|
|
5dfff52b04
|
feat: 后台一键刷新飞书缓存
- feishu_service.py: FeishuKnowledgeService 新增 clear_cache() 方法
- admin_settings.py: 新增 POST /admin/feishu/cache/clear API
- api.ts: 前端封装 clearFeishuCache API
- App.vue: Agent管理页面新增'刷新飞书缓存'按钮
管理员更新飞书文档后，点击按钮即可清空服务端内存缓存，
客户端下次提问时会重新读取最新内容。
|
2026-07-08 17:57:26 +08:00 |
|
|
|
61db0aa9b8
|
feat: 对话摘要+段落按需读取+think标签解析修复
- chat.py: ChatSession 新增 summary/summary_up_to_message_id 字段
- rag_service.py: Prompt 支持历史摘要+最近5轮原文拼接
- model_service.py: 新增 summarize() 方法，复用已启用模型生成摘要
- chat_service.py: 集成对话历史传递，8轮后触发摘要生成
- feishu_service.py: 段落级按需读取，按标题匹配+关键词打分+top5文档
摘要策略：
- 保留最近5轮原文保证追问连续性
- 更早历史由模型压缩成摘要（<=300字）
- 摘要失败不阻断主流程
飞书读取优化：
- 每个知识库最多读5个相关文档
- 每文档提取top10相关段落
- 按原文顺序拼接，不截断
Fix: _extract_openai_answer 重命名为 _extract_gemini_answer
避免同名函数覆盖导致MiniMax模型被错误解析为Gemini响应
|
2026-07-08 17:49:52 +08:00 |
|
|
|
c68af6f806
|
Cache Feishu retrieval data in Redis
|
2026-07-08 17:39:04 +08:00 |
|
|
|
23d5f42972
|
Add Redis-backed chat queue
|
2026-07-08 17:36:20 +08:00 |
|
|
|
812ae80536
|
Add async Feishu retrieval for chat RAG
|
2026-07-08 17:26:10 +08:00 |
|
|
|
72bd2e1e02
|
Use async model streaming for chat
|
2026-07-08 17:20:29 +08:00 |
|
|
|
465a481e7a
|
Stream chat completions from model providers
|
2026-07-08 17:13:53 +08:00 |
|
|
|
6f3bbd3803
|
Implement chat queue fallback
|
2026-07-08 16:57:40 +08:00 |
|
|
|
925ef0238d
|
Support Feishu file node retrieval
|
2026-07-08 12:29:21 +08:00 |
|
|
|
13185ab42b
|
Align user chat with agent model flow
|
2026-07-08 12:18:38 +08:00 |
|
|
|
fb0bbc346b
|
Fix Feishu wiki child node retrieval
|
2026-07-08 12:08:23 +08:00 |
|
|
|
7be5dde1db
|
Support Feishu credentials in system settings
|
2026-07-08 12:03:37 +08:00 |
|
|
|
eda2c5af76
|
Clarify chat retrieval failures
|
2026-07-08 11:55:22 +08:00 |
|
|
|
ac72b6a94a
|
Implement scoped real knowledge retrieval
|
2026-07-08 11:46:09 +08:00 |
|
|
|
29f25dd767
|
Implement student roster login gating
|
2026-07-07 18:45:07 +08:00 |
|
|
|
03b1810a90
|
Implement open knowledge access scope
|
2026-07-07 18:32:01 +08:00 |
|
|
|
da262f4596
|
Separate agent tuning from model setup
|
2026-07-07 13:31:20 +08:00 |
|
|
|
dada4aafcd
|
Add quick LLM provider setup
|
2026-07-07 13:12:27 +08:00 |
|
|
|
55b7c5b0ea
|
Complete admin model management
|
2026-07-07 12:58:26 +08:00 |
|
|
|
08c301c136
|
Complete admin chat records
|
2026-07-07 12:17:23 +08:00 |
|
|
|
62eeb3186c
|
Add V2 admin console
|
2026-07-06 18:13:08 +08:00 |
|
|
|
301661e7b4
|
Add V2 external service providers
|
2026-07-06 18:03:29 +08:00 |
|
|
|
207340973c
|
Add V2 RAG answer skeleton
|
2026-07-06 17:55:22 +08:00 |
|
|
|
17cba7cf61
|
Scaffold V2 backend foundation
|
2026-07-06 17:25:06 +08:00 |
|
