Compare commits

..

2 Commits

Author SHA1 Message Date
CI Bot 5ba8e4b13f style: auto-format with black + isort + prettier [skip ci-format-check]
CI/CD Pipeline / Check push changed paths (pull_request) Has been skipped
CI/CD Pipeline / Dedup Check - skip PR tests when covered by push pipeline (pull_request) Successful in 2s
CI/CD Pipeline / Check if frontend-only change (pull_request) Successful in 2s
CI/CD Pipeline / Build Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / Build Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Build Staging Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Worker Image (pull_request) Has been skipped
CI/CD Pipeline / PR Build API Image (pull_request) Successful in 50s
CI/CD Pipeline / Deploy Staging (Watchtower auto-deploy) (pull_request) Has been skipped
CI/CD Pipeline / PR Build Worker Image (pull_request) Successful in 54s
CI/CD Pipeline / Staging E2E Tests (pull_request) Has been skipped
CI/CD Pipeline / ACR Image Cleanup (pull_request) Has been skipped
CI/CD Pipeline / Staging API Integration Tests (pull_request) Has been skipped
CI/CD Pipeline / Frontend Unit Tests (pull_request) Successful in 1m20s
Preview Deploy / Deploy Preview Environment (pull_request) Successful in 1m39s
CI/CD Pipeline / Integration Tests (pull_request) Successful in 2m10s
CI/CD Pipeline / Unit Tests (pull_request) Successful in 2m6s
CI/CD Pipeline / Frontend Lint (pull_request) Successful in 2m5s
CI/CD Pipeline / PR Build Web Image (pull_request) Successful in 2m6s
CI/CD Pipeline / Validate - Style (pull_request) Failing after 2m17s
CI/CD Pipeline / Validate - Python (mypy + alembic) (pull_request) Successful in 2m28s
PR Automation / Auto Approve on CI Green (pull_request) Successful in 3m18s
AI Code Review / AI Code Review (pull_request) Successful in 6m17s
CI/CD Pipeline / Validate - Security (pull_request) Successful in 7m33s
CI/CD Pipeline / Build Production API Image (pull_request) Has been skipped
CI/CD Pipeline / Build Production Web Image (pull_request) Has been skipped
CI/CD Pipeline / Build Production Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Canary Release to Production (pull_request) Has been skipped
CI/CD Pipeline / Deploy Production (pull_request) Has been skipped
CI/CD Pipeline / Production Browser E2E (pull_request) Has been skipped
CI/CD Pipeline / CI Gate (pull_request) Failing after 4s
PR Automation / Auto Merge on CI Green + Approved (pull_request) Successful in 4m26s
2026-09-10 07:20:56 +00:00
CI Test 3d2e94b884 fix(#1830): 改进异常处理 - 缩小异常类型并添加日志记录
CI/CD Pipeline / Check push changed paths (pull_request) Has been skipped
CI/CD Pipeline / Dedup Check - skip PR tests when covered by push pipeline (pull_request) Successful in 2s
CI/CD Pipeline / Check if frontend-only change (pull_request) Successful in 2s
CI/CD Pipeline / Build Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Frontend Lint (pull_request) Has been skipped
CI/CD Pipeline / Frontend Unit Tests (pull_request) Has been skipped
CI/CD Pipeline / Build Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / PR Build Web Image (pull_request) Has been skipped
CI/CD Pipeline / Build Staging Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / Deploy Staging (Watchtower auto-deploy) (pull_request) Has been skipped
CI/CD Pipeline / Staging E2E Tests (pull_request) Has been skipped
CI/CD Pipeline / Staging API Integration Tests (pull_request) Has been skipped
CI/CD Pipeline / ACR Image Cleanup (pull_request) Has been skipped
CI/CD Pipeline / PR Build API Image (pull_request) Successful in 22s
CI/CD Pipeline / PR Build Worker Image (pull_request) Successful in 59s
CI/CD Pipeline / Integration Tests (pull_request) Successful in 1m37s
CI/CD Pipeline / Unit Tests (pull_request) Successful in 1m59s
Preview Deploy / Deploy Preview Environment (pull_request) Successful in 2m0s
CI/CD Pipeline / Validate - Python (mypy + alembic) (pull_request) Successful in 2m4s
CI/CD Pipeline / Validate - Style (pull_request) Failing after 2m6s
CI/CD Pipeline / Validate - Security (pull_request) Has been cancelled
CI/CD Pipeline / Build Production API Image (pull_request) Has been cancelled
CI/CD Pipeline / Build Production Web Image (pull_request) Has been cancelled
CI/CD Pipeline / Build Production Worker Image (pull_request) Has been cancelled
CI/CD Pipeline / Deploy Production (pull_request) Has been cancelled
CI/CD Pipeline / Production Browser E2E (pull_request) Has been cancelled
CI/CD Pipeline / Canary Release to Production (pull_request) Has been cancelled
CI/CD Pipeline / CI Gate (pull_request) Has been cancelled
AI Code Review / AI Code Review (pull_request) Has been cancelled
PR Automation / Auto Approve on CI Green (pull_request) Has been cancelled
PR Automation / Auto Merge on CI Green + Approved (pull_request) Has been cancelled
2026-09-10 15:18:35 +08:00
41 changed files with 313 additions and 2206 deletions
+10 -1
View File
@@ -196,7 +196,7 @@ jobs:
- name: Run style checks
shell: bash
run: bash scripts/ci/validate_style.sh
- name: Auto-fix formatting (black + isort + ruff)
- name: Auto-fix formatting (black + isort)
if: failure()
shell: sh
env:
@@ -827,6 +827,9 @@ jobs:
CACHE_REF="${REGISTRY}/${{ matrix.cache_name }}:develop"
EXTRA_BUILD_ARGS="APP_VERSION=\"${GITHUB_SHA}\""
if [ "${{ matrix.service }}" = "web" ]; then
EXTRA_BUILD_ARGS="$EXTRA_BUILD_ARGS NGINX_CONF=infra/docker/nginx-staging.conf"
fi
# Worker 与 API/Web 统一走持久 builderci-builder-persist),共享宿主机层缓存
NO_CACHE_FLAG=""
@@ -1023,6 +1026,9 @@ jobs:
CACHE_REF="${REGISTRY}/${{ matrix.cache_name }}:${GITHUB_REF_NAME}"
EXTRA_BUILD_ARGS="APP_VERSION=\"${GITHUB_SHA}\""
if [ "${{ matrix.service }}" = "web" ]; then
EXTRA_BUILD_ARGS="$EXTRA_BUILD_ARGS NGINX_CONF=infra/docker/nginx-staging.conf"
fi
NO_CACHE_FLAG=""
for i in 1 2 3; do
@@ -1561,6 +1567,9 @@ jobs:
CACHE_REF="${REGISTRY}/${{ matrix.cache_name }}:main"
EXTRA_BUILD_ARGS="APP_VERSION=\"${TAG_NAME}\""
if [ "${{ matrix.service }}" = "web" ]; then
EXTRA_BUILD_ARGS="$EXTRA_BUILD_ARGS NGINX_CONF=infra/docker/nginx-production.conf"
fi
# Docker build 带重试:失败自动重试2次,第2次重试加--no-cache
NO_CACHE_FLAG=""
-60
View File
@@ -1,60 +0,0 @@
name: "Debug: Web container v2 (mount conflict)"
on:
push:
branches: [debug/web-crash-v2]
workflow_dispatch:
jobs:
web-diag:
runs-on: runtime-builder
timeout-minutes: 10
steps:
- name: Setup SSH and diagnose
shell: bash
env:
STAGING_SSH_KEY: ${{ secrets.PREVIEW_SSH_KEY }}
run: |
set -x
which ssh || (apt-get update -qq && apt-get install -y -qq openssh-client)
mkdir -p ~/.ssh && chmod 700 ~/.ssh
printf "%s" "$STAGING_SSH_KEY" > ~/.ssh/id_rsa
chmod 600 ~/.ssh/id_rsa
H=47.98.113.167; P=22222
ssh-keyscan -p $P -H $H >> ~/.ssh/known_hosts 2>/dev/null
ssh -p $P -i ~/.ssh/id_rsa -o StrictHostKeyChecking=no root@$H 'bash -s' <<'REMOTE'
set -x
echo "=== Current staging containers ==="
docker ps -a --filter name=xiaoxia-*-staging --format "table {{.Names}}\t{{.Status}}\t{{.Image}}"
echo ""
echo "=== Web container logs (current/current-rolledback) ==="
docker logs xiaoxia-web-staging 2>&1 | tail -40
echo ""
echo "=== Web inspect: env & mounts ==="
docker inspect xiaoxia-web-staging --format 'Entrypoint: {{.Config.Entrypoint}} Cmd: {{.Config.Cmd}}'
docker inspect xiaoxia-web-staging --format '{{range .Config.Env}}{{.}}{{"\n"}}{{end}}' | grep -E "APP_ENV|VERSION"
echo "Mounts:"
docker inspect xiaoxia-web-staging --format '{{range .Mounts}}{{.Type}} {{.Source}} -> {{.Destination}} (rw={{.RW}}){{"\n"}}{{end}}'
echo ""
echo "=== Reproduce: rm on read-only bind mount ==="
docker run --rm --name nginx-ro-test \
-v /var/lib/xiaoxia-saas-staging/nginx-staging.conf:/etc/nginx/conf.d/default.conf:ro \
git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas/xiaoxia-saas-web:387514c \
sh -c '
set -x
echo "Before:"
ls -la /etc/nginx/conf.d/
echo "Try rm (as entrypoint does):"
rm -f /etc/nginx/conf.d/default.conf
echo "rm exitcode=$?"
echo "After rm:"
ls -la /etc/nginx/conf.d/
echo "Test ln:"
ln -s /etc/nginx/nginx-staging.conf /etc/nginx/conf.d/default.conf
echo "ln exitcode=$?"
ls -la /etc/nginx/conf.d/
echo "nginx -t:"
nginx -t 2>&1
' 2>&1
echo ""
echo "=== Also test with NEW fixed image (9c0d4b1 if present) ==="
docker images | grep xiaoxia-saas-web | head -5
REMOTE
View File
+8 -34
View File
@@ -11,7 +11,6 @@
from __future__ import annotations
import logging
from datetime import datetime, timezone
from app.auth import AuthenticatedUser, get_current_user
from app.dependencies import get_db_session
@@ -78,16 +77,10 @@ def create_render_job(
from app.tasks.ai_avatar_render import execute_ai_avatar_render
execute_ai_avatar_render.delay(job.id)
except Exception as exc:
logger.exception("Celery 任务投递失败(创建): job_id=%s err=%s", job.id, exc)
job.status = "failed"
job.error_message = f"任务提交失败:{exc}"
job.updated_at = datetime.now(timezone.utc)
svc.db.commit()
svc.db.refresh(job)
return AiAvatarRenderJobResponse.model_validate(job)
except Exception:
logger.warning("Celery 任务提交失败,渲染任务已创建但未触发执行: %s", job.id)
return AiAvatarRenderJobResponse.model_validate(job)
return job
# ── GET /jobs — 任务列表 ─────────────────────────────────────────────────
@@ -179,16 +172,10 @@ def retry_render_job(
from app.tasks.ai_avatar_render import execute_ai_avatar_render
execute_ai_avatar_render.delay(job.id)
except Exception as exc:
logger.exception("Celery 任务投递失败重试: job_id=%s err=%s", job.id, exc)
job.status = "failed"
job.error_message = f"任务提交失败:{exc}"
job.updated_at = datetime.now(timezone.utc)
svc.db.commit()
svc.db.refresh(job)
return AiAvatarRenderJobResponse.model_validate(job)
except Exception:
logger.warning("Celery 任务提交失败重试任务已重置但未触发执行: %s", job.id)
return AiAvatarRenderJobResponse.model_validate(job)
return job
@@ -211,25 +198,12 @@ def generate_avatar_smart_cover(
if not video_url.startswith(("http://", "https://")):
raise HTTPException(status_code=400, detail="video_url 必须是合法的 HTTP/HTTPS URL")
try:
cover_url = generate_smart_cover(
video_url,
max_frames=body.max_frames,
title_config=getattr(body, "title_config", None),
)
except Exception as exc:
logger.error(
"智能封面生成异常: user=%s video_url=%s err=%s",
current_user.user.id, video_url[:80], exc,
exc_info=True,
)
cover_url = ""
cover_url = generate_smart_cover(video_url, max_frames=body.max_frames)
if not cover_url:
return SmartCoverResponse(
cover_url="",
status="fallback_failed",
message="智能抽帧失败(MediaKit 不可用或抽帧异常),请稍后重试",
)
logger.info("智能封面生成成功: user=%s cover_url=%s", current_user.user.id, cover_url[:120])
logger.info("智能封面生成成功: user=%s", current_user.user.id)
return SmartCoverResponse(cover_url=cover_url, status="completed")
+1 -1
View File
@@ -336,7 +336,7 @@ def get_project_asset_diagnosis(
assets.extend(asset_repository.list_by_library(library.id))
except HTTPException:
raise
except Exception:
except Exception as e:
logger.exception("素材诊断查询失败: project_id=%s", project_id)
# 返回空诊断结果,避免 500
return _build_diagnosis(project_id, [])
+1 -1
View File
@@ -177,7 +177,7 @@ def _cleanup_expired_uploads() -> int:
meta_file.unlink()
cleaned += 1
logger.info(f"Cleaned up expired upload: {upload_id}")
except Exception:
except Exception as e:
logger.exception("Failed to cleanup upload metadata: %s", meta_file)
return cleaned
@@ -107,7 +107,7 @@ def create_variant_plans(
)
if _latest:
source_plan_id = _latest.id
except Exception:
except Exception as e:
logger.exception("[variant-plans] 源 plan 解析失败")
if not source_plan_id:
@@ -121,7 +121,7 @@ def create_variant_plans(
from app.api.routes.generation_tasks import _query_voice_durations
voice_durations = _query_voice_durations(db, voices)
except Exception:
except Exception as e:
logger.exception("[variant-plans] 配音时长查询失败(按占位段长选片)")
voice_durations = [0.0] * request.count
+20 -7
View File
@@ -14,6 +14,7 @@ import logging
from app.auth import AuthenticatedUser, get_current_user
from app.dependencies import (
get_cosyvoice_service,
get_db_session,
get_voice_clone_profile_repository,
)
@@ -23,6 +24,8 @@ from app.services.mediakit_client import MediaKitError
from fastapi import APIRouter, BackgroundTasks, Depends, HTTPException, Query
from sqlalchemy.orm import Session
from packages.application.cosyvoice_service import CosyVoiceError, CosyVoiceService
logger = logging.getLogger(__name__)
router = APIRouter()
@@ -31,11 +34,13 @@ router = APIRouter()
def _get_service(
db: Session = Depends(get_db_session),
voice_clone_repo=Depends(get_voice_clone_profile_repository),
cosyvoice_service: CosyVoiceService = Depends(get_cosyvoice_service),
) -> LipsyncService:
# voice_clone_repo 用于克隆音色 profile 解析
# TTS 合成已移至 Celery 异步任务,无需同步注入 cosyvoice_service
# voice_clone_repo 用于克隆音色 profile 解析cosyvoice_service 用于 TTS 直生
# TTS 合成、音色解析、错误码归一化都在 LipsyncService 内部完成)
return LipsyncService(
db,
cosyvoice_service=cosyvoice_service,
voice_clone_repo=voice_clone_repo,
)
@@ -52,8 +57,8 @@ def create_lipsync_job(
"""提交对口型任务.
#1809/#1822: 前端传 {video_url, voice_id, script_text, speed?, emotion?}
后端创建任务记录(状态 tts_processing),dispatch Celery 异步任务执行 TTS 合成 + MediaKit 提交
也支持直接传 {video_url, audio_url}(同步提交 MediaKit
后端内部解析音色、调 TTS 合成音频、转存 OSS,再提交 MediaKit
也支持直接传 {video_url, audio_url}。
"""
try:
job = svc.create_job(
@@ -70,13 +75,21 @@ def create_lipsync_job(
except ValueError as exc:
# 参数无效(如 voice_id 格式不对、文本过长等)
raise HTTPException(status_code=400, detail=str(exc)) from exc
except CosyVoiceError as exc:
# TTS 合成基础设施失败(API/网络/认证)
raise HTTPException(
status_code=502,
detail={"code": "TTSSynthesisFailed", "message": str(exc)},
) from exc
except MediaKitError as exc:
# 音色无权访问 → 403;参数无效 → 400MediaKit 提交失败 → 502
# TTS 合成失败 / 音色无权访问 → 400/403MediaKit 提交失败 → 502
status_code = 502
if exc.code in ("VoiceForbidden",):
status_code = 403
elif exc.code in ("InvalidInput", "TTSInvalidParam", "VoiceNotReady"):
status_code = 400
elif exc.code == "TTSSynthesisFailed":
status_code = 502
raise HTTPException(
status_code=status_code,
detail={
@@ -174,13 +187,13 @@ def cancel_lipsync_job(
current_user: AuthenticatedUser = Depends(get_current_user),
svc: LipsyncService = Depends(_get_service),
):
"""取消对口型任务(仅 pending/tts_processing/submitted 状态可取消)."""
"""取消对口型任务(仅 pending/submitted 状态可取消)."""
job = svc.cancel_job(job_id, current_user.user.id)
if job is None:
raise HTTPException(status_code=404, detail="任务不存在")
if job.status != "cancelled":
raise HTTPException(
status_code=400,
detail=f"任务状态 {job.status} 不可取消,仅 pending/tts_processing/submitted 可取消",
detail=f"任务状态 {job.status} 不可取消,仅 pending/submitted 可取消",
)
return job
@@ -65,7 +65,7 @@ def _build_asset_analyses(
if url:
video_urls.append(url)
valid_asset_ids.append(aid)
except Exception:
except Exception as e:
logger.exception("获取素材URL失败: asset_id=%s", aid)
if not video_urls:
@@ -176,7 +176,7 @@ def editor_ai_recommend(
)
try:
db.rollback()
except Exception:
except Exception as e:
logger.exception("db rollback failed in ai_recommend")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
@@ -140,7 +140,7 @@ def _build_asset_url_map(
try:
assets = asset_repo.find_by_ids(unique_ids)
asset_map = {a.id: a for a in assets}
except Exception:
except Exception as e:
logger.exception("批量查询素材失败: asset_ids=%s", asset_ids)
return {aid: None for aid in asset_ids if aid}
@@ -155,7 +155,7 @@ def _build_asset_url_map(
result[aid] = None
continue
result[aid] = storage.get_download_url(storage_key, expires_seconds=3600)
except Exception:
except Exception as e:
logger.exception("生成素材签名URL失败: asset_id=%s", aid)
result[aid] = None
@@ -486,7 +486,7 @@ def _get_mediakit_recommendations(
if url:
video_urls.append(url)
valid_asset_ids.append(asset_id)
except Exception:
except Exception as e:
logger.exception("获取素材URL失败: asset_id=%s", asset_id)
if not video_urls:
@@ -987,7 +987,7 @@ def _update_mediakit_recommendations_async( # pragma: no cover
if storage_key and mime.startswith("video/"):
try:
video_url = storage.get_download_url(storage_key)
except Exception:
except Exception as e:
logger.exception("后台任务: 获取素材URL失败: asset_id=%s", asset_id)
# 构建该素材的占用区间列表(排除已更新片段)
@@ -1039,7 +1039,7 @@ def _update_mediakit_recommendations_async( # pragma: no cover
asset_id,
len(scene_changes),
)
except Exception:
except Exception as cache_err:
# 缓存写入失败不影响本次片段更新
logger.exception(
"后台任务: 场景点缓存写入失败: asset_id=%s",
@@ -1123,7 +1123,7 @@ def _update_mediakit_recommendations_async( # pragma: no cover
recommended_start + clip_duration,
plan_id,
)
except Exception:
except Exception as me:
logger.exception(
"后台任务: 同步素材区间记录失败,回滚本次片段更新: clip_id=%s",
clip.id,
@@ -1142,27 +1142,27 @@ def _update_mediakit_recommendations_async( # pragma: no cover
asset_id,
recommended_start,
)
except Exception:
except Exception as ue:
logger.exception("后台任务: 单个片段更新失败: clip_id=%s", clip.id)
try:
db.rollback()
except Exception:
except Exception as e:
pass
continue
logger.info("后台任务完成: plan_id=%s 成功更新 %d 个片段", plan_id, updated_count)
except Exception:
except Exception as e:
# 后台任务失败不影响已创建的片段,静默处理
logger.exception("后台任务异常: plan_id=%s", plan_id)
if db:
try:
db.rollback()
except Exception:
except Exception as e:
pass
finally:
if db:
try:
db.close()
except Exception:
except Exception as e:
pass
+2 -2
View File
@@ -167,7 +167,7 @@ def create_voice_clone(
# P2-3: Celery 调度失败时标记 profile 为 failed,避免永久卡在 processing
try:
workflow.process_clone_failure(profile.id, f"Celery 任务调度失败: {e}")
except Exception:
except Exception as inner_e:
logger.exception("Failed to mark profile as failed after dispatch error")
return _to_response(profile)
@@ -281,7 +281,7 @@ def retry_voice_clone(
# P2-3: Celery 调度失败时标记 profile 为 failed,避免永久卡在 processing
try:
workflow.process_clone_failure(profile.id, f"Celery 任务调度失败: {e}")
except Exception:
except Exception as inner_e:
logger.exception("Failed to mark profile as failed after dispatch error")
return _to_response(profile)
+3 -3
View File
@@ -105,7 +105,7 @@ def _resolve_preset_preview_url(
_preset_preview_cache[voice_id] = (audio_url, time.time())
logger.info("Preset voice preview generated: %s", voice_id)
return audio_url
except Exception:
except Exception as e:
logger.exception("Failed to generate preset voice preview: voice_id=%s", voice_id)
return fallback_url
@@ -126,7 +126,7 @@ def _resolve_all_preset_preview_urls(
for p in presets:
try:
result_map[p.voice_id] = _resolve_preset_preview_url(p.voice_id, p.preview_url, cosyvoice)
except Exception:
except Exception as e:
logger.exception("Failed to resolve preset preview URL: voice_id=%s", p.voice_id)
result_map[p.voice_id] = p.preview_url
return result_map
@@ -732,7 +732,7 @@ def _find_or_create_voice_library_for_extract(*, user_id, project_repository, as
if session is not None:
try:
session.rollback()
except Exception:
except Exception as e:
logger.exception("session rollback failed in _find_or_create_voice_library")
for lib in asset_library_repository.find_by_project(project.id):
kind = lib.kind.value if hasattr(lib.kind, "value") else lib.kind
View File
+1 -4
View File
@@ -110,13 +110,10 @@ class AiAvatarRenderProgressResponse(BaseModel):
class SmartCoverRequest(BaseModel):
"""智能封面请求 — MediaKit 抽帧 + 质量评分选最佳帧 + 可选标题 drawtext 叠加."""
"""智能封面请求 — MediaKit 抽帧 + 质量评分选最佳帧."""
video_url: str = Field(..., description="数字人视频 URL(对口型/渲染成片)")
max_frames: int = Field(5, ge=1, le=10, description="抽帧数量(默认 5")
title_config: Optional[dict[str, Any]] = Field(
None, description="标题配置;传入时在封面上用 drawtext 叠加标题(竖屏 720x1280"
)
class SmartCoverResponse(BaseModel):
+37 -188
View File
@@ -11,55 +11,13 @@
from __future__ import annotations
import logging
import os
import subprocess
import tempfile
import uuid
from pathlib import Path
from typing import Optional
from urllib.parse import urlparse
logger = logging.getLogger(__name__)
# MediaKit 抽帧轮询参数(与 MediaKit API timeout=60s 对齐)
COVER_POLL_INTERVAL = 3.0
COVER_MAX_POLL_ATTEMPTS = 20 # 最多等 60 秒
# 帧图片下载超时(秒)
FRAME_DOWNLOAD_TIMEOUT = 20
# 最佳帧下载超时(用于 persist)
BEST_FRAME_DOWNLOAD_TIMEOUT = 30
# 自家 OSS 私有桶 URL 重签有效期(供 MediaKit GPU worker 拉取)
MEDIAKIT_URL_TTL_SECONDS = 7 * 24 * 3600
def _sign_video_url_for_mediakit(video_url: str) -> str:
"""如果 video_url 是自家 OSS 私有桶 URL,重新签名为长有效期预签名 URL。
MediaKit GPU worker 需要能公网访问 video_url,裸 public_url 在私有桶下会 403。
"""
if not video_url:
return video_url
try:
from packages.shared.storage import get_shared_storage_service
storage = get_shared_storage_service()
public_base = getattr(storage, "public_url", "")
if not isinstance(public_base, str) or not public_base:
return video_url
own_host = urlparse(public_base).netloc.lower()
url_host = urlparse(video_url).netloc.lower()
if own_host and url_host == own_host:
# 是自家 OSS URL,重签 7 天有效期供 MediaKit 拉取
signed = storage.get_download_url(video_url, expires_seconds=MEDIAKIT_URL_TTL_SECONDS)
if signed:
logger.info("[数字人封面] video_url 已重签(自家 OSS 私有桶)")
return signed
except Exception:
logger.warning("[数字人封面] video_url 重签失败,使用原始 URL", exc_info=True)
return video_url
def select_best_cover_frame(video_url: str, *, max_frames: int = 5) -> str:
"""从视频抽取多帧并评分选最佳帧,返回最佳帧的临时 URL.
@@ -73,10 +31,6 @@ def select_best_cover_frame(video_url: str, *, max_frames: int = 5) -> str:
"""
if not video_url:
return ""
# 确保 MediaKit 能访问 video_url(自家 OSS 私有桶需重签)
video_url = _sign_video_url_for_mediakit(video_url)
try:
from packages.shared.cover_frame_scorer import score_frames
from packages.shared.mediakit_client import get_mediakit_client
@@ -86,21 +40,13 @@ def select_best_cover_frame(video_url: str, *, max_frames: int = 5) -> str:
logger.warning("[数字人封面] MediaKit 未配置,无法智能抽帧")
return ""
logger.info(
"[数字人封面] 开始抽帧: video_url=%s max_frames=%d poll_interval=%.1f max_poll=%d",
video_url[:80],
max_frames,
COVER_POLL_INTERVAL,
COVER_MAX_POLL_ATTEMPTS,
)
snapshots = mk.extract_frames(
video_url=video_url,
strategy="SpecifiedFrames",
max_frames=max_frames,
poll_interval=COVER_POLL_INTERVAL,
max_poll_attempts=COVER_MAX_POLL_ATTEMPTS,
max_retries=1,
poll_interval=2.0,
max_poll_attempts=5,
max_retries=0,
)
if not snapshots:
logger.warning("[数字人封面] MediaKit 未返回帧: %s", video_url[:80])
@@ -109,26 +55,24 @@ def select_best_cover_frame(video_url: str, *, max_frames: int = 5) -> str:
if len(snapshots) == 1:
return snapshots[0].get("image_url") or snapshots[0].get("url") or ""
# 使用连接池下载各帧(复用 TCP 连接,减少延迟)
# 下载各帧评分
import httpx
candidates = []
with httpx.Client(timeout=FRAME_DOWNLOAD_TIMEOUT, follow_redirects=True) as client:
for snap in snapshots:
url = snap.get("image_url") or snap.get("url") or ""
if not url:
continue
tmp_path: Optional[str] = None
try:
resp = client.get(url)
resp.raise_for_status()
with tempfile.NamedTemporaryFile(suffix=".jpg", delete=False) as tmp:
tmp.write(resp.content)
tmp_path = tmp.name
candidates.append({"image_path": tmp_path, "url": url})
except Exception as e:
logger.warning("[数字人封面] 帧下载失败,跳过: url=%s err=%s", url[:80], e)
candidates.append({"image_path": None, "url": url, "score": 0.0})
for snap in snapshots:
url = snap.get("image_url") or snap.get("url") or ""
if not url:
continue
tmp_path: Optional[str] = None
try:
resp = httpx.get(url, timeout=15, follow_redirects=True)
resp.raise_for_status()
with tempfile.NamedTemporaryFile(suffix=".jpg", delete=False) as tmp:
tmp.write(resp.content)
tmp_path = tmp.name
candidates.append({"image_path": tmp_path, "url": url})
except Exception:
candidates.append({"image_path": None, "url": url, "score": 0.0})
if not candidates:
return snapshots[0].get("image_url") or snapshots[0].get("url") or ""
@@ -158,79 +102,13 @@ def select_best_cover_frame(video_url: str, *, max_frames: int = 5) -> str:
return ""
def apply_title_to_cover(local_frame: str, *, title_config: dict | None) -> str:
"""用 ffmpeg drawtext 在封面图上叠加标题,返回叠加后图片的本地路径.
ffmpeg 失败时回退返回原始 local_frame。竖屏封面按 720x1280 计算位置。
"""
if not title_config or not isinstance(title_config, dict):
return local_frame
text = (title_config.get("text") or title_config.get("content") or "").strip()
if not text:
return local_frame
enabled = title_config.get("enabled", True)
if not enabled:
return local_frame
try:
from packages.domain.video_filter_builder import build_title_drawtext_filter
drawtext_filter = build_title_drawtext_filter(
title_config,
output_width=720,
output_height=1280,
)
if not drawtext_filter:
return local_frame
base, ext = os.path.splitext(local_frame)
titled_path = f"{base}_titled{ext or '.jpg'}"
cmd = [
"ffmpeg",
"-i",
local_frame,
"-vf",
drawtext_filter,
"-y",
titled_path,
]
logger.info("[数字人封面] 叠加标题: text=%s", text[:30])
result = subprocess.run(
cmd,
capture_output=True,
text=True,
timeout=30,
)
if result.returncode != 0:
logger.warning(
"[数字人封面] drawtext 失败,回退无标题: exit=%s stderr=%s",
result.returncode,
(result.stderr or "")[-300:],
)
return local_frame
if not os.path.exists(titled_path) or os.path.getsize(titled_path) == 0:
logger.warning("[数字人封面] drawtext 输出为空,回退无标题")
return local_frame
return titled_path
except Exception as exc:
logger.warning("[数字人封面] 标题叠加异常,回退无标题: %s", exc, exc_info=True)
return local_frame
def persist_cover_to_oss(
frame_url: str,
*,
job_id: str = "",
prefix: str = "ai-avatar/covers",
title_config: dict | None = None,
) -> str:
def persist_cover_to_oss(frame_url: str, *, job_id: str = "", prefix: str = "ai-avatar/covers") -> str:
"""下载帧图并转存到 OSS,返回公网封面 URL.
Args:
frame_url: MediaKit 返回的临时帧图 URL
job_id: 关联任务 ID(用于 OSS key 命名)
prefix: OSS key 前缀
title_config: 可选标题配置;传入时用 drawtext 叠加标题(竖屏 720x1280
Returns:
OSS 公网 URL;失败回退原始 frame_url
@@ -238,76 +116,47 @@ def persist_cover_to_oss(
if not frame_url:
return ""
tmp_path: Optional[str] = None
titled_path: Optional[str] = None
try:
import httpx
with httpx.Client(timeout=BEST_FRAME_DOWNLOAD_TIMEOUT, follow_redirects=True) as client:
resp = client.get(frame_url)
resp.raise_for_status()
if not resp.content:
logger.warning("[数字人封面] 帧图内容为空: %s", frame_url[:80])
return frame_url
resp = httpx.get(frame_url, timeout=30, follow_redirects=True)
resp.raise_for_status()
if not resp.content:
return frame_url
with tempfile.NamedTemporaryFile(suffix=".jpg", delete=False) as tmp:
tmp.write(resp.content)
tmp_path = tmp.name
with tempfile.NamedTemporaryFile(suffix=".jpg", delete=False) as tmp:
tmp.write(resp.content)
tmp_path = tmp.name
from packages.shared.storage import get_shared_storage_service
storage = get_shared_storage_service()
token = job_id or uuid.uuid4().hex[:12]
cover_key = f"{prefix}/{token}/cover_{uuid.uuid4().hex[:8]}.jpg"
upload_path = apply_title_to_cover(tmp_path, title_config=title_config)
if upload_path != tmp_path:
titled_path = upload_path
public_url = storage.upload_file(
file_or_path=upload_path,
file_or_path=tmp_path,
storage_key=cover_key,
content_type="image/jpeg",
)
logger.info(
"[数字人封面] 封面已转存 OSS: key=%s titled=%s",
cover_key,
bool(titled_path),
)
# 私有桶:返回预签名 URL(前端才能加载)
if public_url:
signed = storage.get_download_url(cover_key, expires_seconds=86400)
return signed
return frame_url
logger.info("[数字人封面] 封面已转存 OSS: key=%s", cover_key)
return public_url or frame_url
except Exception:
logger.warning("[数字人封面] 封面转存 OSS 失败,返回原始 URL", exc_info=True)
return frame_url
finally:
for p in (tmp_path, titled_path):
if p:
try:
Path(p).unlink(missing_ok=True)
except Exception:
pass
if tmp_path:
try:
Path(tmp_path).unlink(missing_ok=True)
except Exception:
pass
def generate_smart_cover(
video_url: str,
*,
job_id: str = "",
max_frames: int = 5,
title_config: dict | None = None,
) -> str:
"""一站式:MediaKit 智能抽帧选最佳 → (可选)drawtext 叠加标题 → 转存 OSS.
def generate_smart_cover(video_url: str, *, job_id: str = "", max_frames: int = 5) -> str:
"""一站式:MediaKit 智能抽帧选最佳 → 转存 OSS,返回封面公网 URL.
供独立封面接口与渲染管线复用。失败返回空字符串。
Args:
video_url: 可公网访问的视频 URL
job_id: 关联任务 ID
max_frames: 抽帧数量
title_config: 可选标题配置;传入时在封面上叠加 drawtext 标题(竖屏 720x1280
"""
best_frame = select_best_cover_frame(video_url, max_frames=max_frames)
if not best_frame:
return ""
return persist_cover_to_oss(best_frame, job_id=job_id, title_config=title_config)
return persist_cover_to_oss(best_frame, job_id=job_id)
+22 -142
View File
@@ -11,7 +11,6 @@ from __future__ import annotations
import logging
import os
import subprocess
import tempfile
import uuid
from datetime import datetime, timezone
@@ -25,6 +24,7 @@ from packages.adapters.sqlalchemy_impl.models import (
ScriptModel,
)
from packages.domain.video_filter_builder import (
build_cover_extract_command,
build_title_drawtext_filter,
)
from packages.shared.storage import get_shared_storage_service
@@ -257,7 +257,7 @@ class AiAvatarRenderService:
with tempfile.TemporaryDirectory() as tmpdir:
output_video_path = os.path.join(tmpdir, "output.mp4")
cmd_list = self._build_ffmpeg_command(
cmd = self._build_ffmpeg_command(
input_video=input_video_path,
b_roll_segments=job.b_roll_segments,
filter_complex=filter_complex,
@@ -265,25 +265,9 @@ class AiAvatarRenderService:
output_path=output_video_path,
)
try:
render_result = subprocess.run(
cmd_list,
capture_output=True,
text=True,
timeout=600,
)
except subprocess.TimeoutExpired as exc:
raise AiAvatarRenderError(
"FFmpeg 渲染超时(600s",
code="FFmpegTimeout",
) from exc
if render_result.returncode != 0:
stderr_tail = (render_result.stderr or "").strip()[-800:]
raise AiAvatarRenderError(
f"FFmpeg 渲染失败,退出码: {render_result.returncode}, stderr: {stderr_tail}",
code="FFmpegFailed",
)
exit_code = os.system(cmd)
if exit_code != 0:
raise AiAvatarRenderError(f"FFmpeg 渲染失败,退出码: {exit_code}", code="FFmpegFailed")
job.progress = 80
self.db.commit()
@@ -292,27 +276,11 @@ class AiAvatarRenderService:
cover_path = ""
if job.cover_config:
cover_path = os.path.join(tmpdir, "cover.jpg")
cover_cmd = self._build_cover_extract_cmd(
cover_config=job.cover_config,
input_video=output_video_path,
output_path=cover_path,
)
try:
cover_result = subprocess.run(
cover_cmd,
capture_output=True,
text=True,
timeout=60,
)
if cover_result.returncode != 0:
logger.warning(
"封面提取失败(非致命),跳过: exit=%s stderr=%s",
cover_result.returncode,
(cover_result.stderr or "")[-300:],
)
cover_path = ""
except Exception as cover_err:
logger.warning("封面提取异常(非致命),跳过: %s", cover_err)
cover_cmd = build_cover_extract_command(job.cover_config, cover_path)
cover_cmd = cover_cmd.replace("INPUT_VIDEO", output_video_path)
cover_exit = os.system(cover_cmd)
if cover_exit != 0:
logger.warning("封面提取失败,跳过: %s", cover_cmd)
cover_path = ""
job.progress = 90
@@ -322,7 +290,7 @@ class AiAvatarRenderService:
output_video_url = self._upload_to_oss(output_video_path, f"ai-avatar/{job_id}/output.mp4")
job.output_video_url = output_video_url
# 封面:优先复用智能剪辑的 MediaKit 抽帧 + 质量评分选最佳帧(支持 drawtext 标题叠加)
# 封面:优先复用智能剪辑的 MediaKit 抽帧 + 质量评分选最佳帧;
# MediaKit 不可用时回退到 FFmpeg 已按 cover_config 抽取的 cover_path
smart_cover_url = ""
if output_video_url:
@@ -331,12 +299,7 @@ class AiAvatarRenderService:
generate_smart_cover,
)
smart_cover_url = generate_smart_cover(
output_video_url,
job_id=job_id,
max_frames=5,
title_config=job.title_config,
)
smart_cover_url = generate_smart_cover(output_video_url, job_id=job_id, max_frames=5)
except Exception:
logger.warning("智能封面(MediaKit)失败,回退 FFmpeg 封面 job_id=%s", job_id, exc_info=True)
@@ -359,52 +322,18 @@ class AiAvatarRenderService:
self.db.commit()
logger.info("渲染任务完成: %s", job_id)
# 7. 自动保存成片记录到成片库
if job.output_video_url:
try:
from packages.adapters.sqlalchemy_impl.generated_video_repository import (
SQLAlchemyGeneratedVideoRepository,
)
from packages.domain.generated_video import GeneratedVideo
clip_name = f"AI数字人_{job_id[:8]}"
clip = GeneratedVideo.create(
project_id=job.project_id,
generation_task_id=job.lipsync_job_id,
name=clip_name,
file_url=job.output_video_url,
user_id=job.user_id,
duration=job.output_duration or 0.0,
thumbnail_url=job.output_cover_url or None,
generation_params={
"source": "ai_avatar_render",
"render_job_id": job.id,
},
)
video_repo = SQLAlchemyGeneratedVideoRepository(self.db)
video_repo.create(clip)
logger.info("成片记录已保存到成片库: clip_id=%s, render_job=%s", clip.id, job_id)
except Exception as clip_err:
logger.warning(
"自动保存成片记录失败(不影响渲染任务状态): render_job=%s, error=%s",
job_id,
clip_err,
)
except AiAvatarRenderError as exc:
job.status = "failed"
job.error_message = str(exc)
job.updated_at = datetime.now(timezone.utc)
self.db.commit()
logger.error("渲染任务失败 [%s]: %s", job_id, exc)
raise
except Exception as exc:
job.status = "failed"
job.error_message = f"渲染异常: {str(exc)}"
job.updated_at = datetime.now(timezone.utc)
self.db.commit()
logger.exception("渲染任务异常 [%s]", job_id)
raise
def _download_video(self, url: str) -> str:
"""下载视频到临时文件."""
@@ -430,73 +359,24 @@ class AiAvatarRenderService:
filter_complex: str,
final_label: Optional[str],
output_path: str,
) -> list[str]:
"""构建 FFmpeg 命令list 形式,shell=False.
根因修复 #1798 P0OSS 预签名 URL 含 `&Expires=...&Signature=...` 特殊字符,
os.system(shell=True) 会把 `&` 解释为后台命令分隔符,导致 -filter_complex 被
当成独立命令报 sh: -filter_complex: not foundexit 127 → Python 32512)。
list + shell=False 彻底规避 shell 转义问题。
"""
cmd: list[str] = ["ffmpeg", "-i", input_video]
) -> str:
"""构建 FFmpeg 命令."""
# 输入文件
inputs = f"-i {input_video}"
for seg in b_roll_segments:
asset_url = seg.get("asset_url", "")
if asset_url:
cmd.extend(["-i", asset_url])
inputs += f" -i {asset_url}"
# 滤镜
if filter_complex and final_label:
cmd.extend(["-filter_complex", filter_complex, "-map", f"[{final_label}]"])
filter_arg = f'-filter_complex "{filter_complex}" -map "[{final_label}]"'
elif filter_complex:
cmd.extend(["-filter_complex", filter_complex])
cmd.extend(
[
"-c:v",
"libx264",
"-preset",
"veryfast",
"-crf",
"23",
"-y",
output_path,
]
)
return cmd
def _build_cover_extract_cmd(
self,
*,
cover_config: dict[str, Any],
input_video: str,
output_path: str,
) -> list[str]:
"""构建封面截帧 FFmpeg 命令(list 形式,shell=False."""
if not cover_config or not isinstance(cover_config, dict):
timestamp = 0.0
width = 0
height = 0
filter_arg = f'-filter_complex "{filter_complex}"'
else:
timestamp = cover_config.get("timestamp", 0.0)
width = cover_config.get("width", 0)
height = cover_config.get("height", 0)
filter_arg = ""
cmd: list[str] = [
"ffmpeg",
"-ss",
str(timestamp),
"-i",
input_video,
"-frames:v",
"1",
]
if width > 0 and height > 0:
vf = (
f"scale={width}:{height}:force_original_aspect_ratio=decrease,"
f"pad={width}:{height}:(ow-iw)/2:(oh-ih)/2"
)
cmd.extend(["-vf", vf])
cmd.extend(["-y", output_path])
return cmd
return f"ffmpeg {inputs} {filter_arg} -c:v libx264 -preset fast -crf 23 -y {output_path}"
def _upload_to_oss(self, local_path: str, oss_key: str) -> str:
"""上传文件到 OSS,返回 URL.
+40 -64
View File
@@ -25,9 +25,6 @@ from app.services.mediakit_client import (
MediaKitError,
get_mediakit_client,
)
# Celery 异步任务:TTS 合成 + MediaKit 提交(#lipsync-speed-optimization
from app.tasks.lipsync_tts import tts_synthesize_and_submit
from sqlalchemy.orm import Session
from packages.adapters.sqlalchemy_impl.models import LipsyncJobModel
@@ -157,31 +154,41 @@ class LipsyncService:
enable_video_loop: bool = False,
project_id: str = "",
) -> LipsyncJobModel:
"""创建对口型任务.
"""创建对口型任务并提交到 MediaKit.
两种输入模式:
- TTS 直生:voice_id + script_textaudio_url 留空)
→ 先创建 DB 记录(状态 tts_processing),再 dispatch Celery 异步任务
执行 TTS 合成 + MediaKit 提交。API 响应 <1s。
- TTS 直生:voice_id + script_textaudio_url 留空),后端先合成音频
- 直接音频:提供 audio_url
→ 同步提交 MediaKit,状态直接设为 submitted。
Raises:
MediaKitError: 参数校验失败或 MediaKit 提交失败(仅直接音频模式)
MediaKitError: TTS 合成或 MediaKit 提交失败
"""
# 0. 输入校验
# 0. TTS 直生模式:先合成音频(在创建 DB 记录之前完成,失败直接抛出)
if not audio_url:
if not (voice_id and script_text):
raise MediaKitError(
"必须提供 audio_url 或 voice_id+script_text",
code="InvalidInput",
)
# TTS 模式:在 HTTP 请求中同步校验音色归属,快速失败
self._resolve_voice_id(voice_id, user_id)
# 预合成:用临时 job_id 命名 OSS 对象
pre_job_id = str(uuid.uuid4())
audio_url = self._synthesize_and_persist_audio(
user_id=user_id,
job_id=pre_job_id,
voice_id=voice_id,
script_text=script_text,
speed=speed,
emotion=emotion,
)
# #1839: 私有桶 OSS 的裸 URL / 即将过期的短预签名会让 MediaKit GPU worker 拉取时 403
# 提交前统一重签长有效期;外部临时 URLCosyVoice/MediaKit)原样透传。
video_url = self._sign_media_url(video_url)
if audio_url:
audio_url = self._sign_media_url(audio_url)
# 1. 创建数据库记录
job_id = str(uuid.uuid4())
is_tts_mode = not bool(audio_url)
job = LipsyncJobModel(
id=job_id,
user_id=user_id,
@@ -193,59 +200,28 @@ class LipsyncService:
script_text=script_text or "",
speed=speed,
emotion=normalize_emotion(emotion),
status="tts_processing" if is_tts_mode else "pending",
status="pending",
)
self.db.add(job)
self.db.flush()
if is_tts_mode:
# 2a. TTS 模式:dispatch Celery 异步任务处理 TTS 合成 + MediaKit 提交
try:
tts_synthesize_and_submit.apply_async(
args=(
job_id,
user_id,
voice_id,
script_text,
speed,
normalize_emotion(emotion),
)
)
except Exception as exc:
# 投递失败时立即把 job 标成 failed 并写入 error_message
# 前端轮询时能直接看到失败原因,不会无限卡在 tts_processing。
logger.exception(
"Celery 任务提交失败,TTS 任务已创建但未触发执行: job_id=%s err=%s",
job_id,
exc,
)
job.status = "failed"
job.error_message = f"Celery 任务投递失败: {exc}"
job.error_code = "AsyncDispatchFailed"
job.updated_at = datetime.now(timezone.utc)
else:
# 2b. 直接音频模式:同步签名并提交 MediaKit
video_url = self._sign_media_url(video_url)
if audio_url:
audio_url = self._sign_media_url(audio_url)
job.audio_url = audio_url
try:
result = self.client.submit_lipsync(
video_url=video_url,
audio_url=audio_url,
enable_video_loop=enable_video_loop,
client_token=job_id,
)
job.mediakit_task_id = result["task_id"]
job.status = "submitted"
job.submitted_at = datetime.now(timezone.utc)
except MediaKitError as exc:
job.status = "failed"
job.error_message = str(exc)
job.error_code = exc.code
logger.error("提交对口型任务失败: %s", exc)
raise
# 3. 提交到 MediaKit
try:
result = self.client.submit_lipsync(
video_url=video_url,
audio_url=audio_url,
enable_video_loop=enable_video_loop,
client_token=job_id, # 幂等控制
)
job.mediakit_task_id = result["task_id"]
job.status = "submitted"
job.submitted_at = datetime.now(timezone.utc)
except MediaKitError as exc:
job.status = "failed"
job.error_message = str(exc)
job.error_code = exc.code
logger.error("提交对口型任务失败: %s", exc)
raise
self.db.commit()
self.db.refresh(job)
@@ -384,12 +360,12 @@ class LipsyncService:
# ── 取消任务 ──────────────────────────────────────────────────────────
def cancel_job(self, job_id: str, user_id: str) -> Optional[LipsyncJobModel]:
"""取消任务(仅 pending/tts_processing/submitted 状态可取消)."""
"""取消任务(仅 pending/submitted 状态可取消)."""
job = self.get_job(job_id, user_id)
if job is None:
return None
if job.status in ("pending", "tts_processing", "submitted"):
if job.status in ("pending", "submitted"):
job.status = "cancelled"
job.updated_at = datetime.now(timezone.utc)
self.db.commit()
-223
View File
@@ -1,223 +0,0 @@
"""AI 数字人对口型 TTS 异步任务 — 将 TTS 合成从 HTTP 请求移至 Celery 后台执行.
优化目标:将 create_job 的 API 响应时间从 6~35s 降到 <1s。
任务流程:
1. 创建新 DB session,加载 job 记录
2. 调用 CosyVoice 合成音频
3. 下载音频并转存到自家 OSS
4. 更新 job 的 audio_url
5. 签名 URL 并提交到 MediaKit
6. 更新 job 状态为 submitted
7. 异常时标记 job 为 failed
注意:使用 @shared_task 而非绑定到某个 celery_app 实例,
确保任务能被 Worker 侧 celery_app 正确注册,同时 API 侧 send_task/apply_async 仍可正常调用。
"""
import io
import logging
from datetime import datetime, timezone
from urllib.parse import urlparse
from celery import shared_task
logger = logging.getLogger(__name__)
# MediaKit 预签名 URL 有效期(7天,秒),与 LipsyncService._sign_media_url 保持一致
_MEDIAKIT_URL_TTL_SECONDS = 7 * 24 * 3600
def _sign_media_url(url: str) -> str:
"""对自家 OSS 私有桶 URL 重签长有效期预签名.
- 自家 OSS URL → 重签 7 天有效期
- 外部临时 URL → 原样透传
- 任何异常降级原样返回,不阻断主流程
"""
if not url:
return url
try:
from packages.shared.storage import get_shared_storage_service
storage = get_shared_storage_service()
public_base = getattr(storage, "public_url", "")
if not isinstance(public_base, str) or not public_base:
return url
own_host = urlparse(public_base).netloc.lower()
host = urlparse(url).netloc.lower()
if not own_host or host != own_host:
return url
signed = storage.get_download_url(url, expires_seconds=_MEDIAKIT_URL_TTL_SECONDS)
return signed or url
except Exception as exc: # noqa: BLE001
logger.warning("[lipsync_tts] URL 重签失败,原样返回: url_prefix=%s err=%s", url[:80], exc)
return url
@shared_task(
bind=True,
name="lipsync_tts.synthesize_and_submit",
max_retries=2,
default_retry_delay=30,
)
def tts_synthesize_and_submit(
self,
job_id: str,
user_id: str,
voice_id: str,
script_text: str,
speed: float,
emotion: str,
):
"""异步执行 TTS 合成 + OSS 转存 + MediaKit 提交.
在 Celery worker 中运行,不阻塞 HTTP 请求。
"""
from app.services.mediakit_client import MediaKitError, get_mediakit_client
from sqlalchemy.orm import Session as DBSession
from packages.adapters.sqlalchemy_impl.models import LipsyncJobModel
from packages.application.cosyvoice_service import CosyVoiceError, CosyVoiceService
from packages.shared.url_security import safe_download_bytes
# SessionLocal 获取:
# - API 容器:app.db.SessionLocal(环境变量完整,导入即建引擎)
# - Worker 容器:worker_app.db.SessionLocalWorker 自己的 settings 初始化引擎)
# API 侧没有 worker_app 模块 → ImportError 直接回退;
# Worker 侧 app.db 会因缺少 API 专有环境变量抛 pydantic ValidationError
# 此时也要回退到 worker_app.db。
try:
from worker_app.db import SessionLocal # type: ignore
except Exception: # noqa: BLE001
from app.db import SessionLocal # type: ignore
db: DBSession = SessionLocal()
try:
job = (
db.query(LipsyncJobModel)
.filter(
LipsyncJobModel.id == job_id,
LipsyncJobModel.user_id == user_id,
)
.first()
)
if job is None:
logger.error("[lipsync_tts] Job not found: job_id=%s", job_id)
return
# 已取消的任务不再处理
if job.status == "cancelled":
logger.info("[lipsync_tts] Job already cancelled, skipping: job_id=%s", job_id)
return
# 1. TTS 合成
try:
cosyvoice = CosyVoiceService()
result = cosyvoice.submit_synthesize_task(
text=script_text,
voice_id=voice_id,
speed=speed,
emotion=emotion,
)
except CosyVoiceError as exc:
logger.error("[lipsync_tts] TTS 合成失败: job_id=%s err=%s", job_id, exc)
job.status = "failed"
job.error_message = f"TTS 合成失败: {exc}"
job.error_code = "TTSSynthesisFailed"
job.updated_at = datetime.now(timezone.utc)
db.commit()
return
except ValueError as exc:
logger.error("[lipsync_tts] TTS 参数错误: job_id=%s err=%s", job_id, exc)
job.status = "failed"
job.error_message = f"TTS 参数错误: {exc}"
job.error_code = "TTSInvalidParam"
job.updated_at = datetime.now(timezone.utc)
db.commit()
return
temp_url = result.get("audio_url", "")
if not temp_url:
logger.error("[lipsync_tts] TTS 未返回音频 URL: job_id=%s", job_id)
job.status = "failed"
job.error_message = "TTS 未返回音频 URL"
job.error_code = "TTSNoAudio"
job.updated_at = datetime.now(timezone.utc)
db.commit()
return
# 2. 下载并转存到自家 OSS
try:
audio_data = safe_download_bytes(
temp_url,
purpose="lipsync_tts_audio",
allowed_mime_types=(
"audio/mpeg",
"audio/mp3",
"audio/wav",
"audio/x-wav", # CosyVoice 部分接口返回 audio/x-wav,与 audio/wav 等价(RIFF/WAVE
"audio/mp4",
"audio/x-m4a",
),
timeout=60.0,
)
from packages.shared.storage import get_shared_storage_service
storage = get_shared_storage_service()
storage_key = f"lipsync-tts/{user_id}/{job_id}.mp3"
permanent_url = storage.upload_file(io.BytesIO(audio_data), storage_key, content_type="audio/mpeg")
logger.info("[lipsync_tts] TTS 音频已转存 OSS: job_id=%s key=%s", job_id, storage_key)
job.audio_url = permanent_url
except Exception as exc:
logger.warning(
"[lipsync_tts] TTS 音频转存 OSS 失败,回退临时 URL: job_id=%s err=%s",
job_id,
exc,
)
job.audio_url = temp_url
db.commit()
# 3. 签名 URL 并提交到 MediaKit(复用模块内 _sign_media_url,避免对 LipsyncService 的耦合)
audio_url = _sign_media_url(job.audio_url)
video_url = _sign_media_url(job.video_url)
client = get_mediakit_client()
try:
mk_result = client.submit_lipsync(
video_url=video_url,
audio_url=audio_url,
enable_video_loop=job.enable_video_loop,
client_token=job_id,
)
job.mediakit_task_id = mk_result["task_id"]
job.status = "submitted"
job.submitted_at = datetime.now(timezone.utc)
logger.info(
"[lipsync_tts] 已提交 MediaKit: job_id=%s task_id=%s",
job_id,
mk_result["task_id"],
)
except MediaKitError as exc:
job.status = "failed"
job.error_message = str(exc)
job.error_code = exc.code
logger.error("[lipsync_tts] 提交 MediaKit 失败: job_id=%s err=%s", job_id, exc)
db.commit()
except Exception:
logger.exception("[lipsync_tts] 未预期的异常: job_id=%s", job_id)
try:
job = db.query(LipsyncJobModel).filter(LipsyncJobModel.id == job_id).first()
if job and job.status not in ("cancelled", "failed", "completed"):
job.status = "failed"
job.error_message = "TTS 异步任务执行异常"
job.error_code = "AsyncTaskError"
job.updated_at = datetime.now(timezone.utc)
db.commit()
except Exception:
logger.exception("[lipsync_tts] 回写失败状态时异常: job_id=%s", job_id)
finally:
db.close()
+3 -20
View File
@@ -549,25 +549,21 @@
/* ── 封面 & 生成 ── */
.aa-cover-preview {
width: 100%;
max-width: 240px;
aspect-ratio: 9/16;
max-height: 160px;
background: #f0f0f5;
border-radius: 12px;
border-radius: 10px;
overflow: hidden;
display: flex;
align-items: center;
justify-content: center;
margin: 0 auto 12px auto;
position: relative;
margin-bottom: 12px;
}
.aa-cover-preview img {
width: 100%;
height: 100%;
aspect-ratio: 9/16;
object-fit: cover;
display: block;
border-radius: 12px;
}
.aa-cover-preview__placeholder {
@@ -575,19 +571,6 @@
color: #8c8ca1;
}
.aa-cover-preview__loading {
position: absolute;
inset: 0;
display: flex;
align-items: center;
justify-content: center;
background: rgba(0, 0, 0, 0.45);
color: #fff;
font-size: 13px;
backdrop-filter: blur(4px);
-webkit-backdrop-filter: blur(4px);
}
.aa-cover-actions {
display: flex;
gap: 8px;
+7 -171
View File
@@ -5,7 +5,6 @@
*/
import React, { useState, useCallback, useEffect, useRef } from "react"
import { message } from "antd"
import { useNavigate } from "react-router-dom"
import "./AiAvatar.css"
import { useAiAvatar } from "./hooks/useAiAvatar"
import { PanelVideoSelector } from "./components/PanelVideoSelector"
@@ -22,7 +21,6 @@ import {
createLipsyncJob,
getLipsyncJob,
submitRender,
getRenderJob,
generateSmartCover,
} from "./api/aiAvatar"
import {
@@ -36,7 +34,6 @@ type PanelKey = "video" | "voice" | "script" | "lipsync" | "title" | "cover"
const AiAvatarPage: React.FC = () => {
const state = useAiAvatar()
const navigate = useNavigate()
const [currentStep, setCurrentStep] = useState<1 | 2>(1)
const [collapsed, setCollapsed] = useState<Record<PanelKey, boolean>>({
video: false,
@@ -55,18 +52,9 @@ const AiAvatarPage: React.FC = () => {
const [lipsyncErrorMessage, setLipsyncErrorMessage] = useState("")
/* ── 智能封面加载态 ── */
const [smartCoverLoading, setSmartCoverLoading] = useState(false)
/* ── 渲染进度弹窗 ── */
const [showRenderModal, setShowRenderModal] = useState(false)
const [renderStatus, setRenderStatus] = useState<"generating" | "completed" | "failed">(
"generating",
)
const [renderProgress, setRenderProgress] = useState(0)
const [renderErrorMessage, setRenderErrorMessage] = useState("")
/* ── 对口型轮询 ── */
const lipsyncTimerRef = useRef<ReturnType<typeof setInterval> | null>(null)
/* ── 渲染进度轮询 ── */
const renderTimerRef = useRef<ReturnType<typeof setInterval> | null>(null)
const togglePanel = useCallback((key: PanelKey) => {
setCollapsed((prev) => ({ ...prev, [key]: !prev[key] }))
@@ -194,60 +182,28 @@ const AiAvatarPage: React.FC = () => {
useEffect(() => {
return () => {
if (lipsyncTimerRef.current) clearInterval(lipsyncTimerRef.current)
if (renderTimerRef.current) clearInterval(renderTimerRef.current)
}
}, [])
/* ── 生成视频(含实时进度轮询) ── */
/* ── 生成视频 ── */
const handleGenerate = useCallback(async () => {
// ② 前置条件提示(#1809
if (!state.lipsyncJob || state.lipsyncJob.status !== "completed") {
message.warning("请先生成对口型视频,待对口型完成后再提交渲染")
return
}
state.setIsGenerating(true)
try {
const job = await submitRender({
await submitRender({
lipsync_job_id: state.lipsyncJob.id,
script_id: state.script?.id,
b_roll_segments: state.bRollSegments.map((seg) => ({
script_segment_index: seg.script_segment_index,
asset_url: seg.asset.file_url || "",
mode: seg.mode,
start_time: seg.start_time,
end_time: seg.end_time,
pip_position: seg.pip_position,
pip_scale: seg.pip_scale,
})) as never,
b_roll_segments: state.bRollSegments as never,
// 字段映射:build_title_drawtext_filter 真实口径 text/font_size/font_color/position/...
title_config: buildTitleConfigPayload(state.titleConfig),
// cover_config:智能封面 cover_url + 截帧 timestamp
cover_config: buildCoverConfigPayload(state.coverConfig, state.coverConfig.smart_cover_url),
})
// 打开渲染进度弹窗,启动轮询
setShowRenderModal(true)
setRenderStatus("generating")
setRenderProgress(job.progress ?? 0)
setRenderErrorMessage("")
if (renderTimerRef.current) clearInterval(renderTimerRef.current)
renderTimerRef.current = setInterval(async () => {
try {
const updated = await getRenderJob(job.id)
setRenderProgress(updated.progress ?? 0)
if (updated.status === "completed") {
if (renderTimerRef.current) clearInterval(renderTimerRef.current)
renderTimerRef.current = null
setRenderStatus("completed")
message.success("视频已生成并保存到成片库")
} else if (updated.status === "failed") {
if (renderTimerRef.current) clearInterval(renderTimerRef.current)
renderTimerRef.current = null
setRenderStatus("failed")
setRenderErrorMessage(updated.error_message || "渲染失败,请重试")
}
} catch (pollErr) {
console.error("[渲染] 轮询失败:", pollErr)
}
}, 3000)
message.success("渲染任务已提交,可在视频管理中查看进度")
} catch (err) {
console.error("渲染任务提交失败:", err)
message.error(err instanceof Error ? err.message : "渲染任务提交失败,请重试")
@@ -257,18 +213,6 @@ const AiAvatarPage: React.FC = () => {
// eslint-disable-next-line react-hooks/exhaustive-deps
}, [state.lipsyncJob, state.script, state.bRollSegments, state.titleConfig, state.coverConfig])
/* ── 关闭渲染进度弹窗 ── */
const handleCancelRender = useCallback(() => {
if (renderTimerRef.current) {
clearInterval(renderTimerRef.current)
renderTimerRef.current = null
}
setShowRenderModal(false)
setRenderStatus("generating")
setRenderProgress(0)
setRenderErrorMessage("")
}, [])
/* ── 智能封面:调后端 MediaKit 选帧接口(#1822 ── */
const handleSmartCover = useCallback(async () => {
// 基于对口型成片抽帧,必须先完成对口型
@@ -408,7 +352,6 @@ const AiAvatarPage: React.FC = () => {
onOpenBRollModal={() => state.setShowBRollModal(true)}
onRemoveBRoll={state.removeBRollSegment}
titleConfig={state.titleConfig}
onTitlePositionChange={(pos) => state.updateTitleConfig(pos)}
/>
<div className="aa-step-btn-row">
<button type="button" className="aa-btn" onClick={handlePrevStep}>
@@ -441,7 +384,6 @@ const AiAvatarPage: React.FC = () => {
<div className="aa-panel__body">
<PanelCoverAndGenerate
coverConfig={state.coverConfig}
titleConfig={state.titleConfig}
onCoverConfigChange={(partial) =>
state.setCoverConfig((prev) => ({ ...prev, ...partial }))
}
@@ -559,112 +501,6 @@ const AiAvatarPage: React.FC = () => {
</div>
</div>
)}
{/* 渲染进度弹窗 */}
{showRenderModal && (
<div className="aa-modal-overlay">
<div className="aa-modal" onClick={(e) => e.stopPropagation()}>
<div className="aa-modal__header">
<span className="aa-modal__title"></span>
<button className="aa-modal__close" onClick={handleCancelRender}>
</button>
</div>
<div
className="aa-modal__body"
style={{
display: "flex",
flexDirection: "column",
alignItems: "center",
padding: "40px 20px",
}}
>
{renderStatus === "generating" && (
<>
<div className="aa-lipsync-spinner" />
<div style={{ marginTop: 20, fontSize: 15, color: "#1a1a2e" }}>
</div>
<div
style={{
marginTop: 16,
fontSize: 32,
fontWeight: 700,
color: "#1890ff",
}}
>
{renderProgress}%
</div>
<div
style={{
marginTop: 12,
width: "80%",
height: 8,
backgroundColor: "#f0f0f0",
borderRadius: 4,
overflow: "hidden",
}}
>
<div
style={{
width: `${renderProgress}%`,
height: "100%",
backgroundColor: "#1890ff",
borderRadius: 4,
transition: "width 0.5s ease",
}}
/>
</div>
<div style={{ marginTop: 12, fontSize: 13, color: "#8c8ca1" }}>
</div>
</>
)}
{renderStatus === "completed" && (
<>
<div style={{ fontSize: 48 }}></div>
<div style={{ marginTop: 16, fontSize: 15, color: "#1a1a2e" }}>
</div>
<button
className="aa-btn"
style={{ marginTop: 16 }}
onClick={() => {
setShowRenderModal(false)
navigate("/app/products")
}}
>
📁
</button>
</>
)}
{renderStatus === "failed" && (
<>
<div style={{ fontSize: 48 }}></div>
<div style={{ marginTop: 16, fontSize: 15, color: "#1a1a2e" }}></div>
{renderErrorMessage && (
<div style={{ marginTop: 8, fontSize: 13, color: "#ff4d4f" }}>
{renderErrorMessage}
</div>
)}
</>
)}
</div>
<div className="aa-modal__footer">
{renderStatus === "generating" && (
<button className="aa-btn aa-btn--danger" onClick={handleCancelRender}>
</button>
)}
{renderStatus !== "generating" && (
<button className="aa-btn" onClick={handleCancelRender}>
</button>
)}
</div>
</div>
</div>
)}
</div>
)
}
+2 -3
View File
@@ -54,7 +54,7 @@ export const createLipsyncJob = async (data: {
}
export const getLipsyncJob = async (id: string): Promise<LipsyncJob> => {
const response = await apiClient.get<LipsyncJob>(`/lipsync/jobs/${id}`, { timeout: 60000 })
const response = await apiClient.get<LipsyncJob>(`/lipsync/jobs/${id}`)
return response.data
}
@@ -66,8 +66,7 @@ export const generateSmartCover = async (
const response = await apiClient.post<{ cover_url: string; status: string; message: string }>(
"/ai-avatar/render/smart-cover",
{ video_url, max_frames },
// smart-cover 链路:下载视频+抽帧+drawtext 加标题+上传 OSS,需要较长时间,120s 超时
{ timeout: 120000 },
{ timeout: 60000 },
)
return response.data
}
@@ -1,18 +1,17 @@
/**
* AI数字人 — 面板5:封面 & 生成
* - 竖屏 9:16 封面预览(从视频截取 / 自定义上传)+ 标题文字实时叠加预览
* - 竖屏 9:16 封面预览(从视频截取 / 自定义上传)
* - 分辨率选择(720p / 1080p / 4K
* - 配置汇总卡片(出镜视频/音色/文案/对口型/B-roll/标题/封面)
* - 渐变紫色生成按钮
*
* 注意:v3 已删除"画面插入模式",本面板不包含该选项。
*/
import React, { useMemo, useRef } from "react"
import type { AiAvatarCoverConfig, AiAvatarTitleConfig } from "../types"
import React, { useRef } from "react"
import type { AiAvatarCoverConfig } from "../types"
interface PanelCoverAndGenerateProps {
coverConfig: AiAvatarCoverConfig
titleConfig: AiAvatarTitleConfig
onCoverConfigChange: (partial: Partial<AiAvatarCoverConfig>) => void
resolution: string
onResolutionChange: (r: string) => void
@@ -48,19 +47,8 @@ const LIPSYNC_STATUS_LABEL: Record<string, { text: string; cls: string }> = {
failed: { text: "失败", cls: "aa-status-badge--failed" },
}
/** 字体名 → CSS font-family 映射(与后端 drawtext 对齐) */
const FONT_FAMILY_MAP: Record<string, string> = {
: "'Noto Sans SC', 'Source Han Sans SC', 'PingFang SC', 'Microsoft YaHei', sans-serif",
: "'Noto Serif SC', 'Source Han Serif SC', 'SimSun', serif",
: "KaiTi, 'STKaiti', serif",
: "'Heiti SC', 'SimHei', 'Microsoft YaHei', sans-serif",
}
const getFontFamily = (font: string): string => FONT_FAMILY_MAP[font] || FONT_FAMILY_MAP["思源黑体"]
const PanelCoverAndGenerate: React.FC<PanelCoverAndGenerateProps> = ({
coverConfig,
titleConfig,
onCoverConfigChange,
resolution,
onResolutionChange,
@@ -98,85 +86,15 @@ const PanelCoverAndGenerate: React.FC<PanelCoverAndGenerateProps> = ({
const canGenerate = summary.lipsyncStatus === "completed" && !isGenerating
/** 封面图实际展示的 url:智能封面 > 自定义上传 > 空 */
const coverUrl =
coverConfig.smart_cover_url || coverConfig.thumbnail_url || coverConfig.upload_url
const hasCoverImage = Boolean(coverUrl)
/** 是否显示标题叠加层:有图、有文字、非加载中 */
const showTitleOverlay =
hasCoverImage && !smartCoverLoading && titleConfig.title.trim().length > 0
/** 计算标题叠加层的 inline 样式 */
const titleOverlayStyle = useMemo<React.CSSProperties>(() => {
const style: React.CSSProperties = {
position: "absolute",
left: "50%",
width: "90%",
transform: "translateX(-50%)",
textAlign: "center",
boxSizing: "border-box",
padding: "0 4px",
wordBreak: "break-word",
whiteSpace: "pre-wrap",
color: titleConfig.color || "#ffffff",
fontSize: `${titleConfig.size}px`,
fontFamily: getFontFamily(titleConfig.font),
fontWeight: titleConfig.bold ? "bold" : "normal",
fontStyle: titleConfig.italic ? "italic" : "normal",
lineHeight: 1.3,
pointerEvents: "none",
}
// 位置
const pos = titleConfig.position || "bottom"
if (pos === "top") {
style.top = "40px"
} else if (pos === "center") {
style.top = "50%"
style.transform = "translate(-50%, -50%)"
} else if (pos === "custom" && titleConfig.pos_x != null && titleConfig.pos_y != null) {
// pos_x/pos_y 是相对预览容器的百分比坐标
style.left = `${titleConfig.pos_x}%`
style.top = `${titleConfig.pos_y}%`
style.transform = "translate(-50%, -50%)"
} else {
style.bottom = "40px"
}
// 描边优先于阴影(二者互斥,与 drawtext 对齐)
if (titleConfig.stroke) {
// 描边宽度按字号估算,保证视觉一致
const strokeWidth = Math.max(1, Math.round(titleConfig.size / 18))
;(style as React.CSSProperties)["WebkitTextStroke"] = `${strokeWidth}px rgba(0,0,0,0.75)`
style.textShadow = "none"
} else if (titleConfig.shadow) {
style.textShadow = "0 2px 8px rgba(0,0,0,0.7), 0 0 2px rgba(0,0,0,0.5)"
} else {
// 默认给轻微阴影保证白字在亮背景可读
style.textShadow = "0 2px 6px rgba(0,0,0,0.6)"
}
return style
}, [titleConfig])
return (
<div className="aa-cover-generate">
{/* 封面预览(竖屏 9:16 */}
<div className="aa-cover-preview">
{hasCoverImage ? (
<img src={coverUrl!} alt="封面预览" draggable={false} />
{coverConfig.thumbnail_url ? (
<img src={coverConfig.thumbnail_url} alt="封面预览" />
) : (
<span className="aa-cover-preview__placeholder"></span>
)}
{/* 智能封面加载遮罩 */}
{smartCoverLoading && <div className="aa-cover-preview__loading"> </div>}
{/* 标题文字叠加层(实时预览,仅前端视觉参考,最终由后端 ffmpeg drawtext 叠加) */}
{showTitleOverlay && (
<div style={titleOverlayStyle} aria-hidden="true">
{titleConfig.title}
</div>
)}
</div>
<div className="aa-cover-actions">
@@ -3,7 +3,6 @@
* B-roll 画面插入 + 对口型视频预览 + 生成/重新生成按钮
* v3.1: 预览容器按 1/2 缩放、标题实时叠加预览
*/
import React, { useRef } from "react"
import type { LipsyncJob, BRollSegment, AiAvatarTitleConfig } from "../types"
interface PanelLipsyncPreviewProps {
@@ -14,8 +13,6 @@ interface PanelLipsyncPreviewProps {
onRemoveBRoll: (id: string) => void
/** 标题配置(实时叠加预览用) */
titleConfig?: AiAvatarTitleConfig
/** 标题位置变更回调(拖拽结束时调用) */
onTitlePositionChange?: (pos: { pos_x: number; pos_y: number }) => void
}
const BROLL_MODE_LABEL: Record<BRollSegment["mode"], string> = {
@@ -36,11 +33,7 @@ export function PanelLipsyncPreview({
onOpenBRollModal,
onRemoveBRoll,
titleConfig,
onTitlePositionChange,
}: PanelLipsyncPreviewProps) {
const titleDragRef = useRef<HTMLDivElement>(null)
const draggingTitleRef = useRef(false)
const previewContainerRef = useRef<HTMLDivElement>(null)
const isGenerating = lipsyncJob?.status === "pending" || lipsyncJob?.status === "processing"
const isDone = lipsyncJob?.status === "completed"
const isFailed = lipsyncJob?.status === "failed"
@@ -63,6 +56,7 @@ export function PanelLipsyncPreview({
fontSize: `${(titleConfig.size || 36) * 0.55}px`, // 预览等比缩
fontWeight: titleConfig.bold ? 700 : 400,
fontStyle: titleConfig.italic ? "italic" : "normal",
whiteSpace: "pre-wrap",
textAlign: "center",
width: "90%",
padding: "4px 8px",
@@ -76,40 +70,6 @@ export function PanelLipsyncPreview({
}
: null
const handleTitlePointerDown = (e: React.PointerEvent<HTMLDivElement>) => {
if (!onTitlePositionChange || !previewContainerRef.current) return
e.preventDefault()
e.stopPropagation()
;(e.target as Element).setPointerCapture(e.pointerId)
draggingTitleRef.current = true
;(e.currentTarget as HTMLDivElement).style.cursor = "grabbing"
}
const handleTitlePointerMove = (e: React.PointerEvent<HTMLDivElement>) => {
if (!draggingTitleRef.current || !previewContainerRef.current) return
e.preventDefault()
e.stopPropagation()
if (titleDragRef.current) {
const rect = previewContainerRef.current.getBoundingClientRect()
const relX = Math.max(0, Math.min(rect.width, e.clientX - rect.left))
const relY = Math.max(0, Math.min(rect.height, e.clientY - rect.top))
const xpct = (relX / rect.width) * 100
const ypct = (relY / rect.height) * 100
titleDragRef.current.style.left = `${xpct}%`
titleDragRef.current.style.top = `${ypct}%`
}
}
const handleTitlePointerUp = (e: React.PointerEvent<HTMLDivElement>) => {
if (!draggingTitleRef.current) return
draggingTitleRef.current = false
if (onTitlePositionChange && previewContainerRef.current) {
const rect = previewContainerRef.current.getBoundingClientRect()
const relX = Math.max(0, Math.min(rect.width, e.clientX - rect.left))
const relY = Math.max(0, Math.min(rect.height, e.clientY - rect.top))
onTitlePositionChange({ pos_x: relX, pos_y: relY })
}
;(e.currentTarget as HTMLDivElement).style.cursor = "grab"
}
return (
<div className="aa-script-lipsync">
{/* ── B-roll 画面 ── */}
@@ -178,31 +138,11 @@ export function PanelLipsyncPreview({
<div className="aa-lipsync-section">
<div className="aa-lipsync-section__title"></div>
<div className="aa-lipsync-preview" ref={previewContainerRef}>
<div className="aa-lipsync-preview">
{isDone && lipsyncJob?.output_video_url ? (
<div style={{ position: "relative", width: "100%", height: "100%" }}>
<video src={lipsyncJob.output_video_url} controls />
{titleOverlayStyle && (
<div
ref={titleDragRef}
style={{
...titleOverlayStyle,
cursor: onTitlePositionChange ? "grab" : "default",
pointerEvents: onTitlePositionChange ? "auto" : "none",
}}
onPointerDown={handleTitlePointerDown}
onPointerMove={handleTitlePointerMove}
onPointerUp={handleTitlePointerUp}
onPointerCancel={handleTitlePointerUp}
>
{titleConfig!.title.split(/[/]/).map((part, i) => (
<span key={i}>
{i > 0 && <br />}
{part}
</span>
))}
</div>
)}
{titleOverlayStyle && <div style={titleOverlayStyle}>{titleConfig!.title}</div>}
</div>
) : isGenerating ? (
<div style={{ width: "80%", textAlign: "center", color: "#fff" }}>
-6
View File
@@ -34,15 +34,9 @@ celery_app.conf.imports = (
"worker_app.tasks.tts_synthesis",
"worker_app.tasks.batch_download",
"worker_app.tasks.duplication_check",
# #1798 AI 数字人渲染:必须在 Worker 实例上注册同名任务,否则消息无人消费(渲染卡 0%)
"worker_app.tasks.ai_avatar_render",
"worker_app.tasks._startup",
"apps.worker.video_processing.dedup",
"worker_app.tasks.cleanup",
# 注意:必须用 app.* 路径,不能用 apps.api.app.* 路径!
# PYTHONPATH=/app/apps/api 下,app.tasks.lipsync_tts 可直接导入且不触发 apps/api/__init__.py
# apps/api/__init__.py 会 from .main import app,级联加载整个 FastAPI 栈,Worker 中不需要且会导致注册失败)
"app.tasks.lipsync_tts",
)
# Celery Beat 定时任务调度
@@ -1,48 +0,0 @@
"""AI 数字人渲染任务 — Worker 侧 Celery 任务注册.
#1798 渲染进度卡在 0% 的根因:渲染任务定义在 API 侧(`app.tasks.ai_avatar_render`),
装饰在 API 自己的 Celery 实例(`xiaoxia-saas-api`)上;而 Worker 用的是
`worker_app.celery_app` 实例,`conf.imports` 从未导入该任务,Worker 的任务
注册表里没有 `ai_avatar_render.execute`,消息被路由到默认 `celery` 队列后
无人消费,任务永远停在 0%
修复:在 Worker 侧用 `worker_app.celery_app` 注册同名任务,直接调用与 API
服务一致的 `AiAvatarRenderService.execute_render` 核心管线(业务逻辑在
`apps.api.app.services`worker 镜像已复制 `apps/api/app`)。任务名保持
`ai_avatar_render.execute`,与 API 生产端 `.delay()` 的消息路由一致;未在
task_routes 显式配置,走默认 `celery` 队列,由 transcode worker 消费。
"""
from __future__ import annotations
import logging
from worker_app.celery_app import celery_app
from worker_app.db import SessionLocal
logger = logging.getLogger(__name__)
@celery_app.task(bind=True, name="ai_avatar_render.execute", max_retries=2)
def execute_ai_avatar_render(self, job_id: str) -> dict:
"""执行 AI 数字人渲染管线(Worker 侧入口).
进度由 service 直接写入 DBAiAvatarRenderJob.progress:
0→5→20→40→80→90→95→100),API 通过轮询 progress 字段展示。
"""
logger.info("开始执行渲染任务: %s", job_id)
self.update_state(state="PROCESSING", meta={"progress": 0, "job_id": job_id})
session = SessionLocal()
try:
from app.services.ai_avatar_render_service import AiAvatarRenderService
service = AiAvatarRenderService(session)
service.execute_render(job_id)
return {"status": "completed", "job_id": job_id}
except Exception as exc:
logger.exception("渲染任务执行异常 [%s]: %s", job_id, exc)
self.update_state(state="FAILED", meta={"progress": 0, "error": str(exc)})
raise
finally:
session.close()
+3 -3
View File
@@ -163,7 +163,7 @@ services:
# context: ../..
# dockerfile: ${WEB_DOCKERFILE:-infra/docker/web.Dockerfile}
# args:
# (NGINX_CONF no longer needed - all configs baked into image)
# NGINX_CONF: ${WEB_NGINX_CONF:-infra/docker/nginx.conf}
container_name: xiaoxia-web-${ENV:-staging}
restart: unless-stopped
@@ -178,12 +178,12 @@ services:
- xiaoxia-net
# =========================================
# Nginx 配置运行时覆盖(双保险:entrypoint 也按 APP_ENV 选择配置)
# Nginx 配置运行时覆盖
# 确保容器使用正确环境的 nginx 配置,即使镜像构建时使用了默认配置
# 注意: 只覆盖 /etc/nginx/conf.d/default.conf,不挂载 /usr/share/nginx/html
# =========================================
environment:
- APP_ENV=${ENV:-staging}
- NGINX_ENV=${ENV:-staging}
volumes:
- ./nginx-${ENV:-staging}.conf:/etc/nginx/conf.d/default.conf:ro
+15 -6
View File
@@ -14,11 +14,11 @@ REGISTRY_TOKEN="${REGISTRY_TOKEN:-}"
ENV_FILE="${ENV_FILE:-/var/lib/xiaoxia-saas-production/.env}"
GENERATED_DIR="${GENERATED_DIR:-/var/lib/xiaoxia-saas-production/generated}"
LEGACY_ASSETS_DIR="${LEGACY_ASSETS_DIR:-/var/lib/xiaoxia-saas-production/legacy-assets}"
REPO_DIR="${REPO_DIR:-/var/lib/xiaoxia-saas-production/repo}"
if [ -z "$IMAGE_TAG" ]; then
echo "ERROR: IMAGE_TAG is required"
exit 1
fi
test -f "$ENV_FILE"
mkdir -p "$GENERATED_DIR"
@@ -31,6 +31,7 @@ if [ -n "$REGISTRY_TOKEN" ]; then
printf %s "$REGISTRY_TOKEN" | docker login "$REGISTRY_HOST" -u "$REGISTRY_USER" --password-stdin 2>/dev/null || {
echo "WARN: docker login failed, will try to pull anyway"
}
fi
# ---- Pull 三镜像 ----
REGISTRY_API="${REGISTRY}/xiaoxia-saas-api:${IMAGE_TAG}"
@@ -65,14 +66,17 @@ if docker inspect xiaoxia-web-production >/dev/null 2>&1; then
if [ -d "$_tmpdir" ] && [ "$(ls -A "$_tmpdir" 2>/dev/null)" ]; then
cp -an "$_tmpdir"/. "$LEGACY_ASSETS_DIR"/ 2>/dev/null || true
echo "Legacy assets backed up: $(ls "$_tmpdir" | wc -l) files"
fi
rm -rf "$_tmpdir"
else
echo "No existing web container, skipping legacy assets backup"
fi
# 清理超过 7 天的旧 assets 文件(避免无限增长)
if [ -d "$LEGACY_ASSETS_DIR" ]; then
find "$LEGACY_ASSETS_DIR" -type f -mtime +7 -delete 2>/dev/null || true
echo "Legacy assets cleanup done (retain 7 days)"
fi
# ---- 确保基础设施容器在运行 ----
echo "Checking infrastructure containers..."
@@ -80,10 +84,12 @@ for c in xiaoxia-postgres-production xiaoxia-redis-production; do
if ! docker inspect "$c" >/dev/null 2>&1; then
echo "ERROR: Required container not found: $c"
exit 1
fi
state=$(docker inspect -f '{{.State.Status}}' "$c")
if [ "$state" != "running" ]; then
echo "ERROR: Container not running: $c ($state)"
exit 1
fi
done
# ---- 确保生产网络存在 ----
@@ -102,6 +108,7 @@ echo "Migrations completed."
echo "Stopping old containers..."
docker rm -f xiaoxia-api-production 2>/dev/null || true
docker rm -f xiaoxia-worker-production 2>/dev/null || true
docker rm -f xiaoxia-web-production 2>/dev/null || true
# ---- 日志配置(所有容器共用) ----
LOG_OPTS="--log-driver json-file --log-opt max-size=50m --log-opt max-file=3"
@@ -159,16 +166,15 @@ docker run -d \
# ---- 启动 Web ----
# Legacy assets 挂载到 /usr/share/nginx/html/assets-legacy/assets/
# nginx 配置中 assets location 有 fallback 逻辑
WEB_VOLUMES=""
LEGACY_VOLUME=""
if [ -d "$LEGACY_ASSETS_DIR" ] && [ "$(ls -A "$LEGACY_ASSETS_DIR" 2>/dev/null)" ]; then
WEB_VOLUMES="-v ${LEGACY_ASSETS_DIR}:/usr/share/nginx/html/assets-legacy/assets:ro"
LEGACY_VOLUME="-v ${LEGACY_ASSETS_DIR}:/usr/share/nginx/html/assets-legacy/assets:ro"
echo "Web container: legacy assets mounted (fallback)"
else
echo "Web container: no legacy assets to mount"
fi
echo "Starting Web container..."
docker rm -f xiaoxia-web-production 2>/dev/null || true
docker run -d \
--name xiaoxia-web-production \
--network xiaoxia-net-production \
@@ -176,8 +182,7 @@ docker run -d \
--restart unless-stopped \
--cpus 0.5 \
--memory 512m \
-e APP_ENV=production \
$WEB_VOLUMES \
$LEGACY_VOLUME \
--health-cmd "wget --spider -q http://127.0.0.1:80" \
--health-interval 30s \
--health-timeout 5s \
@@ -192,6 +197,7 @@ while [ "$i" -lt 40 ]; do
if curl -sf --max-time 5 http://127.0.0.1:8001/health >/dev/null 2>&1; then
echo "API is healthy!"
break
fi
i=$((i + 1))
echo " Waiting... ($i/40)"
sleep 3
@@ -201,6 +207,7 @@ if [ "$i" -ge 40 ]; then
echo "ERROR: API did not become healthy within 120s"
docker logs --tail 50 xiaoxia-api-production
exit 1
fi
# ---- 等待 Web 健康 ----
echo "Waiting for Web to become healthy..."
@@ -209,6 +216,7 @@ while [ "$i" -lt 15 ]; do
if curl -sf --max-time 5 http://127.0.0.1:3002/ >/dev/null 2>&1; then
echo "Web is healthy!"
break
fi
i=$((i + 1))
echo " Waiting... ($i/15)"
sleep 2
@@ -218,6 +226,7 @@ if [ "$i" -ge 15 ]; then
echo "ERROR: Web did not become healthy within 30s"
docker logs --tail 30 xiaoxia-web-production
exit 1
fi
# ---- 清理旧镜像 ----
echo "Cleaning up old images..."
+10 -2
View File
@@ -124,15 +124,23 @@ docker run -d \
"$LOCAL_WORKER"
# ---- 启动 Web ----
# Web 镜像默认打包 production nginx.confstaging 需要挂载 staging 配置
NGINX_CONF="${NGINX_CONF:-${COMPOSE_DIR}/nginx-staging.conf}"
if [ ! -f "$NGINX_CONF" ]; then
echo "WARN: nginx config not found at $NGINX_CONF, using image default"
NGINX_VOLUME=""
else
NGINX_VOLUME="-v ${NGINX_CONF}:/etc/nginx/conf.d/default.conf:ro"
fi
echo "Starting Web container..."
docker rm -f xiaoxia-web-staging 2>/dev/null || true
docker run -d \
--name xiaoxia-web-staging \
--network xiaoxia-net-staging \
-p 127.0.0.1:3001:80 \
--restart unless-stopped \
--label com.centurylinklabs.watchtower.enable=true \
-e APP_ENV=staging \
$NGINX_VOLUME \
--health-cmd "wget --spider -q http://127.0.0.1:80" \
--health-interval 30s \
--health-timeout 5s \
+1
View File
@@ -53,6 +53,7 @@ export API_IMAGE="${API_IMAGE:-${REGISTRY}/xiaoxia-saas-api:dev}"
export WORKER_IMAGE="${WORKER_IMAGE:-${REGISTRY}/xiaoxia-saas-worker:dev}"
# Use staging-specific nginx config (proxy_pass → xiaoxia-api-staging:8000)
export WEB_NGINX_CONF=infra/docker/nginx-staging.conf
if [ "${REBUILD_BACKEND:-0}" = "1" ] || [ "${BUILD_WEB:-0}" = "1" ]; then
if [ "${ALLOW_STAGING_BUILDS:-false}" != "true" ]; then
-49
View File
@@ -1,49 +0,0 @@
#!/bin/sh
# Select nginx config based on APP_ENV (staging/production).
#
# 两种运行模式:
# 1. CI/CD 部署(staging/production):部署脚本通过 `-v 宿主机文件:/etc/nginx/conf.d/default.conf:ro`
# 把宿主机生成的带 resolver/docker upstream 的配置 bind mount 进来,entrypoint 不应改动。
# bind mount 的文件是 readonly 的,rm 会报 EBUSY ("Resource busy"),直接 exec nginx 即可。
# 2. 本地 docker-compose / 直接 `docker run`(无外部挂载):镜像烤入了 nginx-staging.conf 与
# nginx-production.conf 到 /etc/nginx/entrypoint 根据 APP_ENV 把 default.conf 换成正确的 symlink。
#
# 策略:
# - 如果 /etc/nginx/conf.d/default.conf 已经是指向目标 conf 的 symlink,什么都不做;
# - 否则尝试 rm -f 再 ln -srm 失败说明是外部 bind mount(已有正确配置),不阻塞启动;
# - 兜底:只要 conf.d 目录里有 .conf 文件(含 bind mount 来的),就直接启动 nginx。
set -e
NGINX_CONF_DIR="/etc/nginx/conf.d"
TARGET_CONF=""
case "${APP_ENV:-production}" in
staging)
TARGET_CONF="/etc/nginx/nginx-staging.conf"
;;
*)
TARGET_CONF="/etc/nginx/nginx-production.conf"
;;
esac
DEFAULT_CONF="$NGINX_CONF_DIR/default.conf"
# 1. 已经是正确的 symlink:直接启动
if [ -L "$DEFAULT_CONF" ] && [ "$(readlink "$DEFAULT_CONF" 2>/dev/null)" = "$TARGET_CONF" ]; then
exec nginx -g "daemon off;"
fi
# 2. 尝试替换为目标 symlink(无 bind mount 的场景)
# 若 rm 失败(bind mount readonlyEBUSY/EPERM),则认为外部已注入配置,不阻塞。
rm -f "$DEFAULT_CONF" 2>/dev/null || true
if [ -f "$TARGET_CONF" ] && [ ! -e "$DEFAULT_CONF" ]; then
ln -s "$TARGET_CONF" "$DEFAULT_CONF" 2>/dev/null || true
fi
# 3. 兜底:至少要有一个 .conf 文件,否则 nginx 起不来
if ! ls "$NGINX_CONF_DIR"/*.conf >/dev/null 2>&1; then
echo "ERROR: no nginx config found in $NGINX_CONF_DIR (tried $TARGET_CONF and external bind mount)" >&2
exit 1
fi
exec nginx -g "daemon off;"
+3 -7
View File
@@ -1,11 +1,7 @@
FROM git.xiaoxiajianji.com/xiaoxia/base/nginx:alpine AS runner
ARG NGINX_CONF=infra/docker/nginx.conf
WORKDIR /usr/share/nginx/html
COPY apps/web/dist ./
# 将所有 nginx 配置烤入镜像,entrypoint 按 APP_ENV 选择
COPY infra/docker/nginx.conf /etc/nginx/nginx-production.conf
COPY infra/docker/nginx-staging.conf /etc/nginx/nginx-staging.conf
COPY infra/docker/nginx-production.conf /etc/nginx/nginx-production.conf
COPY infra/docker/nginx-entrypoint.sh /docker-entrypoint.sh
RUN chmod +x /docker-entrypoint.sh
COPY ${NGINX_CONF} /etc/nginx/conf.d/default.conf
EXPOSE 80
ENTRYPOINT ["/docker-entrypoint.sh"]
CMD ["nginx", "-g", "daemon off;"]
+3 -7
View File
@@ -28,13 +28,9 @@ RUN --mount=type=cache,target=/app/apps/web/.tscache,sharing=locked \
# Production stage with nginx
FROM git.xiaoxiajianji.com/xiaoxia/base/nginx:alpine AS runner
ARG NGINX_CONF=infra/docker/nginx.conf
WORKDIR /usr/share/nginx/html
COPY --from=builder /app/apps/web/dist ./
# 将所有 nginx 配置烤入镜像,entrypoint 按 APP_ENV 选择
COPY infra/docker/nginx.conf /etc/nginx/nginx-production.conf
COPY infra/docker/nginx-staging.conf /etc/nginx/nginx-staging.conf
COPY infra/docker/nginx-production.conf /etc/nginx/nginx-production.conf
COPY infra/docker/nginx-entrypoint.sh /docker-entrypoint.sh
RUN chmod +x /docker-entrypoint.sh
COPY ${NGINX_CONF} /etc/nginx/conf.d/default.conf
EXPOSE 80
ENTRYPOINT ["/docker-entrypoint.sh"]
CMD ["nginx", "-g", "daemon off;"]
+3 -5
View File
@@ -20,7 +20,7 @@ WORKDIR /app
# 设置 Python 环境变量
ENV PATH="/opt/venv/bin:$PATH"
ENV PYTHONPATH=/app:/app/apps/api:/app/packages
ENV PYTHONPATH=/app:/app/packages
ENV PYTHONUNBUFFERED=1
ENV APP_VERSION=$APP_VERSION
@@ -28,10 +28,8 @@ ENV APP_VERSION=$APP_VERSION
COPY alembic.ini /app/alembic.ini
COPY migrations/ /app/migrations/
COPY packages/ /app/packages/
# PR #1844 起,worker 还需要加载 apps.api.app.tasks.lipsync_tts
# 该 task 依赖 app.services.* 与 app.core.celery_appPYTHONPATH=/app/apps/api 下解析)。
# 为避免后续新增 task 再次漏 COPY,直接把整个 apps/api/app/ 复制进 worker 镜像。
COPY apps/api/app/ /app/apps/api/app/
COPY apps/api/app/config.py /app/apps/api/app/config.py
COPY apps/api/app/core/ /app/apps/api/app/core/
# Worker 启动脚本
COPY infra/docker/entrypoint-worker.sh /usr/local/bin/entrypoint-worker.sh
-1
View File
@@ -1 +0,0 @@
# xiaoxia-saas shared packages namespace
-1
View File
@@ -1 +0,0 @@
# adapter implementations namespace
+6 -43
View File
@@ -1,13 +1,10 @@
#!/usr/bin/env python3
"""CI中自动修复代码格式(Python: black + isort + ruff | Frontend: prettier),并推送回原分支。
"""CI中自动修复代码格式(Python: black + isort | Frontend: prettier),并推送回原分支。
- PR事件:所有PR只要Code Quality因格式问题失败,自动修复并push回源分支
- Push事件(develop/main):自动修复并push回原分支,保持主干格式永远正确
- 防循环:修复commit带 [skip ci-format-check] 标记,检测到该标记则跳过修复
- black/isort/prettier 修格式;ruff check --fix --unsafe-fixes 自动修复
ruff 可修复的 lint 规则(含 F401 未使用 import 等 unsafe fix
- ruff 目标范围与 validate_style.sh 的检查范围对齐:apps packages tests
alembic/scripts 不在 ruff 检查范围内,不做修复)
- 只修格式(black/isort/prettier),ruff逻辑类错误不动
当code quality检查因格式问题失败时触发。
"""
@@ -138,7 +135,7 @@ def get_pr_head_branch(pr_number, api_url, token):
def fix_python(target_py_files, scan_mode):
"""修复 Python 文件 (black 格式化 + isort 排序 + ruff lint 自动修复)"""
"""修复 Python 文件格式 (black + isort)"""
if not target_py_files:
print("没有需要修复的 Python 文件,跳过")
return
@@ -149,48 +146,14 @@ def fix_python(target_py_files, scan_mode):
result = run(f"python3 -m black {target_str}", check=False)
print(result.stdout[-500:] if result.stdout else "")
if result.returncode != 0:
print("black执行失败,但继续尝试isort/ruff", file=sys.stderr)
print("black执行失败,但继续尝试isort", file=sys.stderr)
print()
print("--- isort 排序 ---")
result = run(f"python3 -m isort {target_str}", check=False)
print(result.stdout[-500:] if result.stdout else "")
if result.returncode != 0:
print("isort执行失败,继续尝试ruff", file=sys.stderr)
# ruff lint 自动修复
# 与 validate_style.sh 的检查范围对齐:只修 apps/packages/tests
# alembic 在 pyproject.toml 中被 excludescripts 不在 ruff 检查范围内)
ruff_scopes = ("apps/", "packages/", "tests/")
ruff_files = [f for f in target_py_files if f.startswith(ruff_scopes)]
if scan_mode != "incremental":
ruff_targets = "apps packages tests"
elif ruff_files:
ruff_targets = " ".join(ruff_files)
else:
ruff_targets = ""
if ruff_targets:
# ruff 由 style job 的 requirements-dev.txt 安装;不可用时跳过(不阻断 black/isort 的修复)
avail = run("python3 -m ruff --version", check=False)
if avail.returncode != 0:
print("ruff 不可用,跳过 ruff 自动修复", file=sys.stderr)
else:
print()
print("--- ruff lint 自动修复 (--fix --unsafe-fixes) ---")
# --unsafe-fixes 用于启用 F401(未使用 import)等 ruff 归类为 unsafe 的自动修复;
# 安全性由修复后重跑的完整 CI(单测/构建/staging 健康检查)兜底
result = run(
f"python3 -m ruff check {ruff_targets} --fix --unsafe-fixes",
check=False,
)
print(result.stdout[-1500:] if result.stdout else "")
if result.returncode != 0:
# 可能是仍有不可自动修复的 lint 错误(留待 style check 再次拦截),或修复过程出错
print("ruff 自动修复后仍有未修复项或执行失败,剩余问题由 style check 继续拦截", file=sys.stderr)
else:
print()
print("增量模式且无 apps/packages/tests 范围内的 Python 变更,跳过 ruff 自动修复")
print("isort执行失败", file=sys.stderr)
def fix_frontend(target_fe_files, scan_mode, repo_root):
@@ -371,7 +334,7 @@ def main():
# 提交修复
run("git clean -fd")
run("git add -u")
run('git commit -m "style: auto-format with black + isort + ruff + prettier [skip ci-format-check]"')
run('git commit -m "style: auto-format with black + isort + prettier [skip ci-format-check]"')
# 推送(head_branch已从ensure_git_repo获取)
print(f"\nPR来源分支: {head_branch}")
+35 -103
View File
@@ -132,8 +132,14 @@ def _lipsync_service_with_mocks():
def test_create_job_tts_direct_mode_synthesizes_audio():
svc, client, cosy = _lipsync_service_with_mocks()
with patch("app.services.lipsync_service.tts_synthesize_and_submit") as mock_task:
mock_task.apply_async.return_value = MagicMock(id="celery-task-123")
with (
patch("app.services.lipsync_service.get_shared_storage_service") as storage_patch,
patch("app.services.lipsync_service.safe_download_bytes") as dl_patch,
):
storage = MagicMock()
storage.upload_file.return_value = "https://oss/tts.mp3"
storage_patch.return_value = storage
dl_patch.return_value = b"FAKEAUDIO"
job = svc.create_job(
user_id="user-1",
@@ -144,16 +150,19 @@ def test_create_job_tts_direct_mode_synthesizes_audio():
emotion="兴奋",
)
# v4: TTS 模式下 create_job 返回 tts_processing 状态,dispatch Celery 任务
assert job.status == "tts_processing"
# 调了 TTS 合成,带 speed/emotion
cosy.submit_synthesize_task.assert_called_once()
_, kwargs = cosy.submit_synthesize_task.call_args
assert kwargs["speed"] == 1.2
assert kwargs["emotion"] == "excited"
assert kwargs["voice_id"] == "cosy-v1"
# MediaKit 用合成后的 OSS 音频 URL 提交
_, submit_kwargs = client.submit_lipsync.call_args
assert submit_kwargs["audio_url"] == "https://oss/tts.mp3"
assert submit_kwargs["video_url"] == "https://oss/person.mp4"
# DB 记录了 TTS 字段
assert job.emotion == "excited"
assert job.speed == 1.2
# 不直接调用 CosyVoice(由 Celery 任务处理)
cosy.submit_synthesize_task.assert_not_called()
# 不直接提交 MediaKit(由 Celery 任务处理)
client.submit_lipsync.assert_not_called()
# dispatch 了 Celery 任务
mock_task.apply_async.assert_called_once()
def test_create_job_direct_audio_mode_skips_tts():
@@ -171,25 +180,23 @@ def test_create_job_direct_audio_mode_skips_tts():
def test_create_job_tts_failure_raises():
"""v4: TTS 模式下 create_job 不再同步失败,而是 dispatch Celery 任务。
TTS 合成失败由 Celery 任务内部处理并更新 job 状态。"""
from app.services.mediakit_client import MediaKitError
from packages.application.cosyvoice_service import CosyVoiceError
svc, client, cosy = _lipsync_service_with_mocks()
cosy.submit_synthesize_task.side_effect = CosyVoiceError("Arrearage")
with patch("app.services.lipsync_service.tts_synthesize_and_submit") as mock_task:
mock_task.apply_async.return_value = MagicMock(id="celery-task-456")
job = svc.create_job(
with pytest.raises(MediaKitError) as exc:
svc.create_job(
user_id="user-1",
video_url="https://oss/person.mp4",
voice_id="v-1",
script_text="文本",
)
# create_job 成功返回 tts_processing,不直接调用 TTS
assert job.status == "tts_processing"
cosy.submit_synthesize_task.assert_not_called()
assert exc.value.code == "TTSSynthesisFailed"
# TTS 失败不应提交 MediaKit
client.submit_lipsync.assert_not_called()
mock_task.apply_async.assert_called_once()
# ── refresh 同步中间状态 ────────────────────────────────────────────────
@@ -226,7 +233,7 @@ def test_smart_cover_selects_best_frame_and_persists():
with (
patch("packages.shared.mediakit_client.get_mediakit_client") as mk_patch,
patch("packages.shared.cover_frame_scorer.score_frames") as score_patch,
patch("httpx.Client") as http_client_cls,
patch("httpx.get") as http_get,
patch("packages.shared.storage.get_shared_storage_service") as storage_patch,
):
mk = MagicMock()
@@ -238,107 +245,32 @@ def test_smart_cover_selects_best_frame_and_persists():
{"url": "https://mk/f1.jpg", "score": 90.0, "image_path": cands[1]["image_path"]},
{"url": "https://mk/f0.jpg", "score": 60.0, "image_path": cands[0]["image_path"]},
]
# httpx.Client 连接池 mock
client_instance = MagicMock()
resp = MagicMock()
resp.content = b"IMGDATA"
resp.raise_for_status = MagicMock()
client_instance.get.return_value = resp
client_instance.__enter__ = MagicMock(return_value=client_instance)
client_instance.__exit__ = MagicMock(return_value=False)
http_client_cls.return_value = client_instance
http_get.return_value = resp
storage = MagicMock()
storage.public_url = "https://oss.example.com"
# video_url 不是自家 OSS,不重签
storage.get_download_url.side_effect = lambda url, **kw: f"{url}?signed=1"
storage.upload_file.return_value = "https://oss.example.com/cover.jpg"
storage.upload_file.return_value = "https://oss/cover.jpg"
storage_patch.return_value = storage
url = cov.generate_smart_cover("https://other-host/avatar.mp4", job_id="job-1")
url = cov.generate_smart_cover("https://oss/avatar.mp4", job_id="job-1")
assert "signed=1" in url or url == "https://oss.example.com/cover.jpg"
assert url == "https://oss/cover.jpg"
mk.extract_frames.assert_called_once()
score_patch.assert_called_once()
# 验证使用了增大的轮询参数
call_kwargs = mk.extract_frames.call_args
assert call_kwargs.kwargs.get("poll_interval") == 3.0 or call_kwargs[1].get("poll_interval") == 3.0
assert call_kwargs.kwargs.get("max_poll_attempts") == 20 or call_kwargs[1].get("max_poll_attempts") == 20
def test_smart_cover_returns_empty_when_mediakit_unavailable():
from app.services import ai_avatar_cover_service as cov
with (
patch("packages.shared.mediakit_client.get_mediakit_client") as mk_patch,
patch("packages.shared.storage.get_shared_storage_service") as storage_patch,
):
with patch("packages.shared.mediakit_client.get_mediakit_client") as mk_patch:
mk = MagicMock()
mk.is_available = False
mk_patch.return_value = mk
storage = MagicMock()
storage.public_url = "https://oss.example.com"
storage_patch.return_value = storage
url = cov.generate_smart_cover("https://oss.example.com/avatar.mp4")
url = cov.generate_smart_cover("https://oss/avatar.mp4")
assert url == ""
def test_sign_video_url_resigns_own_oss_url():
"""自家 OSS 私有桶 URL 应被重签为长有效期预签名 URL"""
from app.services.ai_avatar_cover_service import _sign_video_url_for_mediakit
with patch("packages.shared.storage.get_shared_storage_service") as storage_patch:
storage = MagicMock()
storage.public_url = "https://oss.example.com"
storage.get_download_url.return_value = "https://oss.example.com/file.mp4?Expires=xxx&Signature=yyy"
storage_patch.return_value = storage
result = _sign_video_url_for_mediakit("https://oss.example.com/file.mp4")
assert "Signature=yyy" in result
storage.get_download_url.assert_called_once()
def test_sign_video_url_skips_external_url():
"""外部 URL(非自家 OSS)应原样返回,不做重签"""
from app.services.ai_avatar_cover_service import _sign_video_url_for_mediakit
with patch("packages.shared.storage.get_shared_storage_service") as storage_patch:
storage = MagicMock()
storage.public_url = "https://oss.example.com"
storage_patch.return_value = storage
result = _sign_video_url_for_mediakit("https://external-cdn.com/video.mp4")
assert result == "https://external-cdn.com/video.mp4"
storage.get_download_url.assert_not_called()
def test_extract_frames_uses_extended_poll_params():
"""验证 select_best_cover_frame 使用增大后的轮询参数"""
from app.services import ai_avatar_cover_service as cov
with (
patch("packages.shared.mediakit_client.get_mediakit_client") as mk_patch,
patch("packages.shared.storage.get_shared_storage_service") as storage_patch,
):
mk = MagicMock()
mk.is_available = True
mk.extract_frames.return_value = [{"image_url": "https://mk/f0.jpg"}]
mk_patch.return_value = mk
storage = MagicMock()
storage.public_url = "https://oss.example.com"
storage_patch.return_value = storage
cov.select_best_cover_frame("https://other/avatar.mp4", max_frames=3)
call_kwargs = mk.extract_frames.call_args
assert call_kwargs.kwargs.get("poll_interval") == 3.0 or call_kwargs[1].get("poll_interval") == 3.0
assert call_kwargs.kwargs.get("max_poll_attempts") == 20 or call_kwargs[1].get("max_poll_attempts") == 20
assert call_kwargs.kwargs.get("max_retries") == 1 or call_kwargs[1].get("max_retries") == 1
# ── 渲染 script_id 可选(手动文案直生场景)──────────────────────────────
-103
View File
@@ -519,109 +519,6 @@ class TestAiAvatarRenderService:
# 不应执行渲染逻辑
mock_db.commit.assert_not_called()
def test_execute_render_success_creates_clip_record(self):
"""execute_render 完成后自动创建成片记录到成片库."""
from app.services.ai_avatar_render_service import AiAvatarRenderService
mock_db = _make_mock_db()
mock_lipsync_job = _make_mock_lipsync_job(
status="completed",
output_video_url="https://oss/lipsync.mp4",
output_duration=30.0,
)
mock_job = _make_mock_render_job(
job_id="render-ok",
status="pending",
output_video_url="",
output_cover_url="",
output_duration=0.0,
)
mock_filter = MagicMock()
mock_filter.first.side_effect = [mock_job, mock_lipsync_job]
mock_query = MagicMock()
mock_query.filter.return_value = mock_filter
mock_db.query.return_value = mock_query
svc = AiAvatarRenderService(mock_db)
with (
patch.object(svc, "_download_video", return_value="/tmp/video.mp4"),
patch.object(svc, "_upload_to_oss", side_effect=lambda path, key: f"https://oss/{key}"),
patch("subprocess.run") as mock_run,
patch("tempfile.TemporaryDirectory") as tmpdir_mock,
patch(
"app.services.ai_avatar_cover_service.generate_smart_cover", return_value="https://oss/smart_cover.jpg"
),
patch("packages.domain.generated_video.GeneratedVideo.create") as gv_create,
patch(
"packages.adapters.sqlalchemy_impl.generated_video_repository.SQLAlchemyGeneratedVideoRepository"
) as repo_cls,
):
import subprocess as _sp
mock_run.return_value = _sp.CompletedProcess(args=[], returncode=0, stdout="", stderr="")
import tempfile as _tf
tmpdir_mock.return_value.__enter__ = MagicMock(return_value="/tmp/testdir")
tmpdir_mock.return_value.__exit__ = MagicMock(return_value=False)
mock_clip = MagicMock()
mock_clip.id = "clip-001"
gv_create.return_value = mock_clip
mock_repo = MagicMock()
repo_cls.return_value = mock_repo
svc.execute_render("render-ok")
assert mock_job.status == "completed"
gv_create.assert_called_once()
call_kwargs = gv_create.call_args
assert "https://oss/" in call_kwargs.kwargs["file_url"]
assert call_kwargs.kwargs["user_id"] == "user-1"
mock_repo.create.assert_called_once_with(mock_clip)
def test_execute_render_clip_failure_does_not_affect_render(self):
"""成片创建失败不影响渲染任务标记为成功."""
from app.services.ai_avatar_render_service import AiAvatarRenderService
mock_db = _make_mock_db()
mock_lipsync_job = _make_mock_lipsync_job(
status="completed",
output_video_url="https://oss/lipsync.mp4",
output_duration=30.0,
)
mock_job = _make_mock_render_job(
job_id="render-clip-fail",
status="pending",
output_video_url="",
output_cover_url="",
output_duration=0.0,
)
mock_filter = MagicMock()
mock_filter.first.side_effect = [mock_job, mock_lipsync_job]
mock_query = MagicMock()
mock_query.filter.return_value = mock_filter
mock_db.query.return_value = mock_query
svc = AiAvatarRenderService(mock_db)
with (
patch.object(svc, "_download_video", return_value="/tmp/video.mp4"),
patch.object(svc, "_upload_to_oss", side_effect=lambda path, key: f"https://oss/{key}"),
patch("subprocess.run") as mock_run,
patch("tempfile.TemporaryDirectory") as tmpdir_mock,
patch("app.services.ai_avatar_cover_service.generate_smart_cover", side_effect=RuntimeError("DB error")),
):
import subprocess as _sp
mock_run.return_value = _sp.CompletedProcess(args=[], returncode=0, stdout="", stderr="")
tmpdir_mock.return_value.__enter__ = MagicMock(return_value="/tmp/testdir")
tmpdir_mock.return_value.__exit__ = MagicMock(return_value=False)
svc.execute_render("render-clip-fail")
# 即使成片创建失败,渲染任务仍应标记为 completed
assert mock_job.status == "completed"
def test_error_exception_has_code(self):
from app.services.ai_avatar_render_service import AiAvatarRenderError
+54 -54
View File
@@ -208,21 +208,24 @@ class TestLipsyncServiceUnit:
"""Service 层单元测试(纯 mock,不依赖数据库)— #1809 更新."""
def test_create_job_success(self, mock_mediakit, mock_cosyvoice):
"""TTS 直生——v4 异步模式:create_job 只创建 DB 记录 + dispatch Celery 任务."""
"""v3: TTS 直生——service 内部 submit_synthesize_task 合成后转存 OSS,再提交 MediaKit."""
from app.services.lipsync_service import LipsyncService
mock_db = MagicMock()
mock_repo = MagicMock()
mock_repo.get.return_value = None # 预置音色,原样返回 voice_id
with patch(
"app.services.lipsync_service.tts_synthesize_and_submit"
) as mock_task:
mock_task.apply_async.return_value = MagicMock(id="celery-task-123")
with (
patch("app.services.lipsync_service.get_shared_storage_service") as storage_patch,
patch("app.services.lipsync_service.safe_download_bytes") as dl_patch,
):
storage_patch.return_value.upload_file.return_value = "https://my-oss/tts.mp3"
dl_patch.return_value = b"audio-bytes"
svc = LipsyncService(
mock_db,
client=mock_mediakit,
cosyvoice_service=mock_cosyvoice,
voice_clone_repo=mock_repo,
)
@@ -235,23 +238,32 @@ class TestLipsyncServiceUnit:
emotion="兴奋",
)
# TTS 模式:异步返回,状态为 tts_processing
assert job.status == "tts_processing"
assert not job.audio_url # TTS 音频尚未合成(默认空字符串)
# 不直接调用 CosyVoice
mock_cosyvoice.submit_synthesize_task.assert_not_called()
# dispatch 了 Celery 任务
mock_task.apply_async.assert_called_once()
assert job.status == "submitted"
assert job.mediakit_task_id == "mk-task-123"
# TTS 直生走 submit_synthesize_task,带语速/情绪
mock_cosyvoice.submit_synthesize_task.assert_called_once()
_, kwargs = mock_cosyvoice.submit_synthesize_task.call_args
assert kwargs["text"] == "大家好,欢迎来到直播间"
assert kwargs["voice_id"] == "longxiaochun_v3"
assert kwargs["speed"] == 1.2
assert kwargs["emotion"] == "excited" # 兴奋→excited
# job 记录透传字段
assert job.speed == 1.2
assert job.emotion == "excited"
# MediaKit 尚未提交(由 Celery 任务处理)
mock_mediakit.submit_lipsync.assert_not_called()
# MediaKit 用转存后的 OSS audio_url
call_kwargs = mock_mediakit.submit_lipsync.call_args
assert call_kwargs.kwargs["audio_url"] == "https://my-oss/tts.mp3"
def test_create_job_tts_failure(self, mock_mediakit):
"""v4: TTS 模式下 create_job 不再同步失败,而是 dispatch Celery 任务。
TTS 合成失败由 Celery 任务内部处理(见 test_lipsync_speed_optimization.py"""
"""v3: TTS 合成失败时,CosyVoiceError 被包装为 MediaKitError(TTSSynthesisFailed)
在建 DB 记录之前抛出,不提交 MediaKit"""
from app.services.lipsync_service import LipsyncService
from app.services.mediakit_client import MediaKitError
from packages.application.cosyvoice_service import CosyVoiceError
mock_cosyvoice = MagicMock()
mock_cosyvoice.submit_synthesize_task.side_effect = CosyVoiceError("Arrearage 欠费")
mock_db = MagicMock()
mock_repo = MagicMock()
@@ -260,41 +272,38 @@ class TestLipsyncServiceUnit:
svc = LipsyncService(
mock_db,
client=mock_mediakit,
cosyvoice_service=mock_cosyvoice,
voice_clone_repo=mock_repo,
)
with patch(
"app.services.lipsync_service.tts_synthesize_and_submit"
) as mock_task:
mock_task.apply_async.return_value = MagicMock(id="celery-task-456")
job = svc.create_job(
with pytest.raises(MediaKitError) as exc_info:
svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
voice_id="longxiaochun_v3",
script_text="测试文本",
)
assert exc_info.value.code == "TTSSynthesisFailed"
# TTS 模式下 create_job 成功返回,状态为 tts_processing
assert job.status == "tts_processing"
# 不应提交到 MediaKit
mock_mediakit.submit_lipsync.assert_not_called()
mock_task.apply_async.assert_called_once()
def test_create_job_api_failure(self, mock_mediakit):
"""MediaKit 提交失败(直传音频模式同步触发)."""
def test_create_job_api_failure(self, mock_mediakit, mock_cosyvoice):
"""MediaKit 提交失败."""
from app.services.lipsync_service import LipsyncService
from app.services.mediakit_client import MediaKitError
mock_mediakit.submit_lipsync.side_effect = MediaKitError("API 调用失败", code="SubmitFailed")
mock_db = MagicMock()
svc = LipsyncService(mock_db, client=mock_mediakit)
svc = LipsyncService(mock_db, client=mock_mediakit, cosyvoice_service=mock_cosyvoice)
with pytest.raises(MediaKitError, match="API 调用失败"):
svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
audio_url="https://example.com/audio.mp3",
voice_id="longxiaochun_v3",
script_text="测试文本",
)
def test_get_job_delegates_to_db(self, mock_mediakit, mock_cosyvoice):
@@ -425,21 +434,24 @@ class TestLipsyncServiceUnit:
assert result.status == "completed"
def test_create_job_stores_tts_audio_url(self, mock_mediakit, mock_cosyvoice):
"""v4: TTS 模式下 create_job 返回 tts_processing 状态,audio_url 尚未设置(由 Celery 任务处理)."""
"""v3: TTS 直生模式下 job.audio_url 为转存到自家 OSS 的永久地址."""
from app.services.lipsync_service import LipsyncService
mock_db = MagicMock()
mock_repo = MagicMock()
mock_repo.get.return_value = None
with patch(
"app.services.lipsync_service.tts_synthesize_and_submit"
) as mock_task:
mock_task.apply_async.return_value = MagicMock(id="celery-task-789")
with (
patch("app.services.lipsync_service.get_shared_storage_service") as storage_patch,
patch("app.services.lipsync_service.safe_download_bytes") as dl_patch,
):
storage_patch.return_value.upload_file.return_value = "https://my-oss/permanent.mp3"
dl_patch.return_value = b"audio-bytes"
svc = LipsyncService(
mock_db,
client=mock_mediakit,
cosyvoice_service=mock_cosyvoice,
voice_clone_repo=mock_repo,
)
job = svc.create_job(
@@ -449,11 +461,8 @@ class TestLipsyncServiceUnit:
script_text="这是一段测试文本",
)
# TTS 模式下 create_job 返回 tts_processing 状态
assert job.status == "tts_processing"
# audio_url 尚未设置(由 Celery 任务异步处理),模型默认为空字符串
assert not job.audio_url
mock_task.apply_async.assert_called_once()
# job.audio_url 是转存 OSS 后的永久地址
assert job.audio_url == "https://my-oss/permanent.mp3"
def test_create_job_direct_audio_skips_tts(self, mock_mediakit, mock_cosyvoice):
"""v3: 直接音频模式(传 audio_url)不触发 TTS,原样把 audio_url 提交 MediaKit."""
@@ -539,9 +548,9 @@ class TestErrorHandling:
assert exc_info.value.code == "VoiceNotReady"
def test_tts_value_error_mapped_to_invalid_param(self, mock_mediakit):
"""v4: TTS 模式下 create_job 不再同步调用 CosyVoice
而是 dispatch Celery 任务。ValueError 由 Celery 任务内部处理。"""
"""v3: CosyVoice 抛 ValueError(参数无效)被包装为 TTSInvalidParam(路由映射 400."""
from app.services.lipsync_service import LipsyncService
from app.services.mediakit_client import MediaKitError
mock_cosyvoice = MagicMock()
mock_cosyvoice.submit_synthesize_task.side_effect = ValueError("voice_id 为空")
@@ -553,27 +562,18 @@ class TestErrorHandling:
svc = LipsyncService(
mock_db,
client=mock_mediakit,
cosyvoice_service=mock_cosyvoice,
voice_clone_repo=mock_repo,
)
with patch(
"app.services.lipsync_service.tts_synthesize_and_submit"
) as mock_task:
mock_task.apply_async.return_value = MagicMock(id="celery-task-789")
# TTS 模式下 create_job 不再同步失败
job = svc.create_job(
with pytest.raises(MediaKitError) as exc_info:
svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
voice_id="some-voice",
script_text="test",
)
# 确认返回 tts_processing 状态
assert job.status == "tts_processing"
# TTS 合成由 Celery 任务处理,不直接调用 CosyVoice
mock_cosyvoice.submit_synthesize_task.assert_not_called()
mock_task.apply_async.assert_called_once()
assert exc_info.value.code == "TTSInvalidParam"
def test_missing_both_inputs_raises_invalid_input(self, mock_mediakit, mock_cosyvoice):
"""v3: 既无 audio_url 又无 voice_id+script_text 时抛 InvalidInput(路由映射 400."""
@@ -1,290 +0,0 @@
"""AI 数字人口型视频生成速度优化 — 单元测试.
验证两个优化点:
1. FFmpeg 编码 preset 从 fast 改为 veryfast(提速 30~50%
2. TTS 合成从同步改为 Celery 异步任务(API 响应从 6~35s 降到 <1s
Issue: lipsync-speed-optimization
"""
import os
from unittest.mock import MagicMock, patch
import pytest
os.environ.setdefault("JWT_SECRET_KEY", "dev-secret-key-for-testing")
# ═══════════════════════════════════════════════════════════════════════════════
# 优化1: FFmpeg 编码提速 — preset veryfast
# ═══════════════════════════════════════════════════════════════════════════════
class TestFFmpegPresetOptimization:
"""验证 FFmpeg 编码命令从 -preset fast 改为 -preset veryfast."""
def test_preset_is_veryfast(self):
"""_build_ffmpeg_command 输出必须包含 -preset veryfast."""
from app.services.ai_avatar_render_service import AiAvatarRenderService
svc = AiAvatarRenderService.__new__(AiAvatarRenderService)
cmd = svc._build_ffmpeg_command(
input_video="https://example.com/video.mp4",
b_roll_segments=[],
filter_complex="",
final_label=None,
output_path="/tmp/output.mp4",
)
# cmd 现在是 list[str]preset 与值是相邻两个元素
assert "-preset" in cmd, f"期望包含 -preset,实际命令: {cmd}"
preset_idx = cmd.index("-preset")
assert cmd[preset_idx + 1] == "veryfast", f"期望 veryfast,实际: {cmd}"
def test_preset_veryfast_with_filter(self):
"""带滤镜场景下也必须使用 veryfast."""
from app.services.ai_avatar_render_service import AiAvatarRenderService
svc = AiAvatarRenderService.__new__(AiAvatarRenderService)
cmd = svc._build_ffmpeg_command(
input_video="https://example.com/video.mp4",
b_roll_segments=[],
filter_complex="overlay=0:0",
final_label="[v]",
output_path="/tmp/output.mp4",
)
assert "-preset" in cmd
assert cmd[cmd.index("-preset") + 1] == "veryfast"
assert "-filter_complex" in cmd
def test_preset_not_fast(self):
"""确保不再使用旧的 -preset fast."""
from app.services.ai_avatar_render_service import AiAvatarRenderService
svc = AiAvatarRenderService.__new__(AiAvatarRenderService)
cmd = svc._build_ffmpeg_command(
input_video="https://example.com/video.mp4",
b_roll_segments=[],
filter_complex="",
final_label=None,
output_path="/tmp/output.mp4",
)
# 确保是 veryfast 而不是 fast
assert "-preset" in cmd
preset_idx = cmd.index("-preset")
assert cmd[preset_idx + 1] == "veryfast"
# 禁止 fast 单独作为 preset 值(veryfast 包含 "fast" 子串,不影响)
assert cmd[preset_idx + 1] != "fast"
# ═══════════════════════════════════════════════════════════════════════════════
# 优化2: TTS 合成 Celery 异步化
# ═══════════════════════════════════════════════════════════════════════════════
def _make_service_with_mocks():
"""构造 LipsyncService 测试实例及 mock 依赖."""
from app.services.lipsync_service import LipsyncService
db = MagicMock()
client = MagicMock()
client.is_available = True
client.submit_lipsync.return_value = {
"success": True,
"task_id": "mk-1",
"request_id": "req-1",
}
cosy = MagicMock()
cosy.submit_synthesize_task.return_value = {
"audio_url": "https://tts/raw.mp3",
"request_id": "tts-req",
"audio_duration": 3.0,
}
svc = LipsyncService(db, client=client, cosyvoice_service=cosy, voice_clone_repo=MagicMock())
# _resolve_voice_id 默认原样返回(repo.get 返回 None
svc._voice_clone_repo.get.return_value = None
return svc, client, cosy
class TestCreateJobAsyncTTS:
"""验证 TTS 模式改为 Celery 异步后的行为."""
def test_tts_mode_returns_tts_processing_status(self):
"""TTS 模式下 create_job 立即返回,状态为 tts_processing."""
svc, client, cosy = _make_service_with_mocks()
with patch("app.services.lipsync_service.tts_synthesize_and_submit") as mock_task:
mock_task.apply_async = MagicMock()
job = svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
voice_id="longxiaochun_v3",
script_text="大家好",
speed=1.0,
emotion="",
)
assert job.status == "tts_processing"
def test_tts_mode_dispatches_celery_task(self):
"""TTS 模式必须 dispatch Celery 异步任务."""
svc, client, cosy = _make_service_with_mocks()
with patch("app.services.lipsync_service.tts_synthesize_and_submit") as mock_task:
mock_task.apply_async = MagicMock()
svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
voice_id="v-1",
script_text="测试文本",
)
mock_task.apply_async.assert_called_once()
call_kwargs = mock_task.apply_async.call_args
args = call_kwargs.kwargs.get("args") or call_kwargs[1].get("args", call_kwargs[0][0] if call_kwargs[0] else ())
assert args[1] == "user-1" # user_id
assert args[2] == "v-1" # voice_id
assert args[3] == "测试文本" # script_text
def test_tts_mode_celery_dispatch_failure_marks_job_failed(self):
"""Celery dispatch 失败时,job 标为 failed 并写入 error_message,前端轮询能直接看到错误."""
svc, client, cosy = _make_service_with_mocks()
with patch("app.services.lipsync_service.tts_synthesize_and_submit") as mock_task:
mock_task.apply_async = MagicMock(side_effect=Exception("Celery broker down"))
job = svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
voice_id="v-1",
script_text="测试文本",
)
# job 已创建且状态标为 failed
assert job is not None
assert job.status == "failed"
assert "Celery 任务投递失败" in job.error_message
assert job.error_code == "AsyncDispatchFailed"
# MediaKit 未被调用
client.submit_lipsync.assert_not_called()
def test_tts_mode_voice_validation_still_sync(self):
"""TTS 模式下音色校验仍在 HTTP 请求中同步执行."""
from app.services.mediakit_client import MediaKitError
svc, client, cosy = _make_service_with_mocks()
# 模拟音色属于其他用户
other_profile = MagicMock()
other_profile.user_id = "user-other"
svc._voice_clone_repo.get.return_value = other_profile
with patch("app.tasks.lipsync_tts.tts_synthesize_and_submit"):
with pytest.raises(MediaKitError) as exc:
svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
voice_id="clone-profile-id",
script_text="测试",
)
assert exc.value.code == "VoiceForbidden"
def test_tts_mode_missing_input_raises_immediately(self):
"""缺少 voice_id 或 script_text 时立即报错,不 dispatch Celery 任务."""
from app.services.mediakit_client import MediaKitError
svc, client, cosy = _make_service_with_mocks()
with patch("app.tasks.lipsync_tts.tts_synthesize_and_submit") as mock_task:
mock_task.delay = MagicMock()
with pytest.raises(MediaKitError) as exc:
svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
# 缺少 voice_id 和 script_text
)
assert exc.value.code == "InvalidInput"
# Celery 任务未被 dispatch
mock_task.delay.assert_not_called()
# TTS 和 MediaKit 均未调用
cosy.submit_synthesize_task.assert_not_called()
client.submit_lipsync.assert_not_called()
class TestCreateJobDirectAudio:
"""验证直接音频模式不受异步化影响."""
def test_direct_audio_still_submits_synchronously(self):
"""直接音频模式仍然同步提交 MediaKit,状态为 submitted."""
svc, client, cosy = _make_service_with_mocks()
with patch("app.tasks.lipsync_tts.tts_synthesize_and_submit") as mock_task:
mock_task.delay = MagicMock()
job = svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
audio_url="https://example.com/audio.mp3",
)
assert job.status == "submitted"
assert job.mediakit_task_id == "mk-1"
client.submit_lipsync.assert_called_once()
# TTS Celery 任务不应被调用
mock_task.delay.assert_not_called()
def test_direct_audio_skips_tts(self):
"""直接音频模式不调用 CosyVoice TTS."""
svc, client, cosy = _make_service_with_mocks()
job = svc.create_job(
user_id="user-1",
video_url="https://example.com/video.mp4",
audio_url="https://example.com/audio.mp3",
)
cosy.submit_synthesize_task.assert_not_called()
call_kwargs = client.submit_lipsync.call_args
assert call_kwargs.kwargs["audio_url"] == "https://example.com/audio.mp3"
class TestCancelJobTtsProcessing:
"""验证 tts_processing 状态的任务可以被取消."""
def test_cancel_tts_processing(self):
"""tts_processing 状态的任务可以成功取消."""
svc, client, cosy = _make_service_with_mocks()
mock_job = MagicMock()
mock_job.status = "tts_processing"
mock_job.id = "job-1"
svc.get_job = MagicMock(return_value=mock_job)
result = svc.cancel_job("job-1", "user-1")
assert result.status == "cancelled"
def test_cancel_pending_still_works(self):
"""pending 状态仍可取消."""
svc, client, cosy = _make_service_with_mocks()
mock_job = MagicMock()
mock_job.status = "pending"
mock_job.id = "job-1"
svc.get_job = MagicMock(return_value=mock_job)
result = svc.cancel_job("job-1", "user-1")
assert result.status == "cancelled"
def test_cancel_submitted_still_works(self):
"""submitted 状态仍可取消."""
svc, client, cosy = _make_service_with_mocks()
mock_job = MagicMock()
mock_job.status = "submitted"
mock_job.id = "job-1"
svc.get_job = MagicMock(return_value=mock_job)
result = svc.cancel_job("job-1", "user-1")
assert result.status == "cancelled"
-389
View File
@@ -1,389 +0,0 @@
"""AI 数字人口型 TTS Celery 异步任务 — 单元测试.
覆盖 lipsync_tts.py 的全部主要分支:
- Job 不存在/cancelled/正常/异常路径
- TTS 合成、音频下载、OSS 上传、MediaKit 提交
- CosyVoiceError/ValueError/MediaKitError/顶层异常等错误码
"""
import os
import sys
import types
from types import ModuleType
from unittest.mock import MagicMock, patch
os.environ.setdefault("JWT_SECRET_KEY", "dev-secret-key-for-testing")
class _FakeQuery:
"""模拟 SQLAlchemy query.filter().first() 链式调用."""
def __init__(self, job):
self._job = job
def filter(self, *args, **kwargs):
return self
def first(self):
return self._job
def _make_fake_job(**kwargs):
"""构造可 setattr 的 job 记录."""
job = MagicMock()
job.id = kwargs.get("job_id", "job-1")
job.user_id = kwargs.get("user_id", "user-1")
job.status = kwargs.get("status", "tts_processing")
job.audio_url = kwargs.get("audio_url", "")
job.video_url = kwargs.get("video_url", "https://oss/video.mp4")
job.mediakit_task_id = kwargs.get("mediakit_task_id", "")
job.enable_video_loop = kwargs.get("enable_video_loop", False)
job.error_code = ""
job.error_message = ""
job.submitted_at = None
job.updated_at = None
return job
def _build_session(job):
"""构造 mock DB session + factory. 返回 (session, factory)."""
session = MagicMock()
session.query.return_value = _FakeQuery(job)
session.commit = MagicMock()
session.close = MagicMock()
factory = MagicMock(return_value=session)
return session, factory
def _apply_all_patches(
*,
job=None,
cosyvoice_service=None,
cosyvoice_side_effect=None,
cosyvoice_error=None,
download_bytes=b"AUDIO",
download_error=None,
storage=None,
mk_client=None,
mk_submit_return=None,
mk_submit_error=None,
):
"""统一构造测试需要的 patch 列表.
lipsync_tts.run() 在函数体内部懒 import 多个模块,通过 sys.modules 注入
伪造包路径避免真实导入;对存在的模块用 patch() 替换返回值/side_effect。
"""
# SessionLocal 通过懒探测获取(Worker 用 worker_app.dbAPI 用 app.db),
# 测试环境里两个模块都能被真实导入,必须同时 mock 保证用的是 fake session。
fake_app_db = ModuleType("app.db")
fake_worker_db = ModuleType("worker_app.db")
session, factory = _build_session(job)
fake_app_db.SessionLocal = factory
fake_worker_db.SessionLocal = factory
patches = [
patch.dict(sys.modules, {"app.db": fake_app_db, "worker_app.db": fake_worker_db}),
patch(
"app.tasks.lipsync_tts._sign_media_url",
side_effect=lambda url: url + "?signed" if url else url,
),
]
# CosyVoice
if cosyvoice_service is not None:
cosy_instance = cosyvoice_service
else:
cosy_instance = MagicMock()
if cosyvoice_side_effect is not None:
cosy_instance.submit_synthesize_task.side_effect = cosyvoice_side_effect
elif cosyvoice_error is not None:
cosy_instance.submit_synthesize_task.side_effect = cosyvoice_error
else:
cosy_instance.submit_synthesize_task.return_value = {"audio_url": "https://tts/raw.mp3"}
patches.append(patch("packages.application.cosyvoice_service.CosyVoiceService", return_value=cosy_instance))
# safe_download_bytes
if download_error is not None:
patches.append(patch("packages.shared.url_security.safe_download_bytes", side_effect=download_error))
else:
patches.append(patch("packages.shared.url_security.safe_download_bytes", return_value=download_bytes))
# Storage
if storage is None:
storage = MagicMock()
storage.public_url = "https://oss.example.com"
storage.upload_file.return_value = "https://oss.example.com/tts.mp3"
patches.append(patch("packages.shared.storage.get_shared_storage_service", return_value=storage))
# MediaKit client
if mk_client is not None:
patches.append(patch("app.services.mediakit_client.get_mediakit_client", return_value=mk_client))
else:
client = MagicMock()
if mk_submit_error is not None:
client.submit_lipsync.side_effect = mk_submit_error
else:
client.submit_lipsync.return_value = mk_submit_return or {"task_id": "mk-1"}
patches.append(patch("app.services.mediakit_client.get_mediakit_client", return_value=client))
return session, patches
class TestTtsSynthesizeAndSubmit:
"""测试 Celery 任务 tts_synthesize_and_submit.run 的所有分支."""
def test_job_not_found_returns_early(self):
"""Job 不存在 → 日志报错直接返回,不抛异常."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
session, patches = _apply_all_patches(job=None)
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("missing-job", "user-1", "v1", "你好", 1.0, "")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
session.commit.assert_not_called()
session.close.assert_called_once()
def test_cancelled_job_skipped(self):
"""Job 已 cancelled → 跳过不处理,不调用 TTS/MediaKit."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
job = _make_fake_job(status="cancelled")
session, patches = _apply_all_patches(job=job)
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好", 1.0, "")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
# cancelled 不应 commit,不应触发 TTS/MediaKit
session.commit.assert_not_called()
session.close.assert_called_once()
def test_happy_path_tts_to_mediakit(self):
"""完整正常流程:TTS 合成 → OSS 上传 → 签名 → 提交 MediaKit → submitted."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
job = _make_fake_job()
mk_client = MagicMock()
mk_client.submit_lipsync.return_value = {"task_id": "mk-999"}
session, patches = _apply_all_patches(job=job, mk_client=mk_client)
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好世界", 1.0, "happy")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
assert job.status == "submitted"
assert job.mediakit_task_id == "mk-999"
assert job.error_code == ""
mk_client.submit_lipsync.assert_called_once()
call_kwargs = mk_client.submit_lipsync.call_args.kwargs
assert call_kwargs["client_token"] == "job-1"
assert call_kwargs["audio_url"].endswith("?signed")
session.commit.assert_called()
session.close.assert_called_once()
def test_cosyvoice_error_marks_tts_synthesis_failed(self):
"""CosyVoiceError → failed, error_code=TTSSynthesisFailed."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
from packages.application.cosyvoice_service import CosyVoiceError
job = _make_fake_job()
session, patches = _apply_all_patches(job=job, cosyvoice_error=CosyVoiceError("tts boom"))
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好", 1.0, "")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
assert job.status == "failed"
assert job.error_code == "TTSSynthesisFailed"
session.close.assert_called_once()
def test_value_error_marks_tts_invalid_param(self):
"""ValueError(参数错误)→ failed, error_code=TTSInvalidParam."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
job = _make_fake_job()
session, patches = _apply_all_patches(job=job, cosyvoice_side_effect=ValueError("bad param"))
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好", -1.0, "")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
assert job.status == "failed"
assert job.error_code == "TTSInvalidParam"
session.close.assert_called_once()
def test_no_audio_url_marks_tts_no_audio(self):
"""TTS 返回空 audio_url → failed, error_code=TTSNoAudio."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
job = _make_fake_job()
cosy = MagicMock()
cosy.submit_synthesize_task.return_value = {"audio_url": ""}
session, patches = _apply_all_patches(job=job, cosyvoice_service=cosy)
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好", 1.0, "")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
assert job.status == "failed"
assert job.error_code == "TTSNoAudio"
session.close.assert_called_once()
def test_oss_upload_failure_falls_back_to_temp_url(self):
"""OSS 上传失败 → 回退临时 URL,仍然 submitted."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
job = _make_fake_job()
storage = MagicMock()
storage.public_url = "https://oss.example.com"
storage.upload_file.side_effect = RuntimeError("oss down")
mk_client = MagicMock()
mk_client.submit_lipsync.return_value = {"task_id": "mk-7"}
session, patches = _apply_all_patches(job=job, storage=storage, mk_client=mk_client)
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好", 1.0, "")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
# 上传失败后 audio_url 回退为临时 TTS URL,仍继续提交到 MediaKit
assert job.audio_url == "https://tts/raw.mp3"
assert job.status == "submitted"
assert job.mediakit_task_id == "mk-7"
mk_client.submit_lipsync.assert_called_once()
session.close.assert_called_once()
def test_mediakit_error_marks_mediakit_unavailable(self):
"""MediaKit 提交失败 → failed, error_code=MediaKitUnavailable."""
from app.services.mediakit_client import MediaKitError
from app.tasks.lipsync_tts import tts_synthesize_and_submit
job = _make_fake_job()
mk_err = MediaKitError("mk down", code="MediaKitUnavailable")
session, patches = _apply_all_patches(job=job, mk_submit_error=mk_err)
entered = [p.__enter__() for p in patches]
try:
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好", 1.0, "")
finally:
for p in reversed(patches):
p.__exit__(None, None, None)
assert job.status == "failed"
assert job.error_code == "MediaKitUnavailable"
session.close.assert_called_once()
def test_top_level_exception_marks_async_task_error(self):
"""顶层未预期异常 → failed, error_code=AsyncTaskError."""
from app.tasks.lipsync_tts import tts_synthesize_and_submit
job = _make_fake_job()
fake_app_db = ModuleType("app.db")
fake_worker_db = ModuleType("worker_app.db")
session, factory = _build_session(job)
fake_app_db.SessionLocal = factory
fake_worker_db.SessionLocal = factory
# CosyVoiceService 在 __init__ 抛 RuntimeError(非 CosyVoiceError/ValueError
fake_cosy_mod = ModuleType("packages.application.cosyvoice_service")
class _CosyVoiceErrorForTest(Exception):
pass
class _BoomService:
def __init__(self):
raise RuntimeError("top-level boom")
fake_cosy_mod.CosyVoiceError = _CosyVoiceErrorForTest
fake_cosy_mod.CosyVoiceService = _BoomService
with patch.dict(
sys.modules,
{
"app.db": fake_app_db,
"worker_app.db": fake_worker_db,
"packages.application.cosyvoice_service": fake_cosy_mod,
},
):
tts_synthesize_and_submit.run("job-1", "user-1", "v1", "你好", 1.0, "")
assert job.status == "failed"
assert job.error_code == "AsyncTaskError"
session.close.assert_called()
class TestSignMediaUrl:
"""覆盖模块内 _sign_media_url 的所有分支(CI 增量覆盖率需要)."""
def test_empty_url_returns_empty(self):
from app.tasks.lipsync_tts import _sign_media_url
assert _sign_media_url("") == ""
assert _sign_media_url(None) is None
def test_own_oss_url_signed(self):
"""自家 OSS URL → 调用 storage.get_download_url 签名."""
from app.tasks.lipsync_tts import _sign_media_url
fake_storage = MagicMock()
fake_storage.public_url = "https://oss.example.com/"
fake_storage.get_download_url.return_value = "https://oss.example.com/a?sig=xyz"
with patch("packages.shared.storage.get_shared_storage_service", return_value=fake_storage):
result = _sign_media_url("https://oss.example.com/lipsync/a.mp3")
assert result == "https://oss.example.com/a?sig=xyz"
fake_storage.get_download_url.assert_called_once()
def test_external_url_passthrough(self):
"""外部 URL(不是自家 OSS host)→ 原样透传,不签名."""
from app.tasks.lipsync_tts import _sign_media_url
fake_storage = MagicMock()
fake_storage.public_url = "https://oss.example.com/"
with patch("packages.shared.storage.get_shared_storage_service", return_value=fake_storage):
result = _sign_media_url("https://tts.example.com/raw.mp3")
assert result == "https://tts.example.com/raw.mp3"
fake_storage.get_download_url.assert_not_called()
def test_storage_exception_falls_back(self):
"""storage 调用异常 → 降级原样返回,不抛错."""
from app.tasks.lipsync_tts import _sign_media_url
with patch(
"packages.shared.storage.get_shared_storage_service",
side_effect=RuntimeError("storage down"),
):
result = _sign_media_url("https://oss.example.com/a.mp3")
assert result == "https://oss.example.com/a.mp3"
def test_no_public_url_passthrough(self):
"""storage.public_url 为空 → 原样透传."""
from app.tasks.lipsync_tts import _sign_media_url
fake_storage = MagicMock()
fake_storage.public_url = ""
with patch("packages.shared.storage.get_shared_storage_service", return_value=fake_storage):
result = _sign_media_url("https://anything.example.com/a.mp3")
assert result == "https://anything.example.com/a.mp3"
fake_storage.get_download_url.assert_not_called()