Files
dreamweaver/frontend/src/views/VoiceStudio.vue

772 lines
27 KiB
Vue
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
<script setup lang="ts">
import { computed, onBeforeUnmount, onMounted, ref, watch } from 'vue'
import { useRouter } from 'vue-router'
import { api } from '../api/client'
import { useUserStore } from '../stores/user'
import type {
VoiceSessionDetail,
VoiceSessionFinalizeResponse,
VoiceSessionSummary,
VoiceTurnAcceptedResponse,
VoiceTurnSummary,
VoiceTurnUploadAcceptedResponse,
} from '../types/voiceSession'
import BaseButton from '../components/ui/BaseButton.vue'
import BaseCard from '../components/ui/BaseCard.vue'
import BaseSelect from '../components/ui/BaseSelect.vue'
import BaseTextarea from '../components/ui/BaseTextarea.vue'
import LoadingSpinner from '../components/ui/LoadingSpinner.vue'
import EmptyState from '../components/ui/EmptyState.vue'
import {
ArrowPathIcon,
BookOpenIcon,
ExclamationCircleIcon,
MicrophoneIcon,
PaperAirplaneIcon,
SparklesIcon,
StopIcon,
XMarkIcon,
} from '@heroicons/vue/24/outline'
interface ChildProfile {
id: string
name: string
}
interface StoryUniverse {
id: string
name: string
}
const router = useRouter()
const userStore = useUserStore()
const sessions = ref<VoiceSessionSummary[]>([])
const activeSession = ref<VoiceSessionDetail | null>(null)
const profiles = ref<ChildProfile[]>([])
const universes = ref<StoryUniverse[]>([])
const selectedProfileId = ref('')
const selectedUniverseId = ref('')
const textTurnInput = ref('')
const uploadTranscriptHint = ref('')
const loadingSessions = ref(false)
const creatingSession = ref(false)
const loadingSessionDetail = ref(false)
const sendingTurn = ref(false)
const finalizing = ref(false)
const abandoning = ref(false)
const recording = ref(false)
const recordingDurationMs = ref(0)
const error = ref('')
const mediaError = ref('')
const recorderSupported = computed(() => typeof window !== 'undefined' && 'MediaRecorder' in window)
let mediaRecorder: MediaRecorder | null = null
let mediaStream: MediaStream | null = null
let recordingChunks: Blob[] = []
let recordingTimer: number | null = null
let recordingStartedAt = 0
const recordedBlob = ref<Blob | null>(null)
const recordedAudioUrl = ref<string | null>(null)
const profileOptions = computed(() =>
profiles.value.map((profile) => ({ value: profile.id, label: profile.name })),
)
const universeOptions = computed(() =>
universes.value.map((universe) => ({ value: universe.id, label: universe.name })),
)
const activeTurnList = computed(() => activeSession.value?.recent_turns ?? [])
function formatSessionStatus(status: string) {
switch (status) {
case 'draft':
return '待开始'
case 'processing_turn':
return '处理中'
case 'waiting_user':
return '等待下一轮'
case 'finalizing_story':
return '保存中'
case 'completed':
return '已完成'
case 'abandoned':
return '已放弃'
case 'failed':
return '失败'
default:
return status
}
}
function formatTurnStatus(status: string) {
switch (status) {
case 'received':
return '已接收'
case 'transcribing':
return '转写中'
case 'narrative_ready':
return '文本已生成'
case 'audio_ready':
return '语音已生成'
case 'failed':
return '失败'
default:
return status
}
}
function formatIntent(intent: string | null | undefined) {
switch (intent) {
case 'start_story':
return '开启故事'
case 'continue_story':
return '继续讲述'
case 'correct_story':
return '修正走向'
case 'end_story':
return '结束本轮'
case 'save_story':
return '请求保存'
default:
return intent || '未知'
}
}
function formatDate(dateStr: string) {
return new Date(dateStr).toLocaleString('zh-CN', {
month: 'short',
day: 'numeric',
hour: '2-digit',
minute: '2-digit',
})
}
function revokeRecordedAudioUrl() {
if (recordedAudioUrl.value) {
URL.revokeObjectURL(recordedAudioUrl.value)
recordedAudioUrl.value = null
}
}
function clearRecordedAudio() {
revokeRecordedAudioUrl()
recordedBlob.value = null
recordingDurationMs.value = 0
}
async function fetchProfiles() {
if (!userStore.user) return
const data = await api.get<{ profiles: ChildProfile[] }>('/api/profiles')
profiles.value = data.profiles
if (!selectedProfileId.value && profiles.value.length > 0) {
selectedProfileId.value = profiles.value[0].id
}
}
async function fetchUniverses(profileId: string) {
selectedUniverseId.value = ''
if (!profileId) {
universes.value = []
return
}
const data = await api.get<{ universes: StoryUniverse[] }>(`/api/profiles/${profileId}/universes`)
universes.value = data.universes
if (universes.value.length > 0) {
selectedUniverseId.value = universes.value[0].id
}
}
async function loadSessions() {
if (!userStore.user) return
loadingSessions.value = true
try {
sessions.value = await api.get<VoiceSessionSummary[]>('/api/voice-sessions?limit=8')
if (!activeSession.value) {
const resumable = sessions.value.find((item) => item.can_continue)
if (resumable) {
await loadSessionDetail(resumable.id)
}
}
} catch (err) {
error.value = err instanceof Error ? err.message : '会话列表加载失败'
} finally {
loadingSessions.value = false
}
}
async function loadSessionDetail(sessionId: string) {
loadingSessionDetail.value = true
error.value = ''
try {
activeSession.value = await api.get<VoiceSessionDetail>(`/api/voice-sessions/${sessionId}`)
} catch (err) {
error.value = err instanceof Error ? err.message : '会话详情加载失败'
} finally {
loadingSessionDetail.value = false
}
}
async function createSession() {
creatingSession.value = true
error.value = ''
try {
const session = await api.post<VoiceSessionSummary>('/api/voice-sessions', {
child_profile_id: selectedProfileId.value || null,
universe_id: selectedUniverseId.value || null,
target_mode: 'story',
})
await loadSessions()
await loadSessionDetail(session.id)
textTurnInput.value = ''
uploadTranscriptHint.value = ''
clearRecordedAudio()
} catch (err) {
error.value = err instanceof Error ? err.message : '创建语音共创会话失败'
} finally {
creatingSession.value = false
}
}
async function pollTurnResult(sessionId: string, turnId: string) {
const terminalStatuses = new Set(['audio_ready', 'narrative_ready', 'failed'])
for (let attempt = 0; attempt < 20; attempt += 1) {
const turn = await api.get<VoiceTurnSummary>(`/api/voice-sessions/${sessionId}/turns/${turnId}`)
if (terminalStatuses.has(turn.status)) {
return turn
}
await new Promise((resolve) => window.setTimeout(resolve, 600))
}
throw new Error('本轮语音共创处理超时,请刷新后查看最新状态')
}
async function refreshAfterTurn(sessionId: string, turnId: string) {
await pollTurnResult(sessionId, turnId)
await loadSessionDetail(sessionId)
await loadSessions()
}
async function submitTextTurn() {
if (!activeSession.value || !textTurnInput.value.trim()) return
sendingTurn.value = true
error.value = ''
try {
const result = await api.post<VoiceTurnAcceptedResponse>(
`/api/voice-sessions/${activeSession.value.id}/turns/fallback`,
{
transcript_text: textTurnInput.value.trim(),
},
)
textTurnInput.value = ''
await refreshAfterTurn(result.session_id, result.turn_id)
} catch (err) {
error.value = err instanceof Error ? err.message : '文本共创回合提交失败'
} finally {
sendingTurn.value = false
}
}
async function submitRecordedTurn() {
if (!activeSession.value || !recordedBlob.value) return
sendingTurn.value = true
error.value = ''
try {
const formData = new FormData()
formData.append('audio_file', recordedBlob.value, 'voice-turn.webm')
if (recordingDurationMs.value > 0) {
formData.append('duration_ms', String(recordingDurationMs.value))
}
if (uploadTranscriptHint.value.trim()) {
formData.append('transcript_hint', uploadTranscriptHint.value.trim())
}
const result = await api.postForm<VoiceTurnUploadAcceptedResponse>(
`/api/voice-sessions/${activeSession.value.id}/turns`,
formData,
)
clearRecordedAudio()
uploadTranscriptHint.value = ''
await refreshAfterTurn(result.session_id, result.turn_id)
} catch (err) {
error.value = err instanceof Error ? err.message : '录音回合提交失败'
} finally {
sendingTurn.value = false
}
}
async function finalizeSession() {
if (!activeSession.value) return
finalizing.value = true
error.value = ''
try {
const result = await api.post<VoiceSessionFinalizeResponse>(
`/api/voice-sessions/${activeSession.value.id}/finalize`,
{
save_story: true,
generate_cover: true,
generate_final_audio: false,
},
)
await loadSessions()
await loadSessionDetail(activeSession.value.id)
if (result.story_id) {
router.push(`/story/${result.story_id}`)
}
} catch (err) {
error.value = err instanceof Error ? err.message : '保存语音共创故事失败'
} finally {
finalizing.value = false
}
}
async function abandonSession() {
if (!activeSession.value) return
abandoning.value = true
error.value = ''
try {
const summary = await api.post<VoiceSessionSummary>(
`/api/voice-sessions/${activeSession.value.id}/abandon`,
{ reason: '用户在语音共创页主动结束会话' },
)
await loadSessions()
activeSession.value = {
...(activeSession.value as VoiceSessionDetail),
...summary,
}
} catch (err) {
error.value = err instanceof Error ? err.message : '放弃会话失败'
} finally {
abandoning.value = false
}
}
async function startRecording() {
mediaError.value = ''
if (!recorderSupported.value) {
mediaError.value = '当前浏览器不支持录音,请先使用文本共创模式。'
return
}
try {
clearRecordedAudio()
mediaStream = await navigator.mediaDevices.getUserMedia({ audio: true })
mediaRecorder = new MediaRecorder(mediaStream)
recordingChunks = []
mediaRecorder.ondataavailable = (event) => {
if (event.data.size > 0) {
recordingChunks.push(event.data)
}
}
mediaRecorder.onstop = () => {
const blob = new Blob(recordingChunks, { type: mediaRecorder?.mimeType || 'audio/webm' })
recordedBlob.value = blob
revokeRecordedAudioUrl()
recordedAudioUrl.value = URL.createObjectURL(blob)
}
mediaRecorder.start()
recording.value = true
recordingStartedAt = Date.now()
recordingTimer = window.setInterval(() => {
recordingDurationMs.value = Date.now() - recordingStartedAt
}, 200)
} catch (err) {
mediaError.value = err instanceof Error ? err.message : '无法访问麦克风'
}
}
function stopRecording() {
if (!mediaRecorder || !recording.value) return
mediaRecorder.stop()
mediaRecorder = null
recording.value = false
if (recordingTimer) {
window.clearInterval(recordingTimer)
recordingTimer = null
}
if (mediaStream) {
mediaStream.getTracks().forEach((track) => track.stop())
mediaStream = null
}
}
function resetRecording() {
if (recording.value) {
stopRecording()
}
clearRecordedAudio()
}
watch(selectedProfileId, (newId) => {
if (newId) {
void fetchUniverses(newId)
} else {
universes.value = []
selectedUniverseId.value = ''
}
})
onMounted(async () => {
if (!userStore.user) {
await userStore.fetchSession()
}
if (!userStore.user) return
await fetchProfiles()
await loadSessions()
})
onBeforeUnmount(() => {
if (recording.value) {
stopRecording()
}
clearRecordedAudio()
})
</script>
<template>
<div class="max-w-7xl mx-auto space-y-8">
<div class="flex flex-col gap-4 lg:flex-row lg:items-center lg:justify-between">
<div>
<h1 class="text-3xl font-bold gradient-text">语音共创工作台</h1>
<p class="mt-2 text-gray-500">
第一阶段先跑通回合式共创孩子说一句系统接一句再决定是否保存为正式故事
</p>
</div>
<div class="flex flex-wrap gap-3">
<BaseButton variant="secondary" @click="loadSessions" :loading="loadingSessions">
<ArrowPathIcon class="h-5 w-5" />
刷新会话
</BaseButton>
<BaseButton @click="createSession" :loading="creatingSession">
<SparklesIcon class="h-5 w-5" />
开始新会话
</BaseButton>
</div>
</div>
<div v-if="!userStore.user" class="py-12">
<EmptyState
:icon="BookOpenIcon"
title="需要先登录"
description="登录后才能使用语音共创工作台。"
/>
</div>
<template v-else>
<div class="grid grid-cols-1 gap-6 xl:grid-cols-[320px_minmax(0,1fr)]">
<div class="space-y-6">
<BaseCard>
<div class="space-y-4">
<div>
<h2 class="text-lg font-semibold text-gray-900">创建条件</h2>
<p class="mt-1 text-sm text-gray-500">
先选孩子档案与故事宇宙让共创更容易复用现有角色和世界观
</p>
</div>
<BaseSelect
v-model="selectedProfileId"
label="孩子档案"
:options="profileOptions"
placeholder="请选择孩子档案"
/>
<BaseSelect
v-model="selectedUniverseId"
label="故事宇宙"
:options="universeOptions"
placeholder="可选,默认不绑定宇宙"
/>
<BaseButton class="w-full" @click="createSession" :loading="creatingSession">
创建语音共创会话
</BaseButton>
</div>
</BaseCard>
<BaseCard>
<div class="flex items-center justify-between">
<div>
<h2 class="text-lg font-semibold text-gray-900">最近会话</h2>
<p class="mt-1 text-sm text-gray-500">支持恢复最近还在等待下一轮的 session</p>
</div>
<span class="text-xs text-gray-400">{{ sessions.length }} </span>
</div>
<div v-if="loadingSessions" class="py-8">
<LoadingSpinner text="加载会话中..." />
</div>
<div v-else-if="sessions.length === 0" class="pt-6">
<EmptyState
:icon="SparklesIcon"
title="还没有语音共创会话"
description="先创建一个会话,再通过文本或录音开始第一轮故事。"
/>
</div>
<div v-else class="mt-4 space-y-3">
<button
v-for="session in sessions"
:key="session.id"
type="button"
class="w-full rounded-2xl border px-4 py-3 text-left transition-all"
:class="activeSession?.id === session.id
? 'border-purple-300 bg-purple-50'
: 'border-gray-100 bg-white hover:border-gray-300'"
@click="loadSessionDetail(session.id)"
>
<div class="flex items-center justify-between gap-3">
<div class="min-w-0">
<div class="truncate font-medium text-gray-900">
{{ session.working_title || '未命名语音会话' }}
</div>
<div class="mt-1 text-xs text-gray-500">
{{ formatSessionStatus(session.status) }} · {{ session.total_turns }}
</div>
</div>
<div class="text-right text-xs text-gray-400">
{{ formatDate(session.updated_at) }}
</div>
</div>
</button>
</div>
</BaseCard>
</div>
<div class="space-y-6">
<BaseCard v-if="error" class="border border-rose-100 bg-rose-50 text-rose-600">
<div class="flex items-start gap-3">
<ExclamationCircleIcon class="mt-0.5 h-5 w-5 shrink-0" />
<div>{{ error }}</div>
</div>
</BaseCard>
<div v-if="loadingSessionDetail" class="py-16">
<LoadingSpinner text="正在加载会话详情..." />
</div>
<BaseCard v-else-if="activeSession">
<div class="flex flex-col gap-4 lg:flex-row lg:items-start lg:justify-between">
<div>
<div class="flex flex-wrap items-center gap-2">
<h2 class="text-2xl font-semibold text-gray-900">
{{ activeSession.working_title || '语音共创会话' }}
</h2>
<span class="rounded-full bg-purple-100 px-3 py-1 text-xs font-medium text-purple-700">
{{ formatSessionStatus(activeSession.status) }}
</span>
</div>
<p class="mt-2 text-sm text-gray-500">
最近意图{{ formatIntent(activeSession.latest_detected_intent) }} ·
已完成 {{ activeSession.total_turns }}
</p>
</div>
<div class="flex flex-wrap gap-3">
<BaseButton
variant="secondary"
@click="finalizeSession"
:loading="finalizing"
:disabled="!activeSession.can_finalize"
>
<BookOpenIcon class="h-5 w-5" />
保存为正式故事
</BaseButton>
<BaseButton
variant="ghost"
class="text-rose-500 hover:bg-rose-50"
@click="abandonSession"
:loading="abandoning"
:disabled="activeSession.status === 'completed' || activeSession.status === 'abandoned'"
>
<XMarkIcon class="h-5 w-5" />
放弃会话
</BaseButton>
</div>
</div>
<div class="mt-6 grid grid-cols-1 gap-6 xl:grid-cols-[minmax(0,1.2fr)_minmax(0,0.8fr)]">
<div class="space-y-6">
<div class="rounded-2xl border border-gray-100 bg-white p-4">
<div class="flex items-center justify-between">
<h3 class="font-semibold text-gray-900">文本共创回合</h3>
<span class="text-xs text-gray-400">最稳的 fallback 路径</span>
</div>
<div class="mt-4 space-y-4">
<BaseTextarea
v-model="textTurnInput"
label="本轮你想让故事怎么发展"
placeholder="例如:不要让它害怕,我想让它遇见一个新朋友。"
:rows="4"
:max-length="1000"
:disabled="sendingTurn || !activeSession.can_continue"
/>
<BaseButton
@click="submitTextTurn"
:loading="sendingTurn"
:disabled="!activeSession.can_continue || !textTurnInput.trim()"
>
<PaperAirplaneIcon class="h-5 w-5" />
发送文本回合
</BaseButton>
</div>
</div>
<div class="rounded-2xl border border-gray-100 bg-white p-4">
<div class="flex items-center justify-between">
<h3 class="font-semibold text-gray-900">录音共创回合</h3>
<span class="text-xs text-gray-400">已支持上传音频 turn</span>
</div>
<p class="mt-2 text-sm text-gray-500">
当前默认是 demo 转写模式若本地未接真实 ASR可在下方填写转写提示辅助开发验证
</p>
<div class="mt-4 flex flex-wrap gap-3">
<BaseButton
v-if="!recording"
variant="secondary"
@click="startRecording"
:disabled="sendingTurn || !activeSession.can_continue"
>
<MicrophoneIcon class="h-5 w-5" />
开始录音
</BaseButton>
<BaseButton
v-else
variant="danger"
@click="stopRecording"
>
<StopIcon class="h-5 w-5" />
停止录音
</BaseButton>
<BaseButton
variant="ghost"
@click="resetRecording"
:disabled="(!recordedBlob && !recording) || sendingTurn"
>
<XMarkIcon class="h-5 w-5" />
清空录音
</BaseButton>
</div>
<p v-if="recording" class="mt-3 text-sm text-amber-600">
正在录音{{ Math.round(recordingDurationMs / 1000) }}s
</p>
<p v-if="mediaError" class="mt-3 text-sm text-rose-600">
{{ mediaError }}
</p>
<div v-if="recordedAudioUrl" class="mt-4 rounded-2xl border border-gray-100 bg-gray-50 p-4">
<div class="text-sm font-medium text-gray-700">录音预览</div>
<audio class="mt-3 w-full" :src="recordedAudioUrl" controls></audio>
</div>
<div class="mt-4">
<BaseTextarea
v-model="uploadTranscriptHint"
label="开发转写提示(可选)"
placeholder="如果当前环境还是 demo 转写模式,可以把你刚才说的话写在这里。"
:rows="3"
:max-length="1000"
/>
</div>
<div class="mt-4">
<BaseButton
@click="submitRecordedTurn"
:loading="sendingTurn"
:disabled="!activeSession.can_continue || !recordedBlob"
>
<SparklesIcon class="h-5 w-5" />
上传录音回合
</BaseButton>
</div>
</div>
<div class="rounded-2xl border border-gray-100 bg-white p-4">
<div class="flex items-center justify-between">
<h3 class="font-semibold text-gray-900">共创过程</h3>
<span class="text-xs text-gray-400">{{ activeTurnList.length }} 条最近 turn</span>
</div>
<div class="mt-4 space-y-4">
<div
v-for="turn in activeTurnList"
:key="turn.id"
class="rounded-2xl border border-gray-100 bg-gray-50 p-4"
>
<div class="flex flex-wrap items-center gap-2 text-xs text-gray-400">
<span> {{ turn.turn_index }} </span>
<span>·</span>
<span>{{ formatTurnStatus(turn.status) }}</span>
<span>·</span>
<span>{{ formatIntent(turn.detected_intent) }}</span>
<span v-if="turn.transcription_provider">· {{ turn.transcription_provider }}</span>
</div>
<div class="mt-3 text-sm text-gray-800">
<span class="font-medium text-gray-900">孩子</span>
{{ turn.user_transcript || '暂无转写内容' }}
</div>
<div v-if="turn.user_audio_url" class="mt-3">
<audio class="w-full" :src="turn.user_audio_url" controls></audio>
</div>
<div v-if="turn.assistant_text" class="mt-4 text-sm text-gray-700">
<span class="font-medium text-purple-700">织机回应</span>
{{ turn.assistant_text }}
</div>
<div v-if="turn.assistant_audio_url" class="mt-3">
<audio class="w-full" :src="turn.assistant_audio_url" controls></audio>
</div>
<div v-if="turn.error_message" class="mt-3 text-sm text-rose-600">
{{ turn.error_message }}
</div>
</div>
<EmptyState
v-if="activeTurnList.length === 0"
:icon="MicrophoneIcon"
title="从第一轮开始"
description="先发送一句文本或录音,让这个会话拥有第一段故事。"
/>
</div>
</div>
</div>
<div class="space-y-6">
<div class="rounded-2xl border border-gray-100 bg-white p-4">
<h3 class="font-semibold text-gray-900">故事状态快照</h3>
<pre class="mt-4 overflow-x-auto rounded-xl bg-gray-950 p-4 text-xs leading-6 text-emerald-200">{{ JSON.stringify(activeSession.story_state, null, 2) }}</pre>
</div>
<div class="rounded-2xl border border-gray-100 bg-white p-4">
<h3 class="font-semibold text-gray-900">最近事件</h3>
<div class="mt-4 space-y-3">
<div
v-for="event in activeSession.events.slice(-10)"
:key="event.id"
class="rounded-xl border border-gray-100 bg-gray-50 px-3 py-3"
>
<div class="flex items-center justify-between gap-3 text-xs text-gray-400">
<span>{{ event.event_type }}</span>
<span>{{ formatDate(event.created_at) }}</span>
</div>
<div class="mt-1 text-sm font-medium text-gray-800">
{{ event.message || event.status }}
</div>
</div>
</div>
</div>
</div>
</div>
</BaseCard>
<div v-else class="py-16">
<EmptyState
:icon="SparklesIcon"
title="创建或恢复一个语音共创会话"
description="左侧可以直接创建新会话,也可以恢复最近仍在等待下一轮的 session。"
/>
</div>
</div>
</div>
</template>
</div>
</template>