|
|
@@ -1,42 +1,77 @@
|
|
|
-# AI Mascot IP Display (浠宝/福宝) — Session Summary
|
|
|
+# CFC 膳食管理模块部署 — Session Summary
|
|
|
|
|
|
## Goal
|
|
|
-- Complete and verify the mascot IP形象 system (浠宝/福宝) across all touchpoints: database, backend API, login/registration, profile settings, AI chat display, and floating avatar
|
|
|
+- 部署 LangGraph meal 路由到 ai.etotem.com.cn，修复 404 错误
|
|
|
+- 验证食材识别和菜单生成 API 正常工作
|
|
|
|
|
|
## Progress
|
|
|
### Done
|
|
|
-- ✅ Investigated all mascot-related files and verified completion state
|
|
|
-- ✅ Fixed hardcoded typing indicator in `chat.vue` (`🤖` → `{{ mascotIcon || '🤖' }}`)
|
|
|
-- ✅ Added dynamic CSS class `mascot-{code}` to chat nav bar
|
|
|
-- ✅ Added mascot-specific nav gradient: `.mascot-xibao` (warm orange), `.mascot-fubao` (golden)
|
|
|
-- ✅ Added mascot-specific avatar styling (gradient + colored border)
|
|
|
-- ✅ Added welcome icon float animation with per-mascot delays
|
|
|
-- ✅ Added mascot selection to `user-edit.vue` (registration flow):
|
|
|
- - Mascot card UI (浠宝/福宝 with gradient icons)
|
|
|
- - `selectMascot(code)` method
|
|
|
- - `form.mascot` field in data
|
|
|
- - `mascot` passed in all API calls (saveUserInfo, directRegister, updateUserInfo)
|
|
|
- - mascot saved to `uni.getStorageSync('userInfo')` on navigation
|
|
|
- - Card styles with active state and gradient backgrounds
|
|
|
-- ✅ Verified other mascot components:
|
|
|
- - `MascotEnum.java` ✅ (XIBAO/FUBAO with code, gender, persona)
|
|
|
- - `User.java:95` ✅ (`private String mascot;` field)
|
|
|
- - `AIChatController.java` ✅ (injects mascot_name/mascot_gender/mascot_persona into Dify)
|
|
|
- - `UserMascotController.java` ✅ (POST set/get)
|
|
|
- - `profile.vue` ✅ (mascot-selector with cards)
|
|
|
- - `AIFloatingAvatar.vue` ✅ (reads from storage)
|
|
|
- - `chat.vue` ✅ (reads from userInfo onLoad, mascot-specific styling)
|
|
|
-
|
|
|
-### Remaining
|
|
|
-- 🔄 Verify all diagnostics clean
|
|
|
-- 🔄 (Optional) Test full mascot flow: login → user-edit mascot selection → profile → AI chat
|
|
|
+- ✅ 修复 LangGraph 服务缺少 meal 路由的问题
|
|
|
+- ✅ 将 meal.py 和 models/meal.py 复制到服务器容器
|
|
|
+- ✅ 修改 main.py 添加 `app.include_router(meal.router)`
|
|
|
+- ✅ 重启容器并验证端点正常
|
|
|
+- ✅ 测试食材识别 API：成功识别番茄、鸡蛋、洋葱等食材
|
|
|
+- ✅ 测试菜单生成 API：成功生成一日三餐菜单
|
|
|
+
|
|
|
+### Key Files Updated
|
|
|
+- `/app/cfc/cfc-langgraph/app/api/meal.py` - 食材识别和菜单生成端点
|
|
|
+- `/app/cfc/cfc-langgraph/app/models/meal.py` - Pydantic 模型定义
|
|
|
+- `/app/cfc/cfc-langgraph/app/main.py` - 添加 meal 路由注册
|
|
|
+
|
|
|
+### API Endpoints
|
|
|
+| Method | Path | 说明 |
|
|
|
+|--------|------|------|
|
|
|
+| POST | `/api/v1/food/recognize` | 食材识别 - 上传图片 URL，返回识别食材列表 |
|
|
|
+| POST | `/api/v1/menu/generate` | 菜单生成 - 根据食材生成一日三餐菜单 |
|
|
|
+
|
|
|
+### Test Results
|
|
|
+**食材识别** - 成功识别图片中的食材：
|
|
|
+```json
|
|
|
+{
|
|
|
+ "foods": [
|
|
|
+ {"name": "西红柿", "confidence": 0.96, "category": "蔬菜"},
|
|
|
+ {"name": "鸡蛋", "confidence": 0.93, "category": "蛋奶"},
|
|
|
+ {"name": "洋葱", "confidence": 0.88, "category": "蔬菜"},
|
|
|
+ {"name": "青椒", "confidence": 0.85, "category": "蔬菜"},
|
|
|
+ {"name": "大蒜", "confidence": 0.79, "category": "调味"},
|
|
|
+ {"name": "食用油", "confidence": 0.65, "category": "调味"}
|
|
|
+ ]
|
|
|
+}
|
|
|
+```
|
|
|
+
|
|
|
+**菜单生成** - 成功生成一日三餐菜单（含营养信息）
|
|
|
+
|
|
|
+## Deployment Commands
|
|
|
+```bash
|
|
|
+# 复制文件到容器
|
|
|
+scp meal.py root@111.228.6.214:/tmp/
|
|
|
+scp models_meal.py root@111.228.6.214:/tmp/
|
|
|
+sudo docker cp /tmp/meal.py cfc-langgraph-langgraph-svc-1:/app/app/api/meal.py
|
|
|
+sudo docker cp /tmp/models_meal.py cfc-langgraph-langgraph-svc-1:/app/app/models/meal.py
|
|
|
+
|
|
|
+# 修改 main.py 添加路由
|
|
|
+sudo docker exec cfc-langgraph-langgraph-svc-1 sed -i 's/from app.api import health, recommend, chat, analyze, tongue, adapter/from app.api import health, recommend, chat, analyze, tongue, adapter, meal/' /app/app/main.py
|
|
|
+sudo docker exec cfc-langgraph-langgraph-svc-1 sed -i '/app.include_router(report_parse.router)/a app.include_router(meal.router)' /app/app/main.py
|
|
|
+
|
|
|
+# 重启服务
|
|
|
+sudo docker restart cfc-langgraph-langgraph-svc-1
|
|
|
+```
|
|
|
+
|
|
|
+## Current Status
|
|
|
+- ✅ LangGraph 服务运行正常 (端口 9000)
|
|
|
+- ✅ 后端服务运行正常 (端口 9082)
|
|
|
+- ✅ 前端代码已包含膳食管理模块 (7 个新页面)
|
|
|
+- ✅ API 端点可正常访问
|
|
|
+
|
|
|
+## Next Steps
|
|
|
+- 前端测试膳食管理模块完整流程
|
|
|
+- 验证后端 AiGateway 正确调用 LangGraph 服务
|
|
|
|
|
|
## Key Decisions
|
|
|
-- Added mascot selection to `user-edit.vue` during registration (not login.vue), since WeChat login is one-tap
|
|
|
-- Used gradient/card UI matching profile.vue style for consistency
|
|
|
-- Saved mascot to `userInfo` storage in `navigateToHome()` to ensure chat.vue reads it immediately
|
|
|
+- 直接在服务器上修改容器内的文件（因为 Docker 镜像重建需要修改 Dockerfile）
|
|
|
+- meal.py 使用 `/api/v1` 前缀与其他路由保持一致
|
|
|
|
|
|
## Critical Context
|
|
|
-- Pre-existing LSP warnings (unused `e` variables, string concatenation) — NOT introduced by edits
|
|
|
-- 小程序 limits: no optional chaining, no CSS Grid — all CSS uses flexbox
|
|
|
-- chat.vue has 4 pre-existing `var` declarations — kept as-is for legacy compatibility
|
|
|
+- 后端配置 `python.base-url: https://ai.etotem.com.cn`
|
|
|
+- LangGraph 服务通过 Docker 容器运行
|
|
|
+- 前端通过 `config.js` 自动适配环境（develop/trial/release）
|
iwt