feat: improve voice studio alpha recovery flow
This commit is contained in:
@@ -12,6 +12,7 @@ from fastapi import (
|
||||
)
|
||||
from sqlalchemy.ext.asyncio import AsyncSession
|
||||
|
||||
from app.core.config import settings
|
||||
from app.core.deps import require_user
|
||||
from app.core.rate_limiter import check_rate_limit
|
||||
from app.db.database import get_db
|
||||
@@ -34,11 +35,14 @@ from app.services.voice_session_service import (
|
||||
create_voice_turn_from_text_service,
|
||||
create_voice_turn_from_upload_service,
|
||||
finalize_voice_session_service,
|
||||
get_latest_active_voice_session_service,
|
||||
get_voice_session_detail_service,
|
||||
get_voice_turn_audio_service,
|
||||
get_voice_turn_service,
|
||||
get_voice_turn_user_audio_service,
|
||||
list_voice_sessions_service,
|
||||
retry_voice_turn_audio_service,
|
||||
retry_voice_turn_service,
|
||||
)
|
||||
|
||||
router = APIRouter()
|
||||
@@ -68,8 +72,13 @@ async def create_voice_session(
|
||||
|
||||
@router.get("/voice-sessions", response_model=list[VoiceSessionSummaryResponse])
|
||||
async def list_voice_sessions(
|
||||
limit: int = Query(default=8, ge=1, le=20),
|
||||
limit: int = Query(
|
||||
default=settings.voice_session_default_list_limit,
|
||||
ge=1,
|
||||
le=settings.voice_session_max_list_limit,
|
||||
),
|
||||
active_only: bool = Query(default=False),
|
||||
active_first: bool = Query(default=True),
|
||||
user: User = Depends(require_user),
|
||||
db: AsyncSession = Depends(get_db),
|
||||
):
|
||||
@@ -79,9 +88,19 @@ async def list_voice_sessions(
|
||||
db,
|
||||
limit=limit,
|
||||
active_only=active_only,
|
||||
active_first=active_first,
|
||||
)
|
||||
|
||||
|
||||
@router.get("/voice-sessions/active", response_model=VoiceSessionSummaryResponse | None)
|
||||
async def get_latest_active_voice_session(
|
||||
user: User = Depends(require_user),
|
||||
db: AsyncSession = Depends(get_db),
|
||||
):
|
||||
"""Get the latest active voice session for quick resume behavior."""
|
||||
return await get_latest_active_voice_session_service(user.id, db)
|
||||
|
||||
|
||||
@router.get("/voice-sessions/{session_id}", response_model=VoiceSessionDetailResponse)
|
||||
async def get_voice_session(
|
||||
session_id: str,
|
||||
@@ -158,6 +177,21 @@ async def get_voice_turn(
|
||||
return await get_voice_turn_service(session_id, turn_id, user.id, db)
|
||||
|
||||
|
||||
@router.post(
|
||||
"/voice-sessions/{session_id}/turns/{turn_id}/retry",
|
||||
response_model=VoiceTurnAcceptedResponse,
|
||||
status_code=status.HTTP_202_ACCEPTED,
|
||||
)
|
||||
async def retry_voice_turn(
|
||||
session_id: str,
|
||||
turn_id: str,
|
||||
user: User = Depends(require_user),
|
||||
db: AsyncSession = Depends(get_db),
|
||||
):
|
||||
"""Retry one failed voice turn using its saved transcript."""
|
||||
return await retry_voice_turn_service(session_id, turn_id, user.id, db)
|
||||
|
||||
|
||||
@router.get("/voice-sessions/{session_id}/turns/{turn_id}/audio")
|
||||
async def get_voice_turn_audio(
|
||||
session_id: str,
|
||||
@@ -170,6 +204,20 @@ async def get_voice_turn_audio(
|
||||
return Response(content=audio_bytes, media_type="audio/mpeg")
|
||||
|
||||
|
||||
@router.post(
|
||||
"/voice-sessions/{session_id}/turns/{turn_id}/retry-audio",
|
||||
response_model=VoiceTurnSummaryResponse,
|
||||
)
|
||||
async def retry_voice_turn_audio(
|
||||
session_id: str,
|
||||
turn_id: str,
|
||||
user: User = Depends(require_user),
|
||||
db: AsyncSession = Depends(get_db),
|
||||
):
|
||||
"""Retry assistant audio synthesis when one turn only has text output."""
|
||||
return await retry_voice_turn_audio_service(session_id, turn_id, user.id, db)
|
||||
|
||||
|
||||
@router.get("/voice-sessions/{session_id}/turns/{turn_id}/user-audio")
|
||||
async def get_voice_turn_user_audio(
|
||||
session_id: str,
|
||||
|
||||
@@ -82,6 +82,18 @@ class Settings(BaseSettings):
|
||||
"zh",
|
||||
description="Preferred language hint for voice transcription",
|
||||
)
|
||||
voice_turn_max_upload_bytes: int = Field(
|
||||
5 * 1024 * 1024,
|
||||
description="Maximum accepted upload size in bytes for one voice turn audio file",
|
||||
)
|
||||
voice_session_default_list_limit: int = Field(
|
||||
8,
|
||||
description="Default number of recent voice sessions returned to the client",
|
||||
)
|
||||
voice_session_max_list_limit: int = Field(
|
||||
20,
|
||||
description="Maximum number of recent voice sessions returned to the client",
|
||||
)
|
||||
story_audio_cache_ttl_days: int = Field(
|
||||
30,
|
||||
description="TTL in days before cached story audio is pruned",
|
||||
|
||||
@@ -101,6 +101,7 @@ class VoiceSessionSummaryResponse(BaseModel):
|
||||
latest_detected_intent: str | None = None
|
||||
latest_assistant_audio_ready: bool = False
|
||||
last_turn_status: str | None = None
|
||||
transcription_mode_hint: str | None = None
|
||||
can_continue: bool = False
|
||||
can_finalize: bool = False
|
||||
last_error: str | None = None
|
||||
|
||||
@@ -6,9 +6,10 @@ from datetime import datetime, timezone
|
||||
from typing import Any
|
||||
|
||||
from fastapi import HTTPException
|
||||
from sqlalchemy import desc, select
|
||||
from sqlalchemy import case, desc, select
|
||||
from sqlalchemy.ext.asyncio import AsyncSession
|
||||
|
||||
from app.core.config import settings
|
||||
from app.core.logging import get_logger
|
||||
from app.db.models import VoiceSession, VoiceSessionEvent, VoiceTurn
|
||||
from app.schemas.voice_session_schemas import (
|
||||
@@ -134,6 +135,7 @@ def _session_to_summary(
|
||||
session_audio_exists(latest_turn.assistant_audio_path) if latest_turn else False
|
||||
),
|
||||
last_turn_status=latest_turn.status if latest_turn else None,
|
||||
transcription_mode_hint=settings.voice_transcription_mode,
|
||||
can_continue=_session_can_continue(session),
|
||||
can_finalize=_session_can_finalize(session),
|
||||
last_error=session.last_error,
|
||||
@@ -602,17 +604,29 @@ async def list_voice_sessions_service(
|
||||
user_id: str,
|
||||
db: AsyncSession,
|
||||
*,
|
||||
limit: int = 8,
|
||||
limit: int | None = None,
|
||||
active_only: bool = False,
|
||||
active_first: bool = False,
|
||||
) -> list[VoiceSessionSummaryResponse]:
|
||||
query = (
|
||||
select(VoiceSession)
|
||||
.where(VoiceSession.user_id == user_id)
|
||||
.order_by(desc(VoiceSession.updated_at), desc(VoiceSession.created_at))
|
||||
.limit(limit)
|
||||
)
|
||||
resolved_limit = limit or settings.voice_session_default_list_limit
|
||||
resolved_limit = max(1, min(resolved_limit, settings.voice_session_max_list_limit))
|
||||
query = select(VoiceSession).where(VoiceSession.user_id == user_id)
|
||||
if active_only:
|
||||
query = query.where(VoiceSession.status.in_(CONTINUABLE_SESSION_STATUSES))
|
||||
if active_first:
|
||||
query = query.order_by(
|
||||
desc(
|
||||
case(
|
||||
(VoiceSession.status.in_(CONTINUABLE_SESSION_STATUSES), 1),
|
||||
else_=0,
|
||||
)
|
||||
),
|
||||
desc(VoiceSession.updated_at),
|
||||
desc(VoiceSession.created_at),
|
||||
)
|
||||
else:
|
||||
query = query.order_by(desc(VoiceSession.updated_at), desc(VoiceSession.created_at))
|
||||
query = query.limit(resolved_limit)
|
||||
|
||||
sessions = (await db.execute(query)).scalars().all()
|
||||
summaries: list[VoiceSessionSummaryResponse] = []
|
||||
@@ -628,6 +642,30 @@ async def list_voice_sessions_service(
|
||||
return summaries
|
||||
|
||||
|
||||
async def get_latest_active_voice_session_service(
|
||||
user_id: str,
|
||||
db: AsyncSession,
|
||||
) -> VoiceSessionSummaryResponse | None:
|
||||
query = (
|
||||
select(VoiceSession)
|
||||
.where(
|
||||
VoiceSession.user_id == user_id,
|
||||
VoiceSession.status.in_(CONTINUABLE_SESSION_STATUSES),
|
||||
)
|
||||
.order_by(desc(VoiceSession.updated_at), desc(VoiceSession.created_at))
|
||||
.limit(1)
|
||||
)
|
||||
session = (await db.execute(query)).scalar_one_or_none()
|
||||
if session is None:
|
||||
return None
|
||||
latest_turn = await _get_latest_turn(db, session_id=session.id)
|
||||
return _session_to_summary(
|
||||
session,
|
||||
latest_turn=latest_turn,
|
||||
total_turns=session.current_turn_index,
|
||||
)
|
||||
|
||||
|
||||
async def create_voice_session_service(
|
||||
request: VoiceSessionCreateRequest,
|
||||
user_id: str,
|
||||
@@ -766,6 +804,13 @@ async def create_voice_turn_from_upload_service(
|
||||
status_code=409,
|
||||
detail="Voice session is not ready for another turn.",
|
||||
)
|
||||
if not audio_bytes:
|
||||
raise HTTPException(status_code=400, detail="上传音频为空，请重新录音后再试。")
|
||||
if len(audio_bytes) > settings.voice_turn_max_upload_bytes:
|
||||
raise HTTPException(
|
||||
status_code=413,
|
||||
detail="上传音频过大，请缩短单轮录音时长后再试。",
|
||||
)
|
||||
next_turn_index = session.current_turn_index + 1
|
||||
user_audio_path = write_uploaded_user_audio(
|
||||
session_id=session.id,
|
||||
@@ -805,6 +850,91 @@ async def create_voice_turn_from_upload_service(
|
||||
)
|
||||
|
||||
|
||||
async def retry_voice_turn_service(
|
||||
session_id: str,
|
||||
turn_id: str,
|
||||
user_id: str,
|
||||
db: AsyncSession,
|
||||
) -> VoiceTurnAcceptedResponse:
|
||||
turn = await _get_owned_turn(
|
||||
db,
|
||||
session_id=session_id,
|
||||
turn_id=turn_id,
|
||||
user_id=user_id,
|
||||
)
|
||||
if turn.status != "failed":
|
||||
raise HTTPException(status_code=409, detail="Only failed turns can be retried.")
|
||||
if not turn.user_transcript:
|
||||
raise HTTPException(status_code=409, detail="This turn has no transcript to retry.")
|
||||
|
||||
return await create_voice_turn_from_text_service(
|
||||
session_id,
|
||||
VoiceTurnCreateFallbackRequest(
|
||||
transcript_text=turn.user_transcript,
|
||||
duration_ms=turn.user_audio_duration_ms,
|
||||
),
|
||||
user_id,
|
||||
db,
|
||||
)
|
||||
|
||||
|
||||
async def retry_voice_turn_audio_service(
|
||||
session_id: str,
|
||||
turn_id: str,
|
||||
user_id: str,
|
||||
db: AsyncSession,
|
||||
) -> VoiceTurnSummaryResponse:
|
||||
turn = await _get_owned_turn(
|
||||
db,
|
||||
session_id=session_id,
|
||||
turn_id=turn_id,
|
||||
user_id=user_id,
|
||||
)
|
||||
if not turn.assistant_text:
|
||||
raise HTTPException(status_code=409, detail="This turn has no assistant text to speak.")
|
||||
if session_audio_exists(turn.assistant_audio_path):
|
||||
raise HTTPException(status_code=409, detail="Assistant audio already exists for this turn.")
|
||||
|
||||
try:
|
||||
audio_bytes = await text_to_speech(
|
||||
turn.assistant_text,
|
||||
db=db,
|
||||
user_id=user_id,
|
||||
)
|
||||
saved_path = write_session_audio(
|
||||
build_turn_assistant_audio_path(turn.session_id, turn.turn_index),
|
||||
audio_bytes,
|
||||
)
|
||||
turn.assistant_audio_path = saved_path
|
||||
turn.assistant_audio_duration_ms = None
|
||||
if turn.status == "narrative_ready":
|
||||
turn.status = "audio_ready"
|
||||
await db.commit()
|
||||
await db.refresh(turn)
|
||||
await _record_session_event(
|
||||
db,
|
||||
session_id=turn.session_id,
|
||||
turn_id=turn.id,
|
||||
event_type="assistant_audio_retry_succeeded",
|
||||
status="succeeded",
|
||||
message="Assistant audio regenerated for one voice turn.",
|
||||
metadata={"audio_path": saved_path},
|
||||
)
|
||||
except Exception as exc:
|
||||
await _record_session_event(
|
||||
db,
|
||||
session_id=turn.session_id,
|
||||
turn_id=turn.id,
|
||||
event_type="assistant_audio_retry_failed",
|
||||
status="failed",
|
||||
message="Assistant audio retry failed.",
|
||||
metadata={"error": str(exc)},
|
||||
)
|
||||
raise HTTPException(status_code=503, detail="语音补发失败，请稍后再试。") from exc
|
||||
|
||||
return _turn_to_summary(turn)
|
||||
|
||||
|
||||
async def get_voice_turn_service(
|
||||
session_id: str,
|
||||
turn_id: str,
|
||||
|
||||
@@ -2,6 +2,7 @@ from unittest.mock import AsyncMock, patch
|
||||
|
||||
from httpx import ASGITransport, AsyncClient
|
||||
|
||||
from app.core.config import settings
|
||||
from app.db.database import get_db
|
||||
from app.main import app
|
||||
from app.services.adapters.text.models import StoryOutput
|
||||
@@ -343,3 +344,221 @@ async def test_voice_session_list_orders_recent_sessions_first(
|
||||
}
|
||||
finally:
|
||||
app.dependency_overrides.clear()
|
||||
|
||||
|
||||
async def test_voice_session_active_endpoint_returns_latest_active_session(
|
||||
db_session,
|
||||
auth_token,
|
||||
):
|
||||
async def override_get_db():
|
||||
yield db_session
|
||||
|
||||
app.dependency_overrides[get_db] = override_get_db
|
||||
|
||||
with (
|
||||
patch(
|
||||
"app.services.voice_session_service.generate_story_content",
|
||||
new_callable=AsyncMock,
|
||||
) as mock_generate,
|
||||
patch(
|
||||
"app.services.voice_session_service.text_to_speech",
|
||||
new_callable=AsyncMock,
|
||||
) as mock_tts,
|
||||
):
|
||||
mock_generate.return_value = StoryOutput(
|
||||
mode="generated",
|
||||
title="活动会话",
|
||||
story_text="一段活动中的故事。",
|
||||
cover_prompt_suggestion="活动会话封面",
|
||||
)
|
||||
mock_tts.return_value = b"active-audio"
|
||||
|
||||
transport = ASGITransport(app=app)
|
||||
try:
|
||||
async with AsyncClient(transport=transport, base_url="http://test") as client:
|
||||
client.cookies.set("access_token", auth_token)
|
||||
|
||||
response = await client.post("/api/voice-sessions", json={})
|
||||
old_session_id = response.json()["id"]
|
||||
await client.post(
|
||||
f"/api/voice-sessions/{old_session_id}/abandon",
|
||||
json={"reason": "旧会话结束"},
|
||||
)
|
||||
|
||||
response = await client.post("/api/voice-sessions", json={})
|
||||
active_session_id = response.json()["id"]
|
||||
await client.post(
|
||||
f"/api/voice-sessions/{active_session_id}/turns/fallback",
|
||||
json={"transcript_text": "请继续一个新故事"},
|
||||
)
|
||||
|
||||
response = await client.get("/api/voice-sessions/active")
|
||||
assert response.status_code == 200
|
||||
data = response.json()
|
||||
assert data["id"] == active_session_id
|
||||
assert data["can_continue"] is True
|
||||
assert data["status"] == "waiting_user"
|
||||
finally:
|
||||
app.dependency_overrides.clear()
|
||||
|
||||
|
||||
async def test_voice_session_can_retry_failed_turn_from_saved_transcript(
|
||||
db_session,
|
||||
auth_token,
|
||||
):
|
||||
async def override_get_db():
|
||||
yield db_session
|
||||
|
||||
app.dependency_overrides[get_db] = override_get_db
|
||||
|
||||
with (
|
||||
patch(
|
||||
"app.services.voice_session_service.generate_story_content",
|
||||
new_callable=AsyncMock,
|
||||
) as mock_generate,
|
||||
patch(
|
||||
"app.services.voice_session_service.text_to_speech",
|
||||
new_callable=AsyncMock,
|
||||
) as mock_tts,
|
||||
):
|
||||
mock_generate.side_effect = [
|
||||
RuntimeError("provider down"),
|
||||
StoryOutput(
|
||||
mode="generated",
|
||||
title="重试成功",
|
||||
story_text="重试后的故事终于顺利继续了。",
|
||||
cover_prompt_suggestion="重试封面",
|
||||
),
|
||||
]
|
||||
mock_tts.return_value = b"retry-turn-audio"
|
||||
|
||||
transport = ASGITransport(app=app)
|
||||
try:
|
||||
async with AsyncClient(transport=transport, base_url="http://test") as client:
|
||||
client.cookies.set("access_token", auth_token)
|
||||
|
||||
response = await client.post("/api/voice-sessions", json={})
|
||||
session_id = response.json()["id"]
|
||||
|
||||
response = await client.post(
|
||||
f"/api/voice-sessions/{session_id}/turns/fallback",
|
||||
json={"transcript_text": "先讲一个会失败的回合"},
|
||||
)
|
||||
assert response.status_code == 202
|
||||
failed_turn_id = response.json()["turn_id"]
|
||||
|
||||
response = await client.get(
|
||||
f"/api/voice-sessions/{session_id}/turns/{failed_turn_id}"
|
||||
)
|
||||
assert response.status_code == 200
|
||||
assert response.json()["status"] == "failed"
|
||||
|
||||
response = await client.post(
|
||||
f"/api/voice-sessions/{session_id}/turns/{failed_turn_id}/retry"
|
||||
)
|
||||
assert response.status_code == 202
|
||||
retried_turn_id = response.json()["turn_id"]
|
||||
assert retried_turn_id != failed_turn_id
|
||||
|
||||
response = await client.get(
|
||||
f"/api/voice-sessions/{session_id}/turns/{retried_turn_id}"
|
||||
)
|
||||
assert response.status_code == 200
|
||||
retried_turn = response.json()
|
||||
assert retried_turn["status"] == "audio_ready"
|
||||
assert retried_turn["assistant_text"] == "重试后的故事终于顺利继续了。"
|
||||
finally:
|
||||
app.dependency_overrides.clear()
|
||||
|
||||
|
||||
async def test_voice_session_can_retry_missing_assistant_audio(
|
||||
db_session,
|
||||
auth_token,
|
||||
):
|
||||
async def override_get_db():
|
||||
yield db_session
|
||||
|
||||
app.dependency_overrides[get_db] = override_get_db
|
||||
|
||||
with (
|
||||
patch(
|
||||
"app.services.voice_session_service.generate_story_content",
|
||||
new_callable=AsyncMock,
|
||||
) as mock_generate,
|
||||
patch(
|
||||
"app.services.voice_session_service.text_to_speech",
|
||||
new_callable=AsyncMock,
|
||||
) as mock_tts,
|
||||
):
|
||||
mock_generate.return_value = StoryOutput(
|
||||
mode="generated",
|
||||
title="补发语音",
|
||||
story_text="这一轮先有文本，稍后再补发语音。",
|
||||
cover_prompt_suggestion="补发语音封面",
|
||||
)
|
||||
mock_tts.side_effect = [RuntimeError("tts down"), b"recovered-audio"]
|
||||
|
||||
transport = ASGITransport(app=app)
|
||||
try:
|
||||
async with AsyncClient(transport=transport, base_url="http://test") as client:
|
||||
client.cookies.set("access_token", auth_token)
|
||||
|
||||
response = await client.post("/api/voice-sessions", json={})
|
||||
session_id = response.json()["id"]
|
||||
|
||||
response = await client.post(
|
||||
f"/api/voice-sessions/{session_id}/turns/fallback",
|
||||
json={"transcript_text": "请先给我一段只有文本的结果"},
|
||||
)
|
||||
assert response.status_code == 202
|
||||
turn_id = response.json()["turn_id"]
|
||||
|
||||
response = await client.get(
|
||||
f"/api/voice-sessions/{session_id}/turns/{turn_id}"
|
||||
)
|
||||
assert response.status_code == 200
|
||||
turn = response.json()
|
||||
assert turn["status"] == "narrative_ready"
|
||||
assert turn["assistant_audio_ready"] is False
|
||||
|
||||
response = await client.post(
|
||||
f"/api/voice-sessions/{session_id}/turns/{turn_id}/retry-audio"
|
||||
)
|
||||
assert response.status_code == 200
|
||||
retried = response.json()
|
||||
assert retried["status"] == "audio_ready"
|
||||
assert retried["assistant_audio_ready"] is True
|
||||
finally:
|
||||
app.dependency_overrides.clear()
|
||||
|
||||
|
||||
async def test_voice_session_uploaded_audio_respects_size_limit(
|
||||
db_session,
|
||||
auth_token,
|
||||
monkeypatch,
|
||||
):
|
||||
async def override_get_db():
|
||||
yield db_session
|
||||
|
||||
app.dependency_overrides[get_db] = override_get_db
|
||||
monkeypatch.setattr(settings, "voice_turn_max_upload_bytes", 4)
|
||||
|
||||
transport = ASGITransport(app=app)
|
||||
try:
|
||||
async with AsyncClient(transport=transport, base_url="http://test") as client:
|
||||
client.cookies.set("access_token", auth_token)
|
||||
|
||||
response = await client.post("/api/voice-sessions", json={})
|
||||
assert response.status_code == 201
|
||||
session_id = response.json()["id"]
|
||||
|
||||
response = await client.post(
|
||||
f"/api/voice-sessions/{session_id}/turns",
|
||||
files={
|
||||
"audio_file": ("turn.webm", b"12345", "audio/webm"),
|
||||
},
|
||||
data={"transcript_hint": "太长了"},
|
||||
)
|
||||
assert response.status_code == 413
|
||||
finally:
|
||||
app.dependency_overrides.clear()
|
||||
|
||||
@@ -45,6 +45,7 @@ export interface VoiceSessionSummary {
|
||||
latest_detected_intent: string | null
|
||||
latest_assistant_audio_ready: boolean
|
||||
last_turn_status: string | null
|
||||
transcription_mode_hint: string | null
|
||||
can_continue: boolean
|
||||
can_finalize: boolean
|
||||
last_error: string | null
|
||||
|
||||
@@ -1,8 +1,10 @@
|
||||
<script setup lang="ts">
|
||||
import { ref } from 'vue'
|
||||
import { onMounted, ref, watch } from 'vue'
|
||||
import { useRouter } from 'vue-router'
|
||||
import { useI18n } from 'vue-i18n'
|
||||
import { useUserStore } from '../stores/user'
|
||||
import { api } from '../api/client'
|
||||
import type { VoiceSessionSummary } from '../types/voiceSession'
|
||||
import BaseButton from '../components/ui/BaseButton.vue'
|
||||
import LoginDialog from '../components/ui/LoginDialog.vue'
|
||||
import {
|
||||
@@ -25,6 +27,7 @@ function switchLocale(lang: 'en' | 'zh') {
|
||||
|
||||
// ========== 登录对话框状态 ==========
|
||||
const showLoginDialog = ref(false)
|
||||
const activeVoiceSession = ref<VoiceSessionSummary | null>(null)
|
||||
|
||||
// ========== 创作入口 ==========
|
||||
// 旧的创作变量已移除，现在只负责跳转
|
||||
@@ -45,6 +48,26 @@ function openVoiceStudio() {
|
||||
router.push('/voice-studio')
|
||||
}
|
||||
|
||||
function continueVoiceStudio() {
|
||||
if (!activeVoiceSession.value) {
|
||||
openVoiceStudio()
|
||||
return
|
||||
}
|
||||
router.push('/voice-studio')
|
||||
}
|
||||
|
||||
async function loadActiveVoiceSession() {
|
||||
if (!userStore.user) {
|
||||
activeVoiceSession.value = null
|
||||
return
|
||||
}
|
||||
try {
|
||||
activeVoiceSession.value = await api.get<VoiceSessionSummary | null>('/api/voice-sessions/active')
|
||||
} catch {
|
||||
activeVoiceSession.value = null
|
||||
}
|
||||
}
|
||||
|
||||
function scrollToFeatures() {
|
||||
document.getElementById('features')?.scrollIntoView({ behavior: 'smooth' })
|
||||
}
|
||||
@@ -54,6 +77,20 @@ function scrollToFeatures() {
|
||||
// const familiesCount = 5000
|
||||
// const satisfactionCount = 99
|
||||
|
||||
onMounted(async () => {
|
||||
if (!userStore.user) {
|
||||
await userStore.fetchSession()
|
||||
}
|
||||
await loadActiveVoiceSession()
|
||||
})
|
||||
|
||||
watch(
|
||||
() => userStore.user?.id,
|
||||
() => {
|
||||
void loadActiveVoiceSession()
|
||||
},
|
||||
)
|
||||
|
||||
</script>
|
||||
|
||||
<template>
|
||||
@@ -152,6 +189,15 @@ function scrollToFeatures() {
|
||||
<MicrophoneIcon class="h-5 w-5 mr-2" />
|
||||
进入语音共创
|
||||
</BaseButton>
|
||||
<BaseButton
|
||||
v-if="activeVoiceSession"
|
||||
size="lg"
|
||||
variant="ghost"
|
||||
@click="continueVoiceStudio"
|
||||
>
|
||||
<ArrowRightOnRectangleIcon class="h-5 w-5 mr-2" />
|
||||
继续语音共创
|
||||
</BaseButton>
|
||||
<button @click="scrollToFeatures" class="px-6 py-3 rounded-xl font-semibold text-stone-600 bg-white border border-stone-200 hover:border-amber-400 hover:text-amber-700 transition-all shadow-sm">
|
||||
了解更多功能
|
||||
</button>
|
||||
|
||||
@@ -8,6 +8,7 @@ import BaseCard from '../components/ui/BaseCard.vue'
|
||||
import EmptyState from '../components/ui/EmptyState.vue'
|
||||
import LoadingSpinner from '../components/ui/LoadingSpinner.vue'
|
||||
import type { GenerationOpsSummary, GenerationProviderAnalytics } from '../types/generation'
|
||||
import type { VoiceSessionSummary } from '../types/voiceSession'
|
||||
import {
|
||||
getAssetStatusMeta,
|
||||
getGenerationStatusMeta,
|
||||
@@ -40,6 +41,7 @@ const router = useRouter()
|
||||
const stories = ref<StoryItem[]>([])
|
||||
const providerAnalytics = ref<GenerationProviderAnalytics | null>(null)
|
||||
const opsSummary = ref<GenerationOpsSummary | null>(null)
|
||||
const activeVoiceSession = ref<VoiceSessionSummary | null>(null)
|
||||
const loading = ref(true)
|
||||
const error = ref('')
|
||||
const showCreateModal = ref(false)
|
||||
@@ -76,14 +78,16 @@ function buildProviderAnalyticsPath() {
|
||||
|
||||
async function fetchStories() {
|
||||
try {
|
||||
const [storyList, analytics, ops] = await Promise.all([
|
||||
const [storyList, analytics, ops, activeSession] = await Promise.all([
|
||||
api.get<StoryItem[]>('/api/stories'),
|
||||
api.get<GenerationProviderAnalytics>(buildProviderAnalyticsPath()),
|
||||
api.get<GenerationOpsSummary>('/api/generations/ops-summary'),
|
||||
api.get<VoiceSessionSummary | null>('/api/voice-sessions/active').catch(() => null),
|
||||
])
|
||||
stories.value = storyList
|
||||
providerAnalytics.value = analytics
|
||||
opsSummary.value = ops
|
||||
activeVoiceSession.value = activeSession
|
||||
} catch (e) {
|
||||
error.value = e instanceof Error ? e.message : '加载失败'
|
||||
} finally {
|
||||
@@ -112,6 +116,10 @@ function goToCreate() {
|
||||
showCreateModal.value = true
|
||||
}
|
||||
|
||||
function goToVoiceStudio() {
|
||||
router.push('/voice-studio')
|
||||
}
|
||||
|
||||
function getStoryLink(story: StoryItem) {
|
||||
return story.mode === 'storybook' ? `/storybook/view/${story.id}` : `/story/${story.id}`
|
||||
}
|
||||
@@ -200,6 +208,27 @@ watch([selectedWindow, selectedCapability], () => {
|
||||
</div>
|
||||
|
||||
<template v-else>
|
||||
<BaseCard
|
||||
v-if="activeVoiceSession"
|
||||
class="mb-8 border border-purple-100 bg-purple-50/60"
|
||||
padding="lg"
|
||||
>
|
||||
<div class="flex flex-col gap-4 lg:flex-row lg:items-center lg:justify-between">
|
||||
<div>
|
||||
<h2 class="text-xl font-bold text-gray-800">语音共创正在进行中</h2>
|
||||
<p class="mt-2 text-sm leading-6 text-gray-600">
|
||||
最近的语音共创会话仍可继续：
|
||||
{{ activeVoiceSession.working_title || '未命名语音会话' }}，
|
||||
当前状态 {{ activeVoiceSession.status }}，已完成 {{ activeVoiceSession.total_turns }} 轮。
|
||||
</p>
|
||||
</div>
|
||||
<BaseButton @click="goToVoiceStudio">
|
||||
<SparklesIcon class="h-5 w-5 mr-2" />
|
||||
继续语音共创
|
||||
</BaseButton>
|
||||
</div>
|
||||
</BaseCard>
|
||||
|
||||
<BaseCard class="mb-8" padding="lg">
|
||||
<div class="grid grid-cols-2 lg:grid-cols-4 gap-4">
|
||||
<div class="text-center px-4 py-2">
|
||||
|
||||
@@ -47,6 +47,7 @@ const profiles = ref<ChildProfile[]>([])
|
||||
const universes = ref<StoryUniverse[]>([])
|
||||
const selectedProfileId = ref('')
|
||||
const selectedUniverseId = ref('')
|
||||
const sessionFilter = ref<'active' | 'recent'>('active')
|
||||
const textTurnInput = ref('')
|
||||
const uploadTranscriptHint = ref('')
|
||||
const loadingSessions = ref(false)
|
||||
@@ -60,12 +61,14 @@ const recordingDurationMs = ref(0)
|
||||
const error = ref('')
|
||||
const mediaError = ref('')
|
||||
const recorderSupported = computed(() => typeof window !== 'undefined' && 'MediaRecorder' in window)
|
||||
const sessionPollIntervalMs = 1500
|
||||
|
||||
let mediaRecorder: MediaRecorder | null = null
|
||||
let mediaStream: MediaStream | null = null
|
||||
let recordingChunks: Blob[] = []
|
||||
let recordingTimer: number | null = null
|
||||
let recordingStartedAt = 0
|
||||
let sessionPollTimer: number | null = null
|
||||
|
||||
const recordedBlob = ref<Blob | null>(null)
|
||||
const recordedAudioUrl = ref<string | null>(null)
|
||||
@@ -78,6 +81,22 @@ const universeOptions = computed(() =>
|
||||
)
|
||||
|
||||
const activeTurnList = computed(() => activeSession.value?.recent_turns ?? [])
|
||||
const transcriptionModeDescription = computed(() => {
|
||||
switch (activeSession.value?.transcription_mode_hint) {
|
||||
case 'openai':
|
||||
return '当前会尝试使用 OpenAI 语音转写，转写提示可以留空。'
|
||||
case 'disabled':
|
||||
return '当前环境禁用了真实语音转写，请先使用文本共创或填写开发转写提示。'
|
||||
default:
|
||||
return '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。'
|
||||
}
|
||||
})
|
||||
const isSessionProcessing = computed(
|
||||
() =>
|
||||
activeSession.value?.status === 'processing_turn'
|
||||
|| activeSession.value?.last_turn_status === 'received'
|
||||
|| activeSession.value?.last_turn_status === 'transcribing',
|
||||
)
|
||||
|
||||
function formatSessionStatus(status: string) {
|
||||
switch (status) {
|
||||
@@ -182,8 +201,13 @@ 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 params = new URLSearchParams({
|
||||
limit: '8',
|
||||
active_first: 'true',
|
||||
active_only: sessionFilter.value === 'active' ? 'true' : 'false',
|
||||
})
|
||||
sessions.value = await api.get<VoiceSessionSummary[]>(`/api/voice-sessions?${params.toString()}`)
|
||||
if (!activeSession.value && sessionFilter.value === 'recent') {
|
||||
const resumable = sessions.value.find((item) => item.can_continue)
|
||||
if (resumable) {
|
||||
await loadSessionDetail(resumable.id)
|
||||
@@ -196,6 +220,18 @@ async function loadSessions() {
|
||||
}
|
||||
}
|
||||
|
||||
async function loadLatestActiveSession() {
|
||||
if (!userStore.user) return
|
||||
try {
|
||||
const session = await api.get<VoiceSessionSummary | null>('/api/voice-sessions/active')
|
||||
if (session) {
|
||||
await loadSessionDetail(session.id)
|
||||
}
|
||||
} catch {
|
||||
// Ignore active-session bootstrap failures and fall back to normal listing.
|
||||
}
|
||||
}
|
||||
|
||||
async function loadSessionDetail(sessionId: string) {
|
||||
loadingSessionDetail.value = true
|
||||
error.value = ''
|
||||
@@ -208,6 +244,23 @@ async function loadSessionDetail(sessionId: string) {
|
||||
}
|
||||
}
|
||||
|
||||
function stopSessionPolling() {
|
||||
if (sessionPollTimer) {
|
||||
window.clearInterval(sessionPollTimer)
|
||||
sessionPollTimer = null
|
||||
}
|
||||
}
|
||||
|
||||
function startSessionPolling() {
|
||||
if (!activeSession.value?.id || sessionPollTimer) return
|
||||
sessionPollTimer = window.setInterval(() => {
|
||||
if (activeSession.value?.id) {
|
||||
void loadSessionDetail(activeSession.value.id)
|
||||
void loadSessions()
|
||||
}
|
||||
}, sessionPollIntervalMs)
|
||||
}
|
||||
|
||||
async function createSession() {
|
||||
creatingSession.value = true
|
||||
error.value = ''
|
||||
@@ -217,6 +270,7 @@ async function createSession() {
|
||||
universe_id: selectedUniverseId.value || null,
|
||||
target_mode: 'story',
|
||||
})
|
||||
sessionFilter.value = 'active'
|
||||
await loadSessions()
|
||||
await loadSessionDetail(session.id)
|
||||
textTurnInput.value = ''
|
||||
@@ -319,6 +373,39 @@ async function finalizeSession() {
|
||||
}
|
||||
}
|
||||
|
||||
async function retryFailedTurn(turnId: string) {
|
||||
if (!activeSession.value) return
|
||||
sendingTurn.value = true
|
||||
error.value = ''
|
||||
try {
|
||||
const result = await api.post<VoiceTurnAcceptedResponse>(
|
||||
`/api/voice-sessions/${activeSession.value.id}/turns/${turnId}/retry`,
|
||||
)
|
||||
await refreshAfterTurn(result.session_id, result.turn_id)
|
||||
} catch (err) {
|
||||
error.value = err instanceof Error ? err.message : '重试本轮故事失败'
|
||||
} finally {
|
||||
sendingTurn.value = false
|
||||
}
|
||||
}
|
||||
|
||||
async function retryAssistantAudio(turnId: string) {
|
||||
if (!activeSession.value) return
|
||||
sendingTurn.value = true
|
||||
error.value = ''
|
||||
try {
|
||||
await api.post<VoiceTurnSummary>(
|
||||
`/api/voice-sessions/${activeSession.value.id}/turns/${turnId}/retry-audio`,
|
||||
)
|
||||
await loadSessionDetail(activeSession.value.id)
|
||||
await loadSessions()
|
||||
} catch (err) {
|
||||
error.value = err instanceof Error ? err.message : '补发语音失败'
|
||||
} finally {
|
||||
sendingTurn.value = false
|
||||
}
|
||||
}
|
||||
|
||||
async function abandonSession() {
|
||||
if (!activeSession.value) return
|
||||
abandoning.value = true
|
||||
@@ -404,16 +491,34 @@ watch(selectedProfileId, (newId) => {
|
||||
}
|
||||
})
|
||||
|
||||
watch(sessionFilter, () => {
|
||||
void loadSessions()
|
||||
})
|
||||
|
||||
watch(
|
||||
() => isSessionProcessing.value,
|
||||
(processing) => {
|
||||
if (processing) {
|
||||
startSessionPolling()
|
||||
} else {
|
||||
stopSessionPolling()
|
||||
}
|
||||
},
|
||||
{ immediate: true },
|
||||
)
|
||||
|
||||
onMounted(async () => {
|
||||
if (!userStore.user) {
|
||||
await userStore.fetchSession()
|
||||
}
|
||||
if (!userStore.user) return
|
||||
await fetchProfiles()
|
||||
await loadLatestActiveSession()
|
||||
await loadSessions()
|
||||
})
|
||||
|
||||
onBeforeUnmount(() => {
|
||||
stopSessionPolling()
|
||||
if (recording.value) {
|
||||
stopRecording()
|
||||
}
|
||||
@@ -488,6 +593,29 @@ onBeforeUnmount(() => {
|
||||
<span class="text-xs text-gray-400">{{ sessions.length }} 个</span>
|
||||
</div>
|
||||
|
||||
<div class="mt-4 flex gap-2">
|
||||
<button
|
||||
type="button"
|
||||
class="rounded-lg border px-3 py-1.5 text-sm transition-colors"
|
||||
:class="sessionFilter === 'active'
|
||||
? 'border-purple-600 bg-purple-600 text-white'
|
||||
: 'border-gray-200 bg-white text-gray-600 hover:border-gray-400'"
|
||||
@click="sessionFilter = 'active'"
|
||||
>
|
||||
活跃会话
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
class="rounded-lg border px-3 py-1.5 text-sm transition-colors"
|
||||
:class="sessionFilter === 'recent'
|
||||
? 'border-purple-600 bg-purple-600 text-white'
|
||||
: 'border-gray-200 bg-white text-gray-600 hover:border-gray-400'"
|
||||
@click="sessionFilter = 'recent'"
|
||||
>
|
||||
最近全部
|
||||
</button>
|
||||
</div>
|
||||
|
||||
<div v-if="loadingSessions" class="py-8">
|
||||
<LoadingSpinner text="加载会话中..." />
|
||||
</div>
|
||||
@@ -558,6 +686,13 @@ onBeforeUnmount(() => {
|
||||
</p>
|
||||
</div>
|
||||
|
||||
<div
|
||||
v-if="isSessionProcessing"
|
||||
class="rounded-2xl border border-amber-100 bg-amber-50 px-4 py-3 text-sm text-amber-700"
|
||||
>
|
||||
当前会话仍在处理中，页面会自动轮询刷新。
|
||||
</div>
|
||||
|
||||
<div class="flex flex-wrap gap-3">
|
||||
<BaseButton
|
||||
variant="secondary"
|
||||
@@ -614,7 +749,7 @@ onBeforeUnmount(() => {
|
||||
<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。
|
||||
{{ transcriptionModeDescription }}
|
||||
</p>
|
||||
|
||||
<div class="mt-4 flex flex-wrap gap-3">
|
||||
@@ -714,6 +849,28 @@ onBeforeUnmount(() => {
|
||||
<div v-if="turn.assistant_audio_url" class="mt-3">
|
||||
<audio class="w-full" :src="turn.assistant_audio_url" controls></audio>
|
||||
</div>
|
||||
<div class="mt-4 flex flex-wrap gap-3">
|
||||
<BaseButton
|
||||
v-if="turn.status === 'failed'"
|
||||
size="sm"
|
||||
variant="secondary"
|
||||
@click="retryFailedTurn(turn.id)"
|
||||
:disabled="sendingTurn || !activeSession?.can_continue"
|
||||
>
|
||||
<ArrowPathIcon class="h-4 w-4" />
|
||||
重试本轮
|
||||
</BaseButton>
|
||||
<BaseButton
|
||||
v-if="turn.assistant_text && !turn.assistant_audio_ready"
|
||||
size="sm"
|
||||
variant="ghost"
|
||||
@click="retryAssistantAudio(turn.id)"
|
||||
:disabled="sendingTurn"
|
||||
>
|
||||
<ArrowPathIcon class="h-4 w-4" />
|
||||
补发语音
|
||||
</BaseButton>
|
||||
</div>
|
||||
<div v-if="turn.error_message" class="mt-3 text-sm text-rose-600">
|
||||
{{ turn.error_message }}
|
||||
</div>
|
||||
|
||||
Reference in New Issue
Block a user