feat: 功能

frontend/app/web-gold/src/api/kling.js

@@ -2,18 +2,66 @@
  * 可灵数字人 API
  */
 import request from './http'
-import { message } from "ant-design-vue"
 import { MaterialService } from './material'
 
-/**
+// ========== 辅助函数 ==========
- * 显示加载提示
- */
-const showLoading = (text) => message.loading(text, 0)
 
 /**
- * 销毁加载提示
+ * 从视频中提取封面（可选操作）
  */
-const hideLoading = () => message.destroy()
+async function extractVideoCoverOptional(file) {
+  try {
+    const { extractVideoCover } = await import('@/utils/video-cover')
+    const cover = await extractVideoCover(file, {
+      maxWidth: 800,
+      quality: 0.8
+    })
+    return cover.base64
+  } catch {
+    return null
+  }
+}
+
+/**
+ * 执行人脸识别并返回结果
+ */
+async function performFaceIdentification(videoUrl) {
+  const identifyRes = await identifyFace({ video_url: videoUrl })
+  if (identifyRes.code !== 0) {
+    throw new Error(identifyRes.msg || '识别失败')
+  }
+
+  const faceData = identifyRes.data.data?.face_data?.[0]
+  return {
+    sessionId: identifyRes.data.sessionId,
+    faceId: faceData?.face_id || null,
+    startTime: faceData?.start_time || 0,
+    endTime: faceData?.end_time || 0
+  }
+}
+
+/**
+ * 构建标准响应格式
+ */
+function buildIdentifyResponse(fileId, videoUrl, identifyData, isUploadedFile = false) {
+  return {
+    success: true,
+    data: {
+      fileId,
+      videoUrl,
+      sessionId: identifyData.sessionId,
+      faceId: identifyData.faceId,
+      startTime: isUploadedFile
+        ? Math.round(identifyData.startTime * 1000)
+        : identifyData.startTime,
+      endTime: isUploadedFile
+        ? Math.round(identifyData.endTime * 1000)
+        : identifyData.endTime
+    }
+  }
+}
+
+// ========== API 方法 ==========
 
 export function identifyFace(data) {
   return request({
@@ -38,93 +86,46 @@ export function getLipSyncTask(taskId) {
   })
 }
 
+/**
+ * 识别已上传的视频
+ */
 export async function identifyUploadedVideo(videoFile) {
   try {
-    showLoading('正在识别视频中的人脸...')
+    const urlRes = await MaterialService.getVideoPlayUrl(videoFile.fileId)
-    const identifyRes = await identifyFace({ video_url: videoFile.fileUrl })
+    if (urlRes.code !== 0 || !urlRes.data) {
-    hideLoading()
+      throw new Error(urlRes.msg || '获取播放链接失败')
-
-    if (identifyRes.code !== 0) {
-      throw new Error(identifyRes.msg || '识别失败')
     }
 
-    return {
+    const identifyData = await performFaceIdentification(urlRes.data)
-      success: true,
+    return buildIdentifyResponse(videoFile.id, urlRes.data, identifyData, false)
-      data: {
-        fileId: videoFile.id,
-        videoUrl: videoFile.fileUrl,
-        sessionId: identifyRes.data.sessionId,
-        faceId: identifyRes.data.data.face_data[0].face_id || null,
-        startTime: identifyRes.data.data.face_data[0].start_time || 0,
-        endTime: identifyRes.data.data.face_data[0].end_time || 0
-      }
-    }
   } catch (error) {
-    hideLoading()
     throw error
   }
 }
 
+/**
+ * 上传视频并识别
+ */
 export async function uploadAndIdentifyVideo(file) {
+  const coverBase64 = await extractVideoCoverOptional(file)
+
   try {
-    showLoading('正在提取视频封面...')
-    let coverBase64 = null
-    try {
-      const { extractVideoCover } = await import('@/utils/video-cover')
-      const cover = await extractVideoCover(file, {
-        maxWidth: 800,
-        quality: 0.8
-      })
-      coverBase64 = cover.base64
-    } catch (coverError) {
-      // 封面提取失败不影响主流程
-    }
-    hideLoading()
-
-    showLoading('正在上传视频...')
-
-    // 使用useUpload Hook（注意：这里需要在组件中使用，这里先用MaterialService）
-    // TODO: 在组件中集成useUpload Hook
     const uploadRes = await MaterialService.uploadFile(file, 'video', coverBase64, null, null)
-    hideLoading()
-
     if (uploadRes.code !== 0) {
       throw new Error(uploadRes.msg || '上传失败')
     }
 
     const fileId = uploadRes.data
 
-    showLoading('正在生成播放链接...')
     const urlRes = await MaterialService.getVideoPlayUrl(fileId)
-    hideLoading()
-
     if (urlRes.code !== 0) {
       throw new Error(urlRes.msg || '获取播放链接失败')
     }
 
-    const videoUrl = urlRes.data
+    const identifyData = await performFaceIdentification(urlRes.data)
-
+    return buildIdentifyResponse(fileId, urlRes.data, identifyData, true)
-    showLoading('正在识别视频中的人脸...')
-    const identifyRes = await identifyFace({ video_url: videoUrl })
-    hideLoading()
-
-    if (identifyRes.code !== 0) {
-      throw new Error(identifyRes.msg || '识别失败')
-    }
-
-    return {
-      success: true,
-      data: {
-        fileId,
-        videoUrl,
-        sessionId: identifyRes.data.sessionId,
-        faceId: identifyRes.data.data.face_data[0].face_id || null,
-        startTime: identifyRes.data.data.face_data[0].start_time || 0,
-        endTime: identifyRes.data.data.face_data[0].end_time || 0
-      }
-    }
   } catch (error) {
-    hideLoading()
     throw error
   }
 }
+

frontend/app/web-gold/src/components/PipelineProgress.vue

@@ -0,0 +1,112 @@
+<template>
+  <div class="pipeline-progress">
+    <!-- 状态和进度 -->
+    <div class="progress-header">
+      <span class="state-text">{{ stateLabel }}</span>
+      <span v-if="stateDescription" class="state-desc">{{ stateDescription }}</span>
+    </div>
+
+    <a-progress
+      :percent="displayProgress"
+      :status="progressStatus"
+      :stroke-color="progressColor"
+    />
+
+    <!-- 错误时显示 -->
+    <div v-if="isFailed && error" class="error-section">
+      <span class="error-text">{{ error }}</span>
+      <a-button size="small" @click="handleRetry">重试</a-button>
+    </div>
+  </div>
+</template>
+
+<script setup lang="ts">
+import { computed } from 'vue'
+import type { PipelineState } from '@/views/kling/hooks/pipeline/types'
+import { STATE_CONFIG } from '@/views/kling/hooks/pipeline/states'
+
+interface Props {
+  state: PipelineState | string
+  progress: number
+  isBusy: boolean
+  isReady: boolean
+  isFailed: boolean
+  isCompleted: boolean
+  error: string | null
+}
+
+const props = defineProps<Props>()
+
+const emit = defineEmits<{
+  retry: []
+  reset: []
+}>()
+
+const stateConfig = computed(() => STATE_CONFIG[props.state as PipelineState])
+const stateLabel = computed(() => stateConfig.value.label)
+const stateDescription = computed(() => stateConfig.value.description)
+
+const progressStatus = computed(() => {
+  if (props.isFailed) return 'exception'
+  if (props.isCompleted) return 'success'
+  return 'active'
+})
+
+const progressColor = computed(() => {
+  if (props.isFailed) return '#ff4d4f'
+  if (props.isCompleted) return '#52c41a'
+  return '#1890ff'
+})
+
+const displayProgress = computed(() => {
+  if (props.isFailed) return 0
+  return props.progress
+})
+
+function handleRetry() {
+  emit('retry')
+}
+</script>
+
+<style scoped lang="less">
+.pipeline-progress {
+  padding: 12px 16px;
+  background: var(--bg-primary);
+  border-radius: 8px;
+  border: 1px solid var(--border-light);
+  margin-bottom: 16px;
+}
+
+.progress-header {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  margin-bottom: 8px;
+  font-size: 13px;
+
+  .state-text {
+    font-weight: 600;
+    color: var(--text-primary);
+  }
+
+  .state-desc {
+    color: var(--text-secondary);
+  }
+}
+
+.error-section {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  margin-top: 12px;
+  padding: 8px 12px;
+  background: rgba(255, 77, 79, 0.1);
+  border-radius: 6px;
+
+  .error-text {
+    flex: 1;
+    color: #ff4d4f;
+    font-size: 13px;
+  }
+}
+</style>

frontend/app/web-gold/src/composables/useTTS.js

@@ -7,72 +7,97 @@ import { message } from 'ant-design-vue'
 import { VoiceService } from '@/api/voice'
 import { normalizeProviderType, VOICE_PROVIDER_TYPES } from '@/config/voiceConfig'
 
-// 兼容旧代码的导出
+// ========== 常量 ==========
+
+/** 兼容旧代码的导出 */
 const TTS_PROVIDERS = VOICE_PROVIDER_TYPES
 
-const DEFAULT_CONFIG =  {
+/** 默认配置 */
-    apiEndpoint: '/api/tik/voice/tts',
+const DEFAULT_CONFIG = {
-    audioFormat: 'mp3',
+  apiEndpoint: '/api/tik/voice/tts',
-    supportedFormats: ['mp3', 'wav']
+  audioFormat: 'mp3',
+  supportedFormats: ['mp3', 'wav']
 }
 
+/** 最大预览缓存数量 */
+const MAX_PREVIEW_CACHE_SIZE = 50
+
+// ========== 类型定义 ==========
+
+/**
+ * @typedef {Object} AudioData
+ * @property {Blob} blob - 音频 Blob
+ * @property {string} objectUrl - 对象 URL
+ * @property {string} format - 音频格式
+ */
+
 export function useTTS(options = {}) {
   const {
     provider = VOICE_PROVIDER_TYPES.SILICONFLOW,
     customConfig = {}
   } = options
 
-  // 状态管理
+  // 状态管理（移到函数内部，避免模块级状态污染）
   const previewAudioCache = new Map()
-  const MAX_PREVIEW_CACHE_SIZE = 50
   const previewLoadingVoiceId = ref(null)
   const playingPreviewVoiceId = ref(null)
   const ttsText = ref('')
   const speechRate = ref(1.0)
 
-  // 音频实例
+  // 音频实例（移到函数内部）
   let previewAudio = null
   let previewObjectUrl = ''
 
-  // 获取当前供应商配置
+  // ========== 辅助函数 ==========
-  const getProviderConfig = () => {
+
+  function getProviderConfig() {
     return DEFAULT_CONFIG
   }
 
   /**
    * 播放音频预览
-   * @param {string} url 音频URL
+   * @param {string} url - 音频 URL
-   * @param {Object} options 播放选项
+   * @param {Object} playOptions - 播放选项
+   * @param {boolean} [playOptions.revokeOnEnd=false] - 播放结束后是否释放 URL
+   * @param {Function} [playOptions.onEnded] - 播放结束回调
    */
-  const playAudioPreview = (url, options = {}) => {
+  function playAudioPreview(url, playOptions = {}) {
-    if (!url) return message.warning('暂无可试听的音频')
+    if (!url) {
+      message.warning('暂无可试听的音频')
+      return
+    }
 
+    // 停止当前播放
     try {
       previewAudio?.pause?.()
       previewAudio = null
-    } catch (error) {
+    } catch {
+      // 忽略停止播放的错误
     }
 
     const audio = new Audio(url)
-    const cleanup = () => {
+
-      if (options.revokeOnEnd && url.startsWith('blob:')) {
+    function cleanup() {
+      if (playOptions.revokeOnEnd && url.startsWith('blob:')) {
         URL.revokeObjectURL(url)
-        previewObjectUrl === url && (previewObjectUrl = '')
+        if (previewObjectUrl === url) {
+          previewObjectUrl = ''
+        }
       }
       previewAudio = null
-      options.onEnded && options.onEnded()
+      playOptions.onEnded?.()
     }
 
     audio.play()
-      .then(() => {
+      .then(function() {
         previewAudio = audio
         audio.onended = cleanup
-        audio.onerror = () => {
+        audio.onerror = function() {
           cleanup()
           message.error('播放失败')
         }
       })
-      .catch(err => {
+      .catch(function() {
         cleanup()
         message.error('播放失败')
       })
@@ -80,10 +105,10 @@ export function useTTS(options = {}) {
 
   /**
    * 生成预览缓存键
-   * @param {Object} voice 音色对象
+   * @param {Object} voice - 音色对象
    * @returns {string} 缓存键
    */
-  const generatePreviewCacheKey = (voice) => {
+  function generatePreviewCacheKey(voice) {
     const voiceId = voice.voiceId || voice.rawId || voice.id
     const text = ttsText.value.trim()
     const rate = speechRate.value
@@ -92,12 +117,12 @@ export function useTTS(options = {}) {
 
   /**
    * 解码并缓存Base64音频
-   * @param {string} audioBase64 Base64编码的音频数据
+   * @param {string} audioBase64 - Base64 编码的音频数据
-   * @param {string} format 音频格式
+   * @param {string} [format='mp3'] - 音频格式
-   * @param {string} cacheKey 缓存键
+   * @param {string} cacheKey - 缓存键
    * @returns {Promise<Object>} 音频数据
    */
-  const decodeAndCacheBase64 = async (audioBase64, format = 'mp3', cacheKey) => {
+  async function decodeAndCacheBase64(audioBase64, format = 'mp3', cacheKey) {
     const byteCharacters = window.atob(audioBase64)
     const byteNumbers = new Uint8Array(byteCharacters.length)
     for (let i = 0; i < byteCharacters.length; i++) {
@@ -123,10 +148,10 @@ export function useTTS(options = {}) {
 
   /**
    * 播放缓存的音频
-   * @param {Object} audioData 音频数据
+   * @param {Object} audioData - 音频数据
-   * @param {Function} onEnded 播放结束回调
+   * @param {Function} onEnded - 播放结束回调
    */
-  const playCachedAudio = (audioData, onEnded) => {
+  function playCachedAudio(audioData, onEnded) {
     if (previewObjectUrl && previewObjectUrl !== audioData.objectUrl) {
       URL.revokeObjectURL(previewObjectUrl)
     }
@@ -134,7 +159,10 @@ export function useTTS(options = {}) {
 
     playAudioPreview(previewObjectUrl, {
       revokeOnEnd: false,
-      onEnded: () => {
+      onEnded: function() {
+        if (audioData.objectUrl?.startsWith('blob:')) {
+          URL.revokeObjectURL(audioData.objectUrl)
+        }
         onEnded && onEnded()
       }
     })
@@ -143,17 +171,17 @@ export function useTTS(options = {}) {
   /**
    * 重置预览状态
    */
-  const resetPreviewState = () => {
+  function resetPreviewState() {
     previewLoadingVoiceId.value = null
     playingPreviewVoiceId.value = null
   }
 
   /**
    * 提取ID从字符串
-   * @param {string} idStr 包含前缀的ID字符串
+   * @param {string} idStr - 包含前缀的ID字符串
    * @returns {number|null} 提取的ID
    */
-  const extractIdFromString = (idStr) => {
+  function extractIdFromString(idStr) {
     if (typeof idStr !== 'string' || !idStr.startsWith('user-')) return null
     const extractedId = parseInt(idStr.replace('user-', ''))
     return Number.isNaN(extractedId) ? null : extractedId
@@ -161,10 +189,10 @@ export function useTTS(options = {}) {
 
   /**
    * 构建预览参数
-   * @param {Object} voice 音色对象
+   * @param {Object} voice - 音色对象
    * @returns {Object|null} 预览参数
    */
-  const buildPreviewParams = (voice) => {
+  function buildPreviewParams(voice) {
     const configId = voice.rawId || extractIdFromString(voice.id)
     if (!configId) {
       message.error('配音配置无效')
@@ -184,11 +212,11 @@ export function useTTS(options = {}) {
 
   /**
    * 播放音色试听
-   * @param {Object} voice 音色对象
+   * @param {Object} voice - 音色对象
-   * @param {Function} onSuccess 成功回调
+   * @param {Function} onSuccess - 成功回调
-   * @param {Function} onError 错误回调
+   * @param {Function} onError - 错误回调
    */
-  const playVoiceSample = async (voice, onSuccess, onError) => {
+  async function playVoiceSample(voice, onSuccess, onError) {
     if (!voice) return
     if (previewLoadingVoiceId.value === voice.id || playingPreviewVoiceId.value === voice.id) {
       return
@@ -197,7 +225,8 @@ export function useTTS(options = {}) {
       try {
         previewAudio?.pause?.()
         previewAudio = null
-      } catch (error) {
+      } catch {
+        // 忽略错误
       }
     }
 
@@ -225,21 +254,30 @@ export function useTTS(options = {}) {
       if (res.code !== 0) {
         message.error(res.msg || '试听失败')
         resetPreviewState()
-        onError && onError(new Error(res.msg || '试听失败'))
+        onError?.(new Error(res.msg || '试听失败'))
         return
       }
 
       if (res.data?.audioUrl) {
-        playAudioPreview(res.data.audioUrl, { onEnded: resetPreviewState })
+        resetPreviewState()
-        onSuccess && onSuccess(res.data)
+        playAudioPreview(res.data.audioUrl, {
+          revokeOnEnd: true,
+          onEnded: function() {
+            URL.revokeObjectURL(res.data.audioUrl)
+          }
+        })
+        onSuccess?.(res.data)
       } else if (res.data?.audioBase64) {
         const audioData = await decodeAndCacheBase64(res.data.audioBase64, res.data.format, cacheKey)
-        playCachedAudio(audioData, resetPreviewState)
+        resetPreviewState()
-        onSuccess && onSuccess(audioData)
+        playCachedAudio(audioData, function() {
+          URL.revokeObjectURL(audioData.objectUrl)
+        })
+        onSuccess?.(audioData)
       } else {
         message.error('试听失败')
         resetPreviewState()
-        onError && onError(new Error('未收到音频数据'))
+        onError?.(new Error('未收到音频数据'))
       }
     } catch (error) {
       message.error('试听失败')
@@ -250,10 +288,10 @@ export function useTTS(options = {}) {
 
   /**
    * TTS文本转语音
-   * @param {Object} params TTS参数
+   * @param {Object} params - TTS 参数
-   * @returns {Promise<Object>} TTS结果
+   * @returns {Promise<Object>} TTS 结果
    */
-  const synthesize = async (params) => {
+  async function synthesize(params) {
     const providerConfig = getProviderConfig()
 
     const ttsParams = {
@@ -269,25 +307,25 @@ export function useTTS(options = {}) {
 
   /**
    * 设置文本
-   * @param {string} text 要设置的文本
+   * @param {string} text - 要设置的文本
    */
-  const setText = (text) => {
+  function setText(text) {
     ttsText.value = text
   }
 
   /**
    * 设置语速
-   * @param {number} rate 语速倍率
+   * @param {number} rate - 语速倍率
    */
-  const setSpeechRate = (rate) => {
+  function setSpeechRate(rate) {
     speechRate.value = rate
   }
 
   /**
    * 清除音频缓存
    */
-  const clearAudioCache = () => {
+  function clearAudioCache() {
-    previewAudioCache.forEach((audioData) => {
+    previewAudioCache.forEach(function(audioData) {
       URL.revokeObjectURL(audioData.objectUrl)
     })
     previewAudioCache.clear()
@@ -296,14 +334,17 @@ export function useTTS(options = {}) {
   /**
    * 停止当前播放
    */
-  const stopCurrentPlayback = () => {
+  function stopCurrentPlayback() {
     try {
       previewAudio?.pause?.()
       previewAudio = null
-    } catch (error) {
+    } catch {
+      // 忽略错误
     }
   }
 
+  // ========== 返回接口 ==========
+
   return {
     // 状态
     previewLoadingVoiceId,

frontend/app/web-gold/src/views/dh/VoiceCopy.vue

@@ -105,15 +105,15 @@
 </template>
 
 <script setup>
-import { ref, reactive, computed, onMounted, nextTick } from 'vue'
+import { ref, reactive, computed, onMounted } from 'vue'
 import { message, Modal } from 'ant-design-vue'
 import { PlusOutlined, SearchOutlined, UploadOutlined, PlayCircleOutlined } from '@ant-design/icons-vue'
-import { VoiceService } from '@/api/voice'
-import { MaterialService } from '@/api/material'
-import { useUpload } from '@/composables/useUpload'
-import useVoiceText from '@gold/hooks/web/useVoiceText'
 import dayjs from 'dayjs'
 import BasicLayout from '@/layouts/components/BasicLayout.vue'
+import { MaterialService } from '@/api/material'
+import { VoiceService } from '@/api/voice'
+import { useUpload } from '@/composables/useUpload'
+import useVoiceText from '@gold/hooks/web/useVoiceText'
 
 // ========== 常量 ==========
 
@@ -125,10 +125,14 @@ const DEFAULT_FORM_DATA = {
   language: 'zh-CN',
   gender: 'female',
   note: '',
-  text: '', // 音频文本
+  text: '',
-  fileUrl: '' // 文件URL（用于获取音频文本）
+  fileUrl: ''
 }
 
+const MAX_FILE_SIZE = 50 * 1024 * 1024
+const VALID_AUDIO_TYPES = ['audio/mpeg', 'audio/wav', 'audio/wave', 'audio/x-wav', 'audio/aac', 'audio/mp4', 'audio/flac', 'audio/ogg']
+const VALID_AUDIO_EXTENSIONS = ['.mp3', '.wav', '.aac', '.m4a', '.flac', '.ogg']
+
 // ========== 响应式数据 ==========
 const loading = ref(false)
 const submitting = ref(false)
@@ -216,31 +220,31 @@ const loadVoiceList = async () => {
 }
 
 // ========== 搜索和分页 ==========
-const handleSearch = () => {
+function handleSearch() {
   pagination.current = 1
   loadVoiceList()
 }
 
-const handleReset = () => {
+function handleReset() {
   searchParams.name = ''
   pagination.current = 1
   loadVoiceList()
 }
 
-const handleTableChange = (pag) => {
+function handleTableChange(pag) {
   pagination.current = pag.current
   pagination.pageSize = pag.pageSize
   loadVoiceList()
 }
 
 // ========== CRUD 操作 ==========
-const handleCreate = () => {
+function handleCreate() {
   formMode.value = 'create'
   resetForm()
   modalVisible.value = true
 }
 
-const handleEdit = async (record) => {
+async function handleEdit(record) {
   formMode.value = 'edit'
   try {
     const res = await VoiceService.get(record.id)
@@ -252,13 +256,13 @@ const handleEdit = async (record) => {
   modalVisible.value = true
 }
 
-const handleDelete = (record) => {
+function handleDelete(record) {
   Modal.confirm({
     title: '确认删除',
     content: `确定要删除配音「${record.name}」吗？此操作不可恢复。`,
     okButtonProps: { danger: true },
     centered: true,
-    onOk: async () => {
+    onOk: async function() {
       try {
         const res = await VoiceService.delete(record.id)
         if (res.code !== 0) return message.error(res.msg || '删除失败')
@@ -274,7 +278,7 @@ const handleDelete = (record) => {
 }
 
 // ========== 音频播放 ==========
-const handlePlayAudio = (record) => {
+function handlePlayAudio(record) {
   if (record.fileUrl && audioPlayer.value) {
     audioPlayer.value.src = record.fileUrl
     audioPlayer.value.play()
@@ -284,20 +288,20 @@ const handlePlayAudio = (record) => {
 }
 
 // ========== 文件上传 ==========
-const handleBeforeUpload = (file) => {
+function handleBeforeUpload(file) {
-  const MAX_FILE_SIZE = 50 * 1024 * 1024
   if (file.size > MAX_FILE_SIZE) {
     message.error('文件大小不能超过 50MB')
     return false
   }
 
-  const validTypes = ['audio/mpeg', 'audio/wav', 'audio/wave', 'audio/x-wav', 'audio/aac', 'audio/mp4', 'audio/flac', 'audio/ogg']
-  const validExtensions = ['.mp3', '.wav', '.aac', '.m4a', '.flac', '.ogg']
   const fileName = file.name.toLowerCase()
   const fileType = file.type.toLowerCase()
 
-  const isValidType = validTypes.some(type => fileType.includes(type)) ||
+  const isValidType = VALID_AUDIO_TYPES.some(function(type) {
-                      validExtensions.some(ext => fileName.endsWith(ext))
+    return fileType.includes(type)
+  }) || VALID_AUDIO_EXTENSIONS.some(function(ext) {
+    return fileName.endsWith(ext)
+  })
 
   if (!isValidType) {
     message.error('请上传音频文件（MP3、WAV、AAC、M4A、FLAC、OGG）')
@@ -307,7 +311,8 @@ const handleBeforeUpload = (file) => {
   return true
 }
 
-const handleCustomUpload = async (options) => {
+// ========== 文件上传相关 ==========
+async function handleCustomUpload(options) {
   const { file, onSuccess, onError } = options
 
   try {
@@ -315,17 +320,14 @@ const handleCustomUpload = async (options) => {
       fileCategory: 'voice',
       groupId: null,
       coverBase64: null,
-      onStart: () => {},
+      onSuccess: async function(id, fileUrl) {
-      onProgress: () => {},
-      onSuccess: async (id, fileUrl) => {
         formData.fileId = id
-        formData.fileUrl = fileUrl // 保存文件URL
+        formData.fileUrl = fileUrl
         message.success('文件上传成功')
-        // 通过fileId获取播放URL用于语音识别
         await fetchAudioTextById(id)
         onSuccess?.({ code: 0, data: id }, file)
       },
-      onError: (error) => {
+      onError: function(error) {
         const errorMsg = error.message || '上传失败，请稍后重试'
         message.error(errorMsg)
         onError?.(error)
@@ -339,13 +341,10 @@ const handleCustomUpload = async (options) => {
   }
 }
 
-
-
 // 通过fileId获取音频文本
-const fetchAudioTextById = async (fileId) => {
+async function fetchAudioTextById(fileId) {
   if (!fileId) return
   try {
-    // 获取音频播放URL
     const res = await MaterialService.getAudioPlayUrl(fileId)
     if (res.code === 0 && res.data) {
       const rawFileUrl = res.data
@@ -363,39 +362,22 @@ const fetchAudioTextById = async (fileId) => {
   }
 }
 
-// 获取音频文本
+function handleFileListChange(info) {
-const fetchAudioText = async (fileUrl) => {
-  if (!fileUrl) return
-  try {
-    // 阿里云语音识别服务无法访问预签名URL，使用原始URL
-    const rawFileUrl = extractRawUrl(fileUrl)
-    const results = await getVoiceText([{ audio_url: rawFileUrl }])
-    if (results && results.length > 0) {
-      const text = results[0].value
-      formData.text = text
-      if (text) {
-        message.success('音频文本获取成功')
-      }
-    }
-  } catch (error) {
-    console.error('获取音频文本失败:', error)
-  }
-}
-
-const handleFileListChange = (info) => {
   const { fileList: newFileList } = info
   if (newFileList) {
-    fileList.value = newFileList.filter(item => item.status !== 'removed')
+    fileList.value = newFileList.filter(function(item) {
+      return item.status !== 'removed'
+    })
   }
 }
 
-const handleRemoveFile = () => {
+function handleRemoveFile() {
   formData.fileId = null
   fileList.value = []
 }
 
 // ========== 表单操作 ==========
-const handleSubmit = async () => {
+async function handleSubmit() {
   try {
     await formRef.value.validate()
   } catch {
@@ -412,7 +394,7 @@ const handleSubmit = async () => {
         language: formData.language,
         gender: formData.gender,
         note: formData.note,
-        text: formData.text // 传入音频文本
+        text: formData.text
       }
     : {
         id: formData.id,
@@ -443,19 +425,19 @@ const handleSubmit = async () => {
   }
 }
 
-const handleCancel = () => {
+function handleCancel() {
   modalVisible.value = false
   resetForm()
 }
 
-const resetForm = () => {
+function resetForm() {
   Object.assign(formData, { ...DEFAULT_FORM_DATA })
   fileList.value = []
   formRef.value?.resetFields()
 }
 
 // ========== 生命周期 ==========
-onMounted(() => {
+onMounted(function() {
   loadVoiceList()
 })
 </script>

frontend/app/web-gold/src/views/kling/IdentifyFace.vue

@@ -15,10 +15,6 @@
               :show-count="true"
               class="tts-textarea"
             />
-            <div v-if="identifyState.identified && faceDuration > 0" class="text-hint">
-              <span class="hint-icon">💡</span>
-              <span>视频中人脸出现时长约 {{ (faceDuration / 1000).toFixed(1) }} 秒，建议文案不超过 {{ suggestedMaxChars }} 字</span>
-            </div>
           </div>
 
           <!-- 音色选择 -->
@@ -141,58 +137,6 @@
             </div>
           </div>
 
-          <!-- 素材校验结果 -->
-          <div v-if="materialValidation.videoDuration > 0 && materialValidation.audioDuration > 0" class="section">
-            <h3>素材校验</h3>
-            <div class="validation-result" :class="{ 'validation-passed': materialValidation.isValid, 'validation-failed': !materialValidation.isValid }">
-              <div class="validation-status">
-                <span class="status-icon">{{ materialValidation.isValid ? '✅' : '❌' }}</span>
-                <span class="status-text">{{ materialValidation.isValid ? '校验通过' : '校验失败' }}</span>
-              </div>
-
-              <!-- 时长对比进度条 -->
-              <div class="duration-comparison">
-                <div class="duration-bar">
-                  <div class="duration-label">
-                    <span>音频时长</span>
-                    <span class="duration-value">{{ (materialValidation.audioDuration / 1000).toFixed(1) }}s</span>
-                  </div>
-                  <div class="progress-bar audio-bar">
-                    <div
-                      class="progress-fill"
-                      :style="{ width: `${(materialValidation.audioDuration / Math.max(materialValidation.videoDuration, materialValidation.audioDuration)) * 100}%` }"
-                    ></div>
-                  </div>
-                </div>
-
-                <div class="duration-bar">
-                  <div class="duration-label">
-                    <span>视频时长</span>
-                    <span class="duration-value">{{ (materialValidation.videoDuration / 1000).toFixed(1) }}s</span>
-                  </div>
-                  <div class="progress-bar video-bar">
-                    <div
-                      class="progress-fill"
-                      :class="{ 'success': materialValidation.isValid, 'error': !materialValidation.isValid }"
-                      :style="{ width: `${(materialValidation.videoDuration / Math.max(materialValidation.videoDuration, materialValidation.audioDuration)) * 100}%` }"
-                    ></div>
-                  </div>
-                </div>
-              </div>
-
-              <!-- 失败提示和建议 -->
-              <div v-if="!materialValidation.isValid" class="validation-error">
-                <p class="error-message">
-                  视频时长必须大于音频时长才能生成数字人视频
-                </p>
-                <div class="quick-actions">
-                  <a-button size="small" @click="replaceVideo">更换视频</a-button>
-                  <a-button size="small" @click="handleSimplifyScript">精简文案</a-button>
-                </div>
-              </div>
-            </div>
-          </div>
-
           <!-- 配音生成与校验（仅在识别后显示） -->
           <div v-if="identifyState.identified" class="section audio-generation-section">
             <h3>配音生成与校验</h3>
@@ -211,73 +155,67 @@
               </a-button>
             </div>
 
-            <!-- 音频预览（生成后显示） -->
+            <!-- 音频预览 -->
             <div v-if="audioState.generated" class="audio-preview">
               <div class="audio-info">
                 <h4>生成的配音</h4>
                 <div class="duration-info">
                   <span class="label">音频时长：</span>
-                  <span class="value">{{ (audioState.durationMs / 1000).toFixed(1) }} 秒</span>
+                  <span class="value">{{ audioDurationSec }} 秒</span>
                 </div>
                 <div class="duration-info">
                   <span class="label">人脸区间：</span>
-                  <span class="value">{{ (faceDuration / 1000).toFixed(1) }} 秒</span>
+                  <span class="value">{{ faceDurationSec }} 秒</span>
                 </div>
-                <div class="duration-info" :class="{ 'validation-passed': audioState.validationPassed, 'validation-failed': !audioState.validationPassed }">
+                <div class="duration-info" :class="{ 'validation-passed': validationPassed, 'validation-failed': !validationPassed }">
                   <span class="label">校验结果：</span>
                   <span class="value">
-                    {{ audioState.validationPassed ? '✅ 通过' : '❌ 不通过（需至少2秒重合）' }}
+                    {{ validationPassed ? '✅ 通过' : '❌ 不通过（音频时长不能超过人脸时长）' }}
                   </span>
                 </div>
               </div>
 
               <!-- 音频播放器 -->
-              <div class="audio-player">
+              <div v-if="audioUrl" class="audio-player">
-                <audio
+                <audio :src="audioUrl" controls class="audio-element" />
-                  v-if="audioState.generated.audioBase64"
-                  :src="`data:audio/mp3;base64,${audioState.generated.audioBase64}`"
-                  controls
-                  class="audio-element"
-                />
-                <audio
-                  v-else-if="audioState.generated.audioUrl"
-                  :src="audioState.generated.audioUrl"
-                  controls
-                  class="audio-element"
-                />
               </div>
 
               <!-- 重新生成按钮 -->
               <div class="regenerate-row">
-                <a-button
+                <a-button type="link" size="small" @click="generateAudio" :loading="audioState.generating">
-                  type="link"
-                  size="small"
-                  @click="generateAudio"
-                  :loading="audioState.generating"
-                >
                   重新生成
                 </a-button>
               </div>
             </div>
           </div>
 
+          <!-- Pipeline 进度条 -->
+          <PipelineProgress
+            v-if="isPipelineBusy || isPipelineReady || isPipelineFailed || isPipelineCompleted"
+            :state="pipelineState"
+            :progress="pipelineProgress"
+            :is-busy="isPipelineBusy"
+            :is-ready="isPipelineReady"
+            :is-failed="isPipelineFailed"
+            :is-completed="isPipelineCompleted"
+            :error="pipelineError"
+            @retry="retryPipeline"
+            @reset="resetPipeline"
+          />
+
           <!-- 按钮组 -->
           <div class="action-buttons">
             <a-button
               type="primary"
               size="large"
               :disabled="!canGenerate"
+              :loading="isPipelineBusy"
               block
               @click="generateDigitalHuman"
             >
-              生成数字人视频
+              {{ isPipelineBusy ? '处理中...' : '生成数字人视频' }}
             </a-button>
 
-            <!-- 添加提示信息 -->
-            <div v-if="canGenerate && !audioState.validationPassed" class="generate-hint">
-              <span class="hint-icon">⚠️</span>
-              <span>请先生成配音并通过时长校验</span>
-            </div>
           </div>
         </div>
 
@@ -299,6 +237,7 @@ import VideoSelector from '@/components/VideoSelector.vue'
 import VoiceSelector from '@/components/VoiceSelector.vue'
 import ResultPanel from '@/components/ResultPanel.vue'
 import FullWidthLayout from '@/layouts/components/FullWidthLayout.vue'
+import PipelineProgress from '@/components/PipelineProgress.vue'
 
 // Controller Hook
 import { useIdentifyFaceController } from './hooks/useIdentifyFaceController'
@@ -311,6 +250,7 @@ const dragOver = ref(false)
 // Controller 内部直接创建和管理两个子 Hook
 const controller = useIdentifyFaceController()
 
+
 // 解构 controller 以简化模板调用
 const {
   // 语音生成相关
@@ -318,14 +258,11 @@ const {
   speechRate,
   audioState,
   canGenerateAudio,
-  suggestedMaxChars,
   generateAudio,
 
   // 数字人生成相关
   videoState,
   identifyState,
-  materialValidation,
-  faceDuration,
   getVideoPreviewUrl,
 
   // 计算属性
@@ -334,6 +271,21 @@ const {
   textareaPlaceholder,
   speechRateMarks,
   speechRateDisplay,
+  faceDurationSec,
+  audioDurationSec,
+  audioUrl,
+  validationPassed,
+
+  // Pipeline 状态
+  pipelineState,
+  isPipelineBusy,
+  isPipelineReady,
+  isPipelineFailed,
+  isPipelineCompleted,
+  pipelineProgress,
+  pipelineError,
+  retryPipeline,
+  resetPipeline,
 
   // 事件处理方法
   handleVoiceSelect,
@@ -343,7 +295,6 @@ const {
   handleSelectUpload,
   handleSelectFromLibrary,
   handleVideoSelect,
-  handleSimplifyScript,
   handleVideoLoaded,
   replaceVideo,
   generateDigitalHuman,

frontend/app/web-gold/src/views/kling/hooks/pipeline/states.ts

@@ -0,0 +1,124 @@
+/**
+ * @fileoverview 状态机配置 - 状态定义和配置
+ */
+
+import type { PipelineState, StateConfig } from './types'
+
+/**
+ * 状态配置映射表
+ */
+export const STATE_CONFIG: Record<PipelineState, StateConfig> = {
+  idle: {
+    label: '等待开始',
+    progress: 0,
+    description: '请先选择视频并输入文案',
+  },
+  uploading: {
+    label: '上传视频中',
+    progress: 15,
+    description: '正在上传视频文件...',
+  },
+  recognizing: {
+    label: '识别人脸中',
+    progress: 35,
+    description: '正在分析视频中的人脸信息...',
+  },
+  generating: {
+    label: '生成配音中',
+    progress: 55,
+    description: '正在合成语音...',
+  },
+  validating: {
+    label: '校验时长中',
+    progress: 70,
+    description: '正在校验音频与视频时长...',
+  },
+  ready: {
+    label: '准备就绪',
+    progress: 80,
+    description: '校验通过，可以创建数字人视频',
+  },
+  creating: {
+    label: '创建任务中',
+    progress: 95,
+    description: '正在提交数字人视频生成任务...',
+  },
+  completed: {
+    label: '已完成',
+    progress: 100,
+    description: '任务已提交成功',
+  },
+  failed: {
+    label: '失败',
+    progress: 0,
+    description: '操作失败，请重试',
+  },
+}
+
+/**
+ * 状态顺序（用于步骤条显示）
+ */
+export const STATE_ORDER: PipelineState[] = [
+  'idle',
+  'uploading',
+  'recognizing',
+  'generating',
+  'validating',
+  'ready',
+  'creating',
+  'completed',
+]
+
+/**
+ * 忙碌状态（正在执行中的状态）
+ */
+export const BUSY_STATES: PipelineState[] = [
+  'uploading',
+  'recognizing',
+  'generating',
+  'validating',
+  'creating',
+]
+
+/**
+ * 终态（不能再转换的状态）
+ */
+export const TERMINAL_STATES: PipelineState[] = [
+  'completed',
+  'failed',
+]
+
+/**
+ * 获取状态在步骤条中的索引
+ */
+export function getStateIndex(state: PipelineState): number {
+  return STATE_ORDER.indexOf(state)
+}
+
+/**
+ * 获取状态的进度百分比
+ */
+export function getStateProgress(state: PipelineState): number {
+  return STATE_CONFIG[state].progress
+}
+
+/**
+ * 判断是否为忙碌状态
+ */
+export function isBusyState(state: PipelineState): boolean {
+  return BUSY_STATES.includes(state)
+}
+
+/**
+ * 判断是否为终态
+ */
+export function isTerminalState(state: PipelineState): boolean {
+  return TERMINAL_STATES.includes(state)
+}
+
+/**
+ * 判断状态是否可以重试
+ */
+export function canRetryFrom(state: PipelineState): boolean {
+  return state === 'failed'
+}

frontend/app/web-gold/src/views/kling/hooks/pipeline/types.ts

@@ -0,0 +1,126 @@
+/**
+ * @fileoverview 数字人生成流程状态机 - 类型定义
+ */
+
+/**
+ * 状态机所有可能的状态
+ */
+export type PipelineState =
+  | 'idle'          // 空闲
+  | 'uploading'     // 上传视频中
+  | 'recognizing'   // 人脸识别中
+  | 'generating'    // 生成配音中
+  | 'validating'    // 校验时长中
+  | 'ready'         // 准备就绪
+  | 'creating'      // 创建任务中
+  | 'completed'     // 已完成
+  | 'failed'        // 失败
+
+/**
+ * 状态配置
+ */
+export interface StateConfig {
+  /** 状态标签 */
+  label: string
+  /** 进度百分比 */
+  progress: number
+  /** 描述 */
+  description: string
+  /** 图标（可选） */
+  icon?: string
+}
+
+/**
+ * 步骤执行结果
+ */
+export interface StepResult<T = any> {
+  /** 是否成功 */
+  success: boolean
+  /** 返回数据 */
+  data?: T
+  /** 错误信息 */
+  error?: Error
+}
+
+/**
+ * Pipeline 上下文数据
+ */
+export interface PipelineContext {
+  /** 视频文件 */
+  videoFile: File | null
+  /** 已选择的视频 */
+  selectedVideo: any
+  /** 文案内容 */
+  text: string
+  /** 音色 */
+  voice: any
+  /** 语速 */
+  speechRate: number
+  /** 视频文件ID */
+  videoFileId: string | number | null
+  /** 会话ID */
+  sessionId: string
+  /** 人脸ID */
+  faceId: string
+  /** 人脸开始时间 */
+  faceStartTime: number
+  /** 人脸结束时间 */
+  faceEndTime: number
+  /** 音频 Base64 */
+  audioBase64: string
+  /** 音频格式 */
+  audioFormat: string
+  /** 音频时长（毫秒） */
+  audioDurationMs: number
+  /** 视频时长（毫秒） */
+  videoDurationMs: number
+  /** 校验是否通过 */
+  validationPassed: boolean
+}
+
+/**
+ * Pipeline 执行参数
+ */
+export interface PipelineParams {
+  videoFile: File | null
+  selectedVideo: any
+  text: string
+  voice: any
+  speechRate: number
+}
+
+/**
+ * Pipeline 选项配置
+ */
+export interface PipelineOptions {
+  /** 上传视频 */
+  uploadVideo: (file: File) => Promise<string | number>
+  /** 从库中识别 */
+  recognizeFromLibrary: (video: any) => Promise<any>
+  /** 识别已上传视频 */
+  recognizeUploaded: (fileId: string | number) => Promise<any>
+  /** 生成音频 */
+  generateAudio: (text: string, voice: any, speechRate: number) => Promise<{
+    audioBase64: string
+    format?: string
+    durationMs?: number
+  }>
+  /** 创建任务 */
+  createTask: (data: any) => Promise<void>
+}
+
+/**
+ * 状态机执行状态
+ */
+export interface ExecutionState {
+  /** 当前状态 */
+  current: PipelineState
+  /** 历史状态 */
+  history: PipelineState[]
+  /** 上下文数据 */
+  context: Partial<PipelineContext>
+  /** 是否可以继续下一步 */
+  canNext: boolean
+  /** 是否可以重试 */
+  canRetry: boolean
+}

frontend/app/web-gold/src/views/kling/hooks/pipeline/useSimplePipeline.ts

@@ -0,0 +1,287 @@
+/**
+ * @fileoverview 极简状态机 Hook - 数字人生成流程
+ *
+ * 设计理念：
+ * 1. 简单直观 - 用普通 JS/TS 代码，无需学习复杂概念
+ * 2. 易于调试 - 打断点即可查看状态
+ * 3. 功能完整 - 支持状态管理、进度显示、错误处理、重试
+ */
+
+import { ref, computed } from 'vue'
+import { message } from 'ant-design-vue'
+import type { LipSyncTaskData } from '../../types/identify-face'
+import { createLipSyncTask } from '@/api/kling'
+import type {
+  PipelineState,
+  PipelineContext,
+  PipelineParams,
+  PipelineOptions,
+  ExecutionState,
+} from './types'
+import {
+  STATE_CONFIG,
+  getStateIndex,
+  isBusyState,
+  isTerminalState,
+  canRetryFrom,
+} from './states'
+
+/**
+ * 初始上下文
+ */
+const INITIAL_CONTEXT: Partial<PipelineContext> = {
+  videoFile: null,
+  selectedVideo: null,
+  text: '',
+  voice: null,
+  speechRate: 1,
+  videoFileId: null,
+  sessionId: '',
+  faceId: '',
+  faceStartTime: 0,
+  faceEndTime: 0,
+  audioBase64: '',
+  audioFormat: 'mp3',
+  audioDurationMs: 0,
+  videoDurationMs: 0,
+  validationPassed: false,
+}
+
+/**
+ * 极简状态机 Hook
+ */
+export function useSimplePipeline(options: PipelineOptions) {
+  // ========== 状态管理 ==========
+  const state = ref<PipelineState>('idle')
+  const context = ref<Partial<PipelineContext>>({ ...INITIAL_CONTEXT })
+  const error = ref<string | null>(null)
+  const history = ref<PipelineState[]>(['idle'])
+
+  // ========== 计算属性 ==========
+  const stateLabel = computed(() => STATE_CONFIG[state.value].label)
+  const stateDescription = computed(() => STATE_CONFIG[state.value].description)
+  const progress = computed(() => STATE_CONFIG[state.value].progress)
+  const currentStepIndex = computed(() => getStateIndex(state.value))
+  const isBusy = computed(() => isBusyState(state.value))
+  const isReady = computed(() => state.value === 'ready')
+  const isFailed = computed(() => state.value === 'failed')
+  const isCompleted = computed(() => state.value === 'completed')
+  const isTerminal = computed(() => isTerminalState(state.value))
+  const canRetry = computed(() => canRetryFrom(state.value))
+
+  // ========== 内部方法 ==========
+
+  /**
+   * 更新状态
+   */
+  function setState(newState: PipelineState) {
+    const oldState = state.value
+    state.value = newState
+    history.value.push(newState)
+    console.log(`[Pipeline] ${oldState} -> ${newState}`)
+  }
+
+  /**
+   * 设置错误状态
+   */
+  function setError(err: Error | string) {
+    const errorMsg = typeof err === 'string' ? err : err.message
+    error.value = errorMsg
+    setState('failed')
+    message.error(errorMsg)
+  }
+
+  /**
+   * 执行步骤（带错误处理）
+   */
+  async function executeStep<T>(
+    newState: PipelineState,
+    fn: () => Promise<T>
+  ): Promise<T> {
+    setState(newState)
+    try {
+      return await fn()
+    } catch (err) {
+      setError(err as Error)
+      throw err
+    }
+  }
+
+  // ========== 公开方法 ==========
+
+  /**
+   * 运行完整流程（到 ready 状态）
+   */
+  async function run(params: PipelineParams): Promise<void> {
+    // 重置状态
+    reset()
+
+    try {
+      // 保存参数到上下文
+      context.value.videoFile = params.videoFile
+      context.value.selectedVideo = params.selectedVideo
+      context.value.text = params.text
+      context.value.voice = params.voice
+      context.value.speechRate = params.speechRate
+
+      // 步骤1: 上传视频（如果是上传模式）
+      if (params.videoFile && !params.selectedVideo) {
+        const fileId = await executeStep('uploading', () =>
+          options.uploadVideo(params.videoFile!)
+        )
+        context.value.videoFileId = fileId
+      } else if (params.selectedVideo) {
+        context.value.videoFileId = params.selectedVideo.fileId
+      }
+
+      // 步骤2: 识别人脸
+      const recognizeData = params.selectedVideo
+        ? await options.recognizeFromLibrary(params.selectedVideo)
+        : await options.recognizeUploaded(context.value.videoFileId!)
+
+      await executeStep('recognizing', async () => recognizeData)
+
+      context.value.sessionId = recognizeData.sessionId
+      context.value.faceId = recognizeData.faceId
+      context.value.faceStartTime = recognizeData.startTime || 0
+      context.value.faceEndTime = recognizeData.endTime || 0
+      context.value.videoDurationMs = recognizeData.duration || 0
+
+      // 步骤3: 生成音频
+      const audioData = await executeStep('generating', () =>
+        options.generateAudio(params.text, params.voice, params.speechRate)
+      )
+
+      context.value.audioBase64 = audioData.audioBase64
+      context.value.audioFormat = audioData.format || 'mp3'
+      context.value.audioDurationMs = audioData.durationMs || 0
+
+      // 步骤4: 校验时长
+      setState('validating')
+      const videoDurationMs = context.value.videoDurationMs ?? 0
+      if (context.value.audioDurationMs > videoDurationMs) {
+        throw new Error(
+          `校验失败：音频时长(${(context.value.audioDurationMs / 1000).toFixed(1)}秒) 超过人脸时长(${(videoDurationMs / 1000).toFixed(1)}秒)`
+        )
+      }
+      context.value.validationPassed = true
+
+      // 到达 ready 状态
+      setState('ready')
+
+    } catch (err) {
+      // 错误已在 executeStep 中处理
+    }
+  }
+
+  /**
+   * 创建数字人任务（从 ready 状态）
+   */
+  async function createTask(): Promise<void> {
+    if (state.value !== 'ready') {
+      message.warning('请先完成视频识别和音频生成')
+      return
+    }
+
+    try {
+      setState('creating')
+
+      const taskData: LipSyncTaskData = {
+        taskName: `数字人任务_${Date.now()}`,
+        videoFileId: context.value.videoFileId!,
+        inputText: context.value.text!,
+        speechRate: context.value.speechRate!,
+        volume: 0,
+        guidanceScale: 1,
+        seed: 8888,
+        kling_session_id: context.value.sessionId!,
+        kling_face_id: context.value.faceId!,
+        kling_face_start_time: context.value.faceStartTime!,
+        kling_face_end_time: context.value.faceEndTime!,
+        ai_provider: 'kling',
+        voiceConfigId: context.value.voice!.rawId || context.value.voice!.id.match(/[\w-]+$/)?.[0] || context.value.voice!.id,
+        pre_generated_audio: {
+          audioBase64: context.value.audioBase64!,
+          format: context.value.audioFormat!,
+        },
+        sound_end_time: context.value.audioDurationMs!,
+      }
+
+      const res = await createLipSyncTask(taskData)
+      if (res.code !== 0) {
+        throw new Error(res.msg || '任务创建失败')
+      }
+
+      setState('completed')
+      message.success('任务已提交，请在任务中心查看生成进度')
+
+    } catch (err) {
+      setError(err as Error)
+    }
+  }
+
+  /**
+   * 重试（从 failed 状态恢复）
+   */
+  function retry(): void {
+    if (!canRetry.value) {
+      message.warning('当前状态无法重试')
+      return
+    }
+
+    error.value = null
+    // 回到 idle 重新开始
+    setState('idle')
+  }
+
+  /**
+   * 重置到初始状态
+   */
+  function reset(): void {
+    state.value = 'idle'
+    context.value = { ...INITIAL_CONTEXT }
+    error.value = null
+    history.value = ['idle']
+  }
+
+  /**
+   * 获取执行状态（用于调试）
+   */
+  function getExecutionState(): ExecutionState {
+    return {
+      current: state.value,
+      history: [...history.value],
+      context: { ...context.value },
+      canNext: state.value === 'ready',
+      canRetry: canRetry.value,
+    }
+  }
+
+  // ========== 返回 API ==========
+  return {
+    // 状态
+    state,
+    context,
+    error,
+    history,
+
+    // 计算属性
+    stateLabel,
+    stateDescription,
+    progress,
+    currentStepIndex,
+    isBusy,
+    isReady,
+    isFailed,
+    isCompleted,
+    isTerminal,
+    canRetry,
+
+    // 方法
+    run,
+    createTask,
+    retry,
+    reset,
+    getExecutionState,
+  }
+}

frontend/app/web-gold/src/views/kling/hooks/useDigitalHumanGeneration.ts

@@ -1,6 +1,5 @@
 /**
- * @fileoverview useDigitalHumanGeneration Hook - 数字人生成逻辑封装
+ * @fileoverview useDigitalHumanGeneration Hook - 数字人生成逻辑
- * @author Claude Code
  */
 
 import { ref, computed } from 'vue'
@@ -14,18 +13,14 @@ import type {
 import { identifyUploadedVideo } from '@/api/kling'
 import { useUpload } from '@/composables/useUpload'
 
-/**
- * 数字人生成 Hook
- * 独立管理所有状态，不依赖外部状态
- */
 export function useDigitalHumanGeneration(): UseDigitalHumanGeneration {
-  // ==================== 响应式状态 ====================
+  // ========== 状态 ==========
-
   const videoState = ref<VideoState>({
     uploadedVideo: '',
     videoFile: null,
     previewVideoUrl: '',
     selectedVideo: null,
+    fileId: null,
     videoSource: null,
     selectorVisible: false,
   })
@@ -40,24 +35,16 @@ export function useDigitalHumanGeneration(): UseDigitalHumanGeneration {
     videoFileId: null,
   })
 
-  // ==================== Upload Hook ====================
   const { upload } = useUpload()
 
-  // ==================== 计算属性 ====================
+  // ========== 计算属性 ==========
-
+  const faceDuration = computed(function() {
-  /**
-   * 人脸出现时长
-   */
-  const faceDuration = computed(() => {
     return identifyState.value.faceEndTime - identifyState.value.faceStartTime
   })
 
-  // ==================== 核心方法 ====================
+  // ========== 方法 ==========
 
-  /**
+  async function handleFileUpload(file: File): Promise<void> {
-   * 处理视频文件上传
-   */
-  const handleFileUpload = async (file: File): Promise<void> => {
     if (!file.name.match(/\.(mp4|mov)$/i)) {
       message.error('仅支持 MP4 和 MOV')
       return
@@ -68,148 +55,101 @@ export function useDigitalHumanGeneration(): UseDigitalHumanGeneration {
     videoState.value.selectedVideo = null
     videoState.value.previewVideoUrl = ''
     videoState.value.videoSource = 'upload'
-
     resetIdentifyState()
-
-    await performFaceRecognition()
   }
 
-  /**
+  async function handleVideoSelect(video: Video): Promise<void> {
-   * 处理从素材库选择视频
-   */
-  const handleVideoSelect = (video: Video): void => {
     videoState.value.selectedVideo = video
     videoState.value.uploadedVideo = video.fileUrl
     videoState.value.videoFile = null
     videoState.value.videoSource = 'select'
     videoState.value.selectorVisible = false
     resetIdentifyState()
-    identifyState.value.videoFileId = video.id
+    identifyState.value.videoFileId = video.fileId
   }
 
-  /**
+  async function performFaceRecognition(): Promise<void> {
-   * 执行人脸识别
-   */
-  const performFaceRecognition = async (): Promise<void> => {
     const hasUploadFile = videoState.value.videoFile
     const hasSelectedVideo = videoState.value.selectedVideo
 
-    if (!hasUploadFile && !hasSelectedVideo) {
+    if (!hasUploadFile && !hasSelectedVideo) return
-      return
-    }
 
     identifyState.value.identifying = true
 
     try {
-      let res
       if (hasSelectedVideo) {
-        res = await identifyUploadedVideo(hasSelectedVideo)
+        const res = await identifyUploadedVideo(hasSelectedVideo) as { success: boolean; data: { sessionId: string; faceId: string | null; startTime: number; endTime: number } }
-        identifyState.value.videoFileId = hasSelectedVideo.id
+        identifyState.value.videoFileId = hasSelectedVideo.fileId
+
+        identifyState.value.sessionId = res.data.sessionId
+        identifyState.value.faceId = res.data.faceId || ''
+        identifyState.value.faceStartTime = res.data.startTime || 0
+        identifyState.value.faceEndTime = res.data.endTime || 0
       } else {
-        // 处理文件上传（提取封面）
         const file = hasUploadFile!
         let coverBase64 = null
         try {
           const { extractVideoCover } = await import('@/utils/video-cover')
-          const cover = await extractVideoCover(file, {
+          const cover = await extractVideoCover(file, { maxWidth: 800, quality: 0.8 })
-            maxWidth: 800,
-            quality: 0.8
-          })
           coverBase64 = cover.base64
         } catch {
           // 封面提取失败不影响主流程
         }
 
-        // 使用useUpload Hook上传文件
         const fileId = await upload(file, {
           fileCategory: 'video',
-          groupId: null, // 数字人模块不使用groupId
+          groupId: null,
           coverBase64,
-          onStart: () => {},
+          onStart: function() {},
-          onProgress: () => {},
+          onProgress: function() {},
-          onSuccess: () => {
+          onSuccess: function() {},
-            message.success('文件上传成功')
+          onError: function(err: Error) {
-          },
-          onError: (err: Error) => {
             message.error(err.message || '上传失败')
           }
         })
 
-        // 生成播放链接
-        // TODO: 获取播放链接逻辑
-
-        res = {
-          success: true,
-          data: {
-            fileId,
-            videoUrl: '', // TODO: 需要获取实际URL
-            sessionId: '', // TODO: 需要实际识别
-            faceId: null,
-            startTime: 0,
-            endTime: 0
-          }
-        }
         identifyState.value.videoFileId = fileId
+        identifyState.value.sessionId = ''
+        identifyState.value.faceId = ''
+        identifyState.value.faceStartTime = 0
+        identifyState.value.faceEndTime = 0
       }
 
-      identifyState.value.sessionId = res.data.sessionId
-      identifyState.value.faceId = res.data.faceId
-      identifyState.value.faceStartTime = res.data.startTime || 0
-      identifyState.value.faceEndTime = res.data.endTime || 0
       identifyState.value.identified = true
 
-      const durationSec = faceDuration.value / 1000
+      // 识别完成，不显示提示信息
-      const suggestedMaxChars = Math.floor(durationSec * 3.5)
+    } catch (error: unknown) {
-      message.success(`识别完成！人脸出现时长约 ${durationSec.toFixed(1)} 秒，建议文案不超过 ${suggestedMaxChars} 字`)
+      const err = error as Error
-    } catch (error: any) {
+      message.error(err.message || '识别失败')
-      message.error(error.message || '识别失败')
       throw error
     } finally {
       identifyState.value.identifying = false
     }
   }
 
-
+  function resetVideoState(): void {
-  /**
-   * 重置视频状态
-   */
-  const resetVideoState = (): void => {
     videoState.value.uploadedVideo = ''
     videoState.value.videoFile = null
     videoState.value.selectedVideo = null
+    videoState.value.fileId = null
     videoState.value.videoSource = null
     videoState.value.previewVideoUrl = ''
     videoState.value.selectorVisible = false
-
     resetIdentifyState()
   }
 
-  /**
+  function getVideoPreviewUrl(video: Video): string {
-   * 获取视频预览 URL
-   */
-  const getVideoPreviewUrl = (video: Video): string => {
     if (video.coverBase64) {
-      if (!video.coverBase64.startsWith('data:')) {
+      return video.coverBase64.startsWith('data:')
-        return `data:image/jpeg;base64,${video.coverBase64}`
+        ? video.coverBase64
-      }
+        : `data:image/jpeg;base64,${video.coverBase64}`
-      return video.coverBase64
     }
-
+    if (video.previewUrl) return video.previewUrl
-    if (video.previewUrl) {
+    if (video.coverUrl) return video.coverUrl
-      return video.previewUrl
-    }
-
-    if (video.coverUrl) {
-      return video.coverUrl
-    }
-
     return 'data:image/svg+xml;base64,PHN2ZyB3aWR0aD0iMjAwIiBoZWlnaHQ9IjExMCIgdmlld0JveD0iMCAwIDIwMCAxMTAiIGZpbGw9Im5vbmUiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyI+CjxyZWN0IHdpZHRoPSIyMDAiIGhlaWdodD0iMTEwIiBmaWxsPSIjMzc0MTUxIi8+CjxwYXRoIGQ9Ik04NSA0NUwxMTUgNjVMMTA1IDg1TDc1IDc1TDg1IDQ1WiIgZmlsbD0iIzU3MjY1MSIvPgo8L3N2Zz4K'
   }
 
-  /**
+  function resetIdentifyState(): void {
-   * 重置识别状态
-   */
-  const resetIdentifyState = (): void => {
     identifyState.value.identified = false
     identifyState.value.sessionId = ''
     identifyState.value.faceId = ''
@@ -217,14 +157,9 @@ export function useDigitalHumanGeneration(): UseDigitalHumanGeneration {
   }
 
   return {
-    // 响应式状态
     videoState,
     identifyState,
-
-    // 计算属性
     faceDuration,
-
-    // 方法
     handleFileUpload,
     handleVideoSelect,
     performFaceRecognition,

frontend/app/web-gold/src/views/kling/hooks/useIdentifyFaceController.ts

@@ -1,416 +1,329 @@
 /**
- * @fileoverview useIdentifyFaceController Hook - 主控制器 Hook
+ * @fileoverview useIdentifyFaceController Hook - 主控制器
- * @author Claude Code
+ *
+ * 职责：协调语音、视频、Pipeline 各个子模块，提供统一的外部接口
+ *
+ * 模块依赖关系:
+ *   ┌─────────────────────────────────────────────────┐
+ *   │            useIdentifyFaceController              │
+ *   │   ┌──────────────┐  ┌──────────────┐  ┌───────────┐│
+ *   │   │   Voice      │  │  Digital     │  │ Pipeline  ││
+ *   │   │   Generation  │  │  Human       │  │           ││
+ *   │   │              │  │  Generation  │  │           ││
+ *   │   └──────────────┘  └──────────────┘  └───────────┘│
+ *   └─────────────────────────────────────────────────┘
  */
 
-import { ref, computed, watch } from 'vue'
+import { computed } from 'vue'
 import { message } from 'ant-design-vue'
 import type {
   UseIdentifyFaceController,
-  LipSyncTaskData,
+  VoiceMeta,
-  MaterialValidation,
 } from '../types/identify-face'
-// @ts-ignore
-import { createLipSyncTask } from '@/api/kling'
-
-// 导入子 Hooks
 import { useVoiceGeneration } from './useVoiceGeneration'
 import { useDigitalHumanGeneration } from './useDigitalHumanGeneration'
+import { useSimplePipeline } from './pipeline/useSimplePipeline'
+
+// ==================== 常量 ====================
+const SPEECH_RATE_MARKS = { 0.5: '0.5x', 1: '1x', 1.5: '1.5x', 2: '2x' }
+const MAX_TEXT_LENGTH = 4000
 
 /**
- * 识别控制器 Hook - 充当协调器
+ * 主控制器 Hook
- * 内部直接创建和管理两个子 Hook
  */
 export function useIdentifyFaceController(): UseIdentifyFaceController {
-  // ==================== 创建子 Hooks 并解构 ====================
+  // 子 Hooks
+  const voice = useVoiceGeneration()
+  const digitalHuman = useDigitalHumanGeneration()
 
-  // 1. 语音生成 Hook - 解构响应式变量
+  // Pipeline 流程配置（使用新的极简状态机）
-  const {
+  const pipeline = useSimplePipeline({
-    ttsText,
+    uploadVideo: async (_file: File) => {
-    speechRate,
+      // 上传已经在 handleFileUpload 中处理
-    selectedVoiceMeta,
+      // 这里直接返回 fileId
-    audioState,
+      return digitalHuman.identifyState.value.videoFileId || ''
-    canGenerateAudio,
+    },
-    suggestedMaxChars,
+    recognizeFromLibrary: async (video: any) => {
-    generateAudio,
+      await digitalHuman.handleVideoSelect(video)
-    resetAudioState,
+      await digitalHuman.performFaceRecognition()
-  } = useVoiceGeneration()
+      return {
-
+        sessionId: digitalHuman.identifyState.value.sessionId,
-  // 2. 数字人生成 Hook - 解构响应式变量
+        faceId: digitalHuman.identifyState.value.faceId,
-  const {
+        startTime: digitalHuman.identifyState.value.faceStartTime,
-    videoState,
+        endTime: digitalHuman.identifyState.value.faceEndTime,
-    identifyState,
+        duration: digitalHuman.faceDuration.value,
-    faceDuration,
+      }
-    performFaceRecognition,
+    },
-    handleFileUpload,
+    recognizeUploaded: async (_fileId: string | number) => {
-    handleVideoSelect: _handleVideoSelect,
+      await digitalHuman.performFaceRecognition()
-    getVideoPreviewUrl,
+      return {
-    resetVideoState,
+        sessionId: digitalHuman.identifyState.value.sessionId,
-    resetIdentifyState,
+        faceId: digitalHuman.identifyState.value.faceId,
-  } = useDigitalHumanGeneration()
+        startTime: digitalHuman.identifyState.value.faceStartTime,
-
+        endTime: digitalHuman.identifyState.value.faceEndTime,
-  // 3. Controller 统一管理跨 Hook 的状态
+        duration: digitalHuman.faceDuration.value,
-  const materialValidation = ref<MaterialValidation>({
+      }
-    videoDuration: 0,
+    },
-    audioDuration: 0,
+    generateAudio: async (text: string, voiceMeta: any, speechRate: number) => {
-    isValid: false,
+      voice.ttsText.value = text
-    showDetails: false,
+      voice.selectedVoiceMeta.value = voiceMeta
+      voice.speechRate.value = speechRate
+      await voice.generateAudio()
+      const audio = voice.audioState.value.generated!
+      return {
+        audioBase64: audio.audioBase64,
+        format: audio.format || 'mp3',
+        durationMs: voice.audioState.value.durationMs,
+      }
+    },
+    createTask: async () => {
+      // 任务创建在 Pipeline 中处理
+    },
   })
 
-  // 4. 监听音频状态变化，自动触发素材校验
-  watch(
-    () => audioState.value.generated && audioState.value.durationMs > 0,
-    (newVal, oldVal) => {
-      if (newVal && !oldVal) {
-        // 音频生成完成，获取视频时长并校验
-        const videoDurationMs = faceDuration.value || 0
-        const audioDurationMs = audioState.value.durationMs
-
-        if (videoDurationMs > 0) {
-          validateMaterialDuration(videoDurationMs, audioDurationMs)
-        }
-      }
-    },
-    { flush: 'post' }
-  )
-
-  // 5. 监听人脸识别状态变化，更新素材校验的视频时长
-  watch(
-    () => identifyState.value.identified,
-    (newVal, oldVal) => {
-      if (newVal && !oldVal) {
-        // 人脸识别成功，获取视频时长
-        const videoDurationMs = faceDuration.value
-
-        // 如果已有音频，则重新校验
-        if (audioState.value.generated && audioState.value.durationMs > 0) {
-          const audioDurationMs = audioState.value.durationMs
-          validateMaterialDuration(videoDurationMs, audioDurationMs)
-        } else {
-          // 否则只更新视频时长
-          materialValidation.value.videoDuration = videoDurationMs
-        }
-      }
-    },
-    { flush: 'post' }
-  )
   // ==================== 计算属性 ====================
 
-  /**
+  /** 是否可以生成数字人视频 */
-   * 是否可以生成数字人视频（综合检查）
+  const canGenerate = computed((): boolean => {
-   */
+    // Pipeline 运行中禁用
-  const canGenerate = computed(() => {
+    if (pipeline.isBusy.value) return false
-    const hasText = ttsText.value.trim()
-    const hasVoice = selectedVoiceMeta.value
-    const hasVideo = videoState.value.uploadedVideo || videoState.value.selectedVideo
 
-    // 音频校验：只有生成过音频后才需要校验通过
+    const hasText = voice.ttsText.value.trim()
-    const audioValidated = !audioState.value.generated || audioState.value.validationPassed
+    const hasVoice = voice.selectedVoiceMeta.value
-    // 素材校验：只有进行过校验后才需要校验通过
+    const hasVideo = digitalHuman.videoState.value.uploadedVideo || digitalHuman.videoState.value.selectedVideo
-    const materialValidated = materialValidation.value.videoDuration === 0 || materialValidation.value.isValid
+    const hasBasicConfig = hasText && hasVoice && hasVideo
 
-    return !!(hasText && hasVoice && hasVideo && audioValidated && materialValidated)
+    // 未识别时只需要基础配置
+    if (!digitalHuman.identifyState.value.identified) return !!hasBasicConfig
+
+    // 已识别后需要音频生成并通过校验
+    return !!(
+      hasBasicConfig &&
+      voice.audioState.value.generated &&
+      validationPassed.value
+    )
   })
 
-  /**
+  /** 最大文本长度（根据人脸时长动态计算） */
-   * 最大的文本长度
-   */
   const maxTextLength = computed(() => {
-    if (!identifyState.value.identified || faceDuration.value <= 0) {
+    const isIdentified = digitalHuman.identifyState.value.identified
-      return 4000
+    const faceDuration = digitalHuman.faceDuration.value
-    }
+    if (!isIdentified || faceDuration <= 0) return MAX_TEXT_LENGTH
-    return Math.min(4000, Math.floor(suggestedMaxChars.value * 1.2))
+    return Math.min(MAX_TEXT_LENGTH, Math.floor(voice.suggestedMaxChars.value * 1.2))
   })
 
-  /**
+  /** 文本框占位符提示 */
-   * 文本框占位符
-   */
   const textareaPlaceholder = computed(() => {
-    if (identifyState.value.identified && faceDuration.value > 0) {
+    const isIdentified = digitalHuman.identifyState.value.identified
-      return `请输入文案，建议不超过${suggestedMaxChars.value}字以确保与视频匹配`
+    const faceDuration = digitalHuman.faceDuration.value
+    if (isIdentified && faceDuration > 0) {
+      return `请输入文案，建议不超过${voice.suggestedMaxChars.value}字以确保与视频匹配`
     }
     return '请输入你想让角色说话的内容'
   })
 
-  /**
+  /** 语速显示文本 */
-   * 语速标记
+  const speechRateDisplay = computed(() => `${voice.speechRate.value.toFixed(1)}x`)
-   */
+
-  const speechRateMarks = { 0.5: '0.5x', 1: '1x', 1.5: '1.5x', 2: '2x' }
+  /** 人脸时长显示（秒） */
+  const faceDurationSec = computed(() => (digitalHuman.faceDuration.value / 1000).toFixed(1))
+
+  /** 音频时长显示（秒） */
+  const audioDurationSec = computed(() => (voice.audioState.value.durationMs / 1000).toFixed(1))
+
+  /** 是否显示生成提示 */
+  const showGenerateHint = computed(() =>
+    digitalHuman.identifyState.value.identified &&
+    (!voice.audioState.value.generated || !validationPassed.value)
+  )
+
+  /** 音频播放 URL */
+  const audioUrl = computed(() => {
+    const audio = voice.audioState.value.generated
+    if (!audio) return ''
+    return audio.audioBase64 ? `data:audio/mp3;base64,${audio.audioBase64}` : audio.audioUrl || ''
+  })
 
   /**
-   * 语速显示
+   * 校验是否通过（计算属性）
+   * 规则：音频时长 <= 人脸时长（Kling 要求音频不能超过人脸区间）
    */
-  const speechRateDisplay = computed(() => `${speechRate.value.toFixed(1)}x`)
+  const validationPassed = computed(() => {
+    const faceDuration = Number(faceDurationSec.value)
+    const audioDuration = Number(audioDurationSec.value)
+    return audioDuration <= faceDuration
+  })
 
-  // ==================== 业务流程方法 ====================
+  // ==================== 业务方法 ====================
 
   /**
-   * 生成数字人视频
+   * 重置所有状态
    */
-  const generateDigitalHuman = async (): Promise<void> => {
+  function resetAllStates(): void {
+    voice.resetAudioState()
+    digitalHuman.resetVideoState()
+    digitalHuman.resetIdentifyState()
+    pipeline.reset()
+  }
+
+  /**
+   * 生成数字人视频 - 使用新的 Pipeline API
+   */
+  async function generateDigitalHuman(): Promise<void> {
     if (!canGenerate.value) {
       message.warning('请先完成配置')
       return
     }
 
-    const text = ttsText.value.trim()
+    const text = voice.ttsText.value.trim()
+    const voiceMeta = voice.selectedVoiceMeta.value
+
     if (!text) {
       message.warning('请输入文案内容')
       return
     }
 
-    const voice = selectedVoiceMeta.value
+    if (!voiceMeta) {
-    if (!voice) {
       message.warning('请选择音色')
       return
     }
 
     try {
-      // 如果未识别，先进行人脸识别
+      // 运行流程到 ready 状态
-      if (!identifyState.value.identified) {
+      await pipeline.run({
-        const hasUploadFile = videoState.value.videoFile
+        videoFile: digitalHuman.videoState.value.videoFile,
-        const hasSelectedVideo = videoState.value.selectedVideo
+        selectedVideo: digitalHuman.videoState.value.selectedVideo,
+        text,
+        voice: voiceMeta,
+        speechRate: voice.speechRate.value,
+      })
 
-        if (!hasUploadFile && !hasSelectedVideo) {
+      // 如果到达 ready 状态，自动创建任务
-          message.warning('请先选择或上传视频')
+      if (pipeline.isReady.value) {
-          return
+        await pipeline.createTask()
-        }
+        // 任务提交成功后，重置所有状态
-
+        resetAllStates()
-        try {
-          await performFaceRecognition()
-          message.success('人脸识别完成')
-        } catch (error) {
-          return
-        }
       }
-
+    } catch {
-      const videoFileId = identifyState.value.videoFileId
+      // 错误已在 Pipeline 中处理
-
-      const taskData: LipSyncTaskData = {
-        taskName: `数字人任务_${Date.now()}`,
-        videoFileId: videoFileId!,
-        inputText: ttsText.value,
-        speechRate: speechRate.value,
-        volume: 0,
-        guidanceScale: 1,
-        seed: 8888,
-        kling_session_id: identifyState.value.sessionId,
-        kling_face_id: identifyState.value.faceId,
-        kling_face_start_time: identifyState.value.faceStartTime,
-        kling_face_end_time: identifyState.value.faceEndTime,
-        ai_provider: 'kling',
-        voiceConfigId: voice.rawId || extractIdFromString(voice.id),
-      }
-
-      if (!taskData.voiceConfigId) {
-        message.warning('音色配置无效')
-        return
-      }
-
-      // 如果有预生成的音频，添加到任务数据中
-      if (audioState.value.generated && audioState.value.durationMs > 0) {
-        taskData.pre_generated_audio = {
-          audioBase64: audioState.value.generated.audioBase64,
-          format: audioState.value.generated.format || 'mp3',
-        }
-
-        taskData.sound_end_time = audioState.value.durationMs
-      }
-
-      const res = await createLipSyncTask(taskData)
-
-      if (res.code === 0) {
-        message.success('任务已提交到任务中心，请前往查看')
-      } else {
-        throw new Error(res.msg || '任务创建失败')
-      }
-    } catch (error: any) {
-      message.error(error.message || '任务提交失败')
     }
   }
 
   /**
    * 更换视频
    */
-  const replaceVideo = (): void => {
+  function replaceVideo(): void {
-    if (videoState.value.videoSource === 'upload') {
+    digitalHuman.resetVideoState()
-      videoState.value.videoFile = null
+    voice.resetAudioState()
-      videoState.value.uploadedVideo = ''
+    pipeline.reset()
-    } else {
-      videoState.value.selectedVideo = null
-      videoState.value.videoFile = null
-      videoState.value.uploadedVideo = ''
-    }
-
-    // 重置所有状态
-    resetVideoState()
-    resetAudioState()
   }
 
-  /**
+  // ==================== 事件处理方法 ====================
-   * 处理音色选择
+
-   */
+  function handleVoiceSelect(voiceMeta: VoiceMeta): void {
-  const handleVoiceSelect = (voice: any): void => {
+    voice.selectedVoiceMeta.value = voiceMeta
-    selectedVoiceMeta.value = voice
   }
 
-  /**
+  function handleFileSelect(event: Event): void {
-   * 处理文件选择
+    const file = (event.target as HTMLInputElement).files?.[0]
-   */
+    if (file) digitalHuman.handleFileUpload(file)
-  const handleFileSelect = (event: Event): void => {
-    const input = event.target as HTMLInputElement
-    const file = input.files?.[0]
-    if (file) {
-      handleFileUpload(file)
-    }
   }
 
-  /**
+  function handleDrop(event: DragEvent): void {
-   * 处理拖拽上传
-   */
-  const handleDrop = (event: DragEvent): void => {
     event.preventDefault()
     const file = event.dataTransfer?.files[0]
-    if (file) {
+    if (file) digitalHuman.handleFileUpload(file)
-      handleFileUpload(file)
-    }
   }
 
-  /**
+  function triggerFileSelect(): void {
-   * 触发文件选择
+    document.querySelector<HTMLInputElement>('input[type="file"]')?.click()
-   */
-  const triggerFileSelect = (): void => {
-    const fileInput = document.querySelector('input[type="file"]') as HTMLInputElement
-    fileInput?.click()
   }
 
-  /**
+  function handleSelectUpload(): void {
-   * 选择上传模式
+    digitalHuman.videoState.value.videoSource = 'upload'
-   */
+    digitalHuman.videoState.value.selectedVideo = null
-  const handleSelectUpload = (): void => {
+    digitalHuman.resetIdentifyState()
-    videoState.value.videoSource = 'upload'
+    pipeline.reset()
-    videoState.value.selectedVideo = null
-    resetIdentifyState()
   }
 
-  /**
+  function handleSelectFromLibrary(): void {
-   * 从素材库选择
+    digitalHuman.videoState.value.videoSource = 'select'
-   */
+    digitalHuman.videoState.value.videoFile = null
-  const handleSelectFromLibrary = (): void => {
+    digitalHuman.videoState.value.uploadedVideo = ''
-    videoState.value.videoSource = 'select'
+    digitalHuman.videoState.value.selectorVisible = true
-    videoState.value.videoFile = null
+    pipeline.reset()
-    videoState.value.uploadedVideo = ''
-    videoState.value.selectorVisible = true
   }
 
-  /**
+  async function handleVideoSelect(video: any): Promise<void> {
-   * 处理视频选择器选择
+    await digitalHuman.handleVideoSelect(video)
-   */
-  const handleVideoSelect = (video: any): void => {
-    _handleVideoSelect(video)
   }
 
-  /**
+  function handleVideoLoaded(videoUrl: string): void {
-   * 简化文案
+    digitalHuman.videoState.value.previewVideoUrl = videoUrl
-   */
-  const handleSimplifyScript = (): void => {
-    const textarea = document.querySelector('.tts-textarea textarea') as HTMLTextAreaElement
-    if (textarea) {
-      textarea.focus()
-      textarea.scrollIntoView({ behavior: 'smooth', block: 'center' })
-    }
   }
 
-  /**
+  // ==================== UI 工具方法 ====================
-   * 处理视频加载
-   */
-  const handleVideoLoaded = (videoUrl: string): void => {
-    videoState.value.previewVideoUrl = videoUrl
-  }
 
-  // ==================== UI 辅助方法 ====================
+  function formatDuration(seconds: number): string {
-
-  /**
-   * 格式化时长
-   */
-  const formatDuration = (seconds: number): string => {
     if (!seconds) return '--:--'
-    const minutes = Math.floor(seconds / 60)
+    const mins = Math.floor(seconds / 60)
-    const remainingSeconds = Math.floor(seconds % 60)
+    const secs = Math.floor(seconds % 60)
-    return `${String(minutes).padStart(2, '0')}:${String(remainingSeconds).padStart(2, '0')}`
+    return `${String(mins).padStart(2, '0')}:${String(secs).padStart(2, '0')}`
   }
 
-  /**
+  function formatFileSize(bytes: number): string {
-   * 格式化文件大小
-   */
-  const formatFileSize = (bytes: number): string => {
     if (!bytes) return '0 B'
     const units = ['B', 'KB', 'MB', 'GB']
     let size = bytes
-    let unitIndex = 0
+    let idx = 0
-    while (size >= 1024 && unitIndex < units.length - 1) {
+    while (size >= 1024 && idx < units.length - 1) {
       size /= 1024
-      unitIndex++
+      idx++
     }
-    return `${size.toFixed(1)} ${units[unitIndex]}`
+    return `${size.toFixed(1)} ${units[idx]}`
   }
 
-  /**
+  // ==================== 返回接口 ====================
-   * 重置素材校验状态
-   */
-  const resetMaterialValidation = (): void => {
-    materialValidation.value.videoDuration = 0
-    materialValidation.value.audioDuration = 0
-    materialValidation.value.isValid = false
-  }
 
   /**
-   * 验证素材时长
+   * 包装的音频生成方法（延迟识别）
-   * 视频时长必须大于音频时长
+   * 在生成音频前先执行人脸识别
    */
-  const validateMaterialDuration = (videoDurationMs: number, audioDurationMs: number): boolean => {
+  async function generateAudio(): Promise<void> {
-    materialValidation.value.videoDuration = videoDurationMs
+    // 如果有视频但未识别，先执行识别
-    materialValidation.value.audioDuration = audioDurationMs
+    const hasVideo = digitalHuman.videoState.value.uploadedVideo || digitalHuman.videoState.value.selectedVideo
-    materialValidation.value.isValid = videoDurationMs > audioDurationMs
+    if (hasVideo && !digitalHuman.identifyState.value.identified) {
-
+      await digitalHuman.performFaceRecognition()
-    if (!materialValidation.value.isValid) {
-      const videoSec = (videoDurationMs / 1000).toFixed(1)
-      const audioSec = (audioDurationMs / 1000).toFixed(1)
-      message.warning(`素材校验失败：视频时长（${videoSec}s）必须大于音频时长（${audioSec}s）`)
     }
-
+    await voice.generateAudio()
-    return materialValidation.value.isValid
   }
 
   return {
-    // ==================== 语音生成相关 ====================
+    // 语音生成模块
-    ttsText,
+    ttsText: voice.ttsText,
-    speechRate,
+    speechRate: voice.speechRate,
-    selectedVoiceMeta,
+    selectedVoiceMeta: voice.selectedVoiceMeta,
-    audioState,
+    audioState: voice.audioState,
-    canGenerateAudio,
+    canGenerateAudio: voice.canGenerateAudio,
-    suggestedMaxChars,
+    suggestedMaxChars: voice.suggestedMaxChars,
     generateAudio,
-    resetAudioState,
+    resetAudioState: voice.resetAudioState,
 
-    // ==================== 数字人生成相关 ====================
+    // 数字人生成模块
-    videoState,
+    videoState: digitalHuman.videoState,
-    identifyState,
+    identifyState: digitalHuman.identifyState,
-    materialValidation,
+    faceDuration: digitalHuman.faceDuration,
-    faceDuration,
+    performFaceRecognition: digitalHuman.performFaceRecognition,
-    performFaceRecognition,
+    handleFileUpload: digitalHuman.handleFileUpload,
-    handleFileUpload,
+    getVideoPreviewUrl: digitalHuman.getVideoPreviewUrl,
-    getVideoPreviewUrl,
+    resetVideoState: digitalHuman.resetVideoState,
-    resetVideoState,
+    resetIdentifyState: digitalHuman.resetIdentifyState,
-    resetIdentifyState,
-    resetMaterialValidation,
-    validateMaterialDuration,
 
-    // ==================== 业务流程方法 ====================
+    // 业务方法
     generateDigitalHuman,
     replaceVideo,
 
-    // ==================== 事件处理方法 ====================
+    // 事件处理
     handleVoiceSelect,
     handleFileSelect,
     handleDrop,
@@ -418,26 +331,38 @@ export function useIdentifyFaceController(): UseIdentifyFaceController {
     handleSelectUpload,
     handleSelectFromLibrary,
     handleVideoSelect,
-    handleSimplifyScript,
     handleVideoLoaded,
 
-    // ==================== UI 辅助方法 ====================
+    // UI 工具
     formatDuration,
     formatFileSize,
 
-    // ==================== 计算属性 ====================
+    // 计算属性
     canGenerate,
     maxTextLength,
     textareaPlaceholder,
-    speechRateMarks,
+    speechRateMarks: SPEECH_RATE_MARKS,
     speechRateDisplay,
+    faceDurationSec,
+    audioDurationSec,
+    showGenerateHint,
+    audioUrl,
+    validationPassed,
+
+    // Pipeline 状态
+    pipelineState: pipeline.state,
+    pipelineStateLabel: pipeline.stateLabel,
+    pipelineStateDescription: pipeline.stateDescription,
+    isPipelineBusy: pipeline.isBusy,
+    isPipelineReady: pipeline.isReady,
+    isPipelineFailed: pipeline.isFailed,
+    isPipelineCompleted: pipeline.isCompleted,
+    pipelineProgress: pipeline.progress,
+    pipelineCurrentStepIndex: pipeline.currentStepIndex,
+    pipelineError: pipeline.error,
+    runPipeline: pipeline.run,
+    createPipelineTask: pipeline.createTask,
+    retryPipeline: pipeline.retry,
+    resetPipeline: pipeline.reset,
   }
 }
-
-/**
- * 从字符串中提取ID
- */
-function extractIdFromString(str: string): string {
-  const match = str.match(/[\w-]+$/)
-  return match ? match[0] : str
-}

frontend/app/web-gold/src/views/kling/hooks/useVoiceGeneration.ts

@@ -1,6 +1,5 @@
 /**
- * @fileoverview useVoiceGeneration Hook - 语音生成逻辑封装
+ * @fileoverview useVoiceGeneration Hook - 语音生成逻辑
- * @author Claude Code
  */
 
 import { ref, computed } from 'vue'
@@ -11,59 +10,41 @@ import type {
   VoiceMeta,
   AudioData,
 } from '../types/identify-face'
-// @ts-ignore
 import { VoiceService } from '@/api/voice'
 import { DEFAULT_VOICE_PROVIDER } from '@/config/voiceConfig'
 
-/**
+// ========== 常量 ==========
- * 语音生成 Hook
+const DEFAULT_MAX_TEXT_LENGTH = 4000
- * 独立管理所有状态，不依赖外部状态
+const DEFAULT_SPEECH_RATE = 1.0
- */
-export function useVoiceGeneration(): UseVoiceGeneration {
-  // ==================== 响应式状态 ====================
 
+export function useVoiceGeneration(): UseVoiceGeneration {
+  // ========== 状态 ==========
   const ttsText = ref<string>('')
-  const speechRate = ref<number>(1.0)
+  const speechRate = ref<number>(DEFAULT_SPEECH_RATE)
   const selectedVoiceMeta = ref<VoiceMeta | null>(null)
   const audioState = ref<AudioState>({
     generated: null,
     durationMs: 0,
-    validationPassed: false,
     generating: false,
   })
 
-  // ==================== 计算属性 ====================
+  // ========== 计算属性 ==========
-
+  const canGenerateAudio = computed(function() {
-  /**
+    return !!(ttsText.value.trim() && selectedVoiceMeta.value && !audioState.value.generating)
-   * 是否可以生成配音
-   */
-  const canGenerateAudio = computed(() => {
-    const hasText = ttsText.value.trim()
-    const hasVoice = selectedVoiceMeta.value
-    const hasVideo = true // 语音生成不依赖视频状态
-    return !!(hasText && hasVoice && hasVideo && !audioState.value.generating)
   })
 
-  /**
+  const suggestedMaxChars = computed(function() {
-   * 建议的最大字符数（需要从外部传入）
+    return DEFAULT_MAX_TEXT_LENGTH
-   */
-  const suggestedMaxChars = computed(() => {
-    // 默认为 4000，需要从外部设置
-    return 4000
   })
 
-  // ==================== 核心方法 ====================
+  // ========== 方法 ==========
 
-  /**
+  async function generateAudio(): Promise<void> {
-   * 生成配音
-   */
-  const generateAudio = async (): Promise<void> => {
     const voice = selectedVoiceMeta.value
     if (!voice) {
       message.warning('请选择音色')
       return
     }
-
     if (!ttsText.value.trim()) {
       message.warning('请输入文案内容')
       return
@@ -75,7 +56,7 @@ export function useVoiceGeneration(): UseVoiceGeneration {
       const params = {
         inputText: ttsText.value,
         voiceConfigId: voice.rawId || extractIdFromString(voice.id),
-        speechRate: speechRate.value || 1.0,
+        speechRate: speechRate.value || DEFAULT_SPEECH_RATE,
         audioFormat: 'mp3' as const,
         providerType: DEFAULT_VOICE_PROVIDER,
       }
@@ -84,144 +65,92 @@ export function useVoiceGeneration(): UseVoiceGeneration {
 
       if (res.code === 0) {
         const audioData = res.data as AudioData
-
         if (!audioData.audioBase64) {
-          throw new Error('未收到音频数据，无法进行时长解析')
+          throw new Error('未收到音频数据')
         }
 
         audioState.value.generated = audioData
-
+        audioState.value.durationMs = await parseAudioDuration(audioData.audioBase64)
-        try {
+        message.success('配音生成成功！')
-          // 解析音频时长
-          audioState.value.durationMs = await parseAudioDuration(audioData.audioBase64)
-
-          // 验证音频时长
-          validateAudioDuration()
-
-          message.success('配音生成成功！')
-        } catch (error) {
-          message.error('音频解析失败，请重新生成配音')
-          audioState.value.durationMs = 0
-          audioState.value.generated = null
-          audioState.value.validationPassed = false
-        }
       } else {
         throw new Error(res.msg || '配音生成失败')
       }
-    } catch (error: any) {
+    } catch (error: unknown) {
-      message.error(error.message || '配音生成失败')
+      const err = error as Error
+      message.error(err.message || '配音生成失败')
+      audioState.value.generated = null
+      audioState.value.durationMs = 0
     } finally {
       audioState.value.generating = false
     }
   }
 
-  /**
+  async function parseAudioDuration(base64Data: string): Promise<number> {
-   * 解析音频时长
+    const base64 = base64Data.includes(',') ? base64Data.split(',')[1] : base64Data
-   */
+    const binaryString = window.atob(base64)
-  const parseAudioDuration = async (base64Data: string): Promise<number> => {
+    const bytes = new Uint8Array(binaryString.length)
-    return new Promise((resolve, reject) => {
+    for (let i = 0; i < binaryString.length; i++) {
-      try {
+      bytes[i] = binaryString.charCodeAt(i)
-        const base64 = base64Data.includes(',') ? base64Data.split(',')[1] : base64Data
+    }
 
-        const binaryString = window.atob(base64)
+    return new Promise(function(resolve, reject) {
-        const bytes = new Uint8Array(binaryString.length)
+      const blob = new Blob([bytes], { type: 'audio/mp3' })
-        for (let i = 0; i < binaryString.length; i++) {
+      const audio = new Audio()
-          bytes[i] = binaryString.charCodeAt(i)
+      const objectUrl = URL.createObjectURL(blob)
+
+      // 超时机制：5秒后拒绝
+      const timeoutId = setTimeout(function() {
+        URL.revokeObjectURL(objectUrl)
+        reject(new Error('音频时长解析超时'))
+      }, 5000)
+
+      function onLoadedMetadata() {
+        clearTimeout(timeoutId)
+        URL.revokeObjectURL(objectUrl)
+
+        const duration = audio.duration
+        if (!isFinite(duration) || duration <= 0 || isNaN(duration)) {
+          reject(new Error(`音频时长无效: ${duration}，请检查音频格式是否正确`))
+          return
         }
 
-        const blob = new Blob([bytes], { type: 'audio/mp3' })
+        const durationMs = Math.round(duration * 1000)
-        const audio = new Audio()
+        console.log('[parseAudioDuration] 音频时长解析成功:', durationMs, 'ms')
-        const objectUrl = URL.createObjectURL(blob)
+        resolve(durationMs)
-
-        audio.addEventListener('loadedmetadata', () => {
-          URL.revokeObjectURL(objectUrl)
-          const durationMs = Math.round(audio.duration * 1000)
-          resolve(durationMs)
-        })
-
-        audio.addEventListener('error', (error) => {
-          URL.revokeObjectURL(objectUrl)
-          reject(error)
-        })
-
-        audio.src = objectUrl
-        audio.load()
-      } catch (error) {
-        reject(error)
       }
+
+      function onError() {
+        clearTimeout(timeoutId)
+        URL.revokeObjectURL(objectUrl)
+        reject(new Error('音频解析失败，请检查音频格式'))
+      }
+
+      audio.addEventListener('loadedmetadata', onLoadedMetadata)
+      audio.addEventListener('error', onError)
+      audio.src = objectUrl
+      audio.load()
     })
   }
 
-  /**
+  function resetAudioState(): void {
-   * 验证音频与人脸区间的重合时长（外部调用时传入校验参数）
-   */
-  const validateAudioDuration = (
-    faceStartTime: number = 0,
-    faceEndTime: number = 0,
-    minOverlapMs: number = 2000
-  ): boolean => {
-    if (faceStartTime <= 0 || faceEndTime <= 0) {
-      audioState.value.validationPassed = false
-      return false
-    }
-
-    const faceDurationMs = faceEndTime - faceStartTime
-    const audioDuration = audioState.value.durationMs
-
-    const overlapStart = faceStartTime
-    const overlapEnd = Math.min(faceEndTime, faceStartTime + audioDuration)
-    const overlapDuration = Math.max(0, overlapEnd - overlapStart)
-
-    const isValid = overlapDuration >= minOverlapMs
-
-    audioState.value.validationPassed = isValid
-
-    if (!isValid) {
-      const overlapSec = (overlapDuration / 1000).toFixed(1)
-      message.warning(
-        `音频时长（${(audioDuration/1000).toFixed(1)}秒）与人脸区间（${(faceDurationMs/1000).toFixed(1)}秒）不匹配，重合部分仅${overlapSec}秒，至少需要${(minOverlapMs/1000)}秒`
-      )
-    } else {
-      message.success('时长校验通过！')
-    }
-
-    return isValid
-  }
-
-  /**
-   * 重置音频状态
-   */
-  const resetAudioState = (): void => {
     audioState.value.generated = null
     audioState.value.durationMs = 0
-    audioState.value.validationPassed = false
     audioState.value.generating = false
   }
 
   return {
-    // 响应式状态
     ttsText,
     speechRate,
     selectedVoiceMeta,
     audioState,
-
-    // 计算属性
     canGenerateAudio,
     suggestedMaxChars,
-
-    // 方法
     generateAudio,
-    parseAudioDuration,
-    validateAudioDuration,
     resetAudioState,
   }
 }
 
-/**
- * 从字符串中提取ID
- */
 function extractIdFromString(str: string): string {
-  // 尝试从各种格式中提取ID
   const match = str.match(/[\w-]+$/)
   return match ? match[0] : str
 }

frontend/app/web-gold/src/views/kling/types/identify-face.ts

@@ -11,6 +11,7 @@ export interface VideoState {
   videoFile: File | null
   previewVideoUrl: string
   selectedVideo: Video | null
+  fileId: string | number | null
   videoSource: 'upload' | 'select' | null
   selectorVisible: boolean
 }
@@ -20,6 +21,7 @@ export interface VideoState {
  */
 export interface Video {
   id: string | number
+  fileId: string | number
   fileName: string
   fileUrl: string
   fileSize: number
@@ -48,7 +50,6 @@ export interface IdentifyState {
 export interface AudioState {
   generated: AudioData | null
   durationMs: number
-  validationPassed: boolean
   generating: boolean
 }
 
@@ -61,16 +62,6 @@ export interface AudioData {
   format?: string
 }
 
-/**
- * 素材校验接口
- */
-export interface MaterialValidation {
-  videoDuration: number
-  audioDuration: number
-  isValid: boolean
-  showDetails: boolean
-}
-
 /**
  * 音色元数据接口
  */
@@ -97,8 +88,6 @@ export interface UseVoiceGeneration {
 
   // 方法
   generateAudio: () => Promise<void>
-  parseAudioDuration: (base64Data: string) => Promise<number>
-  validateAudioDuration: () => boolean
   resetAudioState: () => void
 }
 
@@ -115,7 +104,7 @@ export interface UseDigitalHumanGeneration {
 
   // 方法
   handleFileUpload: (file: File) => Promise<void>
-  handleVideoSelect: (video: Video) => void
+  handleVideoSelect: (video: Video) => Promise<void>
   performFaceRecognition: () => Promise<void>
   resetVideoState: () => void
   resetIdentifyState: () => void
@@ -140,29 +129,25 @@ export interface UseIdentifyFaceController {
   // ==================== 数字人生成相关 ====================
   videoState: import('vue').Ref<VideoState>
   identifyState: import('vue').Ref<IdentifyState>
-  materialValidation: import('vue').Ref<MaterialValidation>
   faceDuration: import('vue').ComputedRef<number>
   performFaceRecognition: () => Promise<void>
   handleFileUpload: (file: File) => Promise<void>
   getVideoPreviewUrl: (video: Video) => string
   resetVideoState: () => void
   resetIdentifyState: () => void
-  resetMaterialValidation: () => void
-  validateMaterialDuration: (videoDurationMs: number, audioDurationMs: number) => boolean
 
   // ==================== 业务流程方法 ====================
   generateDigitalHuman: () => Promise<void>
   replaceVideo: () => void
 
   // ==================== 事件处理方法 ====================
-  handleVoiceSelect: (voice: VoiceMeta) => void
+  handleVoiceSelect: (voiceMeta: VoiceMeta) => void
   handleFileSelect: (event: Event) => void
   handleDrop: (event: DragEvent) => void
   triggerFileSelect: () => void
   handleSelectUpload: () => void
   handleSelectFromLibrary: () => void
-  handleVideoSelect: (video: Video) => void
+  handleVideoSelect: (video: Video) => Promise<void>
-  handleSimplifyScript: () => void
   handleVideoLoaded: (videoUrl: string) => void
 
   // ==================== 计算属性 ====================
@@ -171,6 +156,32 @@ export interface UseIdentifyFaceController {
   textareaPlaceholder: import('vue').ComputedRef<string>
   speechRateMarks: Record<number, string>
   speechRateDisplay: import('vue').ComputedRef<string>
+  faceDurationSec: import('vue').ComputedRef<string>
+  audioDurationSec: import('vue').ComputedRef<string>
+  showGenerateHint: import('vue').ComputedRef<boolean>
+  audioUrl: import('vue').ComputedRef<string>
+  validationPassed: import('vue').ComputedRef<boolean>
+
+  // ==================== 流程状态 ====================
+  pipelineState: import('vue').Ref<string>
+  pipelineStateLabel: import('vue').ComputedRef<string>
+  pipelineStateDescription: import('vue').ComputedRef<string>
+  isPipelineBusy: import('vue').ComputedRef<boolean>
+  isPipelineReady: import('vue').ComputedRef<boolean>
+  isPipelineFailed: import('vue').ComputedRef<boolean>
+  isPipelineCompleted: import('vue').ComputedRef<boolean>
+  pipelineProgress: import('vue').ComputedRef<number>
+  pipelineError: import('vue').Ref<string | null>
+  runPipeline: (params: {
+    videoFile: File | null
+    selectedVideo: any
+    text: string
+    voice: VoiceMeta
+    speechRate: number
+  }) => Promise<void>
+  createPipelineTask: () => Promise<void>
+  retryPipeline: () => void
+  resetPipeline: () => void
 
   // ==================== UI 辅助方法 ====================
   formatDuration: (seconds: number) => string

yudao-module-tik/src/main/java/cn/iocoder/yudao/module/tik/file/service/TikUserFileServiceImpl.java

@@ -445,12 +445,12 @@ public class TikUserFileServiceImpl implements TikUserFileService {
     }
 
     @Override
-    public String getVideoPlayUrl(Long infraFileId) {
+    public String getVideoPlayUrl(Long fileId) {
         Long userId = SecurityFrameworkUtils.getLoginUserId();
 
-        // 查询文件（根据 infraFileId 字段查询）
+        // 查询文件（根据 fileId 字段查询）
         TikUserFileDO file = userFileMapper.selectOne(new LambdaQueryWrapperX<TikUserFileDO>()
-                .eq(TikUserFileDO::getFileId, infraFileId)
+                .eq(TikUserFileDO::getFileId, fileId)
                 .eq(TikUserFileDO::getUserId, userId));
 
         if (file == null) {
@@ -473,12 +473,12 @@ public class TikUserFileServiceImpl implements TikUserFileService {
     }
 
     @Override
-    public String getAudioPlayUrl(Long infraFileId) {
+    public String getAudioPlayUrl(Long fileId) {
         Long userId = SecurityFrameworkUtils.getLoginUserId();
 
-        // 查询文件（根据 infraFileId 字段查询）
+        // 查询文件（根据 fileId 字段查询）
         TikUserFileDO file = userFileMapper.selectOne(new LambdaQueryWrapperX<TikUserFileDO>()
-                .eq(TikUserFileDO::getFileId, infraFileId)
+                .eq(TikUserFileDO::getFileId, fileId)
                 .eq(TikUserFileDO::getUserId, userId));
 
         if (file == null) {

yudao-module-tik/src/main/java/cn/iocoder/yudao/module/tik/file/vo/app/AppTikUserFileRespVO.java

@@ -14,9 +14,12 @@ import java.time.LocalDateTime;
 @Data
 public class AppTikUserFileRespVO {
 
-    @Schema(description = "文件编号", requiredMode = Schema.RequiredMode.REQUIRED, example = "1")
+    @Schema(description = "文件编号（主键）", requiredMode = Schema.RequiredMode.REQUIRED, example = "1")
     private Long id;
 
+    @Schema(description = "素材文件编号（关联 infra_file.id，用于获取播放URL）", example = "100")
+    private Long fileId;
+
     @Schema(description = "文件名称", requiredMode = Schema.RequiredMode.REQUIRED, example = "test.mp4")
     private String fileName;
 

yudao-module-tik/src/main/java/cn/iocoder/yudao/module/tik/kling/client/KlingClient.java

@@ -52,34 +52,12 @@ public class KlingClient {
         validateRequest(request);
 
         Map<String, Object> payload = buildPayload(request);
-        try {
+        String url = properties.getBaseUrl() + "/klingai/v1/videos/identify-face";
-            String body = objectMapper.writeValueAsString(payload);
+        Request httpRequest = buildPostRequest(url, payload);
-            String url = properties.getBaseUrl() + "/klingai/v1/videos/identify-face";
 
-            Request httpRequest = new Request.Builder()
+        KlingIdentifyFaceResponse response = executeRequest(httpRequest, "identify-face", KlingIdentifyFaceResponse.class);
-                    .url(url)
+        validateSessionId(response.getData() != null ? response.getData().getSessionId() : null, "sessionId");
-                    .addHeader("Authorization", "Bearer " + properties.getApiKey())
+        return response;
-                    .addHeader("Content-Type", "application/json")
-                    .post(RequestBody.create(body.getBytes(StandardCharsets.UTF_8), JSON))
-                    .build();
-
-            try {
-                KlingIdentifyFaceResponse response = executeRequest(httpRequest, "identify-face", KlingIdentifyFaceResponse.class);
-                // 验证sessionId
-                if (StrUtil.isBlank(response.getData() == null ? null : response.getData().getSessionId())) {
-                    throw exception0(LATENTSYNC_SUBMIT_FAILED.getCode(), "可灵返回 sessionId 为空");
-                }
-                return response;
-            } catch (ServiceException ex) {
-                throw ex;
-            } catch (Exception ex) {
-                log.error("[Kling][identify-face exception]", ex);
-                throw exception(LATENTSYNC_SUBMIT_FAILED);
-            }
-        } catch (Exception ex) {
-            log.error("[Kling][build request exception]", ex);
-            throw exception(LATENTSYNC_SUBMIT_FAILED);
-        }
     }
 
     /**
@@ -89,35 +67,13 @@ public class KlingClient {
         validateEnabled();
         validateLipSyncRequest(request);
 
-        try {
+        String url = properties.getBaseUrl() + "/klingai/v1/videos/advanced-lip-sync";
-            String body = objectMapper.writeValueAsString(request);
+        Request httpRequest = buildPostRequest(url, request);
-            log.info("[Kling][create-lip-sync请求体] {}", body);
+        log.info("[Kling][create-lip-sync请求体] {}", request);
-            String url = properties.getBaseUrl() + "/klingai/v1/videos/advanced-lip-sync";
 
-            Request httpRequest = new Request.Builder()
+        KlingLipSyncCreateResponse response = executeRequest(httpRequest, "create-lip-sync", KlingLipSyncCreateResponse.class);
-                    .url(url)
+        validateSessionId(response.getData() != null ? response.getData().getTaskId() : null, "taskId");
-                    .addHeader("Authorization", "Bearer " + properties.getApiKey())
+        return response;
-                    .addHeader("Content-Type", "application/json")
-                    .post(RequestBody.create(body.getBytes(StandardCharsets.UTF_8), JSON))
-                    .build();
-
-            try {
-                KlingLipSyncCreateResponse response = executeRequest(httpRequest, "create-lip-sync", KlingLipSyncCreateResponse.class);
-                // 验证taskId
-                if (StrUtil.isBlank(response.getData() == null ? null : response.getData().getTaskId())) {
-                    throw exception0(LATENTSYNC_SUBMIT_FAILED.getCode(), "可灵返回 taskId 为空");
-                }
-                return response;
-            } catch (ServiceException ex) {
-                throw ex;
-            } catch (Exception ex) {
-                log.error("[Kling][create-lip-sync exception]", ex);
-                throw exception(LATENTSYNC_SUBMIT_FAILED);
-            }
-        } catch (Exception ex) {
-            log.error("[Kling][build request exception]", ex);
-            throw exception(LATENTSYNC_SUBMIT_FAILED);
-        }
     }
 
     /**
@@ -129,29 +85,10 @@ public class KlingClient {
             throw exception0(LATENTSYNC_SUBMIT_FAILED.getCode(), "任务ID不能为空");
         }
 
-        try {
+        String url = properties.getBaseUrl() + "/klingai/v1/videos/advanced-lip-sync/" + taskId;
-            String url = properties.getBaseUrl() + "/klingai/v1/videos/advanced-lip-sync/" + taskId;
+        Request httpRequest = buildGetRequest(url);
 
-            Request httpRequest = new Request.Builder()
+        return executeRequest(httpRequest, "get-lip-sync", KlingLipSyncQueryResponse.class);
-                    .url(url)
-                    .addHeader("Authorization", "Bearer " + properties.getApiKey())
-                    .addHeader("Content-Type", "application/json")
-                    .get()
-                    .build();
-
-            try {
-                KlingLipSyncQueryResponse response = executeRequest(httpRequest, "get-lip-sync", KlingLipSyncQueryResponse.class);
-                return response;
-            } catch (ServiceException ex) {
-                throw ex;
-            } catch (Exception ex) {
-                log.error("[Kling][get-lip-sync exception]", ex);
-                throw exception(LATENTSYNC_SUBMIT_FAILED);
-            }
-        } catch (Exception ex) {
-            log.error("[Kling][build request exception]", ex);
-            throw exception(LATENTSYNC_SUBMIT_FAILED);
-        }
     }
 
     private void validateEnabled() {
@@ -272,10 +209,39 @@ public class KlingClient {
                     throw buildException(responseBody);
                 }
             }
+
+            // 解析响应
+            T result = objectMapper.readValue(responseBody, responseClass);
+
+            // ✅ 检查业务错误码（可灵 API 可能返回 HTTP 200 但 code !== 0）
+            try {
+                JsonNode root = objectMapper.readTree(responseBody);
+                if (root.has("code")) {
+                    int code = root.get("code").asInt();
+                    if (code != 0) {
+                        String message = root.has("message") ? root.get("message").asText() :
+                                       root.has("detail") ? root.get("detail").asText() : "未知错误";
+                        String requestId = root.has("request_id") ? root.get("request_id").asText() : "unknown";
+
+                        log.error("[Kling][{} business error] code={}, message={}, request_id={}", operation, code, message, requestId);
+
+                        throw exception0(LATENTSYNC_SUBMIT_FAILED.getCode(),
+                            String.format("[%s] %s (code: %d, request_id: %s)", operation, message, code, requestId));
+                    }
+                }
+            } catch (ServiceException ex) {
+                throw ex;
+            } catch (Exception ex) {
+                log.warn("[Kling][{} check business code failed, continuing]", operation, ex);
+            }
+
             log.info("[Kling][{} success][responseBody={}]", operation, responseBody);
 
-            return objectMapper.readValue(responseBody, responseClass);
+            return result;
         } catch (Exception ex) {
+            if (ex instanceof ServiceException) {
+                throw (ServiceException) ex;
+            }
             log.error("[Kling][{} exception]", operation, ex);
             throw exception(LATENTSYNC_SUBMIT_FAILED);
         }
@@ -319,4 +285,43 @@ public class KlingClient {
             return exception0(LATENTSYNC_SUBMIT_FAILED.getCode(), body);
         }
     }
+
+    /**
+     * 构建 POST 请求
+     */
+    private Request buildPostRequest(String url, Object payload) {
+        try {
+            String body = objectMapper.writeValueAsString(payload);
+            return new Request.Builder()
+                    .url(url)
+                    .addHeader("Authorization", "Bearer " + properties.getApiKey())
+                    .addHeader("Content-Type", "application/json")
+                    .post(RequestBody.create(body.getBytes(StandardCharsets.UTF_8), JSON))
+                    .build();
+        } catch (Exception ex) {
+            log.error("[Kling][build POST request exception]", ex);
+            throw exception(LATENTSYNC_SUBMIT_FAILED);
+        }
+    }
+
+    /**
+     * 构建 GET 请求
+     */
+    private Request buildGetRequest(String url) {
+        return new Request.Builder()
+                .url(url)
+                .addHeader("Authorization", "Bearer " + properties.getApiKey())
+                .addHeader("Content-Type", "application/json")
+                .get()
+                .build();
+    }
+
+    /**
+     * 验证响应中的 sessionId 不为空
+     */
+    private void validateSessionId(String sessionId, String fieldName) {
+        if (StrUtil.isBlank(sessionId)) {
+            throw exception0(LATENTSYNC_SUBMIT_FAILED.getCode(), "可灵返回 " + fieldName + " 为空");
+        }
+    }
 }

yudao-module-tik/src/main/java/cn/iocoder/yudao/module/tik/voice/dal/dataobject/TikDigitalHumanTaskDO.java

@@ -3,6 +3,7 @@ package cn.iocoder.yudao.module.tik.voice.dal.dataobject;
 import cn.iocoder.yudao.framework.common.enums.CommonStatusEnum;
 import cn.iocoder.yudao.framework.tenant.core.db.TenantBaseDO;
 import com.baomidou.mybatisplus.annotation.KeySequence;
+import com.baomidou.mybatisplus.annotation.TableField;
 import com.baomidou.mybatisplus.annotation.TableId;
 import com.baomidou.mybatisplus.annotation.TableName;
 import lombok.*;
@@ -163,6 +164,7 @@ public class TikDigitalHumanTaskDO extends TenantBaseDO {
     /**
      * 可灵口型同步任务ID（从advanced-lip-sync接口获取）
      */
+    @TableField("kling_task_id")
     private String klingTaskId;
 
 }

yudao-module-tik/src/main/java/cn/iocoder/yudao/module/tik/voice/dal/mysql/TikDigitalHumanTaskMapper.java

@@ -4,9 +4,11 @@ import cn.iocoder.yudao.framework.common.pojo.PageParam;
 import cn.iocoder.yudao.framework.common.pojo.PageResult;
 import cn.iocoder.yudao.framework.mybatis.core.mapper.BaseMapperX;
 import cn.iocoder.yudao.framework.mybatis.core.query.LambdaQueryWrapperX;
+import cn.iocoder.yudao.framework.tenant.core.aop.TenantIgnore;
 import cn.iocoder.yudao.module.tik.voice.dal.dataobject.TikDigitalHumanTaskDO;
 import cn.iocoder.yudao.module.tik.voice.vo.AppTikDigitalHumanPageReqVO;
 import org.apache.ibatis.annotations.Mapper;
+import org.apache.ibatis.annotations.Select;
 
 import java.util.List;
 
@@ -62,4 +64,20 @@ public interface TikDigitalHumanTaskMapper extends BaseMapperX<TikDigitalHumanTa
                 .last("LIMIT " + limit));
     }
 
+    /**
+     * 查询待轮询的可灵任务（状态为PROCESSING且有klingTaskId，最近6小时内）
+     * 使用条件构造器，配合服务类 @TenantIgnore 注解忽略租户限制
+     */
+    @TenantIgnore
+    default List<TikDigitalHumanTaskDO> selectPendingKlingTasks() {
+        return selectList(new LambdaQueryWrapperX<TikDigitalHumanTaskDO>()
+                .eq(TikDigitalHumanTaskDO::getStatus, "PROCESSING")
+                .eq(TikDigitalHumanTaskDO::getAiProvider, "kling")
+                .isNotNull(TikDigitalHumanTaskDO::getKlingTaskId)
+                .ne(TikDigitalHumanTaskDO::getKlingTaskId, "")
+                .apply("create_time >= DATE_SUB(NOW(), INTERVAL 6 HOUR)")
+                .orderByDesc(TikDigitalHumanTaskDO::getCreateTime)
+                .last("LIMIT 50"));
+    }
+
 }

yudao-module-tik/src/main/java/cn/iocoder/yudao/module/tik/voice/job/DigitalHumanTaskStatusSyncJob.java

@@ -23,11 +23,11 @@ public class DigitalHumanTaskStatusSyncJob {
      */
     @Scheduled(fixedDelay = 10000)
     public void syncTaskStatus() {
-        log.debug("开始同步数字人任务状态");
+        log.info("[DigitalHumanTaskStatusSyncJob][开始同步数字人任务状态]");
         try {
             latentsyncPollingService.pollLatentsyncTasks();
         } catch (Exception e) {
-            log.error("同步数字人任务状态失败", e);
+            log.error("[DigitalHumanTaskStatusSyncJob][同步数字人任务状态失败]", e);
         }
     }
 

yudao-module-tik/src/main/java/cn/iocoder/yudao/module/tik/voice/service/LatentsyncPollingService.java

@@ -82,11 +82,14 @@ public class LatentsyncPollingService {
      * 执行轮询任务的具体逻辑
      */
     private void executePollingTasks() {
+        log.info("[executePollingTasks][开始执行轮询任务]");
         try {
             // 轮询Latentsync任务
             List<String> taskIds = getPendingPollingTasks();
+            log.info("[executePollingTasks][获取到Latentsync任务数量={}]", taskIds.size());
+
             if (!taskIds.isEmpty()) {
-                log.debug("[pollLatentsyncTasks][开始轮询Latentsync任务][任务数量={}]", taskIds.size());
+                log.info("[pollLatentsyncTasks][开始轮询Latentsync任务][任务数量={}]", taskIds.size());
 
                 // 逐个处理Latentsync任务
                 for (String taskIdStr : taskIds) {
@@ -100,7 +103,9 @@ public class LatentsyncPollingService {
             }
 
             // 轮询可灵任务
+            log.info("[executePollingTasks][准备轮询可灵任务]");
             pollKlingTasks();
+            log.info("[executePollingTasks][可灵任务轮询完成]");
 
         } catch (Exception e) {
             log.error("[pollLatentsyncTasks][轮询任务异常]", e);
@@ -536,28 +541,19 @@ public class LatentsyncPollingService {
      */
     private void pollKlingTasks() {
         try {
-            // 参考混剪任务实现：添加时间和数量限制，避免并发问题
+            log.info("[pollKlingTasks][开始查询待轮询的可灵任务]");
-            // 1. 时间范围限制：只检查最近6小时内的任务（避免检查历史任务）
-            // 2. 数量限制：每次最多检查50个任务（避免单次查询过多）
-            LocalDateTime startTime = LocalDateTime.now().minusHours(6);
 
-            // 查询有待轮询的可灵任务（状态为PROCESSING且有klingTaskId，限制时间和数量）
+            // 显式忽略租户限制查询待轮询的可灵任务
-            List<TikDigitalHumanTaskDO> klingTasks = taskMapper.selectList(
+            List<TikDigitalHumanTaskDO> klingTasks = cn.iocoder.yudao.framework.tenant.core.util.TenantUtils.executeIgnore(
-                    new cn.iocoder.yudao.framework.mybatis.core.query.LambdaQueryWrapperX<TikDigitalHumanTaskDO>()
+                    () -> taskMapper.selectPendingKlingTasks()
-                            .eq(TikDigitalHumanTaskDO::getStatus, "PROCESSING")
-                            .eq(TikDigitalHumanTaskDO::getAiProvider, "kling")
-                            .isNotNull(TikDigitalHumanTaskDO::getKlingTaskId)
-                            .ne(TikDigitalHumanTaskDO::getKlingTaskId, "")
-                            .ge(TikDigitalHumanTaskDO::getCreateTime, startTime)  // 只检查最近6小时
-                            .orderByDesc(TikDigitalHumanTaskDO::getCreateTime)
-                            .last("LIMIT 50") // 限制数量，避免并发
             );
 
             if (klingTasks.isEmpty()) {
+                log.info("[pollKlingTasks][没有待轮询的可灵任务]");
                 return;
             }
 
-            log.debug("[pollKlingTasks][开始轮询可灵任务][任务数量={}]", klingTasks.size());
+            log.info("[pollKlingTasks][开始轮询可灵任务][任务数量={}]", klingTasks.size());
 
             // 逐个处理可灵任务
             for (TikDigitalHumanTaskDO task : klingTasks) {
@@ -589,53 +585,12 @@ public class LatentsyncPollingService {
             String taskStatus = response.getData().getTaskStatus();
             String taskStatusMsg = response.getData().getTaskStatusMsg();
 
-            log.debug("[pollKlingSingleTask][任务({})状态更新][klingTaskId={}, status={}]",
+            log.info("[pollKlingSingleTask][任务({})状态更新][klingTaskId={}, status={}, msg={}]",
-                    task.getId(), klingTaskId, taskStatus);
+                    task.getId(), klingTaskId, taskStatus, taskStatusMsg);
 
             // 根据状态更新任务
             if ("succeed".equalsIgnoreCase(taskStatus)) {
-                // 任务成功完成
+                handleKlingTaskSucceed(task, response);
-                List<KlingLipSyncVideoVO> videos = response.getData().getTaskResult().getVideos();
-                if (videos != null && !videos.isEmpty()) {
-                    String videoUrl = videos.get(0).getUrl();
-
-                    // 保存视频到OSS（异步处理，轻量化逻辑）
-                    OssSaveResult saveResult = null;
-                    try {
-                        // 保存视频到OSS，避免临时URL过期
-                        saveResult = saveVideoToOss(task, videoUrl);
-                        log.info("[pollKlingSingleTask][任务({})视频已保存到OSS][url={}]", task.getId(), saveResult.getUrl());
-                    } catch (Exception e) {
-                        log.warn("[pollKlingSingleTask][任务({})保存视频失败，使用原URL][error={}]", task.getId(), e.getMessage());
-                        saveResult = new OssSaveResult(videoUrl, 0, null, null); // 降级处理
-                    }
-
-                    // 更新任务状态为成功
-                    TikDigitalHumanTaskDO updateObj = new TikDigitalHumanTaskDO();
-                    updateObj.setId(task.getId());
-                    updateObj.setStatus("SUCCESS");
-                    updateObj.setCurrentStep("finishing");
-                    updateObj.setProgress(100);
-                    updateObj.setResultVideoUrl(saveResult.getUrl());
-                    updateObj.setFinishTime(LocalDateTime.now());
-                    taskMapper.updateById(updateObj);
-
-                    // 缓存结果到Redis（快速回显）
-                    try {
-                        String resultKey = "digital_human:task:result:" + task.getId();
-                        stringRedisTemplate.opsForValue().set(resultKey, saveResult.getUrl(), Duration.ofHours(24));
-                    } catch (Exception e) {
-                        log.warn("[pollKlingSingleTask][任务({})缓存结果失败]", task.getId(), e);
-                    }
-
-                    // 保存结果视频到用户文件表
-                    saveResultVideoToUserFiles(task, saveResult);
-
-                    log.info("[pollKlingSingleTask][任务({})完成][videoUrl={}]", task.getId(), saveResult.getUrl());
-                } else {
-                    log.warn("[pollKlingSingleTask][任务({})成功但无视频结果]", task.getId());
-                }
-
             } else if ("failed".equalsIgnoreCase(taskStatus)) {
                 // 任务失败
                 String errorMsg = "可灵任务执行失败: " + (StrUtil.isNotBlank(taskStatusMsg) ? taskStatusMsg : "未知错误");
@@ -645,7 +600,7 @@ public class LatentsyncPollingService {
             } else if ("submitted".equalsIgnoreCase(taskStatus) || "processing".equalsIgnoreCase(taskStatus)) {
                 // 任务还在处理中，更新进度
                 updateTaskStatus(task.getId(), "PROCESSING", "sync_lip", 70, "口型同步处理中", null);
-                log.debug("[pollKlingSingleTask][任务({})处理中]", task.getId());
+                log.info("[pollKlingSingleTask][任务({})处理中][klingTaskId={}, status={}]", task.getId(), klingTaskId, taskStatus);
 
             } else {
                 log.warn("[pollKlingSingleTask][任务({})未知状态][status={}]", task.getId(), taskStatus);
@@ -687,7 +642,64 @@ public class LatentsyncPollingService {
             updateObj.setFinishTime(LocalDateTime.now());
         }
 
+        // 显式忽略租户限制执行更新操作
+        cn.iocoder.yudao.framework.tenant.core.util.TenantUtils.executeIgnore(
+                () -> taskMapper.updateById(updateObj)
+        );
+    }
+
+    /**
+     * 处理可灵任务成功
+     */
+    private void handleKlingTaskSucceed(TikDigitalHumanTaskDO task, KlingLipSyncQueryResponse response) {
+        List<KlingLipSyncVideoVO> videos = response.getData().getTaskResult().getVideos();
+        if (videos == null || videos.isEmpty()) {
+            log.warn("[pollKlingSingleTask][任务({})成功但无视频结果]", task.getId());
+            return;
+        }
+
+        String videoUrl = videos.get(0).getUrl();
+        OssSaveResult saveResult;
+
+        try {
+            saveResult = saveVideoToOss(task, videoUrl);
+            log.info("[pollKlingSingleTask][任务({})视频已保存到OSS][url={}]", task.getId(), saveResult.getUrl());
+        } catch (Exception e) {
+            log.warn("[pollKlingSingleTask][任务({})保存视频失败，使用原URL][error={}]", task.getId(), e.getMessage());
+            saveResult = new OssSaveResult(videoUrl, 0, null, null);
+        }
+
+        updateTaskSuccess(task.getId(), saveResult.getUrl());
+        cacheTaskResult(task.getId(), saveResult.getUrl());
+        saveResultVideoToUserFiles(task, saveResult);
+
+        log.info("[pollKlingSingleTask][任务({})完成][videoUrl={}]", task.getId(), saveResult.getUrl());
+    }
+
+    /**
+     * 更新任务为成功状态
+     */
+    private void updateTaskSuccess(Long taskId, String videoUrl) {
+        TikDigitalHumanTaskDO updateObj = new TikDigitalHumanTaskDO();
+        updateObj.setId(taskId);
+        updateObj.setStatus("SUCCESS");
+        updateObj.setCurrentStep("finishing");
+        updateObj.setProgress(100);
+        updateObj.setResultVideoUrl(videoUrl);
+        updateObj.setFinishTime(LocalDateTime.now());
         taskMapper.updateById(updateObj);
     }
 
+    /**
+     * 缓存任务结果到 Redis
+     */
+    private void cacheTaskResult(Long taskId, String url) {
+        try {
+            String resultKey = "digital_human:task:result:" + taskId;
+            stringRedisTemplate.opsForValue().set(resultKey, url, Duration.ofHours(24));
+        } catch (Exception e) {
+            log.warn("[cacheTaskResult][任务({})缓存结果失败]", taskId, e);
+        }
+    }
+
 }