mirror of
https://gitcode.com/ageerle/ruoyi-ai.git
synced 2026-03-13 20:53:42 +08:00
affdc5e3a6
1.保存消息和计费逻辑存在耦合 2.修改计费逻辑： 按次计费被阈值限制：旧逻辑把 TIMES 分支放在 totalTokens ≥ 100 的大分支里，导致没到100 token时不扣费，违背“每次调用就扣费”的语义。 token累计不当：TIMES 分支只扣费不处理累计，同时在 totalTokens < 100 时不会进入任何TIMES逻辑，累计会无意义增长。 粒度不稳定：TOKEN 计费一旦达阈值就把 total 全扣完并清零，不利于对账与用户体验。 打印方式：使用 System.out.println，不利于生产追踪。 改动要点 1.新增独立方法 saveMessage(ChatRequest): 只落库。 publishBillingEvent(ChatRequest): 只发布异步计费事件。 保留组合方法 saveMessageAndPublishEvent(ChatRequest) 以便需要一行调用时使用。 调用处已改为“先保存，再发布事件” SseServiceImpl: 先 saveMessage，再 publishBillingEvent。 SSEEventSourceListener: 同上。 DifyServiceImpl: 同上。 2.计费模式分流： TIMES：每次调用直接扣费，不累计。 TOKEN：按阈值（100）批量扣费，保留余数，账单颗粒稳定。 保留余数：total = prev + delta；billable = floor(total/threshold)threshold；remainder = total % threshold。 日志替换：统一使用 log.debug。 结构更清晰、可维护。 所有金额计算统一用 BigDecimal，保留两位小数，RoundingMode.HALF_UP 按次计费：每次直接扣费（BigDecimal），边界转 Double 按 token 计费：按阈值批量结算，保留余数；费用=单价(BigDecimal)×可结算token数 1. 消息分类存储 用户消息：role="user", deductCost=null, totalTokens=本次token数, remark="用户消息" 系统账单：role="system", deductCost=实际扣费, totalTokens=计费token数, remark="TIMES_BILLING/TOKEN_BILLING" 2. 数据流程 用户发送消息 → 预检查余额 → 保存用户消息 → 发布计费事件 → 异步扣费 → 保存账单记录
RuoYi AI
Enterprise-Grade AI Assistant Platform
Production-ready AI platform with deep integration of FastGPT, Coze, DIFY and advanced RAG technology
📖 中文文档 | 📚 Documentation | 🚀 Live Demo | 🐛 Report Bug | 💡 Request Feature
✨ Key Features
🤖 Advanced AI Engine
- Multi-Model Support: OpenAI GPT-4, Azure, ChatGLM, Qwen, ZhipuAI
- AI Platform Integration: Deep integration with FastGPT, Coze, DIFY and other leading AI platforms
- Spring AI MCP Integration: Extensible tool ecosystem with Model Context Protocol
- Streaming Chat: Real-time SSE/WebSocket communication
- AI Copilot: Intelligent code analysis and project scaffolding
🌟 AI Platform Ecosystem
- FastGPT Deep Integration: Native FastGPT API support with knowledge base retrieval, workflow orchestration and context management
- Coze Official SDK: Integration with ByteDance Coze platform official SDK, supporting Bot conversations and streaming responses
- DIFY Full Compatibility: Using DIFY Java Client for app orchestration, workflows and knowledge base management
- Unified Chat Interface: Standardized chat service interface supporting seamless platform switching and load balancing
🧠 Enterprise RAG Solution
- Local Knowledge Base: Langchain4j + BGE-large-zh-v1.5 embeddings
- Vector Database Support: Milvus, Weaviate, Qdrant
- Privacy-First: On-premise deployment with local LLM support
- Ollama & vLLM Compatible: Flexible model deployment options
🎨 Creative AI Tools
- AI Art Generation: DALL·E-3, MidJourney, Stable Diffusion integration
- PPT Creation: Automated slide generation from text input
- Multi-Modal Processing: Text, image, and document understanding
🚀 Quick Start
Live Demo
- User Portal: web.pandarobot.chat (demo/demo123)
- Admin Panel: admin.pandarobot.chat (admin/admin123)
Source Code
| Component | GitHub | Gitee | GitCode |
|---|---|---|---|
| Backend API | ruoyi-ai | ruoyi-ai | ruoyi-ai |
| User Frontend | ruoyi-web | ruoyi-web | ruoyi-web |
| Admin Frontend | ruoyi-admin | ruoyi-admin | ruoyi-admin |
Collaborative Projects
| Project Description | GitHub Repository | Gitee Repository |
|---|---|---|
| Simplified Frontend | ruoyi-element-ai | ruoyi-element-ai |
🛠️ Tech Stack
Core Framework
- Backend: Spring Boot 3.4, Spring AI, Langchain4j
- Database: MySQL 8.0, Redis, Vector Databases (Milvus/Weaviate/Qdrant)
- Frontend: Vue 3, Vben Admin, Naive UI
- Authentication: Sa-Token, JWT
System Components
- File Processing: PDF, Word, Excel parsing, intelligent image analysis
- Real-time Communication: WebSocket real-time communication, SSE streaming
- System Monitoring: Comprehensive logging, performance monitoring, health checks
📚 Documentation
For detailed setup, configuration, and development guides, visit our comprehensive documentation:
🤝 Contributing
We welcome contributions from developers of all skill levels! Whether you're fixing bugs, adding features, or improving documentation, your help is appreciated.
How to Contribute
- Fork the repository
- Create your feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
Please submit PRs to GitHub - they will be synchronized to other platforms automatically.
📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
🙏 Acknowledgments
Special thanks to these amazing open source projects:
- Spring AI Alibaba Copilot - Intelligent coding assistant based on spring-ai-alibaba with MCP protocol integration for project analysis and code generation
- Spring AI - Spring's AI integration framework
- Langchain4j - Java LLM framework
- RuoYi-Vue-Plus - Enterprise development framework
- Vben Admin - Vue admin template
- chatgpt-java - ChatGPT Java SDK
🌐 Ecosystem Partners
- PPIO Cloud - Cost-effective GPU containers and model APIs
💬 Community
 Add Author WeChat Scan to join learning group |
 QQ Group Technical discussion |
⭐ Star this repo • 🍴 Fork it • 📖 中文文档 • 📚 Documentation
Built with ❤️ by the RuoYi AI community