From 4771397301924fd34d9d42e1cd6b5f400e7f2cf5 Mon Sep 17 00:00:00 2001
From: "fengchuanhn@gmail.com" <fengchuanhn@gmail.com>
Date: Mon, 18 May 2026 14:12:57 +0800
Subject: [PATCH] 11

---
 src-tauri/src/commands/fs_util.rs            |  1 +
 src/components/workflow/Step02AudioVideo.vue | 78 +++++++++++++--
 src/services/localAudio.js                   | 20 ++++
 src/services/voiceGenerate.js                | 99 ++++++++++++++++++++
 src/services/voicePreview.js                 | 23 +----
 src/stores/workflow.js                       | 27 ++++++
 src/utils/ttsOptions.js                      | 71 ++++++++++++++
 7 files changed, 292 insertions(+), 27 deletions(-)
 create mode 100644 src/services/localAudio.js
 create mode 100644 src/services/voiceGenerate.js
 create mode 100644 src/utils/ttsOptions.js

diff --git a/src-tauri/src/commands/fs_util.rs b/src-tauri/src/commands/fs_util.rs
index 535fb14..2bf6a29 100644
--- a/src-tauri/src/commands/fs_util.rs
+++ b/src-tauri/src/commands/fs_util.rs
@@ -15,6 +15,7 @@ fn is_allowed_local_media(path: &Path) -> bool {
     let lower = path.to_string_lossy().to_lowercase();
     if lower.contains("aiclient-voice-preview-cache")
         || lower.contains("aiclient-node-pipeline")
+        || lower.contains("aiclient-generated-speech")
     {
         return true;
     }
diff --git a/src/components/workflow/Step02AudioVideo.vue b/src/components/workflow/Step02AudioVideo.vue
index ffd093e..87edbdf 100644
--- a/src/components/workflow/Step02AudioVideo.vue
+++ b/src/components/workflow/Step02AudioVideo.vue
@@ -1,17 +1,53 @@
 <script setup>
-import { computed } from "vue";
+import { computed, ref } from "vue";
 import { storeToRefs } from "pinia";
 import { useWorkflowStore } from "../../stores/workflow.js";
 import VoiceManageDialog from "./VoiceManageDialog.vue";
 
 const workflow = useWorkflowStore();
-const { voiceEmotion, speechRate, voiceLanguage, avatarSelect } = storeToRefs(workflow);
+const {
+  voiceEmotion,
+  speechRate,
+  voiceLanguage,
+  avatarSelect,
+  speechGenerating,
+  generatedAudioSrc,
+  currentAudioId,
+  audioHistory,
+} = storeToRefs(workflow);
+
+const feedback = ref({ severity: "", message: "" });
 
 const voiceButtonLabel = computed(() => workflow.selectedVoiceLabel);
 
+const audioHistoryOptions = computed(() =>
+  audioHistory.value.map((item) => ({
+    label: item.label,
+    value: item.id,
+  })),
+);
+
+const hasAudio = computed(() => Boolean(generatedAudioSrc.value));
+
 function onOpenVoiceManage() {
   workflow.openVoiceManageDialog();
 }
+
+async function onGenerateSpeech() {
+  feedback.value = { severity: "", message: "" };
+  const result = await workflow.generateSpeech();
+  if (result?.message) {
+    feedback.value = {
+      severity: result.ok ? "success" : "error",
+      message: result.message,
+    };
+  }
+}
+
+function onPickAudioHistory() {
+  if (!currentAudioId.value) return;
+  workflow.pickAudioHistory(currentAudioId.value);
+}
 </script>
 
 <template>
@@ -60,14 +96,44 @@ function onOpenVoiceManage() {
           class="w-full"
         />
       </div>
-      <Button label="生成语音" size="small" class="shrink-0" />
+      <Button
+        label="生成语音"
+        size="small"
+        class="shrink-0"
+        :loading="speechGenerating"
+        :disabled="speechGenerating"
+        @click="onGenerateSpeech"
+      />
     </div>
+    <p
+      v-if="feedback.message"
+      class="mt-2 text-sm"
+      :class="feedback.severity === 'error' ? 'text-red-400' : 'text-emerald-400'"
+    >
+      {{ feedback.message }}
+    </p>
     <div class="dashboard-preview-box mt-3">
       <div class="mb-2 flex items-center justify-between gap-2">
         <span class="text-xs text-slate-400">音频预览</span>
-        <Select placeholder="暂无历史记录" disabled size="small" class="w-40" />
+        <Select
+          v-if="audioHistoryOptions.length"
+          v-model="currentAudioId"
+          :options="audioHistoryOptions"
+          option-label="label"
+          option-value="value"
+          placeholder="历史记录"
+          size="small"
+          class="w-48"
+          @update:model-value="onPickAudioHistory"
+        />
+        <Select v-else placeholder="暂无历史记录" disabled size="small" class="w-40" />
       </div>
-      <p class="py-4 text-center text-xs text-gray-500">暂无音频，请生成或选择历史记录</p>
+      <div v-if="hasAudio" class="px-1 py-2">
+        <audio :src="generatedAudioSrc" controls class="h-9 w-full" />
+      </div>
+      <p v-else class="py-4 text-center text-xs text-gray-500">
+        暂无音频，请填写文案后点击「生成语音」
+      </p>
     </div>
     <div class="mt-3">
       <div class="mb-2 text-xs text-slate-400">选择形象</div>
@@ -81,7 +147,7 @@ function onOpenVoiceManage() {
         class="w-full"
       />
     </div>
-    <Button label="生成口播视频" size="small" class="mt-3 w-full" />
+    <Button label="生成口播视频" size="small" class="mt-3 w-full" :disabled="!hasAudio" />
     <div class="mt-2">
       <div class="mb-1 text-xs text-slate-200">视频预览</div>
       <div class="dashboard-aspect-video">
diff --git a/src/services/localAudio.js b/src/services/localAudio.js
new file mode 100644
index 0000000..e0ac6af
--- /dev/null
+++ b/src/services/localAudio.js
@@ -0,0 +1,20 @@
+import { invoke, isTauri } from "@tauri-apps/api/core";
+
+export function mimeFromAudioPath(filePath) {
+  const lower = String(filePath || "").toLowerCase();
+  if (lower.endsWith(".mp3")) return "audio/mpeg";
+  if (lower.endsWith(".ogg")) return "audio/ogg";
+  return "audio/wav";
+}
+
+/**
+ * @param {string} filePath
+ */
+export async function localAudioToPlayableUrl(filePath) {
+  if (!isTauri()) {
+    throw new Error("请在 Tauri 桌面端使用音频播放功能");
+  }
+  const base64 = await invoke("read_local_file_base64", { path: filePath });
+  const mime = mimeFromAudioPath(filePath);
+  return `data:${mime};base64,${base64}`;
+}
diff --git a/src/services/voiceGenerate.js b/src/services/voiceGenerate.js
new file mode 100644
index 0000000..0cb2cf3
--- /dev/null
+++ b/src/services/voiceGenerate.js
@@ -0,0 +1,99 @@
+import { invoke } from "@tauri-apps/api/core";
+import { ensureAppConfigReady } from "../config/videoPipeline.js";
+import { resolveAliyunVoiceId } from "./soundPromptStorage.js";
+import { getVoiceDisplayLabel } from "../utils/voiceLabel.js";
+import { buildTtsInstruction, buildTtsLanguageType } from "../utils/ttsOptions.js";
+import { localAudioToPlayableUrl } from "./localAudio.js";
+
+const VOICE_GENERATE_SCRIPT = "voice_generate.js";
+
+/**
+ * 生成语音（对齐 Electron generateAudio CosyVoice）
+ * @param {ReturnType<typeof import('../stores/workflow.js').useWorkflowStore>} store
+ */
+export async function executeGenerateSpeech(store) {
+  const text = String(store.scriptContent || "").trim();
+  if (!text) {
+    return { ok: false, message: "请先在「视频文案编辑」中填写文案内容" };
+  }
+
+  if (!store.selectedVoiceId) {
+    return { ok: false, message: "请先选择音色" };
+  }
+
+  const cfg = await ensureAppConfigReady();
+  if (!cfg.ok) {
+    return { ok: false, message: cfg.message };
+  }
+
+  const voice = resolveAliyunVoiceId(store.selectedVoiceId);
+  const instruction = buildTtsInstruction(
+    store.voiceLanguage,
+    store.selectedVoiceId,
+    store.voiceEmotion,
+    store.speechRate,
+  );
+  const languageType = buildTtsLanguageType(store.voiceLanguage);
+
+  store.speechGenerating = true;
+
+  try {
+    const result = await invoke("run_nodejs_script", {
+      scriptName: VOICE_GENERATE_SCRIPT,
+      params: {
+        text,
+        voice,
+        instruction,
+        languageType: languageType || null,
+        format: "mp3",
+      },
+    });
+
+    if (!result?.success || !result?.audioPath) {
+      return {
+        ok: false,
+        message: String(result?.error || "语音生成失败，请重试"),
+      };
+    }
+
+    const audioSrc = await localAudioToPlayableUrl(result.audioPath);
+    const label = `${getVoiceDisplayLabel(store.selectedVoiceId)} · ${formatHistoryTime()}`;
+    const entry = {
+      id: `audio_${Date.now()}`,
+      label,
+      audioPath: result.audioPath,
+      audioSrc,
+      createdAt: Date.now(),
+    };
+
+    store.generatedAudioPath = result.audioPath;
+    store.generatedAudioSrc = audioSrc;
+    store.currentAudioId = entry.id;
+    store.audioHistory = [entry, ...store.audioHistory].slice(0, 50);
+
+    return { ok: true, message: "语音生成完成" };
+  } catch (err) {
+    const msg = err instanceof Error ? err.message : String(err || "未知错误");
+    return { ok: false, message: `语音生成失败: ${msg}` };
+  } finally {
+    store.speechGenerating = false;
+  }
+}
+
+function formatHistoryTime() {
+  const d = new Date();
+  const p = (n) => String(n).padStart(2, "0");
+  return `${p(d.getMonth() + 1)}-${p(d.getDate())} ${p(d.getHours())}:${p(d.getMinutes())}`;
+}
+
+/**
+ * @param {ReturnType<typeof import('../stores/workflow.js').useWorkflowStore>} store
+ * @param {string} audioId
+ */
+export function selectAudioHistory(store, audioId) {
+  const item = store.audioHistory.find((h) => h.id === audioId);
+  if (!item) return;
+  store.currentAudioId = item.id;
+  store.generatedAudioPath = item.audioPath;
+  store.generatedAudioSrc = item.audioSrc;
+}
diff --git a/src/services/voicePreview.js b/src/services/voicePreview.js
index e6661a4..c222106 100644
--- a/src/services/voicePreview.js
+++ b/src/services/voicePreview.js
@@ -1,32 +1,13 @@
-import { invoke, isTauri } from "@tauri-apps/api/core";
+import { invoke } from "@tauri-apps/api/core";
 import { ensureAppConfigReady } from "../config/videoPipeline.js";
 import { resolveAliyunVoiceId } from "./soundPromptStorage.js";
+import { localAudioToPlayableUrl } from "./localAudio.js";
 
 const VOICE_PREVIEW_SCRIPT = "voice_preview.js";
 
 /** @type {Map<string, string>} cacheKey → data:audio URL */
 const previewUrlCache = new Map();
 
-function mimeFromPath(filePath) {
-  const lower = String(filePath || "").toLowerCase();
-  if (lower.endsWith(".mp3")) return "audio/mpeg";
-  if (lower.endsWith(".ogg")) return "audio/ogg";
-  return "audio/wav";
-}
-
-/**
- * 将 Node 写入的本地音频转为可在 WebView 播放的 URL（避免 asset.localhost 不可用）
- * @param {string} filePath
- */
-async function localAudioToPlayableUrl(filePath) {
-  if (!isTauri()) {
-    throw new Error("请在 Tauri 桌面端使用试听功能");
-  }
-  const base64 = await invoke("read_local_file_base64", { path: filePath });
-  const mime = mimeFromPath(filePath);
-  return `data:${mime};base64,${base64}`;
-}
-
 /**
  * 音色试听
  * @param {{ listKey: string, title: string, aliyunVoiceId?: string, selectedVoiceId?: string }} opts
diff --git a/src/stores/workflow.js b/src/stores/workflow.js
index 73373e8..1235e28 100644
--- a/src/stores/workflow.js
+++ b/src/stores/workflow.js
@@ -17,6 +17,10 @@ import {
 } from "../config/videoPipeline.js";
 
 import { getVoiceDisplayLabel } from "../utils/voiceLabel.js";
+import {
+  executeGenerateSpeech,
+  selectAudioHistory,
+} from "../services/voiceGenerate.js";
 
 
 
@@ -153,6 +157,17 @@ export const useWorkflowStore = defineStore("workflow", {
 
     avatarSelect: null,
 
+    speechGenerating: false,
+
+    generatedAudioPath: "",
+
+    generatedAudioSrc: "",
+
+    currentAudioId: null,
+
+    /** @type {Array<{ id: string, label: string, audioPath: string, audioSrc: string, createdAt: number }>} */
+    audioHistory: [],
+
 
 
     // 03 视频编辑
@@ -430,6 +445,18 @@ export const useWorkflowStore = defineStore("workflow", {
 
     },
 
+    async generateSpeech() {
+
+      return executeGenerateSpeech(this);
+
+    },
+
+    pickAudioHistory(audioId) {
+
+      selectAudioHistory(this, audioId);
+
+    },
+
     resetAll() {
 
       this.$reset();
diff --git a/src/utils/ttsOptions.js b/src/utils/ttsOptions.js
new file mode 100644
index 0000000..36a2eaa
--- /dev/null
+++ b/src/utils/ttsOptions.js
@@ -0,0 +1,71 @@
+/** TTS 参数构建（对齐 Electron generateAudio CosyVoice 分支） */
+
+const DIALECT_LANGUAGES = new Set([
+  "上海话",
+  "北京话",
+  "四川话",
+  "南京话",
+  "陕西话",
+  "闽南语",
+  "天津话",
+  "粤语",
+]);
+
+const LANGUAGE_TYPE_MAP = {
+  英语: "en",
+  法语: "fr",
+  德语: "de",
+  日语: "ja",
+  韩语: "ko",
+  俄语: "ru",
+  意大利语: "it",
+  西班牙语: "es",
+  葡萄牙语: "pt",
+  上海话: "Chinese",
+  北京话: "Chinese",
+  四川话: "Chinese",
+  南京话: "Chinese",
+  陕西话: "Chinese",
+  闽南语: "Chinese",
+  天津话: "Chinese",
+  粤语: "Chinese",
+};
+
+/**
+ * @param {string} voiceLanguage
+ * @param {string} selectedVoiceId
+ */
+export function buildTtsInstruction(voiceLanguage, selectedVoiceId, voiceEmotion, speechRate) {
+  const lang = String(voiceLanguage || "").trim();
+  const isSystem = String(selectedVoiceId || "").startsWith("sys_");
+
+  let instruction = "请用自然、清晰、适合短视频配音的风格朗读。";
+
+  const emotion = String(voiceEmotion || "").trim();
+  if (emotion && emotion !== "自然") {
+    instruction = `请用${emotion}的情绪，自然、清晰、适合短视频口播的风格朗读。`;
+  }
+
+  if (!isSystem && DIALECT_LANGUAGES.has(lang)) {
+    return `请用${lang}表达。`;
+  }
+  if (lang && lang !== "中文" && lang !== "中文（普通话）") {
+    return `请用${lang}表达。`;
+  }
+
+  const rate = Number(speechRate);
+  if (rate && rate !== 1 && Number.isFinite(rate)) {
+    instruction += `语速约为正常的 ${rate} 倍。`;
+  }
+
+  return instruction;
+}
+
+/** @param {string} voiceLanguage */
+export function buildTtsLanguageType(voiceLanguage) {
+  const lang = String(voiceLanguage || "").trim();
+  if (!lang || lang === "中文" || lang === "中文（普通话）") {
+    return undefined;
+  }
+  return LANGUAGE_TYPE_MAP[lang];
+}