From eae942702cd56d3e20b289262b59a6da957a1e11 Mon Sep 17 00:00:00 2001
From: RainbowBird <git@luoling.moe>
Date: Sun, 7 Jun 2026 01:55:34 +0800
Subject: [PATCH] feat(speech): enhance voice pack handling and add computed
 properties for UI logic

---
 .../src/pages/settings/modules/speech.vue     | 114 +++++++++---------
 .../stage-ui/src/components/scenes/Stage.vue  |   4 +-
 .../src/stores/modules/speech.test.ts         |  46 ++++++-
 .../stage-ui/src/stores/modules/speech.ts     |   7 ++
 4 files changed, 112 insertions(+), 59 deletions(-)

diff --git a/packages/stage-pages/src/pages/settings/modules/speech.vue b/packages/stage-pages/src/pages/settings/modules/speech.vue
index c24ba2db9..2874a0655 100644
--- a/packages/stage-pages/src/pages/settings/modules/speech.vue
+++ b/packages/stage-pages/src/pages/settings/modules/speech.vue
@@ -15,7 +15,7 @@ import {
 import { useAnalytics } from '@proj-airi/stage-ui/composables'
 import { OFFICIAL_SPEECH_PROVIDER_ID } from '@proj-airi/stage-ui/libs/providers/providers/official'
 import { useAiriCardStore, useVoicePacksStore } from '@proj-airi/stage-ui/stores'
-import { useSpeechStore } from '@proj-airi/stage-ui/stores/modules/speech'
+import { useSpeechStore, voicePackForSpeechProvider } from '@proj-airi/stage-ui/stores/modules/speech'
 import { useProvidersStore } from '@proj-airi/stage-ui/stores/providers'
 import {
   FieldCheckbox,
@@ -26,7 +26,7 @@ import {
 } from '@proj-airi/ui'
 import { generateSpeech } from '@xsai/generate-speech'
 import { storeToRefs } from 'pinia'
-import { onMounted, onUnmounted, ref, watch } from 'vue'
+import { computed, onMounted, onUnmounted, ref, watch } from 'vue'
 import { useI18n } from 'vue-i18n'
 import { RouterLink } from 'vue-router'
 
@@ -66,6 +66,12 @@ const audioUrl = ref('')
 const audioPlayer = ref<HTMLAudioElement | null>(null)
 const errorMessage = ref('')
 
+const supportsVoicePackSelection = computed(() => activeSpeechProvider.value === OFFICIAL_SPEECH_PROVIDER_ID)
+const shouldShowVoicePackSection = computed(() =>
+  supportsVoicePackSelection.value
+  && (isLoadingVoicePacks.value || voicePacksError.value != null || voicePacks.value.length > 0),
+)
+
 function createVoicePackVoice(voicePack: VoicePackSnapshot): VoiceInfo {
   return {
     id: voicePack.voiceId,
@@ -186,7 +192,7 @@ async function generateTestSpeech() {
     }
   }
 
-  const voicePack = activeCard.value?.extensions.airi.modules.speech.voicePack
+  const voicePack = voicePackForSpeechProvider(activeSpeechProvider.value, activeCard.value?.extensions.airi.modules.speech.voicePack)
   if (voicePack) {
     model = voicePack.ttsModelId
     if (!voice || voice.id !== voicePack.voiceId)
@@ -318,61 +324,57 @@ function handleDeleteProvider(providerId: string) {
   <div flex="~ col md:row gap-6">
     <div bg="neutral-100 dark:[rgba(0,0,0,0.3)]" rounded-xl p-4 flex="~ col gap-4" class="h-fit w-full md:w-[40%]">
       <div flex="~ col gap-4">
-        <div>
-          <h2 class="text-lg text-neutral-500 md:text-2xl dark:text-neutral-400">
-            {{ t('settings.pages.modules.speech.sections.section.voice-pack.title') }}
-          </h2>
-          <div text="neutral-400 dark:neutral-500">
-            <span>{{ t('settings.pages.modules.speech.sections.section.voice-pack.description') }}</span>
-          </div>
-        </div>
-
-        <div v-if="isLoadingVoicePacks" :class="['flex items-center gap-2', 'text-sm text-neutral-400 dark:text-neutral-500']">
-          <div i-solar:spinner-line-duotone class="animate-spin text-base" />
-          <span>{{ t('settings.pages.modules.speech.sections.section.voice-pack.loading') }}</span>
-        </div>
-
-        <ErrorContainer
-          v-else-if="voicePacksError"
-          :title="t('settings.pages.modules.speech.sections.section.voice-pack.error')"
-          :error="voicePacksError"
-        />
-
-        <div v-else-if="voicePacks.length > 0" :class="['grid grid-cols-1 gap-2']">
-          <button
-            v-for="pack in voicePacks"
-            :key="pack.id"
-            type="button"
-            :class="[
-              'w-full border rounded-lg px-3 py-2 text-left transition-colors',
-              'border-neutral-200 bg-white hover:border-primary-400 dark:border-neutral-800 dark:bg-neutral-900/60 dark:hover:border-primary-500',
-              airiCardStore.activeCard?.extensions.airi.modules.speech.voicePack?.packId === pack.id
-                ? 'border-primary-500 bg-primary-50 dark:border-primary-400 dark:bg-primary-950/30'
-                : '',
-            ]"
-            @click="bindVoicePack(pack)"
-          >
-            <div :class="['flex items-center justify-between gap-3']">
-              <div :class="['min-w-0']">
-                <div :class="['truncate text-sm font-medium text-neutral-700 dark:text-neutral-200']">
-                  {{ pack.name }}
-                </div>
-                <div :class="['truncate text-xs text-neutral-400 dark:text-neutral-500']">
-                  {{ pack.ttsModelId }} / {{ pack.voiceId }}
-                </div>
-              </div>
-              <span :class="['shrink-0 rounded bg-neutral-100 px-2 py-1 text-xs text-neutral-500 dark:bg-neutral-800 dark:text-neutral-400']">
-                {{ formatCostMultiplier(pack.costMultiplier) }}
-              </span>
+        <template v-if="shouldShowVoicePackSection">
+          <div>
+            <h2 class="text-lg text-neutral-500 md:text-2xl dark:text-neutral-400">
+              {{ t('settings.pages.modules.speech.sections.section.voice-pack.title') }}
+            </h2>
+            <div text="neutral-400 dark:neutral-500">
+              <span>{{ t('settings.pages.modules.speech.sections.section.voice-pack.description') }}</span>
             </div>
-          </button>
-        </div>
+          </div>
 
-        <Alert v-else type="info" icon="i-solar:info-circle-line-duotone">
-          <template #title>
-            {{ t('settings.pages.modules.speech.sections.section.voice-pack.empty') }}
-          </template>
-        </Alert>
+          <div v-if="isLoadingVoicePacks" :class="['flex items-center gap-2', 'text-sm text-neutral-400 dark:text-neutral-500']">
+            <div i-solar:spinner-line-duotone class="animate-spin text-base" />
+            <span>{{ t('settings.pages.modules.speech.sections.section.voice-pack.loading') }}</span>
+          </div>
+
+          <ErrorContainer
+            v-else-if="voicePacksError"
+            :title="t('settings.pages.modules.speech.sections.section.voice-pack.error')"
+            :error="voicePacksError"
+          />
+
+          <div v-else-if="voicePacks.length > 0" :class="['grid grid-cols-1 gap-2']">
+            <button
+              v-for="pack in voicePacks"
+              :key="pack.id"
+              type="button"
+              :class="[
+                'w-full border rounded-lg px-3 py-2 text-left transition-colors',
+                'border-neutral-200 bg-white hover:border-primary-400 dark:border-neutral-800 dark:bg-neutral-900/60 dark:hover:border-primary-500',
+                airiCardStore.activeCard?.extensions.airi.modules.speech.voicePack?.packId === pack.id
+                  ? 'border-primary-500 bg-primary-50 dark:border-primary-400 dark:bg-primary-950/30'
+                  : '',
+              ]"
+              @click="bindVoicePack(pack)"
+            >
+              <div :class="['flex items-center justify-between gap-3']">
+                <div :class="['min-w-0']">
+                  <div :class="['truncate text-sm font-medium text-neutral-700 dark:text-neutral-200']">
+                    {{ pack.name }}
+                  </div>
+                  <div :class="['truncate text-xs text-neutral-400 dark:text-neutral-500']">
+                    {{ pack.ttsModelId }} / {{ pack.voiceId }}
+                  </div>
+                </div>
+                <span :class="['shrink-0 rounded bg-neutral-100 px-2 py-1 text-xs text-neutral-500 dark:bg-neutral-800 dark:text-neutral-400']">
+                  {{ formatCostMultiplier(pack.costMultiplier) }}
+                </span>
+              </div>
+            </button>
+          </div>
+        </template>
 
         <div>
           <h2 class="text-lg text-neutral-500 md:text-2xl dark:text-neutral-400">
diff --git a/packages/stage-ui/src/components/scenes/Stage.vue b/packages/stage-ui/src/components/scenes/Stage.vue
index e3d5af490..84fc8d959 100644
--- a/packages/stage-ui/src/components/scenes/Stage.vue
+++ b/packages/stage-ui/src/components/scenes/Stage.vue
@@ -42,7 +42,7 @@ import { useBackgroundStore } from '../../stores/background'
 import { useChatOrchestratorStore } from '../../stores/chat'
 import { useLlmStreamingControlStore } from '../../stores/llm-streaming-control'
 import { useAiriCardStore } from '../../stores/modules'
-import { useSpeechStore } from '../../stores/modules/speech'
+import { useSpeechStore, voicePackForSpeechProvider } from '../../stores/modules/speech'
 import { useProvidersStore } from '../../stores/providers'
 import { useSettings } from '../../stores/settings'
 import { useSpeechOutputControlStore } from '../../stores/speech-output-control'
@@ -424,7 +424,7 @@ const speechPipeline = createSpeechPipeline<AudioBuffer>({
       }
     }
 
-    const voicePack = activeCard.value?.extensions.airi.modules.speech.voicePack
+    const voicePack = voicePackForSpeechProvider(activeSpeechProvider.value, activeCard.value?.extensions.airi.modules.speech.voicePack)
     if (voicePack) {
       model = voicePack.ttsModelId
       if (!voice || voice.id !== voicePack.voiceId)
diff --git a/packages/stage-ui/src/stores/modules/speech.test.ts b/packages/stage-ui/src/stores/modules/speech.test.ts
index 8c3be8d29..4d8f65c52 100644
--- a/packages/stage-ui/src/stores/modules/speech.test.ts
+++ b/packages/stage-ui/src/stores/modules/speech.test.ts
@@ -3,7 +3,7 @@ import { beforeEach, describe, expect, it, vi } from 'vitest'
 
 import { OFFICIAL_SPEECH_PROVIDER_ID, OFFICIAL_SPEECH_STREAMING_PROVIDER_ID } from '../../libs/providers/providers/official'
 import { useProvidersStore } from '../providers'
-import { toSignedPercent, useSpeechStore } from './speech'
+import { toSignedPercent, useSpeechStore, voicePackForSpeechProvider } from './speech'
 
 vi.mock('vue-i18n', () => ({
   useI18n: () => ({
@@ -30,6 +30,50 @@ describe('speech store helpers', () => {
     expect(toSignedPercent(0)).toBe('0%')
   })
 
+  /**
+   * @example
+   * speechStore.resolveVoicePackSpeechInput({ text, voice, providerConfig: { voice: 'plain' } })
+   */
+  it('leaves speech input unchanged when no Voice Pack is configured', () => {
+    const speechStore = useSpeechStore()
+    const voice = {
+      id: 'plain-voice',
+      name: 'Plain Voice',
+      provider: 'openai-compatible-audio-speech',
+      languages: [{ code: 'en-US', title: 'English' }],
+    }
+
+    const request = speechStore.resolveVoicePackSpeechInput({
+      text: 'hello',
+      voice,
+      providerConfig: { voice: 'plain-voice' },
+    })
+
+    expect(request.input).toBe('hello')
+    expect(request.providerConfig).toEqual({ voice: 'plain-voice' })
+  })
+
+  /**
+   * @example
+   * voicePackForSpeechProvider('openai-compatible-audio-speech', voicePack)
+   */
+  it('ignores Voice Pack snapshots for non-official speech providers', () => {
+    const voicePack = {
+      packId: 'vp-1',
+      name: 'Frozen',
+      provider: 'volcengine',
+      model: 'seed-tts-2.0',
+      voiceId: 'voice-a',
+      ttsModelId: 'volcengine/pool-a',
+      params: {},
+      costMultiplier: 1,
+    }
+
+    expect(voicePackForSpeechProvider('openai-compatible-audio-speech', voicePack)).toBeUndefined()
+    expect(voicePackForSpeechProvider(OFFICIAL_SPEECH_PROVIDER_ID, voicePack)).toBe(voicePack)
+    expect(voicePackForSpeechProvider(OFFICIAL_SPEECH_STREAMING_PROVIDER_ID, voicePack)).toBeUndefined()
+  })
+
   /**
    * @example
    * speechStore.resolveVoicePackSpeechInput({ text, voice, params: { rate: '+20%' } })
diff --git a/packages/stage-ui/src/stores/modules/speech.ts b/packages/stage-ui/src/stores/modules/speech.ts
index 2a134f342..8028aef68 100644
--- a/packages/stage-ui/src/stores/modules/speech.ts
+++ b/packages/stage-ui/src/stores/modules/speech.ts
@@ -42,6 +42,13 @@ interface VoicePackSpeechInput {
 
 const voicePackSupportedParams = new Set(['pitch', 'rate', 'volume'])
 
+export function voicePackForSpeechProvider(
+  providerId: string | undefined,
+  voicePack: VoicePackSnapshot | undefined,
+): VoicePackSnapshot | undefined {
+  return providerId === OFFICIAL_SPEECH_PROVIDER_ID ? voicePack : undefined
+}
+
 /**
  * Normalizes a Voice Pack percent-style option.
  *