XEngineers/CamTalk
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
220 Commits 2 Branches 0 Tags
2aa3c98ab6251edfdd838d739b43c92e916495fd
Commit Graph

220 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
hhs
450c2c66c3 feat: 定义 TTS Service 接口（SynthesizeStream + Options/Chunk） 2026-06-13 15:46:48 +08:00
hhs
90f4b907a7 test: 编写 LLM 测试（mock SSE 流，覆盖成功/图片/历史/API 错误/超时/Usage） 2026-06-13 15:45:51 +08:00
hhs
96b4c4899a feat: 实现 OpenAI LLM 服务（SSE 流式解析 + 多模态消息构建）及 System Prompt 定义 2026-06-13 15:44:23 +08:00
hhs
078cfe6278 feat: 定义 LLM Service 接口（ChatStream + Request/Chunk/TokenUsage） 2026-06-13 15:42:14 +08:00
hhs
8f9b0b881e feat: 实现 Deepgram STT 服务及测试（WebSocket mock，覆盖成功/超时/空音频/多轮 final） 2026-06-13 15:41:44 +08:00
hhs
6fcbf023b4 feat: 实现 Deepgram STT 服务（WebSocket 连接 + 音频发送 + 结果接收） 2026-06-13 15:39:39 +08:00
hhs
22784bf421 feat: 定义 STT Service 接口（Recognize + Options） 2026-06-13 15:38:59 +08:00
huanghaosheng
33effe161a Merge pull request '实现会话生命周期管理，让 WS handler 能追踪会话、存储对话历史。' (#35) from feature/phase2 into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/35
 2026-06-13 15:34:04 +08:00
hhs
fe4dedec71 feat: WS handler 接入 SessionManager，main.go 注入依赖 + 健康检查获取活跃会话数
Some checks failed
Backend CI / ci (pull_request) Failing after 15s
Details
Frontend CI / ci (pull_request) Has been cancelled
Details
2026-06-13 15:31:15 +08:00
hhs
26f87ce35a test: 编写 Session Manager 内存实现测试（覆盖 Create/Get/Expire/Destroy/History 上限等） 2026-06-13 15:29:13 +08:00
hhs
ba6d565c0f feat: 实现 Redis 版 SessionManager（Hash + List，TTL 自动刷新） 2026-06-13 15:27:41 +08:00
hhs
f172e6c468 feat: 实现内存版 SessionManager（TTL 30 分钟，历史上限 20 条） 2026-06-13 15:25:23 +08:00
hhs
6b8bd3c554 feat: 定义 SessionManager 接口 + SessionConfigPatch 模型 2026-06-13 15:24:07 +08:00
huanghaosheng
688e13a491 Merge pull request '补充跨域处理方案（Nginx 同源反代 + Vite proxy）' (#33) from feature/phase1 into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/33
 2026-06-13 15:20:06 +08:00
hhs
991ae4834c docs: Phase 1 移除 CORS 中间件任务 
- CORS 由 Nginx 反向代理统一处理，不在后端实现
 2026-06-13 15:18:11 +08:00
hhs
da87ec776b feat: main.go 接入配置 + graceful shutdown + logger 迁移 
- main.go: 接入 config.Load() 替换硬编码 :8080
- main.go: 添加 signal.NotifyContext + http.Server.Shutdown（10s drain）
- main.go: 初始化 Zap 日志，prod 环境切换 Gin release 模式
- ws/handler.go: 所有 log.Printf 替换为 logger.Log 结构化日志
- 修复 .gitignore 排除规则（/server 仅匹配根目录二进制）
- 新增 zap、viper 依赖
 2026-06-13 15:18:03 +08:00
hhs
c523f53724 feat: 实现配置管理（Viper） 
- 新增 internal/config/config.go，定义完整配置结构体
- 支持 YAML 文件 + 环境变量覆盖（CAMTALK_ 前缀）
- 新增 config.yaml 默认配置文件
 2026-06-13 15:17:37 +08:00
hhs
bce3b73480 feat: 实现错误码常量 + WS 错误发送工具 
- 新增 internal/errors/codes.go
- 定义 10 个错误码常量（与接口文档一致）
- 提供 SendWSError 工具函数
 2026-06-13 15:17:31 +08:00
hhs
d5d584c93b feat: 实现 Zap 日志封装 
- 新增 internal/logger/logger.go
- 提供 Init(level, format) 初始化全局 SugaredLogger
- 支持 json/console 两种格式
- 提供 Sync() 刷新缓冲区
 2026-06-13 15:17:26 +08:00
huanghaosheng
b76a71c128 Merge pull request 'feature/base-build' (#32) from feature/base-build into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/32
 2026-06-13 15:02:28 +08:00
hhs
2e165a459c docs: 添加后端分阶段实施计划（8 Phase） 
从基础设施到集成测试的完整实施路线图，涵盖 Session Manager、AI 服务层、
Orchestrator 编排、WS Handler 接入、REST API、Rate Limiter 等模块。
 2026-06-13 15:01:38 +08:00
cfy777
7e2537104d Merge pull request '实现观察模式，支持持续场景监控（US-05/US-10）' (#31) from develop-frontend5 into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/31
 2026-06-13 14:51:57 +08:00
cfy666
f5536005f6 feat: 实现观察模式，支持持续场景监控（US-05/US-10） 
- 新增 useObservationMode Hook：定时 5s 采帧，similarity < 0.85 触发变化回调
- useVisionSession 新增 dialogue/observation 模式切换
- App Footer 加观察模式切换按钮（激活时蓝色脉冲）
- 视频区左上角显示'👁️ 观察中'绿色角标
- 观察模式下画面变化自动发送 query

Co-Authored-By: Claude <noreply@anthropic.com>
 2026-06-13 14:46:42 +08:00
cfy777
e18a2004a0 Merge pull request '实现关键帧检测与成本控制' (#30) from develop-frontend4 into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/30
 2026-06-13 14:40:05 +08:00
cfy666
2c6489b9a6 feat: 实现关键帧检测与成本控制 
- EdgeProcessor：sampleFrame 降采样 160x120 + compareFrames 像素差异对比
- 重复画面跳过：similarity > 0.9 时不发送 query
- 混合采样策略：静默 5s / 用户说话 1s（docs/08-成本控制.md）
- 请求统计：queryCount + totalTokens 实时显示
- App 状态栏显示成本统计

Co-Authored-By: Claude <noreply@anthropic.com>
 2026-06-13 14:38:15 +08:00
cfy777
8dc59d2c09 Merge pull request '实现会话配置与多轮上下文' (#29) from develop-frontend3 into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/29
 2026-06-13 14:24:11 +08:00
cfy666
7e3762b589 feat: 实现会话配置与多轮上下文
All checks were successful
Backend CI / ci (pull_request) Successful in 1m0s
Details
Frontend CI / ci (pull_request) Successful in 1m0s
Details 
- 新增 ConfigPanel 组件：TTS 开关、detail level、语言选择
- 新增 storage.ts：配置持久化到 localStorage
- useVisionSession：连接后自动发送 config，维护最近 10 轮对话历史
- App：齿轮按钮展开配置面板，HD 角标，TTS 播放联动 ttsEnabled

Co-Authored-By: Claude <noreply@anthropic.com>
 2026-06-13 14:19:47 +08:00
hhs
0cafe94f3a docs: 补充跨域处理方案（Nginx 同源反代 + Vite proxy） 
- 02-系统架构: 部署架构改为 Nginx 同源反代模型，新增完整 Nginx 配置和 Vite proxy 配置
- 03-接口文档: 连接管理新增跨域处理小节（生产 Nginx / 开发 Vite proxy / CheckOrigin 策略）
 2026-06-13 14:05:57 +08:00
cfy777
8033807f7c Merge pull request 'feat: 实现 TTS 语音播放，完成 MVP P0 全部用户故事' (#27) from develop-frontend2 into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/27
 2026-06-13 14:02:28 +08:00
cfy666
6a47c4dfbb feat: 实现 TTS 语音播放，完成 MVP P0 全部用户故事 
- 新增 TTSPlayer：收集流式 tts_audio 片段，is_last 时拼接解码播放
- useVisionSession 接入 TTS 播放器，interrupt/stopSession 时停止播放
- App 显示 '🔊 正在播放...' 指示器
- MVP P0 四个用户故事前端全部实现（US-01 ~ US-04）

Co-Authored-By: Claude <noreply@anthropic.com>
 2026-06-13 13:57:36 +08:00
huanghaosheng
0ef7aa657d Merge pull request 'docs' (#26) from docs into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/26
 2026-06-13 13:40:44 +08:00
hhs
360b2a4f50 docs: 补充配置管理规范（Viper + 环境变量）
All checks were successful
Backend CI / ci (pull_request) Successful in 1m5s
Details
Frontend CI / ci (pull_request) Successful in 1m40s
Details 
- 03-接口文档: 新增第六章配置管理（Go 配置结构体、YAML 配置示例、环境变量覆盖规则、Viper 加载代码、启动命令示例）
- 03-接口文档: 原七~九章重新编号为八~十章
- 02-系统架构: 配置管理行补充交叉引用
- README.md: 索引补充配置管理关键词
 2026-06-13 13:31:48 +08:00
hhs
ee557eaa3d docs: 补充 Session Manager 接口规范与 Redis 数据结构 
- 03-接口文档: 新增第五章 Session Manager（Redis Hash+List 数据结构、TTL 策略、接口定义、Handler 集成、MVP 内存实现）
- 03-接口文档: 原五~八章重新编号为六~九章
- 02-系统架构: Session Manager 行补充 Redis 数据结构说明和交叉引用
- README.md: 索引补充 Session Manager 关键词
 2026-06-13 13:24:25 +08:00
hhs
87b6407e2e docs: 补充 AI 编排层接口规范与 tts_audio 音频播放方案 
- 03-接口文档: 新增第三章 AI 服务层接口（STT/LLM/TTS 三个 Service interface + 接入约定）
- 03-接口文档: 新增第四章 AI 编排器（句子级流式并行策略、Orchestrator 实现、错误降级表）
- 03-接口文档: 锁定 tts_audio 音频格式为 audio/mpeg（MP3 24kHz），新增前端 AudioPlayer 播放方案
- 02-系统架构: 更新 Orchestrator 代码为句子级流式并行实现
- README.md: 更新 03-接口文档描述，补充 AI 服务层和编排器关键词
 2026-06-13 13:18:17 +08:00
cfy777
2e222cbe73 Merge pull request '完善端到端对话闭环，增强消息渲染与错误处理;集成 VAD 语音活动检测，打通核心交互链路' (#24) from develop-frontend into develop 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/24
 2026-06-13 11:56:26 +08:00
huanghaosheng
dc5e49d216 Merge pull request 'build/actions' (#22) from build/actions into develop
Some checks failed
Backend CI / ci (pull_request) Successful in 1m5s
Details
Frontend CI / ci (pull_request) Has been cancelled
Details 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/22
 2026-06-13 11:46:47 +08:00
hhs
c7f741943d ci: 移除 health-test 工作流 2026-06-13 11:46:34 +08:00
hhs
8ce5c02730 ci: 统一 workflow 步骤名称为英文 2026-06-13 11:45:08 +08:00
cfy666
734c93a866 feat: 完善端到端对话闭环，增强消息渲染与错误处理 
- ChatPanel：流式光标动画、token/延迟元数据、自动滚动、三种空状态
- Toast 组件：错误码映射为中文友好文案，3 秒自动消失
- useVisionSession：stt_result 流式更新、防重复发送、打断保存未完成内容、结束清理全部状态
- App：断线提示、视频 loading 遮罩、按钮连接中状态
- ChatMessage 增加 latencyMs、model 字段

Co-Authored-By: Claude <noreply@anthropic.com>
 2026-06-13 11:43:16 +08:00
huanghaosheng
fc8b67be1a Merge pull request 'fix: 修正 install node 步骤的 working-directory 并使用阿里云镜像源' (#20) from build/actions into develop
All checks were successful
Backend CI / ci (pull_request) Successful in 1m11s
Details
Frontend CI / ci (pull_request) Successful in 30s
Details 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/20
 2026-06-13 11:41:47 +08:00
hhs
3d7d7d190c fix: 修正 install node 步骤的 working-directory 并使用阿里云镜像源 2026-06-13 11:41:08 +08:00
huanghaosheng
01e3eb421c Merge pull request 'fix: 使用 Alpine 容器镜像解决 checkout action 找不到 node 的问题' (#18) from build/actions into develop
Some checks failed
Backend CI / ci (pull_request) Failing after 42s
Details
Frontend CI / ci (pull_request) Successful in 1m7s
Details 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/18
 2026-06-13 11:39:16 +08:00
hhs
c5c9bf802f fix: 使用 Alpine 容器镜像解决 checkout action 找不到 node 的问题 2026-06-13 11:38:48 +08:00
huanghaosheng
2c9159afad Merge pull request 'fix: 移除 setup-go/setup-node，宿主机已预装 Go 和 Node' (#16) from build/actions into develop
Some checks failed
Backend CI / ci (pull_request) Failing after 0s
Details
Frontend CI / ci (pull_request) Failing after 1s
Details 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/16
 2026-06-13 11:31:18 +08:00
hhs
17c8ced6bd fix: 移除 setup-go/setup-node，宿主机已预装 Go 和 Node 2026-06-13 11:29:41 +08:00
cfy666
9ccd4d6238 feat: 集成 VAD 语音活动检测，打通核心交互链路 
- 重写 useVAD Hook，接入 @ricky0123/vad-web 的 MicVAD
- 支持 onSpeechStart/onSpeechEnd/onVADMisfire 回调
- useVisionSession 中串联摄像头→麦克风→WebSocket→VAD 完整流程
- App.tsx 新增 VAD 加载中和错误状态的 UI 指示
- VAD 参数对齐设计文档：positiveSpeechThreshold=0.5, minSpeechMs=250

Co-Authored-By: Claude <noreply@anthropic.com>
 2026-06-13 11:10:29 +08:00
cfy777
e372c0022d Merge pull request 'fix: 为 uses 的 URL 加引号修复 YAML 解析错误' (#14) from build/actions into develop
Some checks failed
Backend CI / ci (pull_request) Failing after 4s
Details
Frontend CI / ci (pull_request) Failing after 1m14s
Details 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/14
Reviewed-by: cfy777 <3087823110@qq.com>
 2026-06-13 10:48:29 +08:00
hhs
2ce04dcca1 fix: 为 uses 的 URL 加引号修复 YAML 解析错误 2026-06-13 10:46:32 +08:00
cfy777
cc8337bef6 Merge pull request 'build/actions' (#12) from build/actions into develop
Some checks failed
Backend CI / ci (pull_request) Failing after 1m25s
Details
Frontend CI / ci (pull_request) Failing after 1m36s
Details 
Reviewed-on: http://8.161.227.145:3000/XEngineers/CamTalk/pulls/12
Reviewed-by: cfy777 <3087823110@qq.com>
 2026-06-13 10:44:05 +08:00
hhs
8c9a9d6020 ci: 添加 health test 工作流用于验证 Gitea Actions 运行 2026-06-13 10:42:07 +08:00