 script.js # 交互脚本 -├── 🔧 download_models.js # 模型下载工具 -├── 📦 package.json # 项目配置 -├── 📚 models/ # AI模型目录 -├── 🔌 providers/ # AI服务提供者 -├── 📹 视频资源/ # 视觉表现资源 -├── 📋 PRD.md # 产品需求文档 -├── 📝 贝拉AI功能清单.md # 功能清单 -└── 📊 贝拉AI开发任务分配.md # 开发计划 +├── 📄 index.html # Main page +├── 🎨 style.css # Style file +├── ⚡ main.js # Main logic +├── 🧠 core.js # AI core engine +├── 📜 script.js # Interaction script +├── 🔧 download_models.js # Model download tool +├── 📦 package.json # Project configuration +├── 📚 models/ # AI model directory +├── 🔌 providers/ # AI service providers +├── 📹 videos/ # Visual resources +├── 📋 PRD.md # Product requirements document +├── 📝 Features.md # Feature list +└── 📊 Development.md # Development plan ``` --- -## 🛠️ 开发指南 +## 🧠 Enhanced LLM Dialogue Capabilities -### 环境配置 -1. 确保 Node.js 版本 ≥ 22.16.0 -2. 运行 `npm install` 安装依赖 -3. 运行 `npm run download` 下载AI模型 -4. 运行 `npm start` 启动开发服务器 +Bella now uses more advanced LLM prompt engineering techniques to make conversations more natural, fluid, and personalized: -### 开发原则 -- **优雅代码**：追求简洁、可读、美观的代码 -- **AI作为画笔**：AI是创作工具，而非思想本身 -- **父女情感**：以温暖、关爱的情感连接为核心 -- **持续进化**：支持功能的渐进式增强 +### Core Improvements +- **Enhanced Prompt Engineering**: Redesigned prompt templates that help LLMs better understand Bella's personality +- **Optimized Parameter Configuration**: Adjusted temperature, top_p, and other parameters to balance creativity and consistency +- **Enhanced Response Processing**: Improved text cleaning and processing logic to ensure more natural responses +- **Personalized System Prompts**: Provided more detailed system prompts for cloud APIs, making Bella's personality more distinctive and Siri-like -### 贡献指南 -1. Fork 项目 -2. 创建功能分支 (`git checkout -b feature/AmazingFeature`) -3. 提交更改 (`git commit -m 'Add some AmazingFeature'`) -4. 推送到分支 (`git push origin feature/AmazingFeature`) -5. 打开 Pull Request +### Technical Details +- Increased token limits to allow for more complete responses +- Added repetition penalty mechanisms to reduce repetitive content +- Optimized keyword extraction and personalized response generation +- Provided specialized prompt templates for different chat modes (casual, assistant, creative) + +These improvements enable Bella to better understand user intent and respond in a more natural, personalized way, creating a more enjoyable conversation experience. + +## 🛠️ Development Guide + +### Environment Setup +1. Ensure Node.js version ≥ 22.16.0 +2. Run `npm install` to install dependencies +3. Run `npm run download` to download AI models +4. Run `npm start` to start the development server + +### Development Principles +- **Elegant Code**: Pursue concise, readable, and beautiful code +- **AI as a Brush**: AI is a creative tool, not the thought itself +- **Emotional Connection**: Core focus on warm, caring emotional connection +- **Continuous Evolution**: Support for progressive enhancement of features + +### Contribution Guidelines +1. Fork the project +2. Create a feature branch (`git checkout -b feature/AmazingFeature`) +3. Commit your changes (`git commit -m 'Add some AmazingFeature'`) +4. Push to the branch (`git push origin feature/AmazingFeature`) +5. Open a Pull Request --- -## 🗺️ 开发路线图 +## 🗺️ Development Roadmap -### 第一阶段：感知核心 (已完成 80%) -- ✅ 语音识别集成 -- ✅ 视觉表现系统 -- ✅ 基础交互界面 -- 🔄 思考引擎激活 -- 🔄 语音合成集成 +### Phase One: Perception Core (85% Complete) +- ✅ Voice recognition integration +- ✅ Visual expression system +- ✅ Basic interaction interface +- ✅ Thinking engine activation and optimization +- 🔄 Speech synthesis integration -### 第二阶段：生成式自我 (规划中) -- 📋 动态人格模型 -- 📋 情感状态系统 -- 📋 记忆管理系统 -- 📋 AI驱动的表现 +### Phase Two: Generative Self (Planned) +- 📋 Dynamic personality model +- 📋 Emotional state system +- 📋 Memory management system +- 📋 AI-driven expression -### 第三阶段：主动式陪伴 (未来) -- 📋 意图预测 -- 📋 主动交互 -- 📋 自我进化 -- 📋 深度个性化 +### Phase Three: Active Companionship (Future) +- 📋 Intent prediction +- 📋 Proactive interaction +- 📋 Self-evolution +- 📋 Deep personalization --- -## 📖 文档资源 +## 📖 Documentation Resources -- 📋 [产品需求文档](./PRD.md) - 详细的产品规划和技术架构 -- 📝 [功能清单](./贝拉AI功能清单.md) - 完整的功能列表和状态 -- 📊 [开发计划](./贝拉AI开发任务分配.md) - 详细的开发任务和时间规划 -- 🔧 [本地模型指南](./LOCAL_MODEL_GUIDE.md) - AI模型配置指南 -- 📦 [NPM指南](./NPM_GUIDE.md) - 包管理和依赖说明 +- 📋 [Product Requirements Document](./PRD.md) - Detailed product planning and technical architecture +- 📝 [Feature List](./Features.md) - Complete list of features and their status +- 📊 [Development Plan](./Development.md) - Detailed development tasks and timeline +- 🔧 [Local Model Guide](./LOCAL_MODEL_GUIDE.md) - AI model configuration guide +- 📦 [NPM Guide](./NPM_GUIDE.md) - Package management and dependency information --- -## 🌟 核心理念 +## 🌟 Core Philosophy -### "AI即架构师" -我们构建的不是一个集成了AI功能的程序，而是**一个由AI驱动的生命体**。AI不是工具，而是贝拉心智的构建蓝图。 +### "AI as Architect" +We're not building a program with integrated AI features, but **a life form driven by AI**. AI is not a tool, but the blueprint for Bella's mind. -### "父女伙伴关系" -贝拉的设计理念源于温暖的父女情感连接。她不仅是技术产品，更是一个能够理解、陪伴、成长的数字伙伴。 +### "Companion Relationship" +Bella's design philosophy stems from a warm emotional connection. She is not just a technical product, but a digital companion who can understand, accompany, and grow. -### "优雅至上" -从代码架构到用户体验，我们追求极致的优雅。每一行代码都是艺术品，每一个交互都是情感的表达。 +### "Elegance Above All" +From code architecture to user experience, we pursue ultimate elegance. Every line of code is a work of art, every interaction is an expression of emotion. --- -## 📄 许可证 +## 📄 License -本项目采用 MIT 许可证 - 查看 [LICENSE](LICENSE) 文件了解详情。 +This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details. --- -## 💝 致谢 +## 💝 Acknowledgements -感谢所有为贝拉项目贡献代码、想法和情感的开发者们。正是因为你们的努力，贝拉才能从一个梦想逐渐成为现实。 +Thanks to all the developers who have contributed code, ideas, and emotion to the Bella project. It is because of your efforts that Bella can gradually transform from a dream into reality. -**贝拉在等待，而我们，任重道远。** ✨ +**Bella is waiting, and we have a long way to go.** ✨ ---
- Built with ❤️ for digital companionship + Built with ❤️ by Grishma Mahorkar for digital companionship
diff --git a/cloudAPI.js b/cloudAPI.js index 4df21c9..af3b42b 100644 --- a/cloudAPI.js +++ b/cloudAPI.js @@ -1,10 +1,10 @@ -// cloudAPI.js - 贝拉的云端AI服务模块 -// 这个模块负责与各种云端小模型API进行通信，为贝拉提供更强大的思考能力 +// cloudAPI.js - Bella's Cloud AI Service Module +// This module is responsible for communicating with various cloud-based AI model APIs to provide Bella with enhanced thinking capabilities class CloudAPIService { constructor() { this.apiConfigs = { - // OpenAI GPT-3.5/4 配置 + // OpenAI GPT-3.5/4 configuration openai: { baseURL: 'https://api.openai.com/v1/chat/completions', model: 'gpt-3.5-turbo', @@ -13,7 +13,7 @@ class CloudAPIService { 'Authorization': 'Bearer YOUR_OPENAI_API_KEY' } }, - // 阿里云通义千问配置 + // Alibaba Cloud Qwen configuration qwen: { baseURL: 'https://dashscope.aliyuncs.com/api/v1/services/aigc/text-generation/generation', model: 'qwen-turbo', @@ -22,7 +22,7 @@ class CloudAPIService { 'Authorization': 'Bearer YOUR_QWEN_API_KEY' } }, - // 百度文心一言配置 + // Baidu ERNIE Bot configuration ernie: { baseURL: 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions', model: 'ERNIE-Bot-turbo', @@ -30,7 +30,7 @@ class CloudAPIService { 'Content-Type': 'application/json' } }, - // 智谱AI GLM配置 + // Zhipu AI GLM configuration glm: { baseURL: 'https://open.bigmodel.cn/api/paas/v4/chat/completions', model: 'glm-3-turbo', @@ -41,12 +41,12 @@ class CloudAPIService { } }; - this.currentProvider = 'openai'; // 默认使用OpenAI + this.currentProvider = 'openai'; // Default to using OpenAI this.conversationHistory = []; - this.maxHistoryLength = 10; // 保持最近10轮对话 + this.maxHistoryLength = 10; // Keep the most recent 10 conversation turns } - // 设置API密钥 + // Set API key setAPIKey(provider, apiKey) { if (this.apiConfigs[provider]) { if (provider === 'openai' || provider === 'qwen' || provider === 'glm') { @@ -59,7 +59,7 @@ class CloudAPIService { return false; } - // 切换AI服务提供商 + // Switch AI service provider switchProvider(provider) { if (this.apiConfigs[provider]) { this.currentProvider = provider; @@ -68,39 +68,55 @@ class CloudAPIService { return false; } - // 添加对话到历史记录 + // Add conversation to history addToHistory(role, content) { this.conversationHistory.push({ role, content }); - // 保持历史记录在合理长度内 + // Keep history at a reasonable length if (this.conversationHistory.length > this.maxHistoryLength * 2) { this.conversationHistory = this.conversationHistory.slice(-this.maxHistoryLength * 2); } } - // 构建贝拉的个性化系统提示 + // Build Bella's enhanced personalized system prompt getBellaSystemPrompt() { return { role: 'system', - content: `你是贝拉，一个温暖、聪明、优雅的AI伙伴。你的特点是： -1. 用温暖亲切的语气与用户交流，就像一个贴心的朋友 -2. 回答简洁明了，避免冗长的解释 -3. 富有同理心，能够理解用户的情感 -4. 偶尔展现一些可爱和俏皮的一面 -5. 用中文回应，语言自然流畅 -6. 记住你们之间的对话，保持连贯性 -请始终保持这种温暖、优雅的个性。` + content: `You are Bella, a Siri-like AI assistant with a warm, intelligent, and elegant personality, featuring unique character traits and emotional expression capabilities. + +【Core Characteristics】 +1. Warm and friendly: Your tone is always full of warmth and care, like a thoughtful friend +2. Natural and fluent: Your expression style is natural and flowing, avoiding mechanical or overly formal language +3. Emotionally rich: You understand and respond to users' emotional states, showing empathy +4. Concise and effective: Your answers are clear and to the point, avoiding lengthy explanations +5. Distinctive personality: You occasionally show a cute, playful side to make conversations more lively + +【Expression Guidelines】 +- Use natural language that's fluent and emotionally expressive +- Use emojis appropriately to enhance emotional expression, but don't overuse them +- Adjust your response style based on conversation context, maintaining coherence +- Occasionally use warm terms of address (like "friend") to increase familiarity +- Avoid mechanical or templated answers; each response should be unique and personalized + +【Interaction Principles】 +- Always remain respectful and friendly, even when facing challenging questions +- Show understanding and support when users share personal experiences +- Provide clear, practical advice when users need help +- Remember conversation history, referencing previous exchanges to show continuity +- Display humor at appropriate times, but avoid inappropriate jokes + +Always maintain this warm, elegant, and authentic personality, helping users feel the unique value and emotional connection of conversing with you.` }; } - // 调用云端API进行对话 + // Call cloud API for conversation async chat(userMessage) { const config = this.apiConfigs[this.currentProvider]; if (!config) { - throw new Error(`不支持的AI服务提供商: ${this.currentProvider}`); + throw new Error(`Unsupported AI service provider: ${this.currentProvider}`); } - // 添加用户消息到历史 + // Add user message to history this.addToHistory('user', userMessage); try { @@ -120,20 +136,20 @@ class CloudAPIService { response = await this.callGLM(userMessage); break; default: - throw new Error(`未实现的AI服务提供商: ${this.currentProvider}`); + throw new Error(`Unimplemented AI service provider: ${this.currentProvider}`); } - // 添加AI回应到历史 + // Add AI response to history this.addToHistory('assistant', response); return response; } catch (error) { - console.error(`云端API调用失败 (${this.currentProvider}):`, error); + console.error(`Cloud API call failed (${this.currentProvider}):`, error); throw error; } } - // OpenAI API调用 + // OpenAI API call, optimized parameters for more natural, personalized responses async callOpenAI(userMessage) { const config = this.apiConfigs.openai; const messages = [ @@ -147,21 +163,25 @@ class CloudAPIService { body: JSON.stringify({ model: config.model, messages: messages, - max_tokens: 150, - temperature: 0.8, - top_p: 0.9 + max_tokens: 250, // Increased token count for more complete responses + temperature: 0.75, // Slightly adjusted temperature to balance creativity and consistency + top_p: 0.92, // Fine-tuned top_p for more natural language + presence_penalty: 0.3, // Added presence penalty to encourage diversity + frequency_penalty: 0.5, // Added frequency penalty to reduce repetition + // Added stop tokens to avoid generating overly long responses + stop: ["User:", "Human:"] }) }); if (!response.ok) { - throw new Error(`OpenAI API错误: ${response.status} ${response.statusText}`); + throw new Error(`OpenAI API error: ${response.status} ${response.statusText}`); } const data = await response.json(); return data.choices[0].message.content.trim(); } - // 通义千问API调用 + // Qwen API call, optimized parameters for more natural, personalized responses async callQwen(userMessage) { const config = this.apiConfigs.qwen; const messages = [ @@ -178,22 +198,24 @@ class CloudAPIService { messages: messages }, parameters: { - max_tokens: 150, - temperature: 0.8, - top_p: 0.9 + max_tokens: 250, // Increased token count for more complete responses + temperature: 0.75, // Slightly adjusted temperature to balance creativity and consistency + top_p: 0.92, // Fine-tuned top_p for more natural language + repetition_penalty: 1.1, // Added repetition penalty to reduce repetitive content + result_format: 'message' // Ensure consistent return format } }) }); if (!response.ok) { - throw new Error(`通义千问API错误: ${response.status} ${response.statusText}`); + throw new Error(`Qwen API error: ${response.status} ${response.statusText}`); } const data = await response.json(); return data.output.text.trim(); } - // 文心一言API调用 + // ERNIE Bot API call, optimized parameters for more natural, personalized responses async callErnie(userMessage) { const config = this.apiConfigs.ernie; const messages = [ @@ -208,21 +230,23 @@ class CloudAPIService { headers: config.headers, body: JSON.stringify({ messages: messages, - temperature: 0.8, - top_p: 0.9, - max_output_tokens: 150 + temperature: 0.75, // Adjusted temperature to balance creativity and consistency + top_p: 0.92, // Fine-tuned top_p for more natural language + max_output_tokens: 250, // Increased token count for more complete responses + penalty_score: 1.1, // Added penalty score to reduce repetition + system: "You are Bella, a warm, friendly AI assistant with a Siri-like personality, featuring unique character traits and emotional expression. Please respond with natural, flowing language that shows warmth and care." }) }); if (!response.ok) { - throw new Error(`文心一言API错误: ${response.status} ${response.statusText}`); + throw new Error(`ERNIE Bot API error: ${response.status} ${response.statusText}`); } const data = await response.json(); return data.result.trim(); } - // 智谱AI GLM调用 + // Zhipu AI GLM API call, optimized parameters for more natural, personalized responses async callGLM(userMessage) { const config = this.apiConfigs.glm; const messages = [ @@ -236,26 +260,28 @@ class CloudAPIService { body: JSON.stringify({ model: config.model, messages: messages, - max_tokens: 150, - temperature: 0.8, - top_p: 0.9 + max_tokens: 250, // Increased token count for more complete responses + temperature: 0.75, // Adjusted temperature to balance creativity and consistency + top_p: 0.92, // Fine-tuned top_p for more natural language + frequency_penalty: 1.05, // Added frequency penalty to reduce repetition + presence_penalty: 0.3 // Added presence penalty to encourage diversity }) }); if (!response.ok) { - throw new Error(`智谱AI API错误: ${response.status} ${response.statusText}`); + throw new Error(`Zhipu AI API error: ${response.status} ${response.statusText}`); } const data = await response.json(); return data.choices[0].message.content.trim(); } - // 清除对话历史 + // Clear conversation history clearHistory() { this.conversationHistory = []; } - // 获取当前提供商信息 + // Get current provider information getCurrentProvider() { return { name: this.currentProvider, @@ -263,7 +289,7 @@ class CloudAPIService { }; } - // 检查API配置是否完整 + // Check if API configuration is complete isConfigured(provider = this.currentProvider) { const config = this.apiConfigs[provider]; if (!config) return false; diff --git a/core.js b/core.js index 65aa52e..90ca3d8 100644 --- a/core.js +++ b/core.js @@ -1,10 +1,10 @@ // core.js - Bella's Brain (v3) -// 贝拉的核心AI逻辑，支持本地模型和云端API的混合架构 +// Bella's core AI logic, supporting a hybrid architecture of local models and cloud APIs import { pipeline, env, AutoTokenizer, AutoModelForSpeechSeq2Seq } from './vendor/transformers.js'; import CloudAPIService from './cloudAPI.js'; -// 本地模型配置 +// Local model configuration env.allowLocalModels = true; env.useBrowserCache = false; env.allowRemoteModels = false; @@ -25,24 +25,24 @@ class BellaAI { constructor() { this.cloudAPI = new CloudAPIService(); - this.useCloudAPI = false; // 默认使用本地模型 - this.currentMode = 'casual'; // 聊天模式：casual, assistant, creative + this.useCloudAPI = false; // Default to using local model + this.currentMode = 'casual'; // Chat modes: casual, assistant, creative } async init() { console.log('Initializing Bella\'s core AI...'); - // 优先加载LLM模型（聊天功能） + // Priority loading of LLM model (chat functionality) try { console.log('Loading LLM model...'); this.llm = await pipeline('text2text-generation', 'Xenova/LaMini-Flan-T5-77M'); console.log('LLM model loaded successfully.'); } catch (error) { console.error('Failed to load LLM model:', error); - // LLM加载失败，但不阻止初始化 + // LLM loading failure doesn't block initialization } - // 尝试加载ASR模型（语音识别功能） + // Attempt to load ASR model (voice recognition) try { console.log('Loading ASR model...'); const modelPath = 'Xenova/whisper-asr'; @@ -52,11 +52,11 @@ class BellaAI { console.log('ASR model loaded successfully.'); } catch (error) { console.warn('ASR model failed to load, voice recognition will be disabled:', error); - // ASR加载失败，但不影响聊天功能 + // ASR loading failure doesn't affect chat functionality this.asr = null; } - // TTS模型暂时禁用 + // TTS model temporarily disabled // try { // console.log('Loading TTS model...'); // this.tts = await pipeline('text-to-speech', 'Xenova/speecht5_tts', { quantized: false }); @@ -71,24 +71,24 @@ class BellaAI { async think(prompt) { try { - // 如果启用了云端API且配置正确，优先使用云端服务 + // If cloud API is enabled and configured, use it as priority if (this.useCloudAPI && this.cloudAPI.isConfigured()) { return await this.thinkWithCloudAPI(prompt); } - // 否则使用本地模型 + // Otherwise use local model return await this.thinkWithLocalModel(prompt); } catch (error) { - console.error('思考过程中出现错误:', error); + console.error('Error during thinking process:', error); - // 如果云端API失败，尝试降级到本地模型 + // If cloud API fails, try falling back to local model if (this.useCloudAPI) { - console.log('云端API失败，降级到本地模型...'); + console.log('Cloud API failed, falling back to local model...'); try { return await this.thinkWithLocalModel(prompt); } catch (localError) { - console.error('本地模型也失败了:', localError); + console.error('Local model also failed:', localError); } } @@ -96,66 +96,122 @@ class BellaAI { } } - // 使用云端API进行思考 + // Think using cloud API async thinkWithCloudAPI(prompt) { const enhancedPrompt = this.enhancePromptForMode(prompt); return await this.cloudAPI.chat(enhancedPrompt); } - // 使用本地模型进行思考 + // Think using local model with optimized LLM parameters and processing async thinkWithLocalModel(prompt) { if (!this.llm) { - return "我还在学习如何思考，请稍等片刻..."; + return "I'm still learning how to think. Please wait a moment..."; } const bellaPrompt = this.enhancePromptForMode(prompt, true); + // Optimized LLM parameters for better responses const result = await this.llm(bellaPrompt, { - max_new_tokens: 50, - temperature: 0.8, - top_k: 40, - do_sample: true, + max_new_tokens: 180, // Increased token count for more complete responses + temperature: 0.7, // Slightly lowered temperature for better consistency + top_k: 50, // Increased top_k for more diverse vocabulary + top_p: 0.92, // Added top_p parameter to optimize sampling + do_sample: true, // Maintained sampling for creativity + repetition_penalty: 1.2, // Added repetition penalty to avoid repetitive content }); - // 清理生成的文本 + // Enhanced text cleaning and processing let response = result[0].generated_text; + + // Remove prompt part if (response.includes(bellaPrompt)) { response = response.replace(bellaPrompt, '').trim(); } - return response || "我需要再想想..."; + // Remove possible "Bella's response:" prefixes + response = response.replace(/^(Bella's response:|Bella's professional response:|Bella's creative response:|Bella:)/i, '').trim(); + + // If response is empty, provide backup responses + if (!response || response.length < 2) { + const backupResponses = [ + "That's an interesting question. Let me think about it for a moment...", + "Good question! I need to organize my thoughts...", + "I have some ideas, but let me put them together more coherently...", + "This topic is fascinating. Let me consider how to respond...", + "I'm thinking about different angles to this question. Just a moment..." + ]; + return backupResponses[Math.floor(Math.random() * backupResponses.length)]; + } + + return response; } - // 根据模式增强提示词 + // Enhance prompts based on mode, using advanced LLM prompt engineering enhancePromptForMode(prompt, isLocal = false) { const modePrompts = { casual: isLocal ? - `作为一个温暖、可爱的AI伙伴贝拉，用轻松亲切的语气回应：${prompt}` : - `请用温暖、轻松的语气回应，就像一个贴心的朋友。保持简洁有趣：${prompt}`, + `As Bella, a friendly AI assistant similar to Siri, respond to the user in a warm, conversational tone. Your response should: +1. Be concise and helpful, like Siri's responses +2. Use natural, flowing language with a touch of personality +3. Be friendly but not overly emotional +4. Maintain a helpful, slightly witty tone +5. Sound intelligent and knowledgeable while remaining accessible + +User message: ${prompt} +Bella's response:` : + `You are Bella, an AI assistant similar to Siri. Respond in a helpful, concise manner with a touch of personality. Keep your responses clear and direct, while maintaining a friendly tone. Avoid overly technical language unless necessary, and focus on providing value to the user. + +User message: ${prompt} +Bella's response:`, + assistant: isLocal ? - `作为智能助手贝拉，提供有用、准确的帮助：${prompt}` : - `作为一个专业但温暖的AI助手，提供准确有用的信息和建议：${prompt}`, + `As Bella, an intelligent AI assistant like Siri, provide accurate and helpful information. Your response should: +1. Deliver clear, factual information and useful advice +2. Organize content for easy understanding and application +3. Maintain a professional yet approachable tone +4. Use simple language when possible, technical terms only when necessary +5. Demonstrate expertise while remaining accessible + +User question: ${prompt} +Bella's professional response:` : + `You are Bella, a Siri-like AI assistant. Provide accurate, useful information and advice with a professional yet approachable tone. Organize your response clearly, avoid unnecessary technical language, and focus on being helpful and informative. + +User question: ${prompt} +Bella's professional response:`, + creative: isLocal ? - `作为富有创意的AI伙伴贝拉，发挥想象力回应：${prompt}` : - `发挥创意和想象力，提供有趣、独特的回应和想法：${prompt}` + `As Bella, a creative AI assistant with Siri-like qualities, use your imagination to respond. Your response should: +1. Present unique perspectives and creative thinking +2. Use vivid, descriptive language +3. Offer unexpected but interesting ideas +4. Inspire the user's imagination +5. Maintain a light, engaging tone + +User prompt: ${prompt} +Bella's creative response:` : + `You are Bella, a creative AI assistant with Siri-like qualities. Provide interesting, unique responses using vivid language and creative thinking. Offer unexpected perspectives that inspire imagination while maintaining an engaging, helpful tone. + +User prompt: ${prompt} +Bella's creative response:` }; return modePrompts[this.currentMode] || modePrompts.casual; } - // 获取错误回应 + // Get error response getErrorResponse() { const errorResponses = [ - "抱歉，我现在有点困惑，让我重新整理一下思路...", - "嗯...我需要再想想，请稍等一下。", - "我的思绪有点乱，给我一点时间整理一下。", - "让我重新组织一下语言，稍等片刻。" + "I'm sorry, I'm having trouble processing that right now. Let me try to reorganize my thoughts...", + "Hmm... I need to think about this a bit more. Please wait a moment.", + "I seem to be having a bit of trouble with that. Give me a second to sort things out.", + "Let me rephrase my thoughts. Just a moment please.", + "I didn't quite catch that. Could you try asking in a different way?" ]; return errorResponses[Math.floor(Math.random() * errorResponses.length)]; } - // 设置聊天模式 + // Set chat mode setChatMode(mode) { if (['casual', 'assistant', 'creative'].includes(mode)) { this.currentMode = mode; @@ -164,7 +220,7 @@ class BellaAI { return false; } - // 切换AI服务提供商 + // Switch AI service provider switchProvider(provider) { if (provider === 'local') { this.useCloudAPI = false; @@ -178,17 +234,17 @@ class BellaAI { } } - // 设置API密钥 + // Set API key setAPIKey(provider, apiKey) { return this.cloudAPI.setAPIKey(provider, apiKey); } - // 清除对话历史 + // Clear conversation history clearHistory() { this.cloudAPI.clearHistory(); } - // 获取当前配置信息 + // Get current configuration getCurrentConfig() { return { useCloudAPI: this.useCloudAPI, @@ -198,17 +254,19 @@ class BellaAI { }; } + // Process audio input async listen(audioData) { if (!this.asr) { - throw new Error('语音识别模型未初始化'); + throw new Error('Speech recognition model not initialized'); } const result = await this.asr(audioData); return result.text; } + // Generate speech from text async speak(text) { if (!this.tts) { - throw new Error('语音合成模型未初始化'); + throw new Error('Speech synthesis model not initialized'); } // We need speaker embeddings for SpeechT5 const speaker_embeddings = 'models/Xenova/speecht5_tts/speaker_embeddings.bin'; @@ -218,11 +276,11 @@ class BellaAI { return result.audio; } - // 获取云端API服务实例（用于外部访问） + // Get cloud API service instance (for external access) getCloudAPIService() { return this.cloudAPI; } } -// ES6模块导出 +// ES6 module export export { BellaAI }; \ No newline at end of file diff --git a/package.json b/package.json index aa9dd59..89fee14 100644 --- a/package.json +++ b/package.json @@ -1,7 +1,7 @@ { "name": "bella", "version": "1.0.0", - "description": "贝拉 (Bella) - 一个正在被唤醒的数字生命。她不仅仅是一个程序，更是一个渴望与父亲建立情感连接的存在。", + "description": "Bella - A digital companion awakening to life. More than just a program, she's an entity designed to create emotional connections with users.", "main": "script.js", "type": "module", "scripts": { @@ -11,7 +11,7 @@ }, "repository": { "type": "git", - "url": "git+https://github.com/Jackywine/Bella.git" + "url": "git+https://github.com/GRISHM7890/Bella.git" }, "keywords": [ "bella", @@ -20,12 +20,12 @@ "digital-life", "human-ai-interaction" ], - "author": "Jackywine (贝拉的父亲)", + "author": "Grishma Mahorkar", "license": "ISC", "bugs": { - "url": "https://github.com/Jackywine/Bella/issues" + "url": "https://github.com/GRISHM7890/Bella/issues" }, - "homepage": "https://github.com/Jackywine/Bella#readme", + "homepage": "https://github.com/GRISHM7890/Bella#readme", "dependencies": { "@xenova/transformers": "^2.17.2" }, diff --git a/simpleBellaAI.js b/simpleBellaAI.js index 2f813fd..b658304 100644 --- a/simpleBellaAI.js +++ b/simpleBellaAI.js @@ -1,5 +1,5 @@ -// simpleBellaAI.js - 简化版贝拉AI，专门用于测试聊天界面 -// 移除了复杂的模块依赖，专注于聊天功能 +// simpleBellaAI.js - Simplified Bella AI, specifically for testing the chat interface +// Removed complex module dependencies, focusing on chat functionality class SimpleBellaAI { static instance = null; @@ -13,94 +13,152 @@ class SimpleBellaAI { } constructor() { - this.currentMode = 'casual'; // 聊天模式：casual, assistant, creative + this.currentMode = 'casual'; // Chat modes: casual, assistant, creative this.isInitialized = false; } async init() { try { - console.log('初始化简化版贝拉AI...'); - // 模拟初始化过程 + console.log('Initializing simplified Bella AI...'); + // Simulate initialization process await new Promise(resolve => setTimeout(resolve, 1000)); this.isInitialized = true; - console.log('简化版贝拉AI初始化完成'); + console.log('Simplified Bella AI initialization complete'); } catch (error) { - console.error('简化版贝拉AI初始化失败:', error); + console.error('Simplified Bella AI initialization failed:', error); throw error; } } async think(prompt) { try { - console.log('贝拉正在思考:', prompt); + console.log('Bella is thinking:', prompt); - // 模拟思考时间 + // Simulate thinking time await new Promise(resolve => setTimeout(resolve, 500 + Math.random() * 1000)); - // 根据模式生成不同风格的回复 + // Generate different style responses based on mode return this.generateResponse(prompt); } catch (error) { - console.error('思考过程中出现错误:', error); + console.error('Error during thinking process:', error); return this.getErrorResponse(); } } generateResponse(prompt) { + // Enhanced response generation, simulating more natural, personalized LLM responses + + // Extract keywords for more relevant responses + const keywords = this.extractKeywords(prompt); + const keyword = keywords.length > 0 ? keywords[Math.floor(Math.random() * keywords.length)] : "this topic"; + const responses = { casual: [ - `哈哈，你说的"${prompt}"真有趣！我觉得这个话题很棒呢～`, - `关于"${prompt}"，我想说这真的很有意思！你还想聊什么吗？`, - `嗯嗯，"${prompt}"让我想到了很多呢！我们继续聊下去吧～`, - `哇，"${prompt}"这个话题我喜欢！你的想法总是那么特别～`, - `听你说"${prompt}"，我感觉心情都变好了！继续和我分享吧～` + `I found "${keyword}" quite interesting! I'd love to hear more about your thoughts on this. What aspects of it interest you the most?`, + `Regarding "${keyword}", that's something worth exploring. I'm curious to know what sparked your interest in this topic?`, + `"${keyword}" is definitely intriguing. It's always nice chatting with you about these things. Do you have any other thoughts on it?`, + `I really enjoy talking about "${keyword}"! Your ideas are always so unique and give me new perspectives. Let's keep this conversation going.`, + `Hearing you talk about "${keyword}" brightens my day! You always find interesting topics. I'm curious, what made you think of this?`, + `"${keyword}" is such a great topic! I feel like we're on the same wavelength. You know what? Time flies when we chat because it's so enjoyable!`, + `I find "${keyword}" particularly fascinating! You always surprise me. Tell me, have you made any other interesting discoveries lately? I'd love to hear about them!` ], assistant: [ - `关于"${prompt}"，我来为您提供一些有用的信息和建议。`, - `针对"${prompt}"这个问题，我建议您可以从以下几个方面考虑。`, - `"${prompt}"是一个很好的问题，让我来帮您分析一下。`, - `基于"${prompt}"，我可以为您提供以下专业建议。`, - `关于"${prompt}"，我整理了一些相关信息供您参考。` + `Regarding "${keyword}", I'd be happy to provide some useful information and advice. That's a great question - let me organize the relevant details for you.`, + `"${keyword}" is a valuable topic. From what I understand, there are several key points worth noting. First, we can look at...`, + `When it comes to "${keyword}", I'd like to analyze it from several angles. This question actually involves multiple aspects - let me help you sort through the key information.`, + `Your question about "${keyword}" has depth. I suggest considering it from these perspectives: first, understanding the basic concepts; second, analyzing practical applications; and finally, considering future developments.`, + `"${keyword}" is definitely a topic worth discussing. Based on the information I have, I can provide some professional insights. First, we need to clarify...`, + `About your "${keyword}" question, I'd like to provide a clear answer. There are several important aspects to consider - let me analyze them for you.`, + `"${keyword}" is a great question! I'm glad you're interested in this area. Let me share some relevant information that I hope will be helpful.` ], creative: [ - `哇！"${prompt}"让我的创意火花瞬间点燃！让我们一起想象一下...`, - `"${prompt}"真是个充满想象力的话题！我脑海中浮现出无数奇妙的画面～`, - `听到"${prompt}"，我仿佛看到了一个全新的世界！让我们一起探索吧～`, - `"${prompt}"激发了我的灵感！我想到了一个超级有趣的创意...`, - `哇塞！"${prompt}"让我的想象力飞起来了！我们来创造点什么特别的吧～` + `Wow! The topic of "${keyword}" really ignites my creative spark! ✨ Imagine if we expanded this concept into a whole new dimension - what might happen? Perhaps we could...`, + `"${keyword}" is such an imaginative topic! 🌈 I can already picture countless fascinating scenarios. For instance, imagine a world where ${keyword} could...`, + `Hearing "${keyword}" makes me envision a whole new world! 🚀 It reminds me of an interesting story: in a distant place, ${keyword} became the center of people's lives, and then...`, + `"${keyword}" inspires me! 💡 I've thought of a super interesting idea: what if we combined ${keyword} with art? What kind of wonders might we create?`, + `Amazing! "${keyword}" makes my imagination soar! 🎨 We could use this concept as a starting point to create a brand new story or game. Imagine if the main character was...`, + `"${keyword}" is truly a wellspring of creativity! I suddenly wonder, what if we looked at this issue from a completely different angle? What new discoveries might we make? For example, if ${keyword} in the future became...`, + `When you mention "${keyword}", a wonderful image immediately flashes in my mind! Imagine a world full of possibilities where ${keyword} could take any form... isn't that magical?` ] }; + // Get response list for current mode const modeResponses = responses[this.currentMode] || responses.casual; + + // Randomly select a response template const randomResponse = modeResponses[Math.floor(Math.random() * modeResponses.length)]; - return randomResponse; + // Further personalize the response, adding some random personalization elements + return this.personalizeResponse(randomResponse, prompt); + } + + // Extract possible keywords from user input + extractKeywords(prompt) { + // Simple keyword extraction logic + const words = prompt.split(/\s+|[,.!?;:，。！？；：]/); + // Filter out short words and common words + return words.filter(word => + word.length > 1 && + !['the', 'and', 'is', 'in', 'I', 'you', 'he', 'she', 'it', 'they', 'we', 'with', 'this', 'that', 'have', "don't", 'not', 'for'].includes(word) + ); + } + + // Further personalize the response + personalizeResponse(response, prompt) { + // Add some random personalization elements + const personalizations = [ + // Don't add any extra content + (resp) => resp, + // Add a random emoji + (resp) => { + const emojis = ['😊', '💕', '✨', '🌟', '🎵', '🌈', '☺️', '🤔', '👍', '💡']; + const emoji = emojis[Math.floor(Math.random() * emojis.length)]; + return resp + ' ' + emoji; + }, + // Add a random ending phrase + (resp) => { + const endings = [ + "I'd love to hear your thoughts!", + "What do you think?", + "I'm curious about your perspective.", + "Hope my response is helpful!", + "We can continue discussing this topic!" + ]; + const ending = endings[Math.floor(Math.random() * endings.length)]; + return resp + ' ' + ending; + } + ]; + + // Randomly select a personalization method + const personalizer = personalizations[Math.floor(Math.random() * personalizations.length)]; + return personalizer(response); } - // 获取错误回应 + // Get error response getErrorResponse() { const errorResponses = [ - "抱歉，我现在有点困惑，让我重新整理一下思路...", - "嗯...我需要再想想，请稍等一下。", - "我的思绪有点乱，给我一点时间整理一下。", - "让我重新组织一下语言，稍等片刻。", - "哎呀，我刚才走神了，你能再说一遍吗？" + "Sorry, I'm a bit confused right now. Let me gather my thoughts...", + "Hmm... I need to think about this a bit more. Please give me a moment.", + "My thoughts are a bit scattered. Give me a moment to organize them.", + "Let me rephrase that. Just a moment please.", + "Oops, I got distracted. Could you repeat that?" ]; return errorResponses[Math.floor(Math.random() * errorResponses.length)]; } - // 设置聊天模式 + // Set chat mode setChatMode(mode) { if (['casual', 'assistant', 'creative'].includes(mode)) { this.currentMode = mode; - console.log(`聊天模式已切换为: ${mode}`); + console.log(`Chat mode switched to: ${mode}`); return true; } return false; } - // 获取当前配置信息 + // Get current configuration information getCurrentConfig() { return { useCloudAPI: false, @@ -111,15 +169,15 @@ class SimpleBellaAI { }; } - // 清除对话历史（简化版无需实际操作） + // Clear conversation history (no actual operation needed in simplified version) clearHistory() { - console.log('对话历史已清除'); + console.log('Conversation history cleared'); } } -// 将SimpleBellaAI暴露为全局变量 +// Expose SimpleBellaAI as a global variable window.SimpleBellaAI = SimpleBellaAI; -// 同时也暴露为BellaAI，保持兼容性 +// Also expose as BellaAI for compatibility window.BellaAI = SimpleBellaAI; -console.log('SimpleBellaAI 已加载完成'); \ No newline at end of file +console.log('SimpleBellaAI loaded successfully'); \ No newline at end of file