Merge branch 'fix/sse-session-trace-cleanup'

fix(sse): SSE 改为按会话隔离 + 修 trace 租户 bug + 清理冗余
SSE 串台修复： - SseEmitterManager 新增按 sessionId 维度的 connect/sendEvent/disconnect，每个会话一个 SSE 连接，替代原 userId+token 维度（同用户多会话串台） - SseMessageDto 增加 sessionId + eventDto 字段，跨实例按会话路由 - SseTopicListener 优先按 sessionId 路由，回退原 userId/群发逻辑 - SseMessageUtils 增加 sessionId 重载，对话链路全部切换 - ChatServiceFacade / MyMcpClientListener / WorkflowStarter 切到 sessionId - 通知/全局 /sse 端点保留 userId 模式（通知按用户） trace 租户 bug 修复： - trace_run / trace_node 加入 tenant.excludes，监控表跨租户全局可见，绕过异步线程租户上下文不传播导致 node 查不到的问题冗余清理： - 删除零使用的 @TraceNode 注解 + TraceNodeAspect 切面 + 对应单测 - TraceProperties 删除 recordDetail/maxInputLength/maxOutputLength （仅对已删切面生效，编程式埋点用 RagTracePayloadBuilder 不受影响） - TracePayloadUtils 删除 input/output/asString 方法 - TraceConstants 删除无引用的 NODE_METHOD/HTTP/DB/CACHE/TASK/STREAM - TraceContext 删除预留的 copyNodeStack - TraceNodeVo 删除未用的 children 字段 - TraceRecordServiceImpl.getDetail 删除重复 enrich - application.yml 删除 trace.payload 无效配置项 - RagTraceNodeTypes 删除无引用的 NODE_STREAM Co-Authored-By: Claude <noreply@anthropic.com>
2026-09-13 00:14:59 +00:00 · 2026-07-23 15:38:50 +08:00 · 2026-07-23 15:35:10 +08:00 · 2026-07-23 15:10:40 +08:00 · 2026-07-23 14:52:19 +08:00 · 2026-07-23 14:45:43 +08:00
321 changed files with 22945 additions and 2204 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -45,6 +45,7 @@ nbdist/
 *.xml.versionsBackup
 *.swp
 data/
+logs/

 !*/build/*.java
 !*/build/*.html
@@ -52,3 +53,4 @@ data/

 .flattened-pom.xml
 /.claude/settings.local.json
+/docs/docker/milvus/volumes/
--- a/RAG_TEST_REPORT_2026-07-20.md
+++ b/RAG_TEST_REPORT_2026-07-20.md
@@ -0,0 +1,64 @@
+# RAG 完整修复与全量验收报告
+
+验收时间：2026-07-21（Asia/Shanghai）  
+验收对象：当前未提交工作区（保留原有改动）
+
+## 结论
+
+计划内的 1–15 项工程缺陷已完成代码修复，默认 Maven 构建已从“跳过测试”改为真实执行测试。全仓 37 个 reactor 模块测试成功，`ruoyi-chat` 49/49 通过，两个前端生产构建通过，`git diff --check` 通过。
+
+本机已运行 MySQL、Redis、MinIO 和 Weaviate 1.30.0；Milvus/Qdrant 容器以及有效的 embedding/chat/rerank provider 凭证不存在，因此这三项真实 provider/存储引擎冒烟被标记为环境限制，不影响确定性代码验收。
+
+## 1–15 项验收
+
+| # | 状态 | 修复/证据 |
+|---|---|---|
+| 1 | 通过 | Markdown/Java/字符分片的边界、空文档、超长块回归通过。 |
+| 2 | 通过 | Supervisor 每轮仅保留一个 RAG 入口，不再用已含 RAG 的 prompt 重复检索。 |
+| 3 | 通过 | 历史消息进入 Supervisor prompt，检索 query 与最终 prompt 分离。 |
+| 4 | 通过 | `fid` 稳定 ID 贯穿 DB/三种向量库/RRF，融合去重回归通过。 |
+| 5 | 通过 | aiflow vector/hybrid 复用统一检索服务；graph 明确返回不支持，不再伪装为 vector。 |
+| 6 | 通过 | 重解析改为先写新 fid、再清旧向量、最后替换 DB；失败补偿新向量；删片段/附件/库遇向量删除失败即中止。 |
+| 7 | 通过 | embedding/rerank provider 使用 prototype 实例，工厂缓存可按模型刷新，避免跨配置污染。 |
+| 8 | 通过 | `similarityThreshold` 仅用于粗召回；`rerankScoreThreshold` 仅在 rerank 真实成功后生效，回归测试通过。 |
+| 9 | 通过 | 默认配置和 Compose 统一为 Weaviate 1.30.0、`28080:8080`。 |
+| 10 | 通过 | 三种策略均使用 `embedAll`；Weaviate batch objects、Milvus `addAll`、Qdrant `addAll`。 |
+| 11 | 通过 | upload/parse/retrieval 权限保留，parse/retrieval 增加分布式防重复提交，upload 由现有知识库+文件名唯一约束兜底。 |
+| 12 | 通过 | 分隔符使用字面量语义，`|`/`.`/`*` 回归通过。 |
+| 13 | 通过 | hybrid 通道失败可降级到 vector；所有可用通道都失败时抛出明确业务异常。 |
+| 14 | 通过 | Weaviate client 稳定懒加载单例；schema 仅在已存在或创建成功后进入缓存。 |
+| 15 | 通过 | 工厂新增严格 `getStrategy(type)`，知识库 `vectorModel` 优先，空值才回退全局，非法值直接报错。 |
+
+## 其他完成项
+
+- 多知识库并行检索，按 `kid + docId + fid` 去重，统一上限和字符预算。
+- 5 分钟短 TTL 检索缓存，key 覆盖检索参数，知识数据变更主动失效。
+- rerank 仅保留 provider 实际返回的文档。
+- 知识库文档数改为 group-by 查询，消除该 N+1。
+- Milvus/Qdrant/Weaviate 的删 collection/doc/fid 语义对齐；Milvus 删库改为 drop collection。
+- MCP `npx` 根据操作系统解析，支持系统属性/环境变量覆盖。
+- `fid` 非空唯一、`doc_id varchar(32)`、租户/用户索引与可重复执行迁移脚本已提供。
+- 用户端聊天页已接入知识库列表和最小选择器。
+
+## 测试记录
+
+| 检查 | 结果 |
+|---|---|
+| `mvn -Pdev test` | 37/37 reactor 模块 SUCCESS；`ruoyi-chat` 49/49 |
+| `mvn -Pdev -pl ruoyi-modules/ruoyi-aiflow -am -DskipTests compile` | 21/21 SUCCESS |
+| `ruoyi-web: pnpm build` | SUCCESS，2621 modules transformed |
+| `ruoyi-admin: pnpm build` | SUCCESS，10/10 build tasks |
+| `git diff --check` | SUCCESS，无空白错误 |
+| Weaviate/MySQL/Redis/MinIO | Docker 服务运行，Weaviate 1.30.0 映射 28080 |
+| 三向量库 Docker 集成 | SUCCESS；Weaviate 1.30.0、Milvus 2.5.7、Qdrant 1.17.0 真实写入/检索/删除测试 3/3 通过 |
+| 真实 embedding/chat/rerank | 环境限制：当前配置为无效/占位凭证 |
+
+本轮未创建新的 `codex_rag_verify_` 持久化数据；上一轮验收数据已清理，未动现有非测试数据。
+
+## 2026-07-21 三向量库 Docker 补充验收
+
+- 启动并保留 `ruoyi-rag-milvus`、`ruoyi-rag-milvus-etcd`、`ruoyi-rag-milvus-minio`、`ruoyi-rag-qdrant`，四个容器健康检查均为 `healthy`。
+- Milvus 专用 MinIO 仅在 Docker 内网可达，没有占用宿主机 9000/9001；Milvus 映射 19530/9091，Qdrant 映射 6333/6334。
+- `ThreeVectorStoresDockerIT` 使用 32 维确定性 embedding，对三库逐一验证 batch write、vector search、fid delete、docId delete 和 drop collection，3/3 通过。
+- 首轮测试发现 Milvus `autoFlush=false` 导致批量写入后不可立即检索、元数据删除不可立即见；改为写入和删除返回前 flush 后通过。
+- 清理后 Weaviate/Qdrant 的 `CodexRagVerify*` collection 计数均为 0，Milvus collection 也由测试 finally 成功 drop；本轮未写入 MySQL 或 OSS 测试数据。
--- a/README.md
+++ b/README.md
@@ -21,30 +21,22 @@

 *开箱即用的全栈AI平台，支持多智能体协同、Supervisor模式编排、多种决策模式、RAG技术和流程编排能力*

-**[English](README_EN.md)** | **[📖 使用文档](https://doc.pandarobot.chat)** |
-**[🚀 在线体验](https://web.pandarobot.chat)** | **[🐛 问题反馈](https://github.com/ageerle/ruoyi-ai/issues)** | **[💡 功能建议](https://github.com/ageerle/ruoyi-ai/issues)**
+**[English](README_EN.md)** | **[📖 使用文档](https://doc.ruoyiai.chat/)** |
+**[🚀 在线体验](https://web.ruoyiai.chat/)** | **[🐛 问题反馈](https://github.com/ageerle/ruoyi-ai/issues)** | **[💡 功能建议](https://github.com/ageerle/ruoyi-ai/issues)**

 </div>


 ## ✨ 核心亮点

-|     模块     | 现有能力 
-|:----------:|---
-|  **模型管理**  | 多模型接入(OpenAI/DeepSeek/通义/智谱)、多模态理解、Coze/DIFY/FastGPT平台集成 
-|  **知识管理**  | 本地RAG + 向量库(Milvus/Weaviate/Qdrant)  + 文档解析        
-|  **工具管理**  | Mcp协议集成、Skills能力 + 可扩展工具生态                             
-|  **流程编排**  | 可视化工作流设计器、节点拖拽编排、SSE流式执行,目前已经支持模型调用,邮件发送,人工审核等节点  
-|  **多智能体**  | 基于Langchain4j的Agent框架、Supervisor模式编排,支持多种决策模型          
+|    模块     | 现有能力 
+|:---------:|---
+| **模型管理**  | 多模型接入(DeepSeek/智谱/MIMO/百炼/OpenAI)、多模态理解、Coze/DIFY/FastGPT/RAGFlow平台集成
+| **知识管理**  | 本地RAG + 向量库(Milvus/Weaviate/Qdrant)  + 文档解析
+| **工具管理**  | Mcp协议集成、Skills能力 + 可扩展工具生态                             
+| **流程编排**  | 可视化工作流设计器、节点拖拽编排、SSE流式执行,目前已经支持模型调用,邮件发送,人工审核等节点  
+| **智能体管理** | 基于Langchain4j的Agent框架、Supervisor模式编排,支持多种决策模型,可以灵活搭配工具,skills          

-## 🚀 快速体验
-
-### 在线演示
-
-|   平台   | 地址 | 账号 |
-|:------:|---|---|
-|  用户端   | [web.pandarobot.chat](https://web.pandarobot.chat) | admin / admin123 |
-| 管理后台 | [admin.pandarobot.chat](https://admin.pandarobot.chat) | admin / admin123 |

 ### 项目源码

@@ -53,6 +45,9 @@
 | 🔧 后端服务  | [ruoyi-ai](https://github.com/ageerle/ruoyi-ai)       | [ruoyi-ai](https://gitee.com/ageerle/ruoyi-ai)       | [ruoyi-ai](https://gitcode.com/ageerle/ruoyi-ai)       |
 | 🎨 用户前端  | [ruoyi-web](https://github.com/ageerle/ruoyi-web)     | [ruoyi-web](https://gitee.com/ageerle/ruoyi-web)     | [ruoyi-web](https://gitcode.com/ageerle/ruoyi-web)     |
 | 🛠️ 管理后台 | [ruoyi-admin](https://github.com/ageerle/ruoyi-admin) | [ruoyi-admin](https://gitee.com/ageerle/ruoyi-admin) | [ruoyi-admin](https://gitcode.com/ageerle/ruoyi-admin) |
+| 🎬 短剧平台  | [ruoyi-drama](https://github.com/ageerle/ruoyi-drama) | [ruoyi-drama](https://gitee.com/ageerle/ruoyi-drama) | [ruoyi-drama](https://gitcode.com/ageerle/ruoyi-drama)   |
+| 🤖 编程助手  | [ruoyi-copilot](https://github.com/ageerle/ruoyi-copilot) | [ruoyi-copilot](https://gitee.com/ageerle/ruoyi-copilot) | [ruoyi-copilot](https://gitcode.com/ageerle/ruoyi-copilot) |
+| 📱 小程序端  | [ruoyi-uniapp](https://github.com/ageerle/ruoyi-uniapp) | [ruoyi-uniapp](https://gitee.com/ageerle/ruoyi-uniapp) | [ruoyi-uniapp](https://gitcode.com/ageerle/ruoyi-uniapp) |

 ### 合作项目
 | 项目名称           | GitHub 仓库                                             | Gitee 仓库                                             
@@ -66,8 +61,6 @@
 - **数据存储**：MySQL 8.0 + Redis + 向量数据库（Milvus/Weaviate/Qdrant）
 - **前端技术**：Vue 3 + Vben Admin + element-plus-x
 - **安全认证**：Sa-Token + JWT 双重保障
-
-
 - **文档处理**：PDF、Word、Excel 解析，图像智能分析
 - **实时通信**：WebSocket 实时通信，SSE 流式响应
 - **系统监控**：完善的日志体系、性能监控、服务健康检查
@@ -187,7 +180,7 @@ docker-compose -f docker-compose-all.yaml restart [服务名]

 想要深入了解安装部署、功能配置和二次开发？

-**👉 [完整使用文档](https://doc.pandarobot.chat)**
+**👉 [完整使用文档](https://doc.ruoyiai.chat/)**

 ## 🤝 参与贡献

@@ -210,17 +203,29 @@ docker-compose -f docker-compose-all.yaml restart [服务名]
 ## 🙏 特别鸣谢

 感谢以下优秀的开源项目为本项目提供支持：
- [Spring AI Alibaba Copilot](https://github.com/spring-ai-alibaba/copilot) - 基于spring-ai-alibaba
-  的智能编码助手
 - [Langchain4j](https://github.com/langchain4j/langchain4j) - 强大的 Java LLM 开发框架
 - [RuoYi-Vue-Plus](https://gitee.com/dromara/RuoYi-Vue-Plus) - 成熟的企业级快速开发框架
 - [Vben Admin](https://github.com/vbenjs/vue-vben-admin) - 现代化的 Vue 后台管理模板

-## 🌐 生态伙伴

- [PPIO 派欧云](https://ppinfra.com/user/register?invited_by=P8QTUY&utm_source=github_ruoyi-ai) - 提供高性价比的 GPU
-  算力和模型 API 服务
- [优云智算](https://www.compshare.cn/?ytag=GPU_YY-gh_ruoyi) - 万卡RTX40系GPU+海内外主流模型API服务，秒级响应，按量计费，新客免费用。
+## 💎 赞助商
+
+**感谢以下赞助商对本项目的支持：**
+
+<a href="https://www.atlascloud.ai?ref=89F97E">
+  <img src="docs/image/sponsor/atlascloud_banner.png" alt="Atlas Cloud" width="160" height="80">
+</a>
+
+[访问Atlas Cloud官网](https://www.atlascloud.ai?ref=89F97E&utm_source=github&utm_campaign=ruoyi-drama) · [编程计划优惠](https://www.atlascloud.ai/console/coding-plan)
+全模态 AI 推理平台，为开发者提供统一的 AI API，支持视频生成、图像生成和大语言模型。一次接入，即可访问 **300+ 精选模型**。
+
+<a href="https://www.volcengine.com/activity/codingplan?utm_campaign=hw&utm_content=hw&utm_medium=devrel_tool_web&utm_source=OWO&utm_term=ageerle-ruoyi-ai">
+  <img src="docs/image/sponsor/huoshan.png" alt="火山引擎 CodingPlan" width="160" height="80">
+</a>
+
+[注册即领2500万Tokens，立即前往](https://www.volcengine.com/activity/ai618?utm_campaign=hw&utm_content=hw&utm_medium=devrel_tool_web&utm_source=OWO&utm_term=ageerle-ruoyi-ai)
+享字节自研豆包模型+满血版开源 SOTA模型，覆盖文本、VLM、图像生成，全模态一站配齐：Seed-2.1、Seedream-5.0、GLM-5.2、DeepSeek等。不止编程、更能解决 Agent 复杂长程任务！
+

 ## 💬 社区交流

@@ -234,6 +239,11 @@ docker-compose -f docker-compose-all.yaml restart [服务名]
 <em>邀请进群学习</em>
 </td>
 <td align="center">
+<img src="docs/image/wx06.png" alt="微信二维码" width="200" height="200"><br>
+<strong>微信技术交流群</strong><br>
+<em>技术讨论</em>
+</td>
+<td align="center">
 <img src="docs/image/qq.png" alt="QQ群二维码" width="200" height="200"><br>
 <strong>QQ技术交流群</strong><br>
 <em>技术讨论</em>
@@ -247,7 +257,7 @@ docker-compose -f docker-compose-all.yaml restart [服务名]
 ---
 <div align="center">

-**[⭐ 点个Star支持一下](https://github.com/ageerle/ruoyi-ai)** • **[ Fork 开始贡献](https://github.com/ageerle/ruoyi-ai/fork)** • **[📚 English](README_EN.md)** • **[📖 查看完整文档](https://doc.pandarobot.chat)**
+**[⭐ 点个Star支持一下](https://github.com/ageerle/ruoyi-ai)** • **[ Fork 开始贡献](https://github.com/ageerle/ruoyi-ai/fork)** • **[📚 English](README_EN.md)** • **[📖 查看完整文档](https://doc.ruoyiai.chat/)**

 *用 ❤️ 打造，由 RuoYi AI 开源社区维护*

--- a/README_EN.md
+++ b/README_EN.md
@@ -22,8 +22,8 @@

 *An out-of-the-box full-stack AI platform supporting multi-agent collaboration, Supervisor mode orchestration, and multiple decision models, with advanced RAG technology and visual workflow orchestration capabilities*

-**[中文](README.md)** | **[📖 Documentation](https://doc.pandarobot.chat)** |
-**[🚀 Live Demo](https://web.pandarobot.chat)** | **[🐛 Report Issues](https://github.com/ageerle/ruoyi-ai/issues)** | **[💡 Feature Requests](https://github.com/ageerle/ruoyi-ai/issues)**
+**[中文](README.md)** | **[📖 Documentation](https://doc.ruoyiai.chat/)** |
+**[🚀 Live Demo](https://web.ruoyiai.chat/)** | **[🐛 Report Issues](https://github.com/ageerle/ruoyi-ai/issues)** | **[💡 Feature Requests](https://github.com/ageerle/ruoyi-ai/issues)**

 </div>

@@ -34,21 +34,12 @@

 | Module | Current Capabilities |
 |:---:|---|
-| **Model Management** | Multi-model integration (OpenAI/DeepSeek/Tongyi/Zhipu), multi-modal understanding, Coze/DIFY/FastGPT platform integration |
+| **Model Management** | Multi-model integration (OpenAI/DeepSeek/Tongyi/Zhipu/MiniMax), multi-modal understanding, Coze/DIFY/FastGPT platform integration |
 | **Knowledge Base** | Local RAG + Vector DB (Milvus/Weaviate/Qdrant) + Document parsing |
 | **Tool Management** | MCP protocol integration, Skills capability + Extensible tool ecosystem |
 | **Workflow Orchestration** | Visual workflow designer, drag-and-drop node orchestration, SSE streaming execution, currently supports model calls, email sending, manual review nodes |
 | **Multi-Agent** | Agent framework based on Langchain4j, Supervisor mode orchestration, supports multiple decision models |

-## 🚀 Quick Start
-
-### Live Demo
-
-|   Platform   | URL | Account |
-|:------:|---|---|
-|  User Frontend   | [web.pandarobot.chat](https://web.pandarobot.chat) | admin / admin123 |
-| Admin Panel | [admin.pandarobot.chat](https://admin.pandarobot.chat) | admin / admin123 |
-
 ### Project Repositories

 | Module     | GitHub Repository                                             | Gitee Repository                                             | GitCode Repository                                             |
@@ -56,6 +47,9 @@
 | 🔧 Backend  | [ruoyi-ai](https://github.com/ageerle/ruoyi-ai)       | [ruoyi-ai](https://gitee.com/ageerle/ruoyi-ai)       | [ruoyi-ai](https://gitcode.com/ageerle/ruoyi-ai)       |
 | 🎨 User Frontend  | [ruoyi-web](https://github.com/ageerle/ruoyi-web)     | [ruoyi-web](https://gitee.com/ageerle/ruoyi-web)     | [ruoyi-web](https://gitcode.com/ageerle/ruoyi-web)     |
 | 🛠️ Admin Panel | [ruoyi-admin](https://github.com/ageerle/ruoyi-admin) | [ruoyi-admin](https://gitee.com/ageerle/ruoyi-admin) | [ruoyi-admin](https://gitcode.com/ageerle/ruoyi-admin) |
+| 🎬 Drama | [ruoyi-drama](https://github.com/ageerle/ruoyi-drama) | [ruoyi-drama](https://gitee.com/ageerle/ruoyi-drama) | |
+| 🤖 Copilot | [ruoyi-copilot](https://github.com/ageerle/ruoyi-copilot) | [ruoyi-copilot](https://gitee.com/ageerle/ruoyi-copilot) | [ruoyi-copilot](https://gitcode.com/ageerle/ruoyi-copilot) |
+| 📱 Mini-App | [ruoyi-uniapp](https://github.com/ageerle/ruoyi-uniapp) | [ruoyi-uniapp](https://gitee.com/ageerle/ruoyi-uniapp) | [ruoyi-uniapp](https://gitcode.com/ageerle/ruoyi-uniapp) |

 ### Partner Projects
 | Project Name           | GitHub Repository                                             | Gitee Repository |
@@ -69,8 +63,6 @@
 - **Data Storage**: MySQL 8.0 + Redis + Vector Databases (Milvus/Weaviate/Qdrant)
 - **Frontend**: Vue 3 + Vben Admin + element-plus-x
 - **Security**: Sa-Token + JWT dual-layer security
-
-
 - **Document Processing**: PDF, Word, Excel parsing, intelligent image analysis
 - **Real-time Communication**: WebSocket real-time communication, SSE streaming response
 - **System Monitoring**: Comprehensive logging system, performance monitoring, service health checks
@@ -186,11 +178,39 @@ docker-compose -f docker-compose-all.yaml logs -f [service-name]
 docker-compose -f docker-compose-all.yaml restart [service-name]
 ```

+### MiniMax Configuration
+
+The built-in MiniMax provider accepts one API Host value and selects the matching protocol adapter. Use a Base URL from this table:
+
+| Region | OpenAI-compatible Base URL | Anthropic-compatible Base URL |
+| --- | --- | --- |
+| Global | `https://api.minimax.io/v1` | `https://api.minimax.io/anthropic` |
+| China | `https://api.minimaxi.com/v1` | `https://api.minimaxi.com/anthropic` |
+
+For Anthropic-compatible requests, configure the Base URL ending in `/anthropic`. Do not append `/v1` or `/v1/messages`; the provider adapter derives the request path internally.
+
+| Model ID | Total context | Input modalities | Thinking |
+| --- | ---: | --- | --- |
+| `MiniMax-M3` | 1,000,000 tokens | Text, image, video | Adaptive or disabled |
+| `MiniMax-M2.7` | 204,800 tokens | Text | Always on |
+
+Current pay-as-you-go prices are in USD per million tokens:
+
+| Model | Service tier and input range | Input | Output | Cache read | Cache write |
+| --- | --- | ---: | ---: | ---: | ---: |
+| `MiniMax-M3` | Standard, up to 512,000 input tokens | $0.30 | $1.20 | $0.06 | Not listed |
+| `MiniMax-M3` | Standard, over 512,000 input tokens | $0.60 | $2.40 | $0.12 | Not listed |
+| `MiniMax-M3` | Priority, up to 512,000 input tokens | $0.45 | $1.80 | $0.09 | Not listed |
+| `MiniMax-M3` | Priority, over 512,000 input tokens | $0.90 | $3.60 | $0.18 | Not listed |
+| `MiniMax-M2.7` | Standard | $0.30 | $1.20 | $0.06 | $0.375 |
+
+See the [official API overview](https://platform.minimax.io/docs/api-reference/api-overview) and [pay-as-you-go pricing](https://platform.minimax.io/docs/guides/pricing-paygo) for current details.
+
 ## 📚 Documentation

 Want to learn more about installation, deployment, configuration, and secondary development?

-**👉 [Complete Documentation](https://doc.pandarobot.chat)**
+**👉 [Complete Documentation](https://doc.ruoyiai.chat/)**

 ## 🤝 Contributing

@@ -213,23 +233,47 @@ This project is licensed under the **MIT License**. See the [LICENSE](LICENSE) f
 ## 🙏 Acknowledgments

 Thanks to the following excellent open-source projects for their support:
- [Spring AI Alibaba Copilot](https://github.com/spring-ai-alibaba/copilot) - Intelligent coding assistant based on spring-ai-alibaba
 - [Langchain4j](https://github.com/langchain4j/langchain4j) - Powerful Java LLM development framework
 - [RuoYi-Vue-Plus](https://gitee.com/dromara/RuoYi-Vue-Plus) - Mature enterprise-level rapid development framework
 - [Vben Admin](https://github.com/vbenjs/vue-vben-admin) - Modern Vue admin template

-## 🌐 Ecosystem Partners
+## 💎 Sponsors

- [PPIO Cloud](https://ppinfra.com/user/register?invited_by=P8QTUY&utm_source=github_ruoyi-ai) - Provides cost-effective GPU computing and model API services
- [Youyun Intelligent Computing](https://www.compshare.cn/?ytag=GPU_YY-gh_ruoyi) - Thousands of RTX40 series GPUs + mainstream models API services, second-level response, pay-per-use, free for new customers.
+**Thanks to the following sponsors for supporting this project:**

+<a href="https://www.atlascloud.ai?ref=89F97E">
+  <img src="docs/image/sponsor/atlascloud_banner.png" alt="Atlas Cloud" width="160" height="80">
+</a>
+
+[Visit Atlas Cloud](https://www.atlascloud.ai?ref=89F97E) · [Coding Plan Promotion](https://www.atlascloud.ai/console/coding-plan)
+A full-modal AI inference platform that gives developers a unified AI API, supporting video generation, image generation, and LLMs. Connect once to access **300+ curated models**.
+
+<a href="https://www.volcengine.com/activity/codingplan?utm_campaign=hw&utm_content=hw&utm_medium=devrel_tool_web&utm_source=OWO&utm_term=ageerle-ruoyi-ai">
+  <img src="docs/image/sponsor/huoshan.png" alt="Volcengine CodingPlan" width="160" height="80">
+</a>
+
+[Volcengine CodingPlan Developer Program](https://www.volcengine.com/activity/codingplan?utm_campaign=hw&utm_content=hw&utm_medium=devrel_tool_web&utm_source=OWO&utm_term=ageerle-ruoyi-ai)
+Volcengine is ByteDance's cloud and AI service platform. Volcengine Ark provides API access to Doubao LLM, DeepSeek, and more — a one-stop AI development and inference platform for developers.

 ## 💬 Community Chat

 <div align="center">

-**[📱 Join Telegram Group](
-https://t.me/+LqooQAc5HxRmYmE1)**
+<table>
+<tr>
+<td align="center">
+<img src="docs/image/wx.png" alt="WeChat QR Code" width="200" height="200"><br>
+<strong>Scan to add author on WeChat</strong><br>
+<em>Join group for learning</em>
+</td>
+<td align="center">
+<img src="docs/image/qq.png" alt="QQ Group QR Code" width="200" height="200"><br>
+<strong>QQ Tech Exchange Group</strong><br>
+<em>Technical discussion</em>
+</td>
+
+</tr>
+</table>

 </div>

@@ -237,7 +281,7 @@ https://t.me/+LqooQAc5HxRmYmE1)**

 <div align="center">

-**[⭐ Star to Support](https://github.com/ageerle/ruoyi-ai)** • **[Fork to Contribute](https://github.com/ageerle/ruoyi-ai/fork)** • **[📚 中文](README.md)** • **[📖 Complete Documentation](https://doc.pandarobot.chat)**
+**[⭐ Star to Support](https://github.com/ageerle/ruoyi-ai)** • **[Fork to Contribute](https://github.com/ageerle/ruoyi-ai/fork)** • **[📚 中文](README.md)** • **[📖 Complete Documentation](https://doc.ruoyiai.chat/)**

 *Built with ❤️, maintained by the RuoYi AI open-source community*

--- a/docs/docker/milvus/docker-compose.yml
+++ b/docs/docker/milvus/docker-compose.yml
@@ -1,8 +1,6 @@
-version: '3.5'
- 
 services:
  etcd:
-    container_name: milvus-etcd
+    container_name: ruoyi-rag-milvus-etcd
    image: quay.io/coreos/etcd:v3.5.18
    environment:
      - ETCD_AUTO_COMPACTION_MODE=revision
@@ -19,14 +17,11 @@ services:
      retries: 3
 
  minio:
-    container_name: milvus-minio
+    container_name: ruoyi-rag-milvus-minio
    image: minio/minio:RELEASE.2023-03-20T20-16-18Z
    environment:
      MINIO_ACCESS_KEY: minioadmin
      MINIO_SECRET_KEY: minioadmin
-    ports:
-      - "9001:9001"
-      - "9000:9000"
    volumes:
      - ${DOCKER_VOLUME_DIRECTORY:-.}/volumes/minio:/minio_data
    command: minio server /minio_data --console-address ":9001"
@@ -37,7 +32,7 @@ services:
      retries: 3
 
  standalone:
-    container_name: milvus-standalone
+    container_name: ruoyi-rag-milvus
    image: milvusdb/milvus:v2.5.7
    command: ["milvus", "run", "standalone"]
    security_opt:
@@ -61,7 +56,7 @@ services:
      - "minio"
 
  attu:
-    container_name: attu
+    container_name: ruoyi-rag-attu
    image: zilliz/attu:v2.5.7
    environment:
      MILVUS_URL: milvus-standalone:19530
@@ -72,4 +67,4 @@ services:
 
 networks:
  default:
-    name: milvus
+    name: ruoyi-rag-milvus
--- a/docs/docker/qdrant/docker-compose.yml
+++ b/docs/docker/qdrant/docker-compose.yml
@@ -1,12 +1,20 @@
 ---
 services:
  qdrant:
-    image: qdrant/qdrant:latest
+    container_name: ruoyi-rag-qdrant
+    image: qdrant/qdrant:v1.17.0
    ports:
      - 6333:6333
      - 6334:6334
    volumes:
      - qdrant_data:/qdrant/storage
+    healthcheck:
+      test: ["CMD-SHELL", "bash -c 'exec 3<>/dev/tcp/127.0.0.1/6333 && printf \"GET /healthz HTTP/1.1\\r\\nHost: localhost\\r\\nConnection: close\\r\\n\\r\\n\" >&3 && grep -q \"200 OK\" <&3'"]
+      interval: 10s
+      timeout: 5s
+      retries: 12
+      start_period: 10s
+    restart: unless-stopped
 volumes:
  qdrant_data:
 ...
--- a/docs/docker/ruoyi-ai/Dockerfile.backend
+++ b/docs/docker/ruoyi-ai/Dockerfile.backend
@@ -20,6 +20,22 @@ RUN mvn clean package -Pprod -DskipTests
 # 最终运行镜像
 FROM eclipse-temurin:17-jre-alpine

+RUN set -eux; \
+    apk add --no-cache ffmpeg; \
+    for filter in scale pad fps setsar format tpad trim settb setpts xfade \
+        aresample aformat apad atrim asetpts anullsrc concat acrossfade; do \
+        ffmpeg -hide_banner -filters 2>/dev/null | grep -Eq "[[:space:]]${filter}[[:space:]]"; \
+    done; \
+    ffmpeg -hide_banner -h filter=xfade 2>&1 | grep -q dissolve; \
+    ffmpeg -hide_banner -h filter=xfade 2>&1 | grep -q fadeblack; \
+    ffmpeg -hide_banner -h filter=xfade 2>&1 | grep -q slideleft; \
+    ffmpeg -hide_banner -encoders 2>/dev/null | grep -Eq '[[:space:]]libx264[[:space:]]'; \
+    ffmpeg -hide_banner -encoders 2>/dev/null | grep -Eq '[[:space:]]aac[[:space:]]'; \
+    ffprobe -hide_banner -version >/dev/null
+
+ENV FFMPEG_PATH=/usr/bin/ffmpeg \
+    FFPROBE_PATH=/usr/bin/ffprobe
+
 # 设置工作目录
 WORKDIR /app

@@ -27,10 +43,10 @@ WORKDIR /app
 COPY --from=builder /build/ruoyi-admin/target/ruoyi-admin.jar ./ruoyi-admin.jar

 # 创建日志目录
-RUN mkdir -p /ruoyi/server/logs
+RUN mkdir -p /ruoyi/server/logs /ruoyi/server/temp

 # 暴露端口
 EXPOSE 6039

 # 启动命令
-ENTRYPOINT ["java", "-jar", "ruoyi-admin.jar", "--spring.profiles.active=prod"]
+ENTRYPOINT ["java", "-Djava.io.tmpdir=/ruoyi/server/temp", "-jar", "ruoyi-admin.jar", "--spring.profiles.active=prod"]
--- a/docs/docker/ruoyi-ai/Dockerfile.mysql
+++ b/docs/docker/ruoyi-ai/Dockerfile.mysql
@@ -1,16 +1,14 @@
 # 基于官方MySQL 8.0镜像构建自定义镜像
 # 构建命令: docker build -t registry.cn-hangzhou.aliyuncs.com/ruoyi-ai/mysql:v3 -f Dockerfile.mysql .
-FROM mysql:8.0.33
+FROM mysql:8.0

 # 设置时区
 ENV TZ=Asia/Shanghai

-# 复制初始化脚本和SQL文件到镜像中
-COPY docs/script/docker/mysql/init/init-db.sh /docker-entrypoint-initdb.d/init-db.sh
-COPY docs/script/sql/ruoyi-ai-v3_mysql8.sql /docker-entrypoint-initdb.d/ruoyi-ai-v3_mysql8.sql
-
-# 设置脚本可执行权限
-RUN chmod +x /docker-entrypoint-initdb.d/init-db.sh
+# MySQL 官方入口会按文件名顺序执行初始化目录中的 SQL。
+# ruoyi-ai.sql 会导入 MYSQL_DATABASE，snail_job_mysql.sql 会自行创建并切换数据库。
+COPY docs/script/sql/ruoyi-ai.sql /docker-entrypoint-initdb.d/01-ruoyi-ai.sql
+COPY docs/script/sql/snail_job_mysql.sql /docker-entrypoint-initdb.d/02-snail-job.sql

 # MySQL启动参数
 CMD ["--default-authentication-plugin=mysql_native_password", \
@@ -18,4 +16,4 @@ CMD ["--default-authentication-plugin=mysql_native_password", \
     "--collation-server=utf8mb4_general_ci", \
     "--explicit_defaults_for_timestamp=true", \
     "--lower_case_table_names=1", \
-     "--skip-ssl"]
+     "--skip-ssl"]
--- a/docs/docker/ruoyi-ai/docker-compose-all.yaml
+++ b/docs/docker/ruoyi-ai/docker-compose-all.yaml
@@ -12,8 +12,6 @@
 #
 # 镜像仓库地址: crpi-31mraxd99y2gqdgr.cn-beijing.personal.cr.aliyuncs.com/ruoyi_ai

-version: '3.8'
-
 services:
  # ==================== MySQL 数据库 ====================
  mysql:
@@ -30,7 +28,10 @@ services:
    volumes:
      - mysql-data:/var/lib/mysql
    healthcheck:
-      test: ["CMD", "mysqladmin", "ping", "-h", "localhost", "-u", "root", "-proot"]
+      # Force TCP: during first-time initialization MySQL exposes a temporary
+      # socket-only server. A socket ping would mark the service healthy before
+      # the real network listener (used by the backend) is ready.
+      test: ["CMD", "mysqladmin", "ping", "--protocol=tcp", "-h", "127.0.0.1", "-u", "root", "-proot"]
      interval: 15s
      timeout: 10s
      retries: 10
@@ -65,7 +66,7 @@ services:
      - "28080:8080"
    environment:
      QUERY_DEFAULTS_LIMIT: 25
-      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: true
+      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: "true"
      PERSISTENCE_DATA_PATH: /var/lib/weaviate
      DEFAULT_VECTORIZER_MODULE: none
      ENABLE_MODULES: text2vec-cohere,text2vec-huggingface,text2vec-palm,text2vec-openai,generative-openai,generative-cohere,generative-palm,ref2vec-centroid,reranker-cohere,qna-openai
@@ -177,4 +178,4 @@ volumes:
  weaviate-data:
  minio-data:
  logs-data:
-  upload-data:
+  upload-data:
--- a/docs/docker/ruoyi-ai/docker-compose.yaml
+++ b/docs/docker/ruoyi-ai/docker-compose.yaml
@@ -21,8 +21,8 @@ services:
      MYSQL_DATABASE: ruoyi-ai-agent
      TZ: Asia/Shanghai
    volumes:
-      - ./docs/script/docker/mysql/init/init-db.sh:/docker-entrypoint-initdb.d/init-db.sh:ro
-      - ./docs/script/sql/ruoyi-ai-v3_mysql8.sql:/docker-entrypoint-initdb.d/ruoyi-ai-v3_mysql8.sql:ro
+      - ../../script/sql/ruoyi-ai.sql:/docker-entrypoint-initdb.d/01-ruoyi-ai.sql:ro
+      - ../../script/sql/snail_job_mysql.sql:/docker-entrypoint-initdb.d/02-snail-job.sql:ro
      - mysql-data:/var/lib/mysql
    command:
      --default-authentication-plugin=mysql_native_password
@@ -32,7 +32,9 @@ services:
      --lower_case_table_names=1
      --skip-ssl
    healthcheck:
-      test: ["CMD", "mysqladmin", "ping", "-h", "localhost", "-u", "root", "-proot"]
+      # Do not let the socket-only initialization server satisfy the check.
+      # The backend connects over TCP and must wait for that listener.
+      test: ["CMD", "mysqladmin", "ping", "--protocol=tcp", "-h", "127.0.0.1", "-u", "root", "-proot"]
      interval: 15s
      timeout: 10s
      retries: 10
--- a/docs/docker/weaviate/docker-compose.yml
+++ b/docs/docker/weaviate/docker-compose.yml
@@ -5,12 +5,12 @@ services:
      - --host
      - 0.0.0.0
      - --port
-      - '6038'
+      - '8080'
      - --scheme
      - http
-    image: semitechnologies/weaviate:1.19.7
+    image: semitechnologies/weaviate:1.30.0
    ports:
-      - 6038:6038
+      - 28080:8080
      - 50051:50051
    volumes:
      - weaviate_data:/var/lib/weaviate
--- a/docs/image/01.png
+++ b/docs/image/01.png
--- a/docs/image/02.png
+++ b/docs/image/02.png
--- a/docs/image/03.png
+++ b/docs/image/03.png
--- a/docs/image/sponsor/atlascloud_banner.png
+++ b/docs/image/sponsor/atlascloud_banner.png
--- a/docs/image/sponsor/huoshan.png
+++ b/docs/image/sponsor/huoshan.png
--- a/docs/image/wx06.png
+++ b/docs/image/wx06.png
--- a/docs/script/docker/mysql/init/init-db.sh
+++ b/docs/script/docker/mysql/init/init-db.sh
@@ -1,10 +0,0 @@
-#!/bin/bash
-# 数据库初始化脚本
-# 使用 --force 参数确保即使出错也继续执行
-
-echo "开始初始化数据库..."
-
-# 使用 --force 参数忽略错误继续执行
-mysql -uroot -proot ruoyi-ai-agent --force < /docker-entrypoint-initdb.d/ruoyi-ai-v3_mysql8.sql
-
-echo "数据库初始化完成"
--- a/docs/script/install-ffmpeg-windows.ps1
+++ b/docs/script/install-ffmpeg-windows.ps1
@@ -0,0 +1,86 @@
+[CmdletBinding()]
+param(
+    [switch]$Force
+)
+
+$ErrorActionPreference = 'Stop'
+
+function Find-Executable([string]$Name) {
+    $command = Get-Command $Name -ErrorAction SilentlyContinue
+    if ($command) {
+        return $command.Source
+    }
+
+    $roots = @(
+        (Join-Path $env:LOCALAPPDATA 'Microsoft\WinGet\Packages'),
+        (Join-Path $env:USERPROFILE 'scoop\apps\ffmpeg'),
+        'C:\ProgramData\chocolatey\bin',
+        'C:\ffmpeg'
+    )
+    foreach ($root in $roots) {
+        if (-not (Test-Path $root)) { continue }
+        $match = Get-ChildItem -Path $root -Filter "$Name.exe" -Recurse -ErrorAction SilentlyContinue |
+            Select-Object -First 1 -ExpandProperty FullName
+        if ($match) { return $match }
+    }
+    return $null
+}
+
+$ffmpeg = Find-Executable 'ffmpeg'
+$ffprobe = Find-Executable 'ffprobe'
+
+if ($Force -or -not $ffmpeg -or -not $ffprobe) {
+    $winget = Get-Command winget -ErrorAction SilentlyContinue
+    if (-not $winget) {
+        throw '未找到 winget。请先从 Microsoft Store 安装“应用安装程序”，或使用项目 Dockerfile 运行后端。'
+    }
+
+    Write-Host '正在通过 winget 安装 FFmpeg...' -ForegroundColor Cyan
+    & winget install --id Gyan.FFmpeg --exact --source winget `
+        --accept-package-agreements --accept-source-agreements
+    if ($LASTEXITCODE -ne 0) {
+        throw "winget 安装失败，退出码: $LASTEXITCODE"
+    }
+
+    $ffmpeg = Find-Executable 'ffmpeg'
+    $ffprobe = Find-Executable 'ffprobe'
+}
+
+if (-not $ffmpeg -or -not $ffprobe) {
+    throw 'FFmpeg 已安装但未能定位 ffmpeg.exe 或 ffprobe.exe。请重新打开 PowerShell 后再次运行脚本。'
+}
+
+$binDirectory = Split-Path -Parent $ffmpeg
+[Environment]::SetEnvironmentVariable('FFMPEG_PATH', $ffmpeg, 'User')
+[Environment]::SetEnvironmentVariable('FFPROBE_PATH', $ffprobe, 'User')
+
+$userPath = [Environment]::GetEnvironmentVariable('Path', 'User')
+$pathParts = @($userPath -split ';' | Where-Object { $_ })
+if ($pathParts -notcontains $binDirectory) {
+    $newPath = (@($pathParts) + $binDirectory) -join ';'
+    [Environment]::SetEnvironmentVariable('Path', $newPath, 'User')
+}
+
+$env:FFMPEG_PATH = $ffmpeg
+$env:FFPROBE_PATH = $ffprobe
+if (($env:Path -split ';') -notcontains $binDirectory) {
+    $env:Path = "$binDirectory;$env:Path"
+}
+
+Write-Host "ffmpeg:  $ffmpeg" -ForegroundColor Green
+Write-Host "ffprobe: $ffprobe" -ForegroundColor Green
+
+& $ffmpeg -hide_banner -version | Select-Object -First 1
+& $ffprobe -hide_banner -version | Select-Object -First 1
+
+$encoders = & $ffmpeg -hide_banner -encoders 2>&1 | Out-String
+if ($encoders -notmatch '\blibx264\b') {
+    throw '当前 FFmpeg 不包含 libx264 编码器，无法满足短剧合成配置。'
+}
+if ($encoders -notmatch '\baac\b') {
+    throw '当前 FFmpeg 不包含 AAC 编码器，无法满足短剧合成配置。'
+}
+
+Write-Host ''
+Write-Host '安装与配置完成。请完全重启 IntelliJ IDEA 和 ruoyi-ai 后端服务。' -ForegroundColor Yellow
+Write-Host '重启后，Spring 将从 FFMPEG_PATH 和 FFPROBE_PATH 读取绝对路径。'
--- a/docs/script/sql/ruoyi-ai-v3_mysql8.sql
+++ b/docs/script/sql/ruoyi-ai-v3_mysql8.sql
--- a/docs/script/sql/snail_job_mysql.sql
+++ b/docs/script/sql/snail_job_mysql.sql
--- a/docs/script/sql/update/2026-07-20-knowledge-fragment-fid.sql
+++ b/docs/script/sql/update/2026-07-20-knowledge-fragment-fid.sql
@@ -0,0 +1,58 @@
+-- RAG metadata migration (MySQL 8). Safe to execute repeatedly.
+-- 注意：MySQL 8 不支持 ALTER TABLE ... ADD COLUMN IF NOT EXISTS（仅 MariaDB 支持），
+-- 因此列的增量添加统一用 information_schema 守卫 + PREPARE 实现幂等。
+SET @add_file_hash_col = IF(EXISTS(
+    SELECT 1 FROM information_schema.columns WHERE table_schema = DATABASE()
+    AND table_name = 'knowledge_attach' AND column_name = 'file_hash'),
+    'SELECT 1',
+    'ALTER TABLE `knowledge_attach` ADD COLUMN `file_hash` varchar(64) NULL DEFAULT NULL COMMENT ''文件SHA-256摘要'' AFTER `doc_id`');
+PREPARE stmt FROM @add_file_hash_col; EXECUTE stmt; DEALLOCATE PREPARE stmt;
+
+ALTER TABLE `knowledge_attach`
+    MODIFY COLUMN `doc_id` varchar(32) NULL DEFAULT NULL COMMENT '文档ID';
+
+SET @add_file_hash = IF(EXISTS(
+    SELECT 1 FROM information_schema.statistics WHERE table_schema = DATABASE()
+    AND table_name = 'knowledge_attach' AND index_name = 'uk_knowledge_file_hash'),
+    'SELECT 1', 'ALTER TABLE `knowledge_attach` ADD UNIQUE INDEX `uk_knowledge_file_hash` (`knowledge_id`, `file_hash`)');
+PREPARE stmt FROM @add_file_hash; EXECUTE stmt; DEALLOCATE PREPARE stmt;
+
+SET @add_fid_col = IF(EXISTS(
+    SELECT 1 FROM information_schema.columns WHERE table_schema = DATABASE()
+    AND table_name = 'knowledge_fragment' AND column_name = 'fid'),
+    'SELECT 1',
+    'ALTER TABLE `knowledge_fragment` ADD COLUMN `fid` varchar(32) NULL DEFAULT NULL COMMENT ''向量库片段ID'' AFTER `id`');
+PREPARE stmt FROM @add_fid_col; EXECUTE stmt; DEALLOCATE PREPARE stmt;
+
+ALTER TABLE `knowledge_fragment`
+    MODIFY COLUMN `doc_id` varchar(32) NULL DEFAULT NULL COMMENT '文档ID';
+
+UPDATE `knowledge_fragment`
+SET `fid` = LOWER(MD5(CONCAT('knowledge_fragment:', `id`)))
+WHERE `fid` IS NULL OR `fid` = '';
+
+SET @drop_idx_fid = IF(EXISTS(
+    SELECT 1 FROM information_schema.statistics WHERE table_schema = DATABASE()
+    AND table_name = 'knowledge_fragment' AND index_name = 'idx_fid'),
+    'ALTER TABLE `knowledge_fragment` DROP INDEX `idx_fid`', 'SELECT 1');
+PREPARE stmt FROM @drop_idx_fid; EXECUTE stmt; DEALLOCATE PREPARE stmt;
+
+SET @add_uk_fid = IF(EXISTS(
+    SELECT 1 FROM information_schema.statistics WHERE table_schema = DATABASE()
+    AND table_name = 'knowledge_fragment' AND index_name = 'uk_fid'),
+    'SELECT 1', 'ALTER TABLE `knowledge_fragment` ADD UNIQUE INDEX `uk_fid` (`fid`)');
+PREPARE stmt FROM @add_uk_fid; EXECUTE stmt; DEALLOCATE PREPARE stmt;
+
+ALTER TABLE `knowledge_fragment` MODIFY COLUMN `fid` varchar(32) NOT NULL COMMENT '向量库片段ID';
+
+SET @add_tenant_user = IF(EXISTS(
+    SELECT 1 FROM information_schema.statistics WHERE table_schema = DATABASE()
+    AND table_name = 'knowledge_info' AND index_name = 'idx_tenant_user'),
+    'SELECT 1', 'ALTER TABLE `knowledge_info` ADD INDEX `idx_tenant_user` (`tenant_id`, `user_id`)');
+PREPARE stmt FROM @add_tenant_user; EXECUTE stmt; DEALLOCATE PREPARE stmt;
+
+SET @add_tenant_share = IF(EXISTS(
+    SELECT 1 FROM information_schema.statistics WHERE table_schema = DATABASE()
+    AND table_name = 'knowledge_info' AND index_name = 'idx_tenant_share'),
+    'SELECT 1', 'ALTER TABLE `knowledge_info` ADD INDEX `idx_tenant_share` (`tenant_id`, `share`)');
+PREPARE stmt FROM @add_tenant_share; EXECUTE stmt; DEALLOCATE PREPARE stmt;
--- a/docs/script/sql/update/2026-07-21-chat-provider-icon-length.sql
+++ b/docs/script/sql/update/2026-07-21-chat-provider-icon-length.sql
@@ -0,0 +1,7 @@
+-- 加宽 chat_provider.provider_icon 字段 (对应 issue IHPUDA)
+-- 背景：文件系统使用 minio 私有桶时，厂商图标存的是带签名的临时访问 URL，
+--       长度常超过 255，导致「Data too long for column 'provider_icon'」。
+-- MODIFY COLUMN 可重复执行。
+
+ALTER TABLE `chat_provider`
+    MODIFY COLUMN `provider_icon` varchar(1000) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NULL DEFAULT NULL COMMENT '厂商图标';
--- a/docs/script/sql/update/2026-07-21-sys-config-node-template.sql
+++ b/docs/script/sql/update/2026-07-21-sys-config-node-template.sql
@@ -0,0 +1,41 @@
+-- 补充工作流节点消息模板配置 (对应 issue IJX5VV)
+-- 背景：NodeMessageTemplateEnum 依赖以下 9 个 sys_config 键，缺失时
+--       WorkflowMessageUtil.getNodeMessageTemplate 会抛出「请先配置该节点的响应模板」。
+-- 这批配置在历史提交 20d531c0 中存在，SQL 脚本合并重命名时遗失，此处恢复。
+-- 幂等：按 config_key + tenant_id 判重，可重复执行。
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027192921483309058, '000000', 'HTTP请求节点响应模板', 'node.httpRequest.template', '✅ HTTP请求节点：结束响应 - ', 'Y', 103, 1, '2026-02-27 09:23:51', 1, '2026-02-27 09:31:41', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.httpRequest.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027193296990957569, '000000', '文生图节点响应模板', 'node.image.template', '🎨 文生图节点：结束响应 - 图片URL: ', 'Y', 103, 1, '2026-02-27 09:25:20', 1, '2026-02-27 09:31:52', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.image.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027193820393959425, '000000', '发送邮箱节点响应模板', 'node.mailsend.template', '📧 发送邮箱节点：结束响应 - ', 'Y', 103, 1, '2026-02-27 09:27:25', 1, '2026-02-27 09:32:05', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.mailsend.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027194134438277122, '000000', '结束节点响应模板', 'node.end.template', '🔚 流程已执行完毕，如果您有其他需求，请随时重新发起请求。', 'Y', 103, 1, '2026-02-27 09:28:40', 1, '2026-02-27 09:32:53', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.end.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027206492573335554, '000000', '人机交互节点响应模板', 'node.humanFeedback.template', '👤 人机交互节点：等待用户操作 - ', 'Y', 103, 1, '2026-02-27 10:17:46', 1, '2026-02-27 10:17:46', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.humanFeedback.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027208880369647617, '000000', '条件分支节点响应模板', 'node.switch.template', '🔀 条件分支节点：触发 -> 跳转到节点 ', 'Y', 103, 1, '2026-02-27 10:27:15', 1, '2026-02-27 10:35:54', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.switch.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027213914603995137, '000000', '大模型回答节点响应模板', 'node.llmAnswer.template', '🤖 LLM 节点 生成回答：', 'Y', 103, 1, '2026-02-27 10:47:16', 1, '2026-02-27 10:52:40', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.llmAnswer.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027214387000066050, '000000', '关键词提取响应模板', 'node.keywordExtractor.template', '🔑 关键词提取节点 处理完成 ： ', 'Y', 103, 1, '2026-02-27 10:49:08', 1, '2026-02-27 10:52:08', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.keywordExtractor.template' AND `tenant_id` = '000000');
+
+INSERT INTO `sys_config` (`config_id`, `tenant_id`, `config_name`, `config_key`, `config_value`, `config_type`, `create_dept`, `create_by`, `create_time`, `update_by`, `update_time`, `remark`)
+SELECT 2027217577397391361, '000000', '工作流异常响应模板', 'node.exception.template', '🛑 工作流发生异常：', 'N', 103, 1, '2026-02-27 11:01:49', 1, '2026-02-27 11:02:01', NULL
+FROM DUAL WHERE NOT EXISTS (SELECT 1 FROM `sys_config` WHERE `config_key` = 'node.exception.template' AND `tenant_id` = '000000');
--- a/docs/script/sql/update/update-0615-trace.sql
+++ b/docs/script/sql/update/update-0615-trace.sql
@@ -0,0 +1,128 @@
+-- 链路追踪运行记录表
+CREATE TABLE IF NOT EXISTS `trace_run` (
+  `id` bigint NOT NULL COMMENT '主键',
+  `trace_id` varchar(64) NOT NULL COMMENT '链路ID',
+  `trace_name` varchar(128) NOT NULL COMMENT '链路名称',
+  `business_type` varchar(64) NOT NULL COMMENT '业务类型',
+  `business_id` varchar(128) DEFAULT NULL COMMENT '业务ID',
+  `user_id` bigint DEFAULT NULL COMMENT '用户ID',
+  `tenant_id` varchar(20) DEFAULT '000000' COMMENT '租户编号',
+  `status` varchar(32) NOT NULL COMMENT '状态',
+  `start_time` datetime NOT NULL COMMENT '开始时间',
+  `end_time` datetime DEFAULT NULL COMMENT '结束时间',
+  `duration_ms` bigint DEFAULT NULL COMMENT '耗时毫秒',
+  `error_message` varchar(1000) DEFAULT NULL COMMENT '错误摘要',
+  `metadata` text DEFAULT NULL COMMENT '元数据JSON',
+  `create_dept` bigint DEFAULT NULL COMMENT '创建部门',
+  `create_by` bigint DEFAULT NULL COMMENT '创建者',
+  `create_time` datetime DEFAULT NULL COMMENT '创建时间',
+  `update_by` bigint DEFAULT NULL COMMENT '更新者',
+  `update_time` datetime DEFAULT NULL COMMENT '更新时间',
+  `del_flag` char(1) DEFAULT '0' COMMENT '删除标志（0代表存在 2代表删除）',
+  PRIMARY KEY (`id`) USING BTREE,
+  UNIQUE KEY `uk_trace_run_trace_id` (`trace_id`) USING BTREE,
+  KEY `idx_trace_run_business` (`business_type`, `business_id`) USING BTREE,
+  KEY `idx_trace_run_status_time` (`status`, `start_time`) USING BTREE,
+  KEY `idx_trace_run_tenant_time` (`tenant_id`, `start_time`) USING BTREE
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci COMMENT='链路追踪运行记录表' ROW_FORMAT=DYNAMIC;
+
+-- 链路追踪节点记录表
+CREATE TABLE IF NOT EXISTS `trace_node` (
+  `id` bigint NOT NULL COMMENT '主键',
+  `trace_id` varchar(64) NOT NULL COMMENT '链路ID',
+  `node_id` varchar(64) NOT NULL COMMENT '节点ID',
+  `tenant_id` varchar(20) DEFAULT '000000' COMMENT '租户编号',
+  `parent_node_id` varchar(64) DEFAULT NULL COMMENT '父节点ID',
+  `node_name` varchar(128) NOT NULL COMMENT '节点名称',
+  `node_type` varchar(64) NOT NULL COMMENT '节点类型',
+  `depth` int DEFAULT 0 COMMENT '节点深度',
+  `sort_order` int DEFAULT 0 COMMENT '排序',
+  `class_name` varchar(255) DEFAULT NULL COMMENT '类名',
+  `method_name` varchar(128) DEFAULT NULL COMMENT '方法名',
+  `status` varchar(32) NOT NULL COMMENT '状态',
+  `start_time` datetime NOT NULL COMMENT '开始时间',
+  `end_time` datetime DEFAULT NULL COMMENT '结束时间',
+  `duration_ms` bigint DEFAULT NULL COMMENT '耗时毫秒',
+  `error_message` varchar(1000) DEFAULT NULL COMMENT '错误摘要',
+  `input_payload` text DEFAULT NULL COMMENT '输入JSON',
+  `output_payload` text DEFAULT NULL COMMENT '输出JSON',
+  `metadata` text DEFAULT NULL COMMENT '元数据JSON',
+  `create_dept` bigint DEFAULT NULL COMMENT '创建部门',
+  `create_by` bigint DEFAULT NULL COMMENT '创建者',
+  `create_time` datetime DEFAULT NULL COMMENT '创建时间',
+  `update_by` bigint DEFAULT NULL COMMENT '更新者',
+  `update_time` datetime DEFAULT NULL COMMENT '更新时间',
+  `del_flag` char(1) DEFAULT '0' COMMENT '删除标志（0代表存在 2代表删除）',
+  PRIMARY KEY (`id`) USING BTREE,
+  KEY `idx_trace_node_trace_id` (`trace_id`) USING BTREE,
+  KEY `idx_trace_node_parent` (`trace_id`, `parent_node_id`) USING BTREE,
+  KEY `idx_trace_node_time` (`trace_id`, `start_time`) USING BTREE,
+  KEY `idx_trace_node_tenant_time` (`tenant_id`, `start_time`) USING BTREE,
+  KEY `idx_trace_node_type_status` (`node_type`, `status`) USING BTREE
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci COMMENT='链路追踪节点记录表' ROW_FORMAT=DYNAMIC;
+
+-- 兼容已存在的旧 trace 表：CREATE TABLE IF NOT EXISTS 不会给旧表补新字段
+SET @trace_run_add_tenant_sql = (
+  SELECT IF(COUNT(*) = 0,
+    'ALTER TABLE `trace_run` ADD COLUMN `tenant_id` varchar(20) DEFAULT ''000000'' COMMENT ''租户编号'' AFTER `user_id`',
+    'SELECT 1'
+  )
+  FROM information_schema.COLUMNS
+  WHERE TABLE_SCHEMA = DATABASE()
+    AND TABLE_NAME = 'trace_run'
+    AND COLUMN_NAME = 'tenant_id'
+);
+PREPARE trace_run_add_tenant_stmt FROM @trace_run_add_tenant_sql;
+EXECUTE trace_run_add_tenant_stmt;
+DEALLOCATE PREPARE trace_run_add_tenant_stmt;
+
+SET @trace_run_add_tenant_idx_sql = (
+  SELECT IF(COUNT(*) = 0,
+    'ALTER TABLE `trace_run` ADD INDEX `idx_trace_run_tenant_time` (`tenant_id`, `start_time`) USING BTREE',
+    'SELECT 1'
+  )
+  FROM information_schema.STATISTICS
+  WHERE TABLE_SCHEMA = DATABASE()
+    AND TABLE_NAME = 'trace_run'
+    AND INDEX_NAME = 'idx_trace_run_tenant_time'
+);
+PREPARE trace_run_add_tenant_idx_stmt FROM @trace_run_add_tenant_idx_sql;
+EXECUTE trace_run_add_tenant_idx_stmt;
+DEALLOCATE PREPARE trace_run_add_tenant_idx_stmt;
+
+SET @trace_node_add_tenant_sql = (
+  SELECT IF(COUNT(*) = 0,
+    'ALTER TABLE `trace_node` ADD COLUMN `tenant_id` varchar(20) DEFAULT ''000000'' COMMENT ''租户编号'' AFTER `node_id`',
+    'SELECT 1'
+  )
+  FROM information_schema.COLUMNS
+  WHERE TABLE_SCHEMA = DATABASE()
+    AND TABLE_NAME = 'trace_node'
+    AND COLUMN_NAME = 'tenant_id'
+);
+PREPARE trace_node_add_tenant_stmt FROM @trace_node_add_tenant_sql;
+EXECUTE trace_node_add_tenant_stmt;
+DEALLOCATE PREPARE trace_node_add_tenant_stmt;
+
+SET @trace_node_add_tenant_idx_sql = (
+  SELECT IF(COUNT(*) = 0,
+    'ALTER TABLE `trace_node` ADD INDEX `idx_trace_node_tenant_time` (`tenant_id`, `start_time`) USING BTREE',
+    'SELECT 1'
+  )
+  FROM information_schema.STATISTICS
+  WHERE TABLE_SCHEMA = DATABASE()
+    AND TABLE_NAME = 'trace_node'
+    AND INDEX_NAME = 'idx_trace_node_tenant_time'
+);
+PREPARE trace_node_add_tenant_idx_stmt FROM @trace_node_add_tenant_idx_sql;
+EXECUTE trace_node_add_tenant_idx_stmt;
+DEALLOCATE PREPARE trace_node_add_tenant_idx_stmt;
+
+-- 链路追踪监控菜单 & 按钮权限
+INSERT INTO `sys_menu`
+(`menu_id`, `menu_name`, `parent_id`, `order_num`, `path`, `component`, `query_param`, `is_frame`, `is_cache`, `menu_type`, `visible`, `status`, `perms`, `icon`, `create_dept`, `create_by`, `create_time`, `remark`)
+SELECT (SELECT COALESCE(MAX(`menu_id`), 0) + 1 FROM (SELECT `menu_id` FROM `sys_menu`) t), '链路追踪', 2, 7, 'trace', 'monitor/trace/index', '', 1, 0, 'C', '0', '0', 'monitor:trace:list', 'tabler:route', 103, 1, NOW(), '链路追踪监控菜单';
+
+INSERT INTO `sys_menu`
+(`menu_id`, `menu_name`, `parent_id`, `order_num`, `path`, `component`, `query_param`, `is_frame`, `is_cache`, `menu_type`, `visible`, `status`, `perms`, `icon`, `create_dept`, `create_by`, `create_time`, `remark`)
+SELECT (SELECT COALESCE(MAX(`menu_id`), 0) + 1 FROM (SELECT `menu_id` FROM `sys_menu`) t), '链路追踪查询', (SELECT `menu_id` FROM `sys_menu` WHERE `perms` = 'monitor:trace:list' AND `menu_type` = 'C' LIMIT 1), 1, '#', '', '', 1, 0, 'F', '0', '0', 'monitor:trace:query', '#', 103, 1, NOW(), '';
--- a/pom.xml
+++ b/pom.xml
@@ -13,7 +13,7 @@
    <description>全栈式AI开发平台</description>

    <properties>
-        <revision>3.0.0</revision>
+        <revision>3.1.0</revision>
        <spring-boot.version>3.5.8</spring-boot.version>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding>
@@ -54,18 +54,23 @@
        <!-- Jackson XML -->
        <jackson-dataformat-xml.version>2.18.2</jackson-dataformat-xml.version>
        <!-- AI 相关依赖 -->
-        <langchain4j.version>1.13.0</langchain4j.version>
-        <langchain4j.community.version>1.13.0-beta23</langchain4j.community.version>
-        <langgraph4j.version>1.5.3</langgraph4j.version>
+
+        <langchain4j.version>1.17.2</langchain4j.version>
+        <langchain4j.beta.version>1.17.2-beta27</langchain4j.beta.version>
+        <langchain4j.community.version>1.17.0-beta27</langchain4j.community.version>
+        <langgraph4j.version>1.8.20</langgraph4j.version>
+
        <weaviate.version>1.19.6</weaviate.version>
-        <dify.version>1.0.7</dify.version>
+        <dify.version>1.2.7</dify.version>
+        <coze.version>0.4.2</coze.version>
+
        <!-- gRPC 版本 - 解决 Milvus SDK 依赖冲突 -->
        <grpc.version>1.62.2</grpc.version>
        <!-- Apache Commons Compress - 用于POI处理ZIP格式 -->
        <commons-compress.version>1.27.1</commons-compress.version>

-
        <avatar-generator.version>1.1.0</avatar-generator.version>
+        <javassist.version>3.30.2-GA</javassist.version>
        <jsoup.version>1.21.2</jsoup.version>
        <knife4j.version>4.4.0</knife4j.version>
        <swagger-annotations.version>2.2.8</swagger-annotations.version>
@@ -79,7 +84,7 @@
        <maven-surefire-plugin.version>3.5.3</maven-surefire-plugin.version>
        <flatten-maven-plugin.version>1.3.0</flatten-maven-plugin.version>
        <!-- 打包默认跳过测试 -->
-        <skipTests>true</skipTests>
+        <skipTests>false</skipTests>
    </properties>

    <profiles>
@@ -499,20 +504,8 @@
        <resources>
            <resource>
                <directory>src/main/resources</directory>
-                <!-- 关闭过滤 -->
                <filtering>false</filtering>
            </resource>
-            <resource>
-                <directory>src/main/resources</directory>
-                <!-- 引入所有 匹配文件进行过滤 -->
-                <includes>
-                    <include>application*</include>
-                    <include>bootstrap*</include>
-                    <include>banner*</include>
-                </includes>
-                <!-- 启用过滤 即该资源中的变量将会被过滤器中的值替换 -->
-                <filtering>true</filtering>
-            </resource>
        </resources>
    </build>

--- a/ruoyi-admin/Dockerfile
+++ b/ruoyi-admin/Dockerfile
@@ -1,31 +1,59 @@
-# 贝尔实验室 Spring 官方推荐镜像 JDK下载地址 https://bell-sw.com/pages/downloads/
+# FFmpeg-enabled variant of ruoyi-admin/Dockerfile.
 FROM bellsoft/liberica-openjdk-rocky:17.0.16-cds
-#FROM bellsoft/liberica-openjdk-rocky:21.0.8-cds
-#FROM findepi/graalvm:java17-native

 LABEL maintainer="Lion Li"

+# Rocky base repositories do not provide the full codec build. RPM Fusion Free
+# supplies ffmpeg with libx264, while the checks below prevent a reduced build
+# from reaching production unnoticed.
+RUN set -eux; \
+    dnf -y install dnf-plugins-core epel-release; \
+    . /etc/os-release; \
+    rocky_major="${VERSION_ID%%.*}"; \
+    if [ "${rocky_major}" -ge 9 ]; then \
+        dnf config-manager --set-enabled crb; \
+    else \
+        dnf config-manager --set-enabled powertools; \
+    fi; \
+    dnf -y install "https://mirrors.rpmfusion.org/free/el/rpmfusion-free-release-${rocky_major}.noarch.rpm"; \
+    dnf -y install ffmpeg; \
+    for filter in scale pad fps setsar format tpad trim settb setpts xfade \
+        aresample aformat apad atrim asetpts anullsrc concat acrossfade; do \
+        ffmpeg -hide_banner -filters 2>/dev/null | grep -Eq "[[:space:]]${filter}[[:space:]]"; \
+    done; \
+    ffmpeg -hide_banner -h filter=xfade 2>&1 | grep -q dissolve; \
+    ffmpeg -hide_banner -h filter=xfade 2>&1 | grep -q fadeblack; \
+    ffmpeg -hide_banner -h filter=xfade 2>&1 | grep -q slideleft; \
+    ffmpeg -hide_banner -encoders 2>/dev/null | grep -Eq '[[:space:]]libx264[[:space:]]'; \
+    ffmpeg -hide_banner -encoders 2>/dev/null | grep -Eq '[[:space:]]aac[[:space:]]'; \
+    ffprobe -hide_banner -version >/dev/null; \
+    dnf clean all; \
+    rm -rf /var/cache/dnf
+
 RUN mkdir -p /ruoyi/server/logs \
    /ruoyi/server/temp \
    /ruoyi/skywalking/agent

 WORKDIR /ruoyi/server

-ENV SERVER_PORT=8080 SNAIL_PORT=28080 LANG=C.UTF-8 LC_ALL=C.UTF-8 JAVA_OPTS=""
+ENV SERVER_PORT=8080 \
+    SNAIL_PORT=28080 \
+    LANG=C.UTF-8 \
+    LC_ALL=C.UTF-8 \
+    JAVA_OPTS="" \
+    FFMPEG_PATH=/usr/bin/ffmpeg \
+    FFPROBE_PATH=/usr/bin/ffprobe

 EXPOSE ${SERVER_PORT}
-# 暴露 snail job 客户端端口 用于定时任务调度中心通信
 EXPOSE ${SNAIL_PORT}

 ADD ./target/ruoyi-admin.jar ./app.jar

 SHELL ["/bin/bash", "-c"]

-ENTRYPOINT java -Djava.security.egd=file:/dev/./urandom -Dserver.port=${SERVER_PORT} \
+ENTRYPOINT java -Djava.security.egd=file:/dev/./urandom \
+           -Djava.io.tmpdir=/ruoyi/server/temp \
+           -Dserver.port=${SERVER_PORT} \
           -Dsnail-job.port=${SNAIL_PORT} \
-           # 应用名称 如果想区分集群节点监控 改成不同的名称即可
-           #-Dskywalking.agent.service_name=ruoyi-server \
-           #-javaagent:/ruoyi/skywalking/agent/skywalking-agent.jar \
           -XX:+HeapDumpOnOutOfMemoryError -XX:+UseZGC ${JAVA_OPTS} \
           -jar app.jar
-
--- a/ruoyi-admin/src/main/resources/application-dev.yml
+++ b/ruoyi-admin/src/main/resources/application-dev.yml
@@ -8,8 +8,8 @@ spring.boot.admin.client:
    metadata:
      username: ${spring.boot.admin.client.username}
      userpassword: ${spring.boot.admin.client.password}
-  username: @monitor.username@
-  password: @monitor.password@
+  username: ${MONITOR_USERNAME:ruoyi}
+  password: ${MONITOR_PASSWORD:123456}

 --- # mcp配置信息
 mcp:
@@ -27,12 +27,12 @@ snail-job:
  enabled: false
  # 需要在 SnailJob 后台组管理创建对应名称的组,然后创建任务的时候选择对应的组,才能正确分派任务
  group: "ruoyi_group"
-  # SnailJob 接入验证令牌 详见 script/sql/ry_job.sql `sj_group_config` 表
+  # SnailJob 接入验证令牌 详见 docs/script/sql/ruoyi-ai-v3_mysql8.sql `sj_group_config` 表
  token: "SJ_cKqBTPzCsWA3VyuCfFoccmuIEGXjr5KT"
  server:
    host: 127.0.0.1
    port: 17888
-  # 命名空间UUID 详见 script/sql/ry_job.sql `sj_namespace`表`unique_id`字段
+  # 命名空间UUID 详见 docs/script/sql/ruoyi-ai-v3_mysql8.sql `sj_namespace`表`unique_id`字段
  namespace: ${spring.profiles.active}
  # 随主应用端口漂移
  port: 2${server.port}
@@ -60,7 +60,7 @@ spring:
          # rewriteBatchedStatements=true 批处理优化 大幅提升批量插入更新删除性能(对数据库有性能损耗 使用批量操作应考虑性能问题)
          url: jdbc:mysql://127.0.0.1:3306/ruoyi-ai?useUnicode=true&characterEncoding=utf8&zeroDateTimeBehavior=convertToNull&useSSL=true&serverTimezone=GMT%2B8&autoReconnect=true&rewriteBatchedStatements=true&allowPublicKeyRetrieval=true&nullCatalogMeansCurrent=true
          username: root
-          password: root
+          password: 123456
 #        agent:
 #          url: jdbc:mysql://127.0.0.1:3306/test?useUnicode=true&characterEncoding=utf8&zeroDateTimeBehavior=convertToNull&useSSL=true&serverTimezone=GMT%2B8&autoReconnect=true&rewriteBatchedStatements=true&allowPublicKeyRetrieval=true&nullCatalogMeansCurrent=true
 #          # url: jdbc:mysql://localhost:3306/agent_db
--- a/ruoyi-admin/src/main/resources/application-prod.yml
+++ b/ruoyi-admin/src/main/resources/application-prod.yml
@@ -9,8 +9,8 @@ spring.boot.admin.client:
    metadata:
      username: ${spring.boot.admin.client.username}
      userpassword: ${spring.boot.admin.client.password}
-  username: @monitor.username@
-  password: @monitor.password@
+  username: ${MONITOR_USERNAME:ruoyi}
+  password: ${MONITOR_PASSWORD:123456}

 --- # mcp配置信息
 mcp:
@@ -28,12 +28,12 @@ snail-job:
  enabled: false
  # 需要在 SnailJob 后台组管理创建对应名称的组,然后创建任务的时候选择对应的组,才能正确分派任务
  group: "ruoyi_group"
-  # SnailJob 接入验证令牌 详见 script/sql/ry_job.sql `sj_group_config` 表
+  # SnailJob 接入验证令牌 详见 docs/script/sql/ruoyi-ai-v3_mysql8.sql `sj_group_config` 表
  token: "SJ_cKqBTPzCsWA3VyuCfFoccmuIEGXjr5KT"
  server:
    host: 127.0.0.1
    port: 17888
-  # 命名空间UUID 详见 script/sql/ry_job.sql `sj_namespace`表`unique_id`字段
+  # 命名空间UUID 详见 docs/script/sql/ruoyi-ai-v3_mysql8.sql `sj_namespace`表`unique_id`字段
  namespace: ${spring.profiles.active}
  # 随主应用端口漂移
  port: 2${server.port}
--- a/ruoyi-admin/src/main/resources/application.yml
+++ b/ruoyi-admin/src/main/resources/application.yml
@@ -20,6 +20,10 @@ server:
      # 阻塞任务线程池, 当执行类似servlet请求阻塞操作, undertow会从这个线程池中取得线程,它的值设置取决于系统的负载
      worker: 256

+--- # 小程序对话 WebSocket 兜底默认模型（前端未传 model 且无智能体绑定时使用）
+chat:
+  default-model: deepseek-v4-flash
+
 captcha:
  # 是否启用验证码校验
  enable: false
@@ -35,7 +39,7 @@ captcha:
 # 日志配置
 logging:
  level:
-    org.ruoyi: @logging.level@
+    org.ruoyi: ${LOGGING_LEVEL_ORG_RUOYI:info}
    org.springframework: warn
    org.mybatis.spring.mapper: error
    org.apache.fury: warn
@@ -75,7 +79,7 @@ spring:
    # 国际化资源文件路径
    basename: i18n/messages
  profiles:
-    active: @profiles.active@
+    active: ${SPRING_PROFILES_ACTIVE:dev}
  # 文件上传
  servlet:
    multipart:
@@ -125,6 +129,7 @@ security:
    - /*/api-docs/**
    - /warm-flow-ui/config
    - /workflow/run
+    - /coding/**
 # 多租户配置
 tenant:
  # 是否开启
@@ -141,6 +146,9 @@ tenant:
    - sys_client
    - sys_oss_config
    - flow_spel
+    # 链路追踪监控表：运维需跨租户全局查看，且 trace_node 在异步线程写入、租户上下文不传播，故排除租户过滤
+    - trace_run
+    - trace_node

 # MyBatisPlus配置
 # https://baomidou.com/config/
@@ -227,6 +235,15 @@ xss:
  excludeUrls:
    - /system/notice

+--- # 链路追踪配置
+trace:
+  # 是否启用链路追踪，默认 true
+  # 关闭后所有埋点代码会直接透传业务逻辑，不写库、不创建上下文，零性能开销
+  enabled: true
+  payload:
+    # 错误信息最大字符长度（格式: "异常类名: 异常消息"），超过部分会被截断丢弃
+    max-error-length: 1000
+
 --- # 分布式锁 lock4j 全局配置
 lock4j:
  # 获取分布式锁超时时间，默认为 3000 毫秒
@@ -265,7 +282,7 @@ demo:
  # 是否开启演示模式（开启后所有写操作将被拦截）
  enabled: false
  # 提示消息
-  message: "演示模式，不允许进行写操作"
+  message: "演示模式，不允许操作"
  # 排除的路径（这些路径不受演示模式限制）
  excludes:
    - /login
@@ -276,7 +293,9 @@ demo:
    - /chat/send
    - /system/session/**
    - /system/message/**
-
+    - /system/attach/**
+    - /system/fragment/**
+    - /system/info/**
 --- # warm-flow工作流配置
 warm-flow:
  # 是否开启工作流，默认true
@@ -294,11 +313,12 @@ warm-flow:
 vector-store:
  # 向量存储类型 可选(weaviate/milvus/qdrant)
  # 如需修改向量库类型，请修改此配置值!
-  type: milvus
+  # 注意：需与 docker-compose 实际部署的向量库保持一致（当前 compose 内置 weaviate，映射端口 28080）
+  type: weaviate
  # Weaviate配置
  weaviate:
    protocol: http
-    host: 127.0.0.1:6038
+    host: 127.0.0.1:28080
    classname: LocalKnowledge
  # Milvus配置
  milvus:
@@ -311,3 +331,38 @@ vector-store:
    collectionname: LocalKnowledge
    api-key:
    use-tls: false
+
+# 短剧成片合成
+short-drama:
+  composition:
+    ffmpeg-path: ${FFMPEG_PATH:ffmpeg}
+    ffprobe-path: ${FFPROBE_PATH:ffprobe}
+    fps: ${SHORT_DRAMA_COMPOSITION_FPS:30}
+    audio-sample-rate: ${SHORT_DRAMA_COMPOSITION_AUDIO_SAMPLE_RATE:48000}
+    video-codec: ${SHORT_DRAMA_COMPOSITION_VIDEO_CODEC:libx264}
+    audio-codec: ${SHORT_DRAMA_COMPOSITION_AUDIO_CODEC:aac}
+    preset: ${SHORT_DRAMA_COMPOSITION_PRESET:medium}
+    crf: ${SHORT_DRAMA_COMPOSITION_CRF:20}
+    audio-bitrate: ${SHORT_DRAMA_COMPOSITION_AUDIO_BITRATE:192k}
+    max-clips: ${SHORT_DRAMA_COMPOSITION_MAX_CLIPS:100}
+    max-transition-seconds: ${SHORT_DRAMA_COMPOSITION_MAX_TRANSITION_SECONDS:2.0}
+    minimum-output-bytes: ${SHORT_DRAMA_COMPOSITION_MINIMUM_OUTPUT_BYTES:1024}
+    max-process-output-bytes: ${SHORT_DRAMA_COMPOSITION_MAX_PROCESS_OUTPUT_BYTES:1048576}
+    max-source-bytes: ${SHORT_DRAMA_COMPOSITION_MAX_SOURCE_BYTES:536870912}
+    max-total-source-bytes: ${SHORT_DRAMA_COMPOSITION_MAX_TOTAL_SOURCE_BYTES:2147483648}
+    probe-timeout: ${SHORT_DRAMA_COMPOSITION_PROBE_TIMEOUT:30s}
+    process-timeout: ${SHORT_DRAMA_COMPOSITION_PROCESS_TIMEOUT:30m}
+    job-stale-after: ${SHORT_DRAMA_COMPOSITION_JOB_STALE_AFTER:45m}
+    watermark-font-file: ${SHORT_DRAMA_COMPOSITION_WATERMARK_FONT_FILE:}
+    worker-core-size: ${SHORT_DRAMA_COMPOSITION_WORKER_CORE_SIZE:1}
+    worker-max-size: ${SHORT_DRAMA_COMPOSITION_WORKER_MAX_SIZE:2}
+    worker-queue-capacity: ${SHORT_DRAMA_COMPOSITION_WORKER_QUEUE_CAPACITY:8}
+    storage-mode: ${SHORT_DRAMA_COMPOSITION_STORAGE_MODE:local}
+    local-output-directory: ${SHORT_DRAMA_COMPOSITION_LOCAL_OUTPUT_DIRECTORY:logs/short-drama-compositions}
+    download:
+      connect-timeout: ${SHORT_DRAMA_DOWNLOAD_CONNECT_TIMEOUT:30s}
+      call-timeout: ${SHORT_DRAMA_DOWNLOAD_CALL_TIMEOUT:10m}
+      max-redirects: ${SHORT_DRAMA_DOWNLOAD_MAX_REDIRECTS:5}
+      # 生产环境建议配置为视频供应商或 OSS/CDN 域名，多个值用逗号分隔。
+      allowed-hosts: ${SHORT_DRAMA_DOWNLOAD_ALLOWED_HOSTS:}
+      fake-ip-allowed-hosts: ${SHORT_DRAMA_DOWNLOAD_FAKE_IP_ALLOWED_HOSTS:atlas-media.oss-us-west-1.aliyuncs.com}
--- a/ruoyi-common/pom.xml
+++ b/ruoyi-common/pom.xml
@@ -35,6 +35,7 @@
        <module>ruoyi-common-tenant</module>
        <module>ruoyi-common-websocket</module>
        <module>ruoyi-common-sse</module>
+        <module>ruoyi-common-trace</module>
    </modules>

    <artifactId>ruoyi-common</artifactId>
--- a/ruoyi-common/ruoyi-common-bom/pom.xml
+++ b/ruoyi-common/ruoyi-common-bom/pom.xml
@@ -14,7 +14,7 @@
    </description>

    <properties>
-        <revision>3.0.0</revision>
+        <revision>3.1.0</revision>
    </properties>

    <dependencyManagement>
@@ -186,6 +186,13 @@
                <version>${revision}</version>
            </dependency>

+            <!-- 链路追踪模块 -->
+            <dependency>
+                <groupId>org.ruoyi</groupId>
+                <artifactId>ruoyi-common-trace</artifactId>
+                <version>${revision}</version>
+            </dependency>
+
        </dependencies>
    </dependencyManagement>

--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/bo/chat/ChatModelBo.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/bo/chat/ChatModelBo.java
@@ -1,10 +1,12 @@
 package org.ruoyi.common.chat.domain.bo.chat;

 import io.github.linpeilie.annotations.AutoMapper;
+import jakarta.validation.constraints.NotBlank;
 import jakarta.validation.constraints.NotNull;
 import lombok.Data;
 import lombok.EqualsAndHashCode;
 import org.ruoyi.common.chat.entity.chat.ChatModel;
+import org.ruoyi.common.core.validate.AddGroup;
 import org.ruoyi.common.core.validate.EditGroup;
 import org.ruoyi.common.mybatis.core.domain.BaseEntity;

@@ -28,16 +30,19 @@ public class ChatModelBo extends BaseEntity {
    /**
     * 模型分类
     */
+    @NotBlank(message = "模型分类不能为空", groups = { AddGroup.class, EditGroup.class })
    private String category;

    /**
     * 模型名称
     */
+    @NotBlank(message = "模型名称不能为空", groups = { AddGroup.class, EditGroup.class })
    private String modelName;

    /**
     * 模型供应商
     */
+    @NotBlank(message = "模型供应商不能为空", groups = { AddGroup.class, EditGroup.class })
    private String providerCode;

    /**
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/bo/chat/ModelBatchKeyBo.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/bo/chat/ModelBatchKeyBo.java
@@ -0,0 +1,30 @@
+package org.ruoyi.common.chat.domain.bo.chat;
+
+import jakarta.validation.constraints.NotBlank;
+import lombok.Data;
+import lombok.EqualsAndHashCode;
+import org.ruoyi.common.mybatis.core.domain.BaseEntity;
+
+/**
+ * 按厂商批量更新密钥业务对象
+ *
+ * @author ageerle
+ * @date 2025-12-14
+ */
+@Data
+@EqualsAndHashCode(callSuper = true)
+public class ModelBatchKeyBo extends BaseEntity {
+
+    /**
+     * 模型供应商
+     */
+    @NotBlank(message = "厂商编码不能为空")
+    private String providerCode;
+
+    /**
+     * 密钥
+     */
+    @NotBlank(message = "密钥不能为空")
+    private String apiKey;
+
+}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/dto/request/ChatRequest.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/dto/request/ChatRequest.java
@@ -21,9 +21,19 @@ import java.util.List;
@Data
 public class ChatRequest {

-    @NotEmpty(message = "传入的模型不能为空")
+    /**
+     * 模型名称。
+     * 智能体模式下可缺省：传 agentId 时后端按智能体绑定的模型解析，此字段仅作回退。
+     */
    private String model;

+    /**
+     * 智能体ID。传入时后端按智能体配置解析模型/工具/技能/知识库/提示词/是否深度思考。
+     */
+    @JsonSerialize(using = ToStringSerializer.class)
+    @JSONField(serializeUsing = String.class)
+    private Long agentId;
+
    /**
     * 对话消息
     */
@@ -35,16 +45,6 @@ public class ChatRequest {
     */
    private WorkFlowRunner workFlowRunner;

-    /**
-     * 人机交互信息体
-     */
-    private ReSumeRunner reSumeRunner;
-
-    /**
-     * 是否为人机交互用户继续输入
-     */
-    private Boolean isResume = false;
-
    /**
     * 是否启用工作流
     */
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/dto/request/ReSumeRunner.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/domain/dto/request/ReSumeRunner.java
@@ -1,19 +0,0 @@
-package org.ruoyi.common.chat.domain.dto.request;
-
-import lombok.Data;
-
-/**
- * 人机交互输入信息
- */
-@Data
-public class ReSumeRunner {
-    /**
-     * 运行节点UUID
-     */
-    private String runtimeUuid;
-
-    /**
-     * 人机交互用户输入信息
-     */
-    private String feedbackContent;
-}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/audio/AudioContext.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/audio/AudioContext.java
@@ -0,0 +1,44 @@
+package org.ruoyi.common.chat.entity.audio;
+
+import jakarta.validation.constraints.NotBlank;
+import jakarta.validation.constraints.NotNull;
+import lombok.Builder;
+import lombok.Data;
+import org.ruoyi.common.chat.domain.vo.chat.ChatModelVo;
+
+/**
+ * Text-to-speech context.
+ */
+@Data
+@Builder
+public class AudioContext {
+
+    @NotNull(message = "模型配置不能为空")
+    private ChatModelVo chatModelVo;
+
+    @NotBlank(message = "输入文本不能为空")
+    private String input;
+
+    private String voice;
+
+    private String responseFormat;
+
+    private Double speed;
+
+    private String instructions;
+
+    /** Atlas seed-audio 参考资源：[{speaker, audioUrl, audioData, imageData}]，speaker 为音色名 */
+    private java.util.List<java.util.Map<String, String>> references;
+
+    /** Atlas seed-audio 采样率 */
+    private Integer sampleRate;
+
+    /** Atlas seed-audio 音调调整 (-12~12) */
+    private Integer pitchRate;
+
+    /** Atlas seed-audio 语速调整 (-50~100) */
+    private Integer speechRate;
+
+    /** Atlas seed-audio 响度调整 (-50~100) */
+    private Integer loudnessRate;
+}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/image/ImageContext.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/image/ImageContext.java
@@ -42,4 +42,9 @@ public class ImageContext {
    @Min(value = 0, message = "随机数种子不能小于0")
    @Max(value = 2147483647, message = "随机数种子不能大于2147483647")
    private Integer seed;
+
+    /**
+     * 参考图 URL（图生图用）
+     */
+    private String image;
 }
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/media/MediaGenerationResponse.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/media/MediaGenerationResponse.java
@@ -0,0 +1,31 @@
+package org.ruoyi.common.chat.entity.media;
+
+import lombok.Builder;
+import lombok.Data;
+
+/**
+ * Unified media generation response.
+ */
+@Data
+@Builder
+public class MediaGenerationResponse {
+
+    private String type;
+
+    private String mimeType;
+
+    private String url;
+
+    private String b64Json;
+
+    private String dataUrl;
+
+    private String id;
+
+    private String status;
+
+    private String rawResponse;
+
+    /** 末帧图片 URL（Atlas return_last_frame=true 时返回，用于下一镜首帧承接） */
+    private String lastFrameUrl;
+}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/video/VideoContext.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/entity/video/VideoContext.java
@@ -0,0 +1,47 @@
+package org.ruoyi.common.chat.entity.video;
+
+import jakarta.validation.constraints.NotBlank;
+import jakarta.validation.constraints.NotNull;
+import lombok.Builder;
+import lombok.Data;
+import org.ruoyi.common.chat.domain.vo.chat.ChatModelVo;
+
+/**
+ * Text-to-video context.
+ */
+@Data
+@Builder
+public class VideoContext {
+
+    @NotNull(message = "模型配置不能为空")
+    private ChatModelVo chatModelVo;
+
+    @NotBlank(message = "提示词不能为空")
+    private String prompt;
+
+    private String size;
+
+    private Integer seconds;
+
+    private String quality;
+
+    /** 参考图 URL（图生视频模式，单图） */
+    private String imageUrl;
+
+    /** 多参考图 URL（多参考图生视频模式，配合 @imageN prompt 使用） */
+    private java.util.List<String> referenceImages;
+
+    /** 参考音频 URL 列表（对白口型对齐模式） */
+    private java.util.List<String> referenceAudios;
+
+    /** 是否让模型生成同步音频（环境音/动效） */
+    private Boolean generateAudio;
+
+    /** 是否要求返回末帧，用于下一镜首帧承接 */
+    private Boolean returnLastFrame;
+
+    /** 上一镜末帧 URL（同场景连续镜头首帧承接用，作为额外参考图传入） */
+    private String lastFrameUrl;
+
+    private String videoId;
+}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/factory/AudioServiceFactory.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/factory/AudioServiceFactory.java
@@ -0,0 +1,40 @@
+package org.ruoyi.common.chat.factory;
+
+import org.ruoyi.common.chat.service.audio.IAudioGenerationService;
+import org.springframework.beans.BeansException;
+import org.springframework.context.ApplicationContext;
+import org.springframework.context.ApplicationContextAware;
+import org.springframework.stereotype.Component;
+
+import java.util.Map;
+import java.util.concurrent.ConcurrentHashMap;
+
+/**
+ * Text-to-speech service factory.
+ */
+@Component
+public class AudioServiceFactory implements ApplicationContextAware {
+
+    private final Map<String, IAudioGenerationService> audioServiceMap = new ConcurrentHashMap<>();
+
+    @Override
+    public void setApplicationContext(ApplicationContext applicationContext) throws BeansException {
+        Map<String, IAudioGenerationService> serviceMap = applicationContext.getBeansOfType(IAudioGenerationService.class);
+        for (IAudioGenerationService service : serviceMap.values()) {
+            if (service != null) {
+                audioServiceMap.put(service.getProviderName(), service);
+            }
+        }
+    }
+
+    public IAudioGenerationService getOriginalService(String providerCode) {
+        IAudioGenerationService service = audioServiceMap.get(providerCode);
+        if (service == null && "custom_api".equals(providerCode)) {
+            service = audioServiceMap.get("openai");
+        }
+        if (service == null) {
+            throw new IllegalArgumentException("不支持的语音模型供应商: " + providerCode);
+        }
+        return service;
+    }
+}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/factory/ImageServiceFactory.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/factory/ImageServiceFactory.java
@@ -37,6 +37,9 @@ public class ImageServiceFactory implements ApplicationContextAware {
     */
    public IImageGenerationService getOriginalService(String category) {
        IImageGenerationService service = imageSerivceMap.get(category);
+        if (service == null && "custom_api".equals(category)) {
+            service = imageSerivceMap.get("openai");
+        }
        if (service == null) {
            throw new IllegalArgumentException("不支持的模型类别: " + category);
        }
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/factory/VideoServiceFactory.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/factory/VideoServiceFactory.java
@@ -0,0 +1,40 @@
+package org.ruoyi.common.chat.factory;
+
+import org.ruoyi.common.chat.service.video.IVideoGenerationService;
+import org.springframework.beans.BeansException;
+import org.springframework.context.ApplicationContext;
+import org.springframework.context.ApplicationContextAware;
+import org.springframework.stereotype.Component;
+
+import java.util.Map;
+import java.util.concurrent.ConcurrentHashMap;
+
+/**
+ * Text-to-video service factory.
+ */
+@Component
+public class VideoServiceFactory implements ApplicationContextAware {
+
+    private final Map<String, IVideoGenerationService> videoServiceMap = new ConcurrentHashMap<>();
+
+    @Override
+    public void setApplicationContext(ApplicationContext applicationContext) throws BeansException {
+        Map<String, IVideoGenerationService> serviceMap = applicationContext.getBeansOfType(IVideoGenerationService.class);
+        for (IVideoGenerationService service : serviceMap.values()) {
+            if (service != null) {
+                videoServiceMap.put(service.getProviderName(), service);
+            }
+        }
+    }
+
+    public IVideoGenerationService getOriginalService(String providerCode) {
+        IVideoGenerationService service = videoServiceMap.get(providerCode);
+        if (service == null && "custom_api".equals(providerCode)) {
+            service = videoServiceMap.get("openai");
+        }
+        if (service == null) {
+            throw new IllegalArgumentException("不支持的视频模型供应商: " + providerCode);
+        }
+        return service;
+    }
+}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/audio/IAudioGenerationService.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/audio/IAudioGenerationService.java
@@ -0,0 +1,15 @@
+package org.ruoyi.common.chat.service.audio;
+
+import jakarta.validation.Valid;
+import org.ruoyi.common.chat.entity.audio.AudioContext;
+import org.ruoyi.common.chat.entity.media.MediaGenerationResponse;
+
+/**
+ * Public text-to-speech service.
+ */
+public interface IAudioGenerationService {
+
+    MediaGenerationResponse generateSpeech(@Valid AudioContext audioContext);
+
+    String getProviderName();
+}
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/chat/IChatModelService.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/chat/IChatModelService.java
@@ -73,4 +73,13 @@ public interface IChatModelService {
     * @return 是否删除成功
     */
    Boolean deleteWithValidByIds(Collection<Long> ids, Boolean isValid);
+
+    /**
+     * 按厂商批量更新密钥
+     *
+     * @param providerCode 厂商编码
+     * @param apiKey       密钥
+     * @return 是否更新成功
+     */
+    Boolean updateApiKeyByProvider(String providerCode, String apiKey);
 }
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/image/IImageGenerationService.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/image/IImageGenerationService.java
@@ -2,6 +2,8 @@ package org.ruoyi.common.chat.service.image;

 import jakarta.validation.Valid;
 import org.ruoyi.common.chat.entity.image.ImageContext;
+import org.ruoyi.common.chat.entity.media.MediaGenerationResponse;
+import org.ruoyi.common.chat.domain.vo.chat.ChatModelVo;

 /**
 * 公共文生图接口
@@ -9,10 +11,31 @@ import org.ruoyi.common.chat.entity.image.ImageContext;
 public interface IImageGenerationService {

    /**
-     * 根据文字生成图片
+     * 根据文字生成图片（同步阻塞，返回 URL）
     */
    String generateImage(@Valid ImageContext imageContext);

+    /**
+     * 异步启动图片生成（返回 prediction 信息，用于轮询进度）
+     * 默认回退到同步模式
+     */
+    default MediaGenerationResponse startImageGeneration(@Valid ImageContext imageContext) {
+        String url = generateImage(imageContext);
+        return MediaGenerationResponse.builder()
+            .type("image")
+            .url(url)
+            .status("completed")
+            .build();
+    }
+
+    /**
+     * 上传图生图/图生视频使用的临时媒体文件。
+     * 不支持上传的供应商保持默认异常，避免静默回退到错误存储。
+     */
+    default String uploadMedia(ChatModelVo model, byte[] content, String fileName, String contentType) {
+        throw new UnsupportedOperationException("当前图片供应商不支持临时媒体上传");
+    }
+
    /**
     * 获取服务提供商名称
     */
--- a/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/video/IVideoGenerationService.java
+++ b/ruoyi-common/ruoyi-common-chat/src/main/java/org/ruoyi/common/chat/service/video/IVideoGenerationService.java
@@ -0,0 +1,17 @@
+package org.ruoyi.common.chat.service.video;
+
+import jakarta.validation.Valid;
+import org.ruoyi.common.chat.entity.media.MediaGenerationResponse;
+import org.ruoyi.common.chat.entity.video.VideoContext;
+
+/**
+ * Public text-to-video service.
+ */
+public interface IVideoGenerationService {
+
+    MediaGenerationResponse generateVideo(@Valid VideoContext videoContext);
+
+    MediaGenerationResponse retrieveVideo(@Valid VideoContext videoContext);
+
+    String getProviderName();
+}
--- a/ruoyi-common/ruoyi-common-core/src/main/java/org/ruoyi/common/core/config/ThreadPoolConfig.java
+++ b/ruoyi-common/ruoyi-common-core/src/main/java/org/ruoyi/common/core/config/ThreadPoolConfig.java
@@ -10,6 +10,7 @@ import org.springframework.boot.context.properties.EnableConfigurationProperties
 import org.springframework.context.annotation.Bean;
 import org.springframework.core.task.VirtualThreadTaskExecutor;

+import org.springframework.scheduling.concurrent.ThreadPoolTaskExecutor;
 import java.util.concurrent.*;

 /**
@@ -22,6 +23,12 @@ import java.util.concurrent.*;
@EnableConfigurationProperties(ThreadPoolProperties.class)
 public class ThreadPoolConfig {

+    private final ThreadPoolProperties properties;
+
+    public ThreadPoolConfig(ThreadPoolProperties properties) {
+        this.properties = properties;
+    }
+
    /**
     * 核心线程数 = cpu 核心数 + 1
     */
@@ -54,6 +61,22 @@ public class ThreadPoolConfig {
        return scheduledThreadPoolExecutor;
    }

+    /**
+     * 知识库解析专用异步线程池
+     */
+    @Bean(name = "knowledgeParseExecutor")
+    public ThreadPoolTaskExecutor knowledgeParseExecutor() {
+        ThreadPoolTaskExecutor executor = new ThreadPoolTaskExecutor();
+        executor.setCorePoolSize(core);
+        executor.setMaxPoolSize(core * 2);
+        executor.setQueueCapacity(properties.getQueueCapacity());
+        executor.setKeepAliveSeconds(properties.getKeepAliveSeconds());
+        executor.setThreadNamePrefix("knowledge-parse-pool-");
+        executor.setRejectedExecutionHandler(new ThreadPoolExecutor.CallerRunsPolicy());
+        executor.initialize();
+        return executor;
+    }
+
    /**
     * 销毁事件
     * 停止线程池
--- a/ruoyi-common/ruoyi-common-core/src/main/java/org/ruoyi/common/core/service/OssService.java
+++ b/ruoyi-common/ruoyi-common-core/src/main/java/org/ruoyi/common/core/service/OssService.java
@@ -1,8 +1,11 @@
 package org.ruoyi.common.core.service;

+import jakarta.servlet.http.HttpServletResponse;
 import org.ruoyi.common.core.domain.dto.OssDTO;
 import org.springframework.web.multipart.MultipartFile;

+import java.io.File;
+import java.io.IOException;
 import java.util.List;

 /**
@@ -33,5 +36,19 @@ public interface OssService {
     */
    OssDTO uploadFile(MultipartFile file);

+    /**
+     * Upload a server-side file without buffering the complete payload in memory.
+     */
+    OssDTO uploadFile(File file);
+
+    /**
+     * Stream an object as an authenticated attachment response.
+     */
+    void downloadFile(Long ossId, HttpServletResponse response) throws IOException;
+
+    /**
+     * Delete one object and its sys_oss metadata.
+     */
+    Boolean deleteFile(Long ossId);

 }
--- a/ruoyi-common/ruoyi-common-oss/src/main/java/org/ruoyi/common/oss/core/OssClient.java
+++ b/ruoyi-common/ruoyi-common-oss/src/main/java/org/ruoyi/common/oss/core/OssClient.java
@@ -310,7 +310,7 @@ public class OssClient {
            client.deleteObject(
                x -> x.bucket(properties.getBucketName())
                    .key(removeBaseUrl(path))
-                    .build());
+                    .build()).join();
        } catch (Exception e) {
            throw new OssException("删除文件失败，请检查配置信息:[" + e.getMessage() + "]");
        }
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
ageerle	afaa86ef6e	Merge branch 'fix/sse-session-trace-cleanup'	2026-07-23 15:38:50 +08:00
ageerle	d1a820728f	fix(sse): SSE 改为按会话隔离 + 修 trace 租户 bug + 清理冗余 SSE 串台修复： - SseEmitterManager 新增按 sessionId 维度的 connect/sendEvent/disconnect，每个会话一个 SSE 连接，替代原 userId+token 维度（同用户多会话串台） - SseMessageDto 增加 sessionId + eventDto 字段，跨实例按会话路由 - SseTopicListener 优先按 sessionId 路由，回退原 userId/群发逻辑 - SseMessageUtils 增加 sessionId 重载，对话链路全部切换 - ChatServiceFacade / MyMcpClientListener / WorkflowStarter 切到 sessionId - 通知/全局 /sse 端点保留 userId 模式（通知按用户） trace 租户 bug 修复： - trace_run / trace_node 加入 tenant.excludes，监控表跨租户全局可见，绕过异步线程租户上下文不传播导致 node 查不到的问题冗余清理： - 删除零使用的 @TraceNode 注解 + TraceNodeAspect 切面 + 对应单测 - TraceProperties 删除 recordDetail/maxInputLength/maxOutputLength （仅对已删切面生效，编程式埋点用 RagTracePayloadBuilder 不受影响） - TracePayloadUtils 删除 input/output/asString 方法 - TraceConstants 删除无引用的 NODE_METHOD/HTTP/DB/CACHE/TASK/STREAM - TraceContext 删除预留的 copyNodeStack - TraceNodeVo 删除未用的 children 字段 - TraceRecordServiceImpl.getDetail 删除重复 enrich - application.yml 删除 trace.payload 无效配置项 - RagTraceNodeTypes 删除无引用的 NODE_STREAM Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-23 15:35:10 +08:00
ageerle	11bb1dba0f	Merge PR #308 : 通用链路追踪模块与 RAG 对话全链路埋点	2026-07-23 15:10:40 +08:00
Obltv	345d2819d8	fix: 修复链路追踪输出记录及关闭追踪时的额外开销	2026-07-23 14:52:19 +08:00
Obltv	36e3cf8c52	fix: 修复链路追踪输出记录及关闭追踪时的额外开销	2026-07-23 14:45:43 +08:00
Obltv	06ff95cdc3	Merge upstream/main into feature/common-trace-rag-chat	2026-07-23 14:30:38 +08:00
ageerle	9adeee6c83	fix(short-drama): 修复角色形象跨分镜不一致 1. findRefImageIndex 补上 appearance 参数，与 findStoryboardReferenceImages 的调用对齐。之前多形象角色（换装/青年老年）的 @imageN 绑定查的是默认形象 URL，和实际参考图列表对不上，导致 prompt 里没有 @imageN 标记 2. buildEnrichedVideoPrompt 角色段追加 visualDescription，文字 prompt 也承载外貌信息（发型/服装/体态），不能全靠参考图 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-21 16:06:06 +08:00
ageerle	e42d35f3bf	fix: 工作流实体保留字转义 + 千问向量模型厂商映射 1. WorkflowRuntimeNode/WorkflowRuntime: input/output/status 字段加反引号，修复 MySQL 8.0 保留字导致 MyBatis-Plus SQL 解析异常 (#307) 2. EmbeddingModelFactory: 新增 providerCode→BeanName 映射， qianwen→alibailian，修复千问向量模型找不到 Bean 的问题 (#301) Closes #307 #301 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-21 15:43:36 +08:00
ageerle	a69d1f51b9	feat: 打通 thinking 流式链路 + SqlAgent 表白名单校验 1. OllamaServiceImpl 补 .think() + .returnThinking()，对齐其他 provider (#306) 2. ChatServiceFacade.createCombinedHandler 新增 onPartialThinking， reasoning 内容经 SseMessageUtils.sendReasoning 流式推到前端 (#303) 3. SqlAgent 系统提示词追加无表时禁止执行 SQL 的规则 (#279) 4. ExecuteSqlQueryTool 代码层硬拦截：空白名单直接拒绝，有白名单则校验 SQL 中 FROM/JOIN 引用的表名 (#279) Closes #303 #306 #279 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-21 15:31:32 +08:00
ageerle	911ae6cb1f	fix: 修复 issue IJX5VV/IJPYWL/IHPUDA - IJX5VV: 补全工作流 9 个节点消息模板 sys_config 配置（SQL 脚本合并重命名时遗失），附幂等增量脚本；此前节点执行会抛「请先配置该节点的响应模板」 - IJPYWL: 向量模型未配置维度时，Milvus/Qdrant 回退默认 1024，避免 Integer 拆箱 NPE - IHPUDA: chat_provider.provider_icon 由 varchar(255) 加宽至 varchar(1000)，容纳 minio 私有桶签名 URL，附增量脚本 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-21 15:28:28 +08:00
ageerle	9a2f326e42	fix(rag): 合并后修正-升级脚本兼容 MySQL 8、重解析按 docId 清理旧向量 1. 升级脚本 2026-07-20-knowledge-fragment-fid.sql 使用了 MySQL 8 不支持的 ADD COLUMN IF NOT EXISTS（MariaDB 语法），在存量库执行会直接语法报错。改为与脚本内索引守卫一致的 information_schema 判断 + PREPARE 方式，保持幂等。 2. parse() 重新解析时按 fragment.fid 删除旧向量：存量数据的 fid 为迁移脚本回填的 MD5 值，与向量库中实际存储的随机 fid 不一致，首次重解析无法命中旧向量，导致新旧向量重复累积。改为写入新向量前先按 docId 清理（三种向量库策略均支持），写入失败时仍补偿删除新 fid。 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-21 14:33:28 +08:00
ageerle@163.com	46a8d6b552	Merge pull request #315 from MuSan-Li/update/2026-07-19-fixes 修复RAG相关逻辑优化:统一文档切割配置、向量数据生命周期、统一检索服务(阈值/重排/缓存)、按知识库路由向量库、重新解析支持	2026-07-21 14:33:11 +08:00
evo	cf5423f69f	fix(rag): 统一文档切割配置并支持重新解析	2026-07-21 12:18:19 +08:00
evo	177375a7b7	docs(rag): 更新完整修复与验收报告	2026-07-21 09:35:56 +08:00
evo	016bdbdd94	test(rag): 增加三向量库 Docker 生命周期测试	2026-07-21 09:35:56 +08:00
evo	41e0dcb528	chore(rag): 对齐默认向量配置并启用测试	2026-07-21 09:35:56 +08:00
evo	787e447d09	fix(aiflow): 接入统一知识库检索服务	2026-07-21 09:35:33 +08:00
evo	9c092ae3bb	fix(chat): 统一单轮检索并完善多库上下文	2026-07-21 09:35:26 +08:00
evo	42bc8fea95	fix(rag): 完善向量路由批量写入与数据生命周期	2026-07-21 09:35:18 +08:00
evo	c4fc1e6fcf	fix(rag): 隔离模型实例并修正缓存刷新	2026-07-21 09:35:07 +08:00
evo	6f6d0893ec	fix(rag): 统一检索阈值重排降级与缓存语义	2026-07-21 09:34:56 +08:00
evo	a6a55202a3	fix(rag): 修复文本分片边界与字面量分隔符	2026-07-21 09:34:47 +08:00
ageerle@163.com	24bee53f9e	feat: 新增 ChitChatAgent 闲聊兜底子 Agent 并接入 supervisor 简单任务无专业子 Agent 可用时由闲聊 Agent 兜底，避免 supervisor 空转返回空内容；通过 supervisorContext 约束仅在问候/闲聊场景路由到该 Agent。 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-20 16:08:48 +08:00
evo	afc9d3d17d	fix: 修复本地启动、数据库初始化与构建问题 (#314 ) * fix: 修复本地启动、数据库初始化与构建问题 * fix: 修复pr的一些问题提交	2026-07-19 13:42:43 +08:00
evo	97ce15d116	fix: 修复pr的一些问题提交	2026-07-19 13:30:09 +08:00
evo	879fe48945	fix: 修复本地启动、数据库初始化与构建问题	2026-07-19 12:20:22 +08:00
ageerle@163.com	5d834d924a	feat: 编程助手按聊天模型过滤并补充 MCP 工具 thinking 事件，README 补充 copilot/uniapp 仓库 - CodingController.models 按 category=chat 过滤，编程对话只展示聊天模型 - 各 MCP 文件/命令工具在执行前推送 thinking 提示事件 - README/README_EN 项目源码表新增 ruoyi-copilot、ruoyi-uniapp 三端链接 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-17 14:47:18 +08:00
ageerle@163.com	fece90b307	feat: 新增短剧音频/编码模块、MCP 文件工具与 WebSocket 支持 - 短剧: 新增 ShortDramaAudio 实体/Mapper/BO/VO，扩展 ShortDramaServiceImpl 合成逻辑与 FfmpegFilterGraphBuilder 滤镜图 - MCP: 新增 WriteFileTool/DeleteFileTool/ExecuteCommandTool，重构 EditFileTool/ReadFileTool/ListDirectoryTool - 编码: 新增 coding 模块（CodingAgent/WorkspaceService/SSE 事件通道） - Atlas: 新增音频生成实现，扩展音视频/媒体实体与预测服务 - WebSocket: 新增公众号聊天 WebSocket 处理器与握手拦截 - 忽略 logs/ 目录，避免合成产物入库 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-17 11:52:53 +08:00
ageerle@163.com	06d110ed16	@docs: 修正 snailjob 注释里失效的 ry_job.sql 引用实际建表脚本在 docs/script/sql/ruoyi-ai-v3_mysql8.sql（sj_ 前缀表），仓库中并无 ry_job.sql。关联 issue #311。 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-15 18:25:55 +08:00
ageerle@163.com	dfcadda2bb	@style: 修正 QdrantVectorStoreStrategy 缩进为空格 createSchema 调用与文件其余代码保持 4 空格缩进统一，避免 checkstyle/spotless 报错。 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-15 18:00:07 +08:00
Jack Jiang	3c9546ded9	fix: 修复 Qdrant 初始化创建集合问题 (#298 )	2026-07-15 17:58:12 +08:00
Octopus	a3d82e092e	Refresh MiniMax models and protocol support (#313 ) Co-authored-by: octo-patch <266937838+octo-patch@users.noreply.github.com>	2026-07-15 17:43:51 +08:00
ageerle@163.com	96a1b741df	fix: 兼容 Atlas 预测结果查询 404，避免前端轮询刷屏 Atlas /model/prediction/{id} 对已过期或不存在的任务返回 404 （常见于更换 API Key 后残留的旧任务）。原实现直接抛异常，前端轮询反复触发错误提示。改为 404 时返回 status=failed 的响应，让上层走正常失败分支落库，前端识别后终止轮询。 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-15 11:17:20 +08:00
ageerle@163.com	e7b7d1d084	更新README.md Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-15 10:52:15 +08:00
ageerle@163.com	efc30c3ac8	docs: README 新增短剧服务 ruoyi-drama 源码链接 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-14 22:30:10 +08:00
ageerle@163.com	2ae13aafd4	feat: 新增短剧/媒体生成能力并脱敏数据库脚本 - 新增音频/视频/图片生成服务与 Atlas、OpenAI 实现 - 新增短剧脚本生成、分镜与 ffmpeg 合成模块 - 新增 Coze/Dify 聊天 provider 及 Agent 配置 - 脱敏 SQL 脚本：移除真实 DeepSeek API Key，掩码 COS 桶名 - 清理过期的 SQL update 脚本与 chat 文档 Co-Authored-By: Claude <noreply@anthropic.com>	2026-07-14 21:49:03 +08:00
Obltv	e925050207	feat: 新增通用链路追踪模块与 RAG 对话全链路埋点 - 新增 ruoyi-common-trace 模块：TraceContext 上下文、TraceNodeAspect 注解切面、数据库双表持久化 - ChatServiceFacade / KnowledgeRetrievalServiceImpl 完成 RAG 对话 retrieval → rerank → llm-call 全链路采集 - 新增 TraceController 提供列表/详情/节点查询 API 及数据统计 - 支持 trace.enabled 开关、payload 长度截断、错误安全降级 - TraceNodeTemplate 模板封装标准节点生命周期，减少重复样板代码	2026-06-16 10:14:35 +08:00
ageerle@163.com	7aa67cfc2f	docs: 移除赞助商区块引用样式及多余换行 - 去掉 blockquote 引用块样式消除左侧竖线 - 去掉多余 <br> 避免链接文字断行 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-06-10 18:19:04 +08:00
ageerle@163.com	6d38e2bf25	docs: 赞助商图片改用本地路径并新增火山引擎 CodingPlan - Atlas Cloud 图片从远程链接改为 docs/image/sponsor/ 本地路径 - 新增火山引擎 CodingPlan 开发者计划赞助商（中英文 README 同步） Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-06-10 18:14:46 +08:00
ageerle@163.com	8ed5e61f63	docs: 添加微信技术交流群二维码 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-06-05 10:25:18 +08:00
ageerle@163.com	c3c89f8cc8	docs: 调整中文版赞助商图标大小 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-06-02 11:44:42 +08:00
ageerle@163.com	a3a6bdbc55	docs: 调整英文版赞助商图标大小与中文版一致 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-06-02 11:43:37 +08:00
ageerle@163.com	07d9fc1f45	docs: 同步英文版 README 与中文版保持一致 - 移除英文版 Quick Start/Live Demo 章节 - 移除 Ecosystem Partners 章节，合并到 Sponsors - 更新赞助商格式为图片横幅样式 - 社区交流改为微信/QQ 二维码（替换 Telegram） - 移除 Acknowledgments 中的 Spring AI Alibaba Copilot - 修复 Technical Architecture 多余空行 - 更新文档链接域名（pandarobot -> ruoyiai） Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-06-02 11:38:48 +08:00
ageerle@163.com	73a60e3591	docs: 更新 Atlas Cloud 赞助链接添加推荐参数 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-05-21 14:37:14 +08:00
ageerle@163.com	c79848c654	feat: 替换 PPIO 为 Atlas Cloud 并增强知识库功能 - 将 PPIO 服务提供商替换为 Atlas Cloud - 更新 README 赞助商信息 - 删除 PPIOEmbeddingProvider 和相关测试 - 知识库新增重排序、混合检索、相似度阈值等字段 - 更新 SQL 初始化脚本 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-05-21 11:18:59 +08:00
ageerle@163.com	ec092a11c3	docs: 添加 web coding 快速部署说明在 README 中新增使用 web coding 急速部署的章节，包含在线体验链接和部署截图展示 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-04-30 17:45:16 +08:00
ageerle	9a7b727413	Merge pull request #293 from RobustH/main 升级RAG模块	2026-04-23 09:15:02 +08:00
RobustH	b8d16b7669	feat(rag): 对接用户端用户知识库对话，集成知识库配置应用	2026-04-23 00:52:53 +08:00
RobustH	058a4aee2a	feat(rag): 新增测试配置应用的功能	2026-04-21 22:54:11 +08:00
RobustH	1b50c7f9f1	fix(rag): 修复合并重复，重排模型新增硅基流动供应商	2026-04-21 22:41:00 +08:00
RobustH	e7f53fd55f	Merge remote-tracking branch 'origin/main' # Conflicts: # ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/domain/bo/knowledge/KnowledgeInfoBo.java # ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/domain/entity/knowledge/KnowledgeInfo.java # ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/domain/vo/knowledge/KnowledgeInfoVo.java # ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/service/chat/impl/ChatServiceFacade.java	2026-04-21 21:41:51 +08:00
ageerle	07bdc5e585	Merge pull request #292 from yangzhen233/feature/rerank-model Feature/rerank model	2026-04-20 21:06:27 +08:00
yangzhen	e1b8a5f011	新增千问3重排序模型，并附带新增sql文件	2026-04-20 16:07:02 +08:00
杨振	80ca76ea37	添加重排序功能	2026-04-20 01:02:09 +08:00
wangle	2c6ff66830	fix: 修正application.yml演示模式message缩进 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-04-19 13:42:39 +08:00
wangle	4f79a66559	feat: 添加小米MiMo、DeepSeek、自定义厂商等provider支持 - 新增小米MiMo服务实现类(MiMoServiceImpl) - ChatModeType添加XIAOMI枚举 - 更新SQL初始化脚本，新增多家厂商(provider)和模型数据 - 添加2026-04-19数据库更新脚本 - application.yml演示模式排除路径增加attach/fragment/info接口 - 删除独立的minimax_provider.sql(数据已合并到主SQL) Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-04-19 13:42:05 +08:00
wangle	22883b4334	Merge branch pr-280: 添加MiniMax作为LLM提供商解决冲突: - README: 保留Qdrant向量库信息 + 合并MiniMax模型接入 - pom.xml: 保留spring-boot-starter-test + 添加MiniMax测试依赖 - ChatModeType: 保留CUSTOM_API + 新增MINIMAX枚举 - MinimaxServiceImpl: 保留MyChatModelListener监听器 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-04-19 11:47:04 +08:00
wangle	081da6d18d	feat: 添加MiniMax作为LLM提供商，合并PR#280并补充监听合并PR#280的MiniMax provider实现，解决与main分支的冲突，并在MinimaxServiceImpl中补充MyChatModelListener监听，与其他provider保持一致。 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>	2026-04-17 18:31:53 +08:00
ageerle	74eb5b2530	Merge pull request #290 from yangzhen233/feature/zhipu-embedding 添加智谱向量模型实现	2026-04-17 18:24:24 +08:00
ageerle	b0328fe0ef	Merge pull request #291 from xiaonieli7/main fate：增加自定义模型，调整前端模型选择下拉框	2026-04-17 18:16:59 +08:00
Administrator	2ee0aae57e	fate：增加自定义模型，调整前端模型选择下拉框	2026-04-17 08:31:40 +08:00
杨振	d9c3de660a	添加智谱向量模型实现	2026-04-16 21:18:11 +08:00
RobustH	ccbf5c9520	feat(rag): 知识库检索测试新增混合检索	2026-04-14 23:18:29 +08:00
ageerle	c4f7c1f5d0	Merge pull request #283 from MrWws/fix/docker-compose-all.yaml fix: 修复运行docker-compose-all.yaml报错的问题	2026-04-14 14:36:14 +08:00
RobustH	1208c46cca	feat(rag): 集成硅基流动、阿里百炼重排模型并全方位增强检索测试体验	2026-04-14 01:40:28 +08:00
RobustH	06a63c377e	feat: 新增检索测试相关接口 - 实现向量 L2 归一化，统一 Milvus/Qdrant/Weaviate 检索评分为 [0, 1] 空间	2026-04-13 23:33:56 +08:00
RobustH	0fa25032a3	feat(knowledge): 优化知识库文件状态枚举为"未解析，解析中，解析成功，解析失败"，支持异步线程池解析文档	2026-04-13 00:15:01 +08:00
RobustH	28ad29d6ed	feat(knowledge): 完善知识库及附件统计功能并修复分块数统计问题	2026-04-12 18:38:32 +08:00
MrWws	b9097b4989	fix: 修复运行docker-compose-all.yaml报错的问题	2026-03-31 22:59:26 +08:00
octopus	5d14eb20af	feat: add MiniMax as first-class LLM provider Add MiniMax AI as the 7th LLM provider, supporting chat (M2.7, M2.5, M2.5-highspeed) and embedding (embo-01) models via OpenAI-compatible API. Changes: - Add MINIMAX enum to ChatModeType - Add MinimaxServiceImpl chat provider (OpenAI-compat streaming) - Add MinimaxEmbeddingProvider for vector embeddings - Add SQL migration for provider and model registration - Add 14 unit tests + 3 integration tests - Update README/README_EN with MiniMax in provider list	2026-03-21 16:14:19 +08:00