feat: AI 场景路由、ASR 服务及前后端全链路同步

- 新增 AI 场景路由控制器和管理接口 - 新增 ASR 语音识别服务及前后端集成 - 同步 AI Runtime 客户端到 Web/小程序/Life-Script - 完善 AI 配置测试修复和管理后台路由配置 - 新增数据库迁移脚本 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-23 13:25:21 +08:00
parent d77090aa5e
commit 89fc42819d
72 changed files with 4584 additions and 383 deletions
@@ -0,0 +1,64 @@
+# Mini Program ASR Service Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Deploy a private Chinese ASR service on `101.200.208.45` and wire the mini program voice orb to transcribe speech into the wish input.
+
+**Architecture:** A local FastAPI ASR service listens on `127.0.0.1:19120` and uses FunASR/SenseVoiceSmall when available. The Java backend exposes `/api/asr/transcribe` as the authenticated upload proxy, and the mini program records audio with `uni.getRecorderManager()` then uploads it through the existing request service.
+
+**Tech Stack:** FunASR/SenseVoiceSmall, Python 3.11, FastAPI, Spring Boot multipart upload, uni-app recorder/upload APIs.
+
+---
+
+### Task 1: ASR Service
+
+**Files:**
+- Create: `backend-single/asr-service/app.py`
+- Create: `backend-single/asr-service/requirements.txt`
+- Create: `backend-single/asr-service/emotion-museum-asr.service`
+- Create: `backend-single/asr-service/README.md`
+
+- [x] Add a FastAPI service with `/health` and `/transcribe`.
+- [x] Accept an uploaded audio file, save it under `/tmp/emotion-museum-asr`, run the ASR model, and return JSON with `success`, `text`, `language`, `durationMs`, `engine`, and `errorMessage`.
+- [x] Keep the service bound to `127.0.0.1:19120`.
+
+### Task 2: Java Backend Proxy
+
+**Files:**
+- Create: `backend-single/src/main/java/com/emotion/dto/response/asr/AsrTranscribeResponse.java`
+- Create: `backend-single/src/main/java/com/emotion/service/AsrService.java`
+- Create: `backend-single/src/main/java/com/emotion/service/impl/AsrServiceImpl.java`
+- Create: `backend-single/src/main/java/com/emotion/controller/AsrController.java`
+- Modify: `backend-single/src/main/resources/application.yml`
+- Modify: `backend-single/src/main/resources/application-prod.yml`
+
+- [x] Add `emotion.asr` config for `enabled`, `engine-url`, `max-file-size`, and `allowed-types`.
+- [x] Validate upload presence, size, and suffix.
+- [x] Forward multipart audio to the local ASR service.
+- [x] Return a normal `Result<AsrTranscribeResponse>` to the mini program.
+
+### Task 3: Mini Program Recording
+
+**Files:**
+- Create: `mini-program/src/services/asr.js`
+- Modify: `mini-program/src/pages/main/ScriptView.vue`
+
+- [x] Use `uni.getRecorderManager()` to start recording on press.
+- [x] Stop recording on release/cancel.
+- [x] Upload the recorded temp file to `/asr/transcribe`.
+- [x] Fill `wishText` with recognized text and keep the current visual theme.
+- [x] Track success/failure analytics.
+
+### Task 4: Deploy And Verify
+
+**Commands:**
+- `npm run build:mp-weixin`
+- `mvn -DskipTests package`
+- Deploy backend jar to `101.200.208.45`.
+- Install/start `emotion-museum-asr`.
+- Check `curl http://127.0.0.1:19120/health`.
+
+**Acceptance:**
+- Pressing and releasing the mini program voice orb sends audio to backend ASR.
+- Recognized Chinese text appears in the wish input.
+- Server ASR and TTS services both remain active.
@@ -0,0 +1,226 @@
+# AI 配置管理接口测试修复实施计划
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** 删除废弃的旧 AI 配置页面，在新系统测试对话框中增加非流式测试功能
+
+**Architecture:** 路由已正确指向新系统 `AiRoutingList.vue`，只需删除旧文件残留并增强测试对话框。后端接口 `testAiRuntime` 已存在，只需在 API 层和 UI 层连接。
+
+**Tech Stack:** Vue 3, Element Plus, TypeScript
+
+---
+
+### 前置确认：检查生产数据库配置
+
+- [ ] 在生产数据库上执行以下查询，确认 Dify provider 和 endpoint 状态
+
+```sql
+-- 检查 Dify provider
+SELECT id, provider_code, provider_name, provider_type, base_url, is_enabled 
+FROM t_ai_provider WHERE provider_code = 'dify_default';
+
+-- 检查 Dify endpoints
+SELECT id, endpoint_code, endpoint_name, provider_id, api_path, support_stream, is_enabled 
+FROM t_ai_endpoint_config WHERE endpoint_code LIKE 'dify.%';
+
+-- 检查场景绑定
+SELECT id, scene_code, scene_name, endpoint_id, is_enabled, required_stream 
+FROM t_ai_scene_binding WHERE is_enabled = 1;
+```
+
+- [ ] 如果 Dify provider `is_enabled = 0`，执行：
+```sql
+UPDATE t_ai_provider SET is_enabled = 1 WHERE provider_code = 'dify_default';
+```
+
+- [ ] 如果 Dify endpoints `is_enabled = 0`，执行：
+```sql
+UPDATE t_ai_endpoint_config SET is_enabled = 1 WHERE endpoint_code LIKE 'dify.%';
+```
+
+### Task 1: 删除废弃的旧 AI 配置页面
+
+**Files:**
+- Delete: `web-admin/src/views/aiconfig/AiConfigList.vue`
+
+- [ ] 删除旧页面文件
+
+```bash
+rm web-admin/src/views/aiconfig/AiConfigList.vue
+```
+
+### Task 2: 在 API 层导出非流式测试函数
+
+**Files:**
+- Modify: `web-admin/src/api/aiconfig.ts`
+
+- [ ] 确认 `testAiRuntime` 已存在（已确认在第 263 行）
+
+```typescript
+export function testAiRuntime(data: AiRuntimeRequest) {
+  return request({ url: '/ai/runtime/test', method: 'post', data })
+}
+```
+
+已存在，无需修改。但需要确认 `AiRoutingList.vue` 中是否 import 了这个函数。
+
+检查 `web-admin/src/views/aiconfig/AiRoutingList.vue` 的 import 语句（当前第 269-281 行）：
+- 当前 import 了 `streamAiRuntime`，**没有 import** `testAiRuntime`
+- 需要在 import 列表中添加 `testAiRuntime`
+
+### Task 3: 在测试对话框中增加非流式测试按钮
+
+**Files:**
+- Modify: `web-admin/src/views/aiconfig/AiRoutingList.vue`
+
+- [ ] 在 import 中添加 `testAiRuntime`
+
+修改第 269-281 行的 import 语句：
+
+```typescript
+import {
+  deleteAiEndpoint,
+  deleteAiProvider,
+  deleteAiScene,
+  listAiCallLogs,
+  listAiEndpoints,
+  listAiProviders,
+  listAiScenes,
+  saveAiEndpoint,
+  saveAiProvider,
+  saveAiScene,
+  streamAiRuntime,
+  testAiRuntime          // ← 新增
+} from '@/api/aiconfig'
+```
+
+- [ ] 在测试对话框 footer 中增加「非流式测试」按钮
+
+找到第 257-261 行的 `<template #footer>`，修改为：
+
+```vue
+<template #footer>
+  <el-button @click="testDialog = false">关闭</el-button>
+  <el-button :loading="testing" @click="submitNonStreamTest">非流式测试</el-button>
+  <el-button type="primary" :loading="testing" @click="submitRuntimeTest">流式测试</el-button>
+</template>
+```
+
+- [ ] 添加 `nonStreamResult` 响应式变量
+
+在 `testResult` 声明后面（第 296 行）添加：
+
+```typescript
+const nonStreamResult = ref<AiRuntimeTestResponse | null>(null)
+```
+
+- [ ] 实现 `submitNonStreamTest` 函数
+
+在 `submitRuntimeTest` 函数（第 451 行）之前添加：
+
+```typescript
+async function submitNonStreamTest() {
+  let inputs: Record<string, any>
+  try {
+    inputs = JSON.parse(testInputsJson.value || '{}')
+  } catch (error) {
+    ElMessage.error('入参 JSON 格式不正确')
+    return
+  }
+  testing.value = true
+  try {
+    const res = await testAiRuntime({ sceneCode: testForm.sceneCode, inputs })
+    nonStreamResult.value = res.data as AiRuntimeTestResponse
+    if (nonStreamResult.value.status === 'success') {
+      ElMessage.success('非流式测试成功')
+    } else {
+      ElMessage.error(`测试失败: ${nonStreamResult.value.errorMessage || nonStreamResult.value.errorCode}`)
+    }
+    await loadAll()
+  } catch (error: any) {
+    nonStreamResult.value = {
+      sceneCode: testForm.sceneCode,
+      status: 'failed',
+      errorMessage: error?.message || '非流式测试失败'
+    } as AiRuntimeTestResponse
+    ElMessage.error(error?.message || '非流式测试失败')
+  } finally {
+    testing.value = false
+  }
+}
+```
+
+- [ ] 在测试对话框中展示非流式测试结果
+
+在 `submitRuntimeTest` 相关的 `<el-alert>` 和 `<pre>` 展示块（第 250-256 行）之前，添加非流式测试结果的展示：
+
+```vue
+<el-alert
+  v-if="nonStreamResult"
+  :type="nonStreamResult.status === 'success' ? 'success' : 'error'"
+  :title="nonStreamResult.status === 'success' ? '非流式测试成功' : '非流式测试失败'"
+  show-icon
+  :closable="false"
+  style="margin-bottom: 12px;"
+/>
+<pre v-if="nonStreamResult" class="test-output">{{ nonStreamResult.output || nonStreamResult.errorMessage || '暂无输出' }}</pre>
+
+<el-alert
+  v-if="testResult"
+  :type="testResult.status === 'success' ? 'success' : 'error'"
+  :title="testResult.status === 'success' ? '流式测试成功' : '流式测试失败'"
+  show-icon
+/>
+<pre v-if="testResult" class="test-output">{{ testResult.output || testResult.errorMessage || '暂无输出' }}</pre>
+```
+
+- [ ] 在打开测试对话框时清空非流式结果
+
+在 `openRuntimeTest` 函数（第 400-404 行）中添加：
+
+```typescript
+function openRuntimeTest() {
+  testForm.sceneCode = scenes.value.find(item => item.isEnabled === 1 && item.endpointId)?.sceneCode || scenes.value[0]?.sceneCode || ''
+  testResult.value = null
+  nonStreamResult.value = null    // ← 新增
+  testDialog.value = true
+}
+```
+
+- [ ] 在 TypeScript 类型定义中确保 `AiRuntimeTestResponse` 包含所有必要字段
+
+检查 `web-admin/src/types/aiconfig.ts` 中的 `AiRuntimeTestResponse` 定义，确保包含：
+
+```typescript
+export interface AiRuntimeTestResponse {
+  sceneCode: string
+  status: 'success' | 'failed'
+  output?: string
+  streamChunks?: number
+  durationMs?: number
+  errorCode?: string
+  errorMessage?: string
+}
+```
+
+### Task 4: 浏览器验证
+
+- [ ] 启动 web-admin 开发服务器
+
+```bash
+cd web-admin && npm run dev
+```
+
+- [ ] 在浏览器中访问管理后台 `http://localhost:5174/emotion-museum-admin/`
+
+- [ ] 导航到「AI 配置管理」页面，确认旧页面已删除、新页面正常显示
+
+- [ ] 在「场景绑定」Tab 中点击「流式测试」按钮，确认测试对话框正常打开
+
+- [ ] 确认测试对话框中有两个按钮：「非流式测试」和「流式测试」
+
+- [ ] 选择一个场景（如 `script_generate`），输入入参 `{ "prompt": "请用一句中文回复测试成功。" }`，点击「非流式测试」，确认返回成功结果
+
+- [ ] 点击「流式测试」，确认流式输出正常
+
+- [ ] 确认浏览器 Console 中没有任何错误
@@ -0,0 +1,58 @@
+# AI Routing Admin Runtime Fix Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Make the AI configuration center match the existing admin theme, use Chinese copy everywhere, show existing provider/workflow bindings from the database, and expose a single user-facing streaming runtime entry point driven by scene code plus JSON inputs.
+
+**Architecture:** Keep the existing provider/endpoint/scene/log model. The admin page uses the current Element Plus dark glass theme and existing list-page conventions. Runtime calls accept `sceneCode` and `inputs` JSON, enrich inputs with the logged-in user context on the backend, resolve the scene binding from the database, then stream provider output back to the user.
+
+**Tech Stack:** Spring Boot, MyBatis-Plus, FastJSON, Vue 3, Element Plus, uni-app chunked streaming.
+
+---
+
+### Task 1: Admin Page Theme And Chinese Copy
+
+**Files:**
+- Modify: `web-admin/src/views/aiconfig/AiRoutingList.vue`
+
+- [x] Replace English page title, tabs, table columns, dialogs, buttons, alerts, empty states, and confirm text with Chinese.
+- [x] Replace the standalone white panel styles with the project page-header, glass card, table card, and dark Element Plus table styles.
+- [x] Add summary cards for providers, endpoints, scenes, and logs so the page matches existing management pages.
+
+### Task 2: User Runtime Request Shape
+
+**Files:**
+- Modify: `backend-single/src/main/java/com/emotion/dto/request/ai/AiRuntimeRequest.java`
+- Modify: `backend-single/src/main/java/com/emotion/controller/AiRoutingController.java`
+- Modify: `backend-single/src/main/java/com/emotion/service/impl/AiRuntimeServiceImpl.java`
+
+- [x] Accept user calls as `sceneCode` plus `inputs` JSON.
+- [x] Keep user identity server-side by reading `UserContextHolder`.
+- [x] Enrich runtime inputs with safe user context keys before provider invocation.
+- [x] Keep SSE stream events as the only formal output path for user-facing AI calls.
+
+### Task 3: Seed Existing Configurations Into Routing Tables
+
+**Files:**
+- Modify: `sql/2026-05-22-ai-scene-routing.sql`
+
+- [x] Insert a Coze provider from existing `t_ai_config` rows when missing.
+- [x] Insert endpoint rows from current enabled Coze workflow configurations.
+- [x] Bind existing scenes to known workflow configs where a safe mapping exists.
+- [x] Keep unknown scenes visible but disabled until explicitly bound in the admin page.
+
+### Task 4: Verify End To End
+
+**Commands:**
+- `mvn test`
+- `mvn -DskipTests clean package`
+- `npm run build`
+- `npm run build:mp-weixin`
+- Remote SQL migration execution.
+- Remote health checks for backend and ASR.
+
+- [x] Backend tests pass.
+- [x] Backend package succeeds.
+- [x] Admin build succeeds.
+- [x] Mini program build succeeds.
+- [x] Remote routing tables contain provider, endpoints, and scene bindings.
@@ -0,0 +1,66 @@
+# AI Runtime Client Sync Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Synchronize AI calls in web, web-admin, life-script, and mini-program so user-facing AI output uses the backend scene routing runtime and streams text progressively.
+
+**Architecture:** The backend remains the single trust boundary for provider credentials, scene bindings, user context enrichment, and workflow selection. Browser and mini-program clients call `/ai/runtime/stream` with `sceneCode` and `inputs`, while existing WebSocket chat is bridged through `AiRuntimeService` so chat also follows the configured `chat` scene. Existing business save APIs remain responsible for persistence after a stream completes.
+
+**Tech Stack:** Spring Boot, STOMP WebSocket, FastJSON, Vue 3, React/Vite, uni-app chunked request, Element Plus.
+
+---
+
+### Task 1: Backend Chat Bridge To Runtime
+
+**Files:**
+- Modify: `backend-single/src/main/java/com/emotion/service/impl/WebSocketServiceImpl.java`
+- Modify: `backend-single/src/main/java/com/emotion/dto/websocket/WebSocketMessage.java`
+
+- [x] Route chat generation through `AiRuntimeService` with `sceneCode=chat`.
+- [x] Push `start`, `delta`, `done`, and `error` events over the existing user WebSocket topic.
+- [x] Save the final assistant message after `done` using the accumulated stream output.
+- [x] Preserve the existing user-message persistence and conversation ID behavior.
+
+### Task 2: Shared Runtime Clients
+
+**Files:**
+- Create: `web/src/services/aiRuntime.ts`
+- Create: `life-script/src/services/aiRuntime.js`
+- Modify: `mini-program/src/services/aiRuntime.js`
+- Modify: `web-admin/src/api/aiconfig.ts`
+- Modify: `web-admin/src/views/aiconfig/AiRoutingList.vue`
+
+- [x] Add browser SSE client helpers that parse unified AI stream events.
+- [x] Keep mini-program chunk parsing but localize error messages and flush trailing frames.
+- [x] Add a web-admin stream test helper and make the test dialog render output progressively.
+
+### Task 3: Client Scene Migration
+
+**Files:**
+- Modify: `life-script/src/services/ai.js`
+- Modify: `life-script/src/views/ScriptView.jsx`
+- Modify: `life-script/src/views/TimelineView.jsx`
+- Modify: `life-script/src/views/PathView.jsx`
+- Modify: `mini-program/src/pages/main/ScriptView.vue`
+- Inspect: `web/src/stores/chat.ts`
+
+- [x] Remove direct OpenRouter calls and hard-coded client API keys.
+- [x] Route `script_generate`, `life_healing`, and path generation through `streamAiScene`.
+- [x] Ensure pages append `delta` content instead of waiting for complete text.
+- [x] Keep final persistence through existing business save APIs after stream completion.
+
+### Task 4: Verification
+
+**Commands:**
+- `mvn test`
+- `mvn -DskipTests clean package`
+- `npm run build` in `web`
+- `npm run build` in `web-admin`
+- `npm run build` in `life-script`
+- `npm run build:mp-weixin` in `mini-program`
+
+- [x] Backend tests pass.
+- [x] Backend package succeeds.
+- [x] All four frontend builds succeed.
+- [x] Source scan shows no client-side external AI provider key or direct Dify/Coze/OpenRouter call.
+- [x] Database scene bindings are enabled with streaming endpoints for `chat`, `script_generate`, `short_story_generate`, `diary_summary`, `emotion_summary`, `emotion_analysis`, and `life_healing`.