fix(e2e): Playwright chromium禁用GPU，避免无显示环境下GPU初始化失败导致浏览器不稳定

fix(api+ci): DELETE 204响应体全修复 + isort/black整理 + CI环境兼容 + 测试修复

.env.example

@@ -3,6 +3,7 @@
 # ==================== 应用配置 ====================
 APP_NAME=小虾 SaaS
 APP_BASE_URL=http://localhost:3000
+APP_ENV=development
 
 # ==================== 数据库配置 ====================
 DATABASE_URL=postgresql://xiaoxia_user:your_password@localhost:5432/xiaoxia_saas
@@ -35,7 +36,8 @@ ENVIRONMENT=development
 DEBUG=true
 
 # ==================== CORS 配置 ====================
-CORS_ORIGINS=["http://localhost:3000","http://localhost:5173"]
+# 逗号分隔的域名列表（Settings 读取 CORS_ORIGINS_RAW）
+CORS_ORIGINS_RAW=http://localhost:3000,http://localhost:5173
 
 # ==================== 阿里云 OSS 配置 ====================
 OSS_ENDPOINT=oss-cn-hangzhou.aliyuncs.com
@@ -49,6 +51,7 @@ OSS_BUCKET_NAME=xiaoxia-autocut
 #       cosyvoice-v3-plus (高质量，系统音色少)
 #       cosyvoice-v3.5-flash / cosyvoice-v3.5-plus (仅支持克隆/设计音色，无系统音色)
 # 音色: v3系列系统音色带 _v3 后缀，如 longxiaochun_v3, longxiaoxia_v3, longanyang (无后缀)
+# 注意：COSYVOICE_* 变量由 packages/shared/config.py 的 SharedSettings 读取
 COSYVOICE_API_KEY=your-cosyvoice-api-key
 COSYVOICE_BASE_URL=https://dashscope.aliyuncs.com/api/v1
 COSYVOICE_MODEL=cosyvoice-v3-flash

.env.production.example

@@ -5,7 +5,6 @@ APP_ENV=production
 ENVIRONMENT=production
 DEBUG=false
 USE_IN_MEMORY_DB=false
-LOG_LEVEL=WARNING
 
 # ==================== 数据库（必须修改）====================
 DATABASE_URL=postgresql://prod_user:CHANGE_THIS_PASSWORD@db-prod:5432/xiaoxia_prod

.gitea/workflows/ci-cd.yml

@@ -19,6 +19,10 @@ on:
 permissions:
   contents: read
 
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
 jobs:
   validate:
     name: Validate Code Quality And Tests
@@ -158,57 +162,102 @@ jobs:
             python3 scripts/check_migration_safety.py --allow-medium-risk
           fi
 
-      - name: Debug coverage paths
+  unit-tests:
+    name: Unit Tests
+    runs-on: host
+    timeout-minutes: 8
+
+    env:
+      USE_IN_MEMORY_DB: "true"
+      OSS_ACCESS_KEY_ID: placeholder
+      OSS_ACCESS_KEY_SECRET: placeholder
+      OSS_BUCKET_NAME: xiaoxia-autocut
+      OSS_ENDPOINT: oss-cn-hangzhou.aliyuncs.com
+
+    steps:
+      - name: Checkout code
+        shell: sh
+        env:
+          GITHUB_TOKEN: ${{ github.token }}
+        run: |
+          set -eu
+          python3 - <<'PY'
+          import io, os, tarfile, time, urllib.request, urllib.error
+          url = f"{os.environ['GITHUB_API_URL']}/repos/{os.environ['GITHUB_REPOSITORY']}/archive/{os.environ['GITHUB_SHA']}.tar.gz"
+          request = urllib.request.Request(url, headers={"Authorization": f"token {os.environ['GITHUB_TOKEN']}"})
+          last_err = None
+          for attempt in range(5):
+              try:
+                  with urllib.request.urlopen(request, timeout=120) as response:
+                      archive = response.read()
+                  break
+              except urllib.error.HTTPError as e:
+                  last_err = e
+                  if e.code >= 500 and attempt < 4:
+                      wait = 2 ** attempt
+                      print(f"Checkout HTTP {e.code}, retrying in {wait}s (attempt {attempt+1}/5)...")
+                      time.sleep(wait)
+                      continue
+                  raise
+              except Exception as e:
+                  last_err = e
+                  if attempt < 4:
+                      wait = 2 ** attempt
+                      print(f"Checkout error: {e}, retrying in {wait}s (attempt {attempt+1}/5)...")
+                      time.sleep(wait)
+                      continue
+                  raise
+          else:
+              raise last_err
+          with tarfile.open(fileobj=io.BytesIO(archive), mode='r:gz') as tar:
+              root_prefix = tar.getmembers()[0].name.split('/', 1)[0] + '/'
+              for member in tar.getmembers():
+                  name = member.name
+                  if name == root_prefix[:-1]:
+                      continue
+                  if name.startswith(root_prefix):
+                      member.name = name[len(root_prefix):]
+                  if member.name:
+                      tar.extract(member, '.')
+          PY
+
+      - name: Install ffmpeg
         shell: sh
         run: |
           set +e
-          echo "=== PWD ==="
-          pwd
-          echo "=== check source dirs ==="
-          ls -d apps/api/app packages
-          echo "=== python import check ==="
-          python3 - <<'PY'
-          import sys, os
-          os.environ["PYTHONPATH"] = f"{os.getcwd()}/apps/api:{os.getcwd()}"
-          sys.path.insert(0, f"{os.getcwd()}/apps/api")
-          sys.path.insert(0, os.getcwd())
-          print(f"cwd: {os.getcwd()}")
-          print(f"sys.path[:5]: {sys.path[:5]}")
-          try:
-              import app
-              print(f"app.__file__: {app.__file__}")
-          except Exception as e:
-              print(f"import app failed: {e}")
-          try:
-              import packages
-              print(f"packages.__file__: {packages.__file__}")
-          except Exception as e:
-              print(f"import packages failed: {e}")
-          PY
-          echo "=== coverage debug ==="
-          python3 - <<'PY'
-          import os, sys
-          sys.path.insert(0, f"{os.getcwd()}/apps/api")
-          sys.path.insert(0, os.getcwd())
-          import coverage
-          cov = coverage.Coverage(source=["apps/api/app", "packages"])
-          print(f"source: {cov.config.source}")
-          for src in cov.config.source or []:
-              abspath = os.path.abspath(src)
-              print(f"  {src} -> {abspath} exists={os.path.exists(src)}")
-              if os.path.isdir(src):
-                  pyfiles = []
-                  for root, dirs, files in os.walk(src):
-                      for f in files:
-                          if f.endswith('.py'):
-                              pyfiles.append(os.path.join(root, f))
-                  print(f"    .py files: {len(pyfiles)}")
-          PY
+          if command -v ffmpeg > /dev/null 2>&1; then
+            echo "ffmpeg already installed: $(ffmpeg -version | head -1)"
+            exit 0
+          fi
+          if command -v apt-get > /dev/null 2>&1; then
+            apt-get update -qq && apt-get install -y -qq ffmpeg
+          elif command -v yum > /dev/null 2>&1; then
+            yum install -y -q epel-release 2>/dev/null
+            yum install -y -q ffmpeg 2>/dev/null
+            if [ $? -ne 0 ] && command -v dnf > /dev/null 2>&1; then
+              dnf install -y -q --nogpgcheck https://download1.rpmfusion.org/free/el/rpmfusion-free-release-$(rpm -E %rhel).noarch.rpm 2>/dev/null
+              dnf install -y -q ffmpeg 2>/dev/null
+            fi
+          elif command -v dnf > /dev/null 2>&1; then
+            dnf install -y -q ffmpeg 2>/dev/null
+          fi
+          if command -v ffmpeg > /dev/null 2>&1; then
+            echo "ffmpeg installed successfully: $(ffmpeg -version | head -1)"
+          else
+            echo "Warning: ffmpeg installation failed or not available, some tests may be skipped"
+          fi
 
-      - name: Run unit tests
+      - name: Install dependencies
+        shell: sh
+        run: |
+          set -eu
+          python3 -m pip install -q -r requirements-base.txt
+          python3 -m pip install -q -r requirements.txt
+          python3 -m pip install -q -r requirements-dev.txt
+          pytest --version
+
+      - name: Run unit tests with coverage
         shell: sh
-        env:
-          USE_IN_MEMORY_DB: "true"
         run: |
           set -eu
           PYTHONPATH="$PWD/apps/api:$PWD" python3 -m coverage run \
@@ -220,16 +269,16 @@ jobs:
           python3 -m coverage xml -o coverage.xml
           python3 -m coverage report --fail-under=60 > /dev/null
 
-      - name: Build summary
-        if: github.ref == 'refs/heads/develop' || github.ref == 'refs/heads/main'
+      - name: CI failure notification
+        if: failure()
         shell: sh
+        env:
+          GITEA_TOKEN: ${{ secrets.GITEA_TOKEN }}
+          CI_WEBHOOK_URL: ${{ secrets.CI_WEBHOOK_URL }}
         run: |
-          set -eu
-          echo "Build completed successfully!"
-          echo "Branch: ${GITHUB_REF_NAME}"
-          echo "Commit: ${GITHUB_SHA}"
-          # 输出最终覆盖率
-          python3 scripts/ci_coverage_summary.py
+          set +e
+          FAILED_JOB="Unit Tests" python3 scripts/ci_notify_failure.py
+
 
   integration-tests:
     name: Integration Tests
@@ -241,6 +290,10 @@ jobs:
     env:
       DATABASE_URL: postgresql+psycopg://postgres:postgres@127.0.0.1:5432/xiaoxia_saas
       USE_IN_MEMORY_DB: "false"
+      OSS_ACCESS_KEY_ID: placeholder
+      OSS_ACCESS_KEY_SECRET: placeholder
+      OSS_BUCKET_NAME: xiaoxia-autocut
+      OSS_ENDPOINT: oss-cn-hangzhou.aliyuncs.com
 
     steps:
       - name: Checkout code
@@ -306,6 +359,32 @@ jobs:
           python3 -m pip install -q -r requirements-dev.txt
           pytest --version
 
+      - name: Install ffmpeg
+        shell: sh
+        run: |
+          set +e
+          if command -v ffmpeg > /dev/null 2>&1; then
+            echo "ffmpeg already installed: $(ffmpeg -version | head -1)"
+            exit 0
+          fi
+          if command -v apt-get > /dev/null 2>&1; then
+            apt-get update -qq && apt-get install -y -qq ffmpeg
+          elif command -v yum > /dev/null 2>&1; then
+            yum install -y -q epel-release 2>/dev/null
+            yum install -y -q ffmpeg 2>/dev/null
+            if [ $? -ne 0 ] && command -v dnf > /dev/null 2>&1; then
+              dnf install -y -q --nogpgcheck https://download1.rpmfusion.org/free/el/rpmfusion-free-release-$(rpm -E %rhel).noarch.rpm 2>/dev/null
+              dnf install -y -q ffmpeg 2>/dev/null
+            fi
+          elif command -v dnf > /dev/null 2>&1; then
+            dnf install -y -q ffmpeg 2>/dev/null
+          fi
+          if command -v ffmpeg > /dev/null 2>&1; then
+            echo "ffmpeg installed successfully: $(ffmpeg -version | head -1)"
+          else
+            echo "Warning: ffmpeg installation failed or not available, some tests may be skipped"
+          fi
+
       - name: Start Redis
         shell: sh
         run: |
@@ -375,7 +454,7 @@ jobs:
         shell: sh
         run: |
           set -eu
-          pip install -q pytest-rerunfailures
+          python3 -m pip install -q pytest-rerunfailures
           PYTHONPATH="$PWD/apps/api:$PWD" python3 -m coverage run --append \
             --source=apps/api/app,packages \
             --omit="*/migrations/*,*/tests/*,*/test_*.py,*/site-packages/*" \
@@ -563,13 +642,22 @@ jobs:
             -w /workspace/apps/web \
             docker.m.daocloud.io/library/node:20 \
             sh -lc 'npx vitest run src/test'
+
+      - name: Notify CI failure
+        if: failure()
+        shell: sh
+        run: |
+          set +e
+          echo "=== CI 失败通知 ==="
+          FAILED_JOB="Frontend Lint" python3 scripts/ci_notify_failure.py
+
   deploy-staging:
     name: Build & Push Staging (Watchtower auto-deploy)
     runs-on: saas
     timeout-minutes: 30
     needs: [validate, frontend-lint]
 
-    if: github.ref_name == 'main' || github.ref_name == 'develop' || startsWith(github.ref_name, 'feature/')
+    if: github.event_name == 'push' && (github.ref_name == 'main' || github.ref_name == 'develop')
 
     steps:
     - name: Checkout code
@@ -694,6 +782,23 @@ jobs:
         echo "Branch: ${GITHUB_REF_NAME}"
         echo "Commit: ${GITHUB_SHA}"
 
+    - name: Notify CI success
+      if: success()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 成功通知 ==="
+        SUCCESS_JOB="Staging部署成功" python3 scripts/ci_notify_success.py
+
+    - name: Notify CI failure
+      if: failure()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 失败通知 ==="
+        FAILED_JOB="Build & Push Staging (Watchtower auto-deploy)" python3 scripts/ci_notify_failure.py
+
+
 
   staging-e2e:
     name: Staging E2E Tests
@@ -762,6 +867,15 @@ jobs:
           git.xiaoxiajianji.com/xiaoxia/base/playwright:v1.45.0-jammy \
           sh -lc "npm ci && npx playwright test --reporter=line --project=chromium e2e/auth.spec.ts e2e/auth-guard.spec.ts e2e/core-upload.spec.ts e2e/core-generation.spec.ts"
 
+    - name: Notify CI failure
+      if: failure()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 失败通知 ==="
+        FAILED_JOB="Staging E2E Tests" python3 scripts/ci_notify_failure.py
+
+
   staging-api-tests:
     name: Staging API Integration Tests
     runs-on: saas
@@ -827,6 +941,15 @@ jobs:
           git.xiaoxiajianji.com/xiaoxia/base/playwright:v1.45.0-jammy \
           sh -lc 'npm ci && npx playwright test --reporter=line e2e/test_auth.spec.ts e2e/test_asset.spec.ts e2e/test_project.spec.ts'
 
+    - name: Notify CI failure
+      if: failure()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 失败通知 ==="
+        FAILED_JOB="Staging API Integration Tests" python3 scripts/ci_notify_failure.py
+
+
 
   build-production-runtime-images:
     name: Build Production Runtime Images
@@ -907,6 +1030,15 @@ jobs:
         echo "Disk usage after cleanup:"
         df -h / | tail -1
 
+    - name: Notify CI failure
+      if: failure()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 失败通知 ==="
+        FAILED_JOB="Build Production Runtime Images" python3 scripts/ci_notify_failure.py
+
+
   deploy-production:
     name: Deploy Production
     runs-on: saas
@@ -972,6 +1104,23 @@ jobs:
 
         echo "$DEPLOY_B64" | base64 -d | ssh -p 22222 -i "$key_path" "$production_user@$production_host" "IMAGE_TAG='${GITHUB_REF_NAME}' REGISTRY_TOKEN='${REGISTRY_TOKEN}' sh"
 
+    - name: Notify CI success
+      if: success()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 成功通知 ==="
+        SUCCESS_JOB="生产部署成功" python3 scripts/ci_notify_success.py
+
+    - name: Notify CI failure
+      if: failure()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 失败通知 ==="
+        FAILED_JOB="Deploy Production" python3 scripts/ci_notify_failure.py
+
+
   production-e2e:
     name: Production Browser E2E
     runs-on: saas
@@ -1040,3 +1189,12 @@ jobs:
           -w /workspace/apps/web \
           git.xiaoxiajianji.com/xiaoxia/base/playwright:v1.45.0-jammy \
           sh -lc 'npm ci && npx playwright test --reporter=line --project=chromium e2e/auth.spec.ts e2e/auth-guard.spec.ts e2e/core-upload.spec.ts e2e/core-generation.spec.ts e2e/core-titles.spec.ts'
+
+    - name: Notify CI failure
+      if: failure()
+      shell: sh
+      run: |
+        set +e
+        echo "=== CI 失败通知 ==="
+        FAILED_JOB="Production Browser E2E" python3 scripts/ci_notify_failure.py
+

apps/api/app/__init__.py

@@ -1 +0,0 @@
-"""API application package."""

apps/api/app/api/__init__.py

@@ -1 +0,0 @@
-"""API package."""

apps/api/app/api/router.py

@@ -4,17 +4,14 @@ from app.api.routes.assets import router as assets_router
 from app.api.routes.auth import router as auth_router
 from app.api.routes.chunked_upload import router as chunked_upload_router
 from app.api.routes.classification_jobs import router as classification_jobs_router
-from app.api.routes.dashboard import router as dashboard_router
 from app.api.routes.duplication import router as duplication_router
 from app.api.routes.edit_plans import router as edit_plans_router
-from app.api.routes.edit_templates import router as edit_templates_router
-from app.api.routes.generated_videos import router as generated_videos_router
+from app.api.routes.feature_flags import router as feature_flags_router
 from app.api.routes.generation_tasks import router as generation_tasks_router
 from app.api.routes.health import router as health_check_router
 from app.api.routes.ingest_jobs import router as ingest_jobs_router
-from app.api.routes.jobs import router as jobs_router
+from app.api.routes.internal_render import router as internal_render_router
 from app.api.routes.projects import router as projects_router
-from app.api.routes.recipes import router as recipes_router
 from app.api.routes.subscription import router as subscription_router
 from app.api.routes.tags import router as tags_router
 from app.api.routes.task_center import router as task_center_router
@@ -87,15 +84,6 @@ api_router.include_router(
     prefix="/generation",
     tags=["Generation"],
 )
-api_router.include_router(
-    jobs_router,
-    tags=["Job"],
-)
-api_router.include_router(
-    generated_videos_router,
-    prefix="/generated-videos",
-    tags=["GeneratedVideo"],
-)
 api_router.include_router(
     titles_router,
     prefix="/titles",
@@ -121,26 +109,11 @@ api_router.include_router(
     prefix="/subscription",
     tags=["Subscription"],
 )
-api_router.include_router(
-    recipes_router,
-    prefix="/recipes",
-    tags=["Recipe"],
-)
 api_router.include_router(
     templates_router,
     prefix="/templates",
     tags=["Template"],
 )
-api_router.include_router(
-    dashboard_router,
-    prefix="/dashboard",
-    tags=["Dashboard"],
-)
-api_router.include_router(
-    edit_templates_router,
-    prefix="/edit-templates",
-    tags=["EditTemplate"],
-)
 api_router.include_router(
     edit_plans_router,
     prefix="/edit-plans",
@@ -151,3 +124,11 @@ api_router.include_router(
     prefix="/tts",
     tags=["TTS"],
 )
+api_router.include_router(
+    feature_flags_router,
+    tags=["Internal"],
+)
+api_router.include_router(
+    internal_render_router,
+    tags=["Internal"],
+)

apps/api/app/api/routes/_helpers.py

@@ -0,0 +1,48 @@
+"""路由层共享辅助函数 — 消除跨文件重复定义。"""
+
+from typing import Any
+
+from fastapi import HTTPException, status
+
+from packages.application import GetProjectUseCase
+from packages.ports.user_repository import UserRepository
+
+
+def check_project_access(project_id: str, user_id: str, project_repository) -> None:
+    """检查用户是否有项目访问权限。
+
+    合并自 asset_libraries.py / edit_plans.py 的同名函数。
+    - 空 project_id 直接放行（兼容 edit_plans 中 project_id 可选的场景）
+    - 错误信息使用中文，与项目其他路由保持一致
+    """
+    if not project_id or not project_id.strip():
+        return
+    project = project_repository.find_by_id(project_id)
+    if project is None:
+        raise HTTPException(status_code=404, detail="项目不存在")
+    if not project.can_access(user_id):
+        raise HTTPException(status_code=403, detail="无权访问该项目")
+
+
+def get_user_plan(user_id: str, user_repository: UserRepository) -> str:
+    """获取用户的订阅计划名称。"""
+    user = user_repository.find_by_id(user_id)
+    if user is None:
+        return "free"
+    return getattr(user, "subscription_plan", "free") or "free"
+
+
+def require_project_and_library(
+    project_id: str,
+    library_id: str,
+    project_repository: Any,
+    asset_library_repository: Any,
+) -> None:
+    """Verify project and asset library exist."""
+    project = GetProjectUseCase(project_repository).execute(project_id)
+    if project is None:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Project not found")
+
+    libraries = asset_library_repository.find_by_project(project_id)
+    if not any(item.id == library_id for item in libraries):
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Asset library not found")

apps/api/app/api/routes/asset_libraries.py

@@ -12,7 +12,7 @@ from app.schemas.asset_library import (
     EnsureDefaultLibraryRequest,
     ListAssetLibrariesResponse,
 )
-from fastapi import APIRouter, Depends, HTTPException, Query, status
+from fastapi import APIRouter, Depends, HTTPException, Query, Response, status
 
 from packages.application import (
     CreateAssetLibraryCommand,
@@ -22,18 +22,11 @@ from packages.application import (
 )
 from packages.domain import AssetLibrary, AssetLibraryKind
 
+from ._helpers import check_project_access
+
 router = APIRouter()
 
 
-def _check_project_access(project_id: str, user_id: str, project_repository) -> None:
-    """检查用户是否有项目访问权限"""
-    project = project_repository.find_by_id(project_id)
-    if project is None:
-        raise HTTPException(status_code=404, detail=f"Project {project_id} not found")
-    if not project.can_access(user_id):
-        raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail="Access denied to project")
-
-
 def _to_asset_library_response(item) -> AssetLibraryResponse:
     return AssetLibraryResponse(
         id=item.id,
@@ -153,7 +146,7 @@ def ensure_default_library(
     return _to_asset_library_response(created)
 
 
-@router.delete("/{library_id}", status_code=status.HTTP_204_NO_CONTENT)
+@router.delete("/{library_id}", status_code=status.HTTP_204_NO_CONTENT, response_class=Response)
 def delete_asset_library(
     library_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -168,7 +161,7 @@ def delete_asset_library(
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="素材库不存在")
 
     # 权限校验：检查用户是否有项目访问权限
-    _check_project_access(library.project_id, authenticated_user.user.id, project_repository)
+    check_project_access(library.project_id, authenticated_user.user.id, project_repository)
 
     # 删除库内所有素材（无 FK 级联，需手动清理）
     assets_in_library = asset_repository.find_by_library(library_id)

apps/api/app/api/routes/assets.py

@@ -1,6 +1,7 @@
 import logging
 from typing import Any, Optional
 
+from app.api.routes._helpers import check_project_access
 from app.auth import AuthenticatedUser, get_current_user
 from app.core.storage import get_storage_service
 from app.dependencies import (
@@ -19,7 +20,7 @@ from app.schemas.asset import (
     UpdateAssetReviewRequest,
 )
 from app.schemas.tag import TagAssetsRequest
-from fastapi import APIRouter, Depends, HTTPException, Query
+from fastapi import APIRouter, Depends, HTTPException, Query, Response
 
 from packages.application import (
     CreateAssetCommand,
@@ -72,15 +73,6 @@ def _to_asset_response(item, storage_service=None) -> AssetResponse:
     )
 
 
-def _check_project_access(project_id: str, user_id: str, project_repository) -> None:
-    """检查用户是否有项目访问权限"""
-    project = project_repository.find_by_id(project_id)
-    if project is None:
-        raise HTTPException(status_code=404, detail=f"Project {project_id} not found")
-    if not project.can_access(user_id):
-        raise HTTPException(status_code=403, detail="Access denied to project")
-
-
 @router.get("", response_model=ListAssetsResponse)
 def list_assets(
     library_id: Optional[str] = Query(None),
@@ -136,7 +128,7 @@ def list_assets(
             library = asset_library_repository.get(library_id)
             if library is None:
                 raise HTTPException(status_code=404, detail=f"AssetLibrary {library_id} not found")
-            _check_project_access(library.project_id, user_id, project_repository)
+            check_project_access(library.project_id, user_id, project_repository)
             if ft:
                 items = asset_repository.find_by_library_and_file_type(library_id, ft, skip=skip, limit=limit)
                 total = asset_repository.count_by_project(library.project_id) if not kind else len(items)
@@ -152,7 +144,7 @@ def list_assets(
 
         # 模式2：指定 project_id
         if project_id:
-            _check_project_access(project_id, user_id, project_repository)
+            check_project_access(project_id, user_id, project_repository)
             if ft:
                 # 无直接方法，加载后按 file_type 过滤（仍比全量加载好）
                 all_items = asset_repository.find_by_project(project_id)
@@ -210,13 +202,13 @@ def list_assets(
         library = asset_library_repository.get(library_id)
         if library is None:
             raise HTTPException(status_code=404, detail=f"AssetLibrary {library_id} not found")
-        _check_project_access(library.project_id, user_id, project_repository)
+        check_project_access(library.project_id, user_id, project_repository)
         if kind:
             all_items = asset_repository.find_by_library_and_file_type(library_id, kind_to_file_type[kind])
         else:
             all_items = asset_repository.find_by_library(library_id)
     elif project_id:
-        _check_project_access(project_id, user_id, project_repository)
+        check_project_access(project_id, user_id, project_repository)
         all_items = asset_repository.find_by_project(project_id)
     else:
         try:
@@ -262,7 +254,7 @@ def update_asset_review_status(
     item = asset_repository.get(asset_id)
     if item is None:
         raise HTTPException(status_code=404, detail=f"Asset {asset_id} not found")
-    _check_project_access(item.project_id, authenticated_user.user.id, project_repository)
+    check_project_access(item.project_id, authenticated_user.user.id, project_repository)
     _apply_asset_review_status(item, request.review_status)
     updated = asset_repository.update(item)
     return _to_asset_response(updated)
@@ -286,7 +278,7 @@ def batch_delete_assets(
             failed_ids.append(asset_id)
             continue
         try:
-            _check_project_access(item.project_id, user_id, project_repository)
+            check_project_access(item.project_id, user_id, project_repository)
             deleted_ids.append(asset_id)
         except HTTPException:
             failed_ids.append(asset_id)
@@ -307,7 +299,7 @@ def get_asset(
     item = asset_repository.find_by_id(asset_id)
     if item is None:
         raise HTTPException(status_code=404, detail=f"Asset {asset_id} not found")
-    _check_project_access(item.project_id, authenticated_user.user.id, project_repository)
+    check_project_access(item.project_id, authenticated_user.user.id, project_repository)
     return _to_asset_response(item)
 
 
@@ -322,7 +314,7 @@ def update_asset(
     item = asset_repository.find_by_id(asset_id)
     if item is None:
         raise HTTPException(status_code=404, detail=f"Asset {asset_id} not found")
-    _check_project_access(item.project_id, authenticated_user.user.id, project_repository)
+    check_project_access(item.project_id, authenticated_user.user.id, project_repository)
 
     # 合并可修改字段
     if request.name is not None:
@@ -336,7 +328,7 @@ def update_asset(
     return _to_asset_response(updated)
 
 
-@router.delete("/{asset_id}", status_code=204)
+@router.delete("/{asset_id}", status_code=204, response_class=Response)
 def delete_asset(
     asset_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -346,7 +338,7 @@ def delete_asset(
     item = asset_repository.find_by_id(asset_id)
     if item is None:
         raise HTTPException(status_code=404, detail=f"Asset {asset_id} not found")
-    _check_project_access(item.project_id, authenticated_user.user.id, project_repository)
+    check_project_access(item.project_id, authenticated_user.user.id, project_repository)
     asset_repository.delete(asset_id)
 
 
@@ -363,7 +355,7 @@ def tag_asset(
     item = asset_repository.find_by_id(asset_id)
     if item is None:
         raise HTTPException(status_code=404, detail=f"Asset {asset_id} not found")
-    _check_project_access(item.project_id, authenticated_user.user.id, project_repository)
+    check_project_access(item.project_id, authenticated_user.user.id, project_repository)
     for tag_id in request.tag_ids:
         tag = tag_repository.get(tag_id)
         if tag is None:
@@ -375,7 +367,7 @@ def tag_asset(
     return _to_asset_response(updated)
 
 
-@router.delete("/{asset_id}/tags/{tag_id}", status_code=204)
+@router.delete("/{asset_id}/tags/{tag_id}", status_code=204, response_class=Response)
 def untag_asset(
     asset_id: str,
     tag_id: str,
@@ -387,7 +379,7 @@ def untag_asset(
     item = asset_repository.find_by_id(asset_id)
     if item is None:
         raise HTTPException(status_code=404, detail=f"Asset {asset_id} not found")
-    _check_project_access(item.project_id, authenticated_user.user.id, project_repository)
+    check_project_access(item.project_id, authenticated_user.user.id, project_repository)
     item.remove_tag(tag_id)
     asset_repository.update(item)
 

apps/api/app/api/routes/auth.py

@@ -105,7 +105,7 @@ async def register(
     request: RegisterRequest,
     user_repository: UserRepository = Depends(get_user_repository),
     email_service=Depends(get_auth_email_service),
-):
+) -> RegisterResponse:
     use_case = RegisterUserUseCase(
         user_repository=user_repository,
         base_url=settings.APP_BASE_URL,
@@ -136,7 +136,7 @@ async def login(
     request: LoginRequest,
     user_repository: UserRepository = Depends(get_user_repository),
     session_store=Depends(get_auth_session_store),
-):
+) -> LoginResponse:
     use_case = LoginUseCase(
         user_repository=user_repository,
         session_store=session_store,
@@ -162,7 +162,7 @@ async def refresh(
     request: RefreshRequest,
     user_repository: UserRepository = Depends(get_user_repository),
     session_store=Depends(get_auth_session_store),
-):
+) -> LoginResponse:
     use_case = RefreshTokenUseCase(
         user_repository=user_repository,
         session_store=session_store,
@@ -194,7 +194,7 @@ def _verify_email_token(token: str, user_repository: UserRepository) -> MessageR
 async def verify_email(
     token: str,
     user_repository: UserRepository = Depends(get_user_repository),
-):
+) -> MessageResponse:
     return _verify_email_token(token, user_repository)
 
 
@@ -202,16 +202,16 @@ async def verify_email(
 async def verify_email_post(
     request: VerifyEmailRequestModel,
     user_repository: UserRepository = Depends(get_user_repository),
-):
+) -> MessageResponse:
     return _verify_email_token(request.token, user_repository)
 
 
-@router.post("/password/forgot", response_model=MessageResponse, status_code=status.HTTP_202_ACCEPTED)
+@router.post("/forgot-password", response_model=MessageResponse, status_code=status.HTTP_202_ACCEPTED)
 async def forgot_password(
     request: PasswordResetRequestModel,
     user_repository: UserRepository = Depends(get_user_repository),
     email_service=Depends(get_auth_email_service),
-):
+) -> MessageResponse:
     success, error = RequestPasswordResetUseCase(
         user_repository=user_repository,
         base_url=settings.APP_BASE_URL,
@@ -223,11 +223,11 @@ async def forgot_password(
     return MessageResponse(message="如果账户存在，密码重置邮件已发送")
 
 
-@router.post("/password/reset", response_model=MessageResponse)
+@router.post("/reset-password", response_model=MessageResponse)
 async def reset_password(
     request: ResetPasswordModel,
     user_repository: UserRepository = Depends(get_user_repository),
-):
+) -> MessageResponse:
     success, error = ResetPasswordUseCase(user_repository=user_repository).execute(
         ResetPasswordRequest(token=request.token, new_password=request.new_password)
     )
@@ -241,9 +241,8 @@ async def reset_password(
 async def logout(
     credentials: HTTPAuthorizationCredentials | None = Depends(bearer_scheme),
     current_user: AuthenticatedUser = Depends(get_current_user),
-):
+) -> MessageResponse:
     """登出 - 将当前 token 加入黑名单"""
-    from fastapi.security import HTTPAuthorizationCredentials, HTTPBearer
 
     if credentials:
         try:
@@ -258,7 +257,7 @@ async def logout(
 @router.get("/me", response_model=CurrentUserResponse)
 async def get_current_user_info(
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
-):
+) -> CurrentUserResponse:
     user = authenticated_user.user
     return CurrentUserResponse(
         user_id=user.id,
@@ -324,7 +323,7 @@ def _get_internal_api_keys() -> list[str]:
             if content:
                 return [k.strip() for k in content.split(",") if k.strip()]
     except Exception:
-        logger.debug("Failed to read internal API keys from file", exc_info=True)
+        logger.warning("无法读取内部 API 密钥文件，仅依赖环境变量配置", exc_info=True)
 
     return []
 
@@ -355,7 +354,7 @@ async def wechat_sync(
     request: WechatSyncRequest,
     user_repository: UserRepository = Depends(get_user_repository),
     _: bool = Depends(_verify_internal_api_key),
-):
+) -> WechatSyncResponse:
     """
     微信同步登录/注册（系统级内部接口）
 

apps/api/app/api/routes/chunked_upload.py

@@ -13,8 +13,8 @@ from pathlib import Path
 from typing import Any
 from uuid import uuid4
 
+from app.api.routes._helpers import require_project_and_library
 from app.auth import AuthenticatedUser, get_current_user
-from app.config import get_settings
 from app.core.celery_app import celery_app
 from app.core.storage import OSSStorageService, get_storage_service
 from app.dependencies import (
@@ -113,22 +113,6 @@ def _atomic_check_and_record(upload_id: str, chunk_index: int) -> bool:
             fcntl.flock(f.fileno(), fcntl.LOCK_UN)
 
 
-def _require_project_and_library(
-    project_id: str,
-    library_id: str,
-    project_repository: Any,
-    asset_library_repository: Any,
-) -> None:
-    """Verify project and asset library exist"""
-    project = GetProjectUseCase(project_repository).execute(project_id)
-    if project is None:
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Project not found")
-
-    libraries = asset_library_repository.find_by_project(project_id)
-    if not any(item.id == library_id for item in libraries):
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Asset library not found")
-
-
 def _load_upload_meta(upload_id: str) -> dict[str, Any]:
     """Load upload metadata"""
     meta_path = _get_upload_meta_path(upload_id)
@@ -206,7 +190,6 @@ async def init_chunked_upload(
     asset_library_repository: Any = Depends(get_asset_library_repository),
 ) -> ChunkedUploadInitResponse:
     """Initialize chunked upload"""
-    settings = get_settings()
 
     # Validate file size
     if request.file_size > MAX_FILE_SIZE:
@@ -221,7 +204,7 @@ async def init_chunked_upload(
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Project not found")
 
     # Verify asset library
-    _require_project_and_library(
+    require_project_and_library(
         request.project_id,
         request.library_id,
         project_repository,

apps/api/app/api/routes/dashboard.py

@@ -1,92 +0,0 @@
-from typing import Any
-
-from app.auth import AuthenticatedUser, get_current_user
-from app.dependencies import (
-    get_asset_repository,
-    get_generation_task_repository,
-    get_project_repository,
-    get_title_library_repository,
-    get_voice_library_repository,
-)
-from app.schemas.dashboard import DashboardOverviewResponse, RecentTaskItem, SubscriptionInfo
-from fastapi import APIRouter, Depends
-
-router = APIRouter()
-
-
-def _status_value(status) -> str:
-    return status.value if hasattr(status, "value") else str(status)
-
-
-def _generation_step(status: str) -> str:
-    if status == "pending":
-        return "等待 Worker 执行"
-    if status == "running":
-        return "正在生成成片"
-    if status == "completed":
-        return "生成完成"
-    if status == "failed":
-        return "生成失败"
-    return status
-
-
-@router.get("/overview", response_model=DashboardOverviewResponse)
-def get_dashboard_overview(
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-    asset_repository: Any = Depends(get_asset_repository),
-    generation_task_repository: Any = Depends(get_generation_task_repository),
-    title_library_repository: Any = Depends(get_title_library_repository),
-    voice_library_repository: Any = Depends(get_voice_library_repository),
-) -> DashboardOverviewResponse:
-    """Dashboard 概览：用户级汇总数据。"""
-    user_id = authenticated_user.user.id
-
-    # 获取用户可访问的所有 project
-    projects = project_repository.find_accessible_projects(user_id)
-    project_ids = [p.id for p in projects]
-
-    # 素材统计
-    total_assets = asset_repository.count_by_project_ids(project_ids)
-    used_storage_bytes = asset_repository.sum_storage_by_project_ids(project_ids)
-
-    # 标题库 / 配音库统计
-    total_titles = title_library_repository.count_by_user(user_id)
-    total_voices = voice_library_repository.count_by_user(user_id)
-
-    # 生成任务统计
-    total_tasks = generation_task_repository.count_by_user(user_id)
-
-    # 最近任务（SQL 层 LIMIT 5）
-    recent = generation_task_repository.list_recent_by_user(user_id, limit=5)
-    recent_tasks = []
-    for task in recent:
-        s = _status_value(task.status)
-        recent_tasks.append(
-            RecentTaskItem(
-                id=task.id,
-                task_type="generation",
-                status=s,
-                current_step=_generation_step(s),
-                error_message=task.error_message or "",
-                updated_at=task.completed_at or task.started_at or task.created_at,
-            )
-        )
-
-    # 订阅信息
-    user = authenticated_user.user
-    subscription = SubscriptionInfo(
-        plan=getattr(user, "subscription_plan", "free") or "free",
-        is_active=getattr(user, "subscription_status", "") == "active",
-    )
-
-    return DashboardOverviewResponse(
-        total_assets=total_assets,
-        used_storage_bytes=used_storage_bytes,
-        total_titles=total_titles,
-        total_voices=total_voices,
-        total_tasks=total_tasks,
-        total_products=len(projects),
-        subscription=subscription,
-        recent_tasks=recent_tasks,
-    )

apps/api/app/api/routes/duplication.py

@@ -239,7 +239,7 @@ def get_duplication_detail(
     return _to_detail_response(record)
 
 
-@router.delete("/records/{record_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
+@router.delete("/records/{record_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_duplication_record(
     record_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -257,7 +257,7 @@ def delete_duplication_record(
 
     use_case = DeleteDuplicationRecordUseCase(duplication_repository)
     use_case.execute(record_id)
-    return Response(status_code=204)
+    return
 
 
 @router.post("/records/{record_id}/retry", response_model=DuplicationUploadResponse)

apps/api/app/api/routes/edit_plans.py

@@ -6,12 +6,11 @@ RESTful CRUD for EditPlan:
 - POST   /api/v1/edit-plans            创建
 - PUT    /api/v1/edit-plans/{id}       更新（含状态机流转）
 - DELETE /api/v1/edit-plans/{id}       删除
-- POST   /api/v1/edit-plans/{id}/generate          触发剪辑渲染生成（任务 2.05）
-- GET    /api/v1/edit-plans/{id}/generation-status  查询生成进度（任务 2.05）
-- POST   /api/v1/edit-plans/{id}/ai-recommend       AI 推荐片段方案（任务 3.09）
-- POST   /api/v1/edit-plans/{id}/generate-cover     AI 生成封面（任务 3.09）
-- GET    /api/v1/edit-plans/{id}/timeline           时间线场景数据
-- POST   /api/v1/edit-plans/generate-from-template  基于模板+素材自动生成剪辑计划
+
+拆分模块（各自独立 router，由本文件 include_router 聚合）:
+- edit_plans_generation.py   生成相关（generate / generation-status / generations）
+- edit_plans_ai.py           AI 推荐 & 封面（ai-recommend / generate-cover）
+- edit_plans_timeline.py     时间线 & 模板生成（timeline / generate-from-template）
 
 业务逻辑委托给 EditPlanService 服务层。
 """
@@ -23,37 +22,18 @@ from datetime import datetime
 from typing import Any, List, Optional
 
 from app.auth import AuthenticatedUser, get_current_user
-from app.core.celery_app import celery_app
-from app.core.task_enqueue import GLOBAL_PENDING_LIMIT, USER_PENDING_LIMIT
-from app.dependencies import get_asset_library_repository, get_asset_repository, get_db_session, get_project_repository
+from app.dependencies import get_db_session, get_project_repository
 from app.schemas.generation_task import GenerationTaskResponse
-from app.services import EditPlanService, PlanGeneratorService
-from fastapi import APIRouter, Depends, HTTPException, Query, status
+from app.services import EditPlanService
+from fastapi import APIRouter, Depends, HTTPException, Query, Response, status
 from pydantic import BaseModel, Field
 from sqlalchemy.orm import Session
 
-from packages.adapters.sqlalchemy_impl.asset_library_repository import (
-    SQLAlchemyAssetLibraryRepository,
-)
-from packages.adapters.sqlalchemy_impl.asset_repository import (
-    SQLAlchemyAssetRepository,
-)
-from packages.adapters.sqlalchemy_impl.generation_task_repository import (
-    SQLAlchemyGenerationTaskRepository,
-)
-from packages.adapters.sqlalchemy_impl.template_clip_config_repository import (
-    SQLAlchemyTemplateClipConfigRepository,
-)
-from packages.adapters.sqlalchemy_impl.template_repository import (
-    SQLAlchemyTemplateRepository,
-)
-from packages.application.generation_tasks import (
-    CreateGenerationTaskCommand,
-    CreateGenerationTaskUseCase,
-)
 from packages.domain.config_schemas import normalize_plan_config
 from packages.domain.edit_plan import EditPlan, EditPlanStatus
 
+from ._helpers import check_project_access
+
 logger = logging.getLogger(__name__)
 
 router = APIRouter()
@@ -247,17 +227,6 @@ class GenerateFromTemplateResponse(BaseModel):
 # ── Helpers ───────────────────────────────────────────────────────────────────
 
 
-def _check_project_access(project_id: str, user_id: str, project_repository: Any) -> None:
-    """校验用户对项目的访问权限（参照 assets.py 的 can_access 模式）"""
-    if not project_id or not project_id.strip():
-        return
-    project = project_repository.find_by_id(project_id)
-    if project is None:
-        raise HTTPException(status_code=404, detail="项目不存在")
-    if not project.can_access(user_id):
-        raise HTTPException(status_code=403, detail="无权访问该项目")
-
-
 def _to_response(p: EditPlan) -> EditPlanResponse:
     return EditPlanResponse(
         id=p.id,
@@ -273,7 +242,7 @@ def _to_response(p: EditPlan) -> EditPlanResponse:
     )
 
 
-# ── Routes ────────────────────────────────────────────────────────────────────
+# ── CRUD Routes ───────────────────────────────────────────────────────────────
 
 
 @router.get("", response_model=EditPlanListResponse)
@@ -311,7 +280,7 @@ def list_plans(
 
     # 项目鉴权：如果指定了 project_id，校验用户是否有权访问
     if project_id:
-        _check_project_access(project_id, current_user.user.id, project_repository)
+        check_project_access(project_id, current_user.user.id, project_repository)
 
     skip = (page - 1) * page_size
     plans = svc.list_plans(
@@ -353,7 +322,7 @@ def get_plan(
         )
     # 项目鉴权
     if plan.project_id:
-        _check_project_access(plan.project_id, current_user.user.id, project_repository)
+        check_project_access(plan.project_id, current_user.user.id, project_repository)
     return _to_response(plan)
 
 
@@ -369,7 +338,7 @@ def create_plan(
     project_id = (body.project_id or "").strip()
     # 项目鉴权
     if project_id:
-        _check_project_access(project_id, current_user.user.id, project_repository)
+        check_project_access(project_id, current_user.user.id, project_repository)
     svc = EditPlanService(db)
     # 标准化 config，填充 cover/title/subtitle/bgm 默认值
     normalized_config = normalize_plan_config(body.config)
@@ -411,7 +380,7 @@ def update_plan(
     if existing is None:
         raise HTTPException(status_code=404, detail=f"剪辑计划不存在: {plan_id}")
     if existing.project_id:
-        _check_project_access(existing.project_id, current_user.user.id, project_repository)
+        check_project_access(existing.project_id, current_user.user.id, project_repository)
 
     # 基础字段更新
     try:
@@ -453,7 +422,7 @@ def update_plan(
     return _to_response(result)
 
 
-@router.delete("/{plan_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
+@router.delete("/{plan_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_plan(
     plan_id: str,
     db: Session = Depends(get_db_session),
@@ -465,7 +434,7 @@ def delete_plan(
     # 项目鉴权
     existing = svc.get_plan(plan_id)
     if existing and existing.project_id:
-        _check_project_access(existing.project_id, current_user.user.id, project_repository)
+        check_project_access(existing.project_id, current_user.user.id, project_repository)
     deleted = svc.delete_plan(plan_id)
     if not deleted:
         raise HTTPException(
@@ -479,753 +448,12 @@ def delete_plan(
     )
 
 
-# ── 生成相关端点（任务 2.05） ─────────────────────────────────────────────────
+# ── Include sub-routers (拆分模块) ────────────────────────────────────────────
 
+from .edit_plans_ai import router as ai_router
+from .edit_plans_generation import router as generation_router
+from .edit_plans_timeline import router as timeline_router
 
-@router.post("/{plan_id}/generate", response_model=EditPlanGenerateResponse)
-def generate_plan(
-    plan_id: str,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-    asset_library_repo: Any = Depends(get_asset_library_repository),
-    asset_repo: Any = Depends(get_asset_repository),
-) -> EditPlanGenerateResponse:
-    """触发剪辑计划渲染生成
-
-    前置条件：计划状态必须为 editing，且至少有一个片段。
-    流程：
-    1. 验证计划状态为 editing
-    2. 将 pending 片段标记为 ready
-    3. 创建 GenerationTask
-    4. 调度 Celery 任务 worker.render_edit_plan
-    5. 将计划状态流转为 rendering
-    """
-    svc = EditPlanService(db)
-    # 项目鉴权
-    plan_check = svc.get_plan(plan_id)
-    if plan_check is None:
-        raise HTTPException(status_code=404, detail=f"剪辑计划不存在: {plan_id}")
-    if plan_check.project_id:
-        _check_project_access(plan_check.project_id, current_user.user.id, project_repository)
-
-    # ── 自动兜底 1: draft → editing ──────────────────────────────────────
-    if plan_check.status == EditPlanStatus.DRAFT:
-        logger.info("自动兜底: plan=%s draft→editing", plan_id)
-        svc.transition_status(plan_id, EditPlanStatus.EDITING)
-
-    # ── 自动兜底 2: 无片段 + 有 template_id → 从模板复制片段配置 ──────────
-    existing_clips = svc.count_clips(plan_id)
-    if existing_clips == 0 and plan_check.template_id:
-        logger.info(
-            "自动兜底: plan=%s 无片段，从模板 %s 复制片段配置",
-            plan_id,
-            plan_check.template_id,
-        )
-        # 优先从新模型 template_clip_configs 读取，若无则回退到旧模型 template_segments
-        clip_config_repo = SQLAlchemyTemplateClipConfigRepository(db)
-        configs = clip_config_repo.list_by_template(plan_check.template_id)
-        if configs:
-            for cfg in configs:
-                svc.create_clip(
-                    plan_id=plan_id,
-                    clip_type=cfg.clip_type.value if hasattr(cfg.clip_type, "value") else cfg.clip_type,
-                    order=cfg.order,
-                    template_clip_config_id=cfg.id,
-                    duration=cfg.default_duration,
-                    transition_effect=(
-                        cfg.transition_effect.value
-                        if hasattr(cfg.transition_effect, "value")
-                        else cfg.transition_effect
-                    ),
-                )
-            logger.info("自动兜底: plan=%s 从新模型 template_clip_configs 复制了 %d 个片段", plan_id, len(configs))
-        else:
-            # 回退到旧模型 template_segments
-            tpl_repo = SQLAlchemyTemplateRepository(db)
-            segments = tpl_repo.list_segments(plan_check.template_id)
-            for seg in segments:
-                avg_duration = (seg.duration_min + seg.duration_max) / 2
-                svc.create_clip(
-                    plan_id=plan_id,
-                    clip_type="main",  # 旧模型无结构角色，统一为主体片段
-                    order=seg.segment_order,
-                    duration=avg_duration,
-                    config={
-                        "material_type": seg.material_type or "",
-                        "template_segment_id": seg.id,
-                    },
-                )
-            logger.info("自动兜底: plan=%s 从旧模型 template_segments 复制了 %d 个片段", plan_id, len(segments))
-
-    # ── 自动兜底 3: 为没有素材的片段分配素材 ──────────────────────────
-    # 如果 plan.config.asset_ids 有素材，但 clips 没有 asset_id，自动按顺序分配
-    all_clips = svc.list_clips(plan_id)
-    clips_without_asset = [c for c in all_clips if not c.asset_id]
-    config_asset_ids = (plan_check.config or {}).get("asset_ids", [])
-    material_mode = (plan_check.config or {}).get("material_mode", "manual")
-
-    if clips_without_asset and config_asset_ids:
-        logger.info(
-            "自动兜底3: plan=%s 为 %d 个无素材片段分配 %d 个指定素材",
-            plan_id,
-            len(clips_without_asset),
-            len(config_asset_ids),
-        )
-        for i, clip in enumerate(clips_without_asset):
-            asset_idx = i % len(config_asset_ids)
-            svc.assign_asset(clip.id, config_asset_ids[asset_idx])
-        logger.info("自动兜底3: plan=%s 素材分配完成", plan_id)
-        clips_without_asset = []  # 已分配完
-
-    # ── 自动兜底 4: 自动素材模式 → 从项目默认视频素材库选取 ────────────
-    if clips_without_asset and material_mode == "auto" and plan_check.project_id:
-        import random
-
-        logger.info(
-            "自动兜底4: plan=%s 自动素材模式，从项目素材库选取素材 (%d 个片段需要)",
-            plan_id,
-            len(clips_without_asset),
-        )
-        # 找到项目的视频素材库
-        libs = asset_library_repo.find_by_project(plan_check.project_id)
-        video_lib = None
-        for lib in libs:
-            lib_kind = lib.kind.value if hasattr(lib.kind, "value") else lib.kind
-            if lib_kind == "video":
-                video_lib = lib
-                break
-
-        if video_lib:
-            assets = asset_repo.find_by_library(video_lib.id)
-            # 筛选 ready 状态的视频素材
-            ready_videos = [
-                a
-                for a in assets
-                if (a.status.value if hasattr(a.status, "value") else a.status) == "ready"
-                and a.mime_type
-                and a.mime_type.startswith("video")
-            ]
-            if ready_videos:
-                # 随机选取，按片段数轮询分配
-                random.shuffle(ready_videos)
-                for i, clip in enumerate(clips_without_asset):
-                    asset = ready_videos[i % len(ready_videos)]
-                    svc.assign_asset(clip.id, asset.id)
-                logger.info(
-                    "自动兜底4: plan=%s 从素材库 %s 分配了 %d 个素材给 %d 个片段",
-                    plan_id,
-                    video_lib.name,
-                    len(ready_videos),
-                    len(clips_without_asset),
-                )
-            else:
-                logger.warning("自动兜底4: plan=%s 素材库无可用视频素材", plan_id)
-        else:
-            logger.warning("自动兜底4: plan=%s 项目无视频素材库", plan_id)
-
-    # 检查是否可生成
-    try:
-        can_gen, reason = svc.can_generate(plan_id)
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(exc),
-        )
-    if not can_gen:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=reason,
-        )
-
-    # 核心生成流程：捕获异常返回明确错误信息，避免裸 500
-    try:
-        # 将 pending 片段标记为 ready
-        clip_count = svc.mark_clips_ready(plan_id)
-
-        # 创建 GenerationTask
-        gen_task_repo = SQLAlchemyGenerationTaskRepository(db)
-
-        # 队列限流预检查（repository 不支持计数时跳过）
-        user_id = current_user.user.id
-        try:
-            has_count = hasattr(gen_task_repo, "count_pending_by_user") and hasattr(
-                gen_task_repo, "count_pending_total"
-            )
-            if has_count:
-                user_pending = gen_task_repo.count_pending_by_user(user_id)
-                global_pending = gen_task_repo.count_pending_total()
-                if user_pending >= USER_PENDING_LIMIT:
-                    raise HTTPException(
-                        status_code=429,
-                        detail=f"您的待处理任务过多（当前 {user_pending}/{USER_PENDING_LIMIT}），请等待完成后再提交",
-                    )
-                if global_pending >= GLOBAL_PENDING_LIMIT:
-                    raise HTTPException(
-                        status_code=503,
-                        detail="系统繁忙，请稍后再试",
-                    )
-        except HTTPException:
-            raise
-        except Exception as e:
-            logger.warning("[队列限流] 剪辑计划限流检查失败，跳过: %s", e)
-
-        gen_task_use_case = CreateGenerationTaskUseCase(gen_task_repo)
-        plan = svc.get_plan_or_raise(plan_id)
-        gen_task = gen_task_use_case.execute(
-            CreateGenerationTaskCommand(
-                project_id="",
-                template_id=plan.template_id,
-                created_by_user_id=current_user.user.id,
-                source_edit_plan_id=plan_id,
-            )
-        )
-
-        # 将 generation_task_id 存入 plan config
-        svc.update_plan_config(plan_id, {"generation_task_id": gen_task.id})
-
-        # 流转状态为 rendering
-        svc.transition_status(plan_id, EditPlanStatus.RENDERING)
-
-        # 调度 Celery 任务
-        celery_app.send_task("worker.render_edit_plan", args=[plan_id])
-
-        # 获取最新状态
-        updated_plan = svc.get_plan_or_raise(plan_id)
-
-        logger.info(
-            "触发剪辑计划生成: plan_id=%s gen_task_id=%s clips=%d by user=%s",
-            plan_id,
-            gen_task.id,
-            clip_count,
-            current_user.user.id,
-        )
-
-        return EditPlanGenerateResponse(
-            plan_id=plan_id,
-            plan_status=updated_plan.status.value if hasattr(updated_plan.status, "value") else updated_plan.status,
-            generation_task_id=gen_task.id,
-            clip_count=clip_count,
-        )
-    except HTTPException:
-        # 已处理的 HTTP 异常直接透传
-        raise
-    except Exception as exc:
-        logger.exception("触发剪辑计划生成失败: plan_id=%s", plan_id)
-        # 尝试将计划标记为失败（RENDERING → FAILED 是合法的状态流转）
-        try:
-            svc.transition_status(plan_id, EditPlanStatus.FAILED)
-        except Exception:
-            logger.warning("标记计划失败状态时异常: plan_id=%s", plan_id)
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail="生成失败，请稍后重试",
-        )
-
-
-@router.get(
-    "/{plan_id}/generation-status",
-    response_model=EditPlanGenerationStatusResponse,
-)
-def get_generation_status(
-    plan_id: str,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-) -> EditPlanGenerationStatusResponse:
-    """查询剪辑计划生成进度
-
-    返回计划状态、关联的 GenerationTask ID、以及每个片段的状态。
-    """
-    svc = EditPlanService(db)
-    try:
-        gen_status = svc.get_generation_status(plan_id)
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(exc),
-        )
-
-    plan = gen_status["plan"]
-    # 项目鉴权
-    if plan.project_id:
-        _check_project_access(plan.project_id, current_user.user.id, project_repository)
-    clips = gen_status["clips"]
-
-    clip_items = [
-        ClipStatusItem(
-            clip_id=c.id,
-            clip_type=c.clip_type,
-            order=c.order,
-            status=c.status.value if hasattr(c.status, "value") else c.status,
-            asset_id=c.asset_id or "",
-            text_content=c.text_content or "",
-            duration=c.duration,
-        )
-        for c in clips
-    ]
-
-    return EditPlanGenerationStatusResponse(
-        plan_id=plan_id,
-        plan_status=plan.status.value if hasattr(plan.status, "value") else plan.status,
-        generation_task_id=gen_status["generation_task_id"],
-        clips=clip_items,
-    )
-
-
-@router.get(
-    "/{plan_id}/generations",
-    response_model=EditPlanGenerationsResponse,
-)
-def list_plan_generations(
-    plan_id: str,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-) -> EditPlanGenerationsResponse:
-    """查询剪辑计划关联的所有生成记录
-
-    返回该剪辑计划触发的所有 GenerationTask，按创建时间倒序。
-    """
-    svc = EditPlanService(db)
-    # 验证计划存在 + 项目鉴权
-    plan = svc.get_plan_or_raise(plan_id)
-    if plan.project_id:
-        _check_project_access(plan.project_id, current_user.user.id, project_repository)
-
-    gen_task_repo = SQLAlchemyGenerationTaskRepository(db)
-    tasks = gen_task_repo.list_by_source_edit_plan(plan_id)
-    items = [
-        GenerationTaskResponse(
-            id=t.id,
-            project_id=t.project_id,
-            asset_library_id=t.asset_library_id,
-            strategy_id=t.strategy_id,
-            voice_library_id=t.voice_library_id,
-            template_id=t.template_id,
-            asset_ids=t.asset_ids,
-            title_ids=t.title_ids,
-            voice_ids=t.voice_ids,
-            source_edit_plan_id=t.source_edit_plan_id or "",
-            status=t.status.value if hasattr(t.status, "value") else t.status,
-            progress=t.progress,
-            result_count=t.result_count,
-            error_message=t.error_message,
-        )
-        for t in tasks
-    ]
-    return EditPlanGenerationsResponse(items=items, total=len(items))
-
-
-# ── AI 推荐 & 封面生成端点（任务 3.09） ────────────────────────────────────────
-
-
-@router.post(
-    "/{plan_id}/ai-recommend",
-    response_model=AIRecommendResponse,
-)
-def ai_recommend_clips(
-    plan_id: str,
-    body: AIRecommendRequest,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-) -> AIRecommendResponse:
-    """AI 推荐片段方案
-
-    调用 AI 服务分析素材，自动生成片段编排方案并写入剪辑计划。
-
-    流程：
-    1. 验证计划存在且状态为 draft/editing
-    2. 调用 AI 推荐服务（当前为 stub，后续接入真实 AI）
-    3. 清除计划现有片段，按推荐方案重新创建
-    4. 更新计划 config（cover/title/subtitle/bgm）和 total_duration
-    5. 返回推荐方案详情
-
-    前端对接：
-    - 请求体只需传 asset_ids（必填），editing_mode 和 target_duration 可选
-    - 返回的 clips 可直接渲染到时间线
-    - 返回的 config 包含推荐的封面/标题/字幕/BGM 配置
-    """
-    svc = EditPlanService(db)
-
-    # 验证计划存在
-    try:
-        plan = svc.get_plan_or_raise(plan_id)
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(exc),
-        )
-
-    # 项目鉴权
-    if plan.project_id:
-        _check_project_access(plan.project_id, current_user.user.id, project_repository)
-
-    # 验证状态：只允许 draft 或 editing
-    plan_status = plan.status.value if hasattr(plan.status, "value") else plan.status
-    if plan_status not in ("draft", "editing"):
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail="当前计划状态不支持AI推荐，请先创建或编辑计划后再试",
-        )
-
-    # 调用 AI 推荐服务（同步调用 stub，后续改为 Celery 异步）
-    from apps.worker.worker_app.tasks.ai_tasks import run_ai_recommend
-
-    result = run_ai_recommend(
-        plan_id=plan_id,
-        template_id=plan.template_id,
-        asset_ids=body.asset_ids,
-        editing_mode=body.editing_mode,
-        target_duration=body.target_duration,
-    )
-
-    # ── 事务保护：清除 → 重建 → 更新 必须在同一逻辑事务中 ──
-    # TODO: 当前各 repo 方法内部 commit()，无法真正回滚。
-    #       后续重构 repo 为 flush() 模式后，此处改为统一 commit。
-    try:
-        # 清除现有片段
-        svc.delete_all_clips(plan_id)
-
-        # 按推荐方案创建新片段
-        for clip_data in result["clips"]:
-            svc.create_clip(
-                plan_id=plan_id,
-                clip_type=clip_data["clip_type"],
-                order=clip_data["order"],
-                text_content=clip_data.get("text_content", ""),
-                duration=clip_data["duration"],
-                transition_effect=clip_data.get("transition_effect", "cut"),
-                asset_id=clip_data.get("asset_id", ""),
-                start_time=clip_data.get("start_time", 0.0),
-                config=clip_data.get("config", {}),
-            )
-
-        # 更新计划 config 和 total_duration
-        normalized_config = normalize_plan_config(result.get("config", {}))
-        svc.update_plan(
-            plan_id,
-            config=normalized_config,
-            total_duration=result["total_duration"],
-        )
-    except Exception as exc:
-        logger.exception("AI 推荐写入失败，plan_id=%s 数据可能不一致", plan_id)
-        # 尝试回滚未提交的变更
-        try:
-            db.rollback()
-        except Exception:
-            pass
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail="AI推荐结果保存失败，请稍后重试",
-        )
-
-    logger.info(
-        "AI 推荐片段方案: plan_id=%s clips=%d duration=%.1f by user=%s",
-        plan_id,
-        len(result["clips"]),
-        result["total_duration"],
-        current_user.user.id,
-    )
-
-    return AIRecommendResponse(
-        plan_id=plan_id,
-        clips=[
-            AIRecommendClipItem(
-                clip_type=c["clip_type"],
-                order=c["order"],
-                text_content=c.get("text_content", ""),
-                duration=c["duration"],
-                transition_effect=c.get("transition_effect", "cut"),
-                asset_id=c.get("asset_id", ""),
-                start_time=c.get("start_time", 0.0),
-                config=c.get("config", {}),
-            )
-            for c in result["clips"]
-        ],
-        config=normalized_config,
-        total_duration=result["total_duration"],
-        confidence=result["confidence"],
-    )
-
-
-@router.post(
-    "/{plan_id}/generate-cover",
-    response_model=GenerateCoverResponse,
-)
-def generate_cover(
-    plan_id: str,
-    body: GenerateCoverRequest,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-) -> GenerateCoverResponse:
-    """AI 生成封面
-
-    调用 AI 服务从视频中选帧或生成封面图，并更新计划 config.cover。
-
-    流程：
-    1. 验证计划存在
-    2. 调用 AI 封面生成服务（当前为 stub，后续接入真实 AI）
-    3. 更新 plan.config["cover"] 为生成的封面数据
-    4. 返回封面数据
-
-    前端对接：
-    - cover_type=ai_frame: AI 智能选帧（默认）
-    - cover_type=manual: 手动选帧，需传 frame_time
-    - cover_type=upload: 用户上传，接口返回空 image_url，前端自行上传后更新
-    - cover_type=ai_regenerate: AI 重新生成
-    """
-    svc = EditPlanService(db)
-
-    # 验证计划存在
-    try:
-        plan = svc.get_plan_or_raise(plan_id)
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(exc),
-        )
-
-    # 项目鉴权
-    if plan.project_id:
-        _check_project_access(plan.project_id, current_user.user.id, project_repository)
-
-    # 调用 AI 封面生成服务
-    from apps.worker.worker_app.tasks.ai_tasks import run_generate_cover
-
-    cover_data = run_generate_cover(
-        plan_id=plan_id,
-        asset_ids=body.asset_ids,
-        cover_type=body.cover_type,
-        frame_time=body.frame_time,
-    )
-
-    # 更新 plan.config["cover"]
-    current_config = dict(plan.config)
-    current_config["cover"] = cover_data
-    normalized = normalize_plan_config(current_config)
-    svc.update_plan_config(plan_id, {"cover": normalized["cover"]})
-
-    logger.info(
-        "AI 封面生成: plan_id=%s type=%s by user=%s",
-        plan_id,
-        body.cover_type,
-        current_user.user.id,
-    )
-
-    return GenerateCoverResponse(
-        plan_id=plan_id,
-        cover=cover_data,
-    )
-
-
-# ── Timeline / Scene 端点（P2-6） ─────────────────────────────────────────────
-
-
-class TimelineSceneResponse(BaseModel):
-    """时间线场景"""
-
-    scene: str = Field(..., description="场景描述")
-    time: str = Field(..., description='时间范围，如 "0:00 - 0:05"')
-    duration: float = Field(..., ge=0, description="时长（秒）")
-    color: str = Field(..., description="展示颜色")
-    clip_id: str = Field(default="", description="关联的片段 ID")
-    clip_type: str = Field(default="", description="片段类型")
-
-
-class TimelineResponse(BaseModel):
-    """时间线响应"""
-
-    plan_id: str
-    total_duration: float
-    scenes: List[TimelineSceneResponse]
-
-
-# clip_type → 颜色映射
-_CLIP_TYPE_COLORS = {
-    "intro": "#6366f1",
-    "title": "#6366f1",
-    "product": "#818cf8",
-    "showcase": "#10b981",
-    "scene": "#10b981",
-    "subtitle": "#f59e0b",
-    "text": "#f59e0b",
-    "cta": "#ef4444",
-    "outro": "#ef4444",
-    "voiceover": "#8b5cf6",
-    "transition": "#64748b",
-}
-
-_DEFAULT_COLOR = "#6366f1"
-
-
-def _format_time(seconds: float) -> str:
-    """将秒数格式化为 M:SS"""
-    m = int(seconds) // 60
-    s = int(seconds) % 60
-    return f"{m}:{s:02d}"
-
-
-def _clip_type_to_scene_label(clip_type: str, text_content: str) -> str:
-    """根据 clip_type 和 text_content 生成场景描述"""
-    type_labels = {
-        "intro": "开场",
-        "title": "标题",
-        "product": "产品展示",
-        "showcase": "场景展示",
-        "scene": "场景",
-        "subtitle": "字幕",
-        "text": "文字",
-        "cta": "结尾 CTA",
-        "outro": "结尾",
-        "voiceover": "配音",
-        "transition": "转场",
-    }
-    label = type_labels.get(clip_type, clip_type or "片段")
-    if text_content:
-        # 截取前 20 个字符作为副标题
-        short = text_content[:20].strip()
-        if short:
-            return f"{label} - {short}"
-    return label
-
-
-@router.get(
-    "/{plan_id}/timeline",
-    response_model=TimelineResponse,
-)
-def get_plan_timeline(
-    plan_id: str,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-) -> TimelineResponse:
-    """获取剪辑计划的时间线场景数据
-
-    返回按计划片段排序的时间线场景列表，供前端 GeneratePage 渲染使用。
-    """
-    svc = EditPlanService(db)
-    plan = svc.get_plan_or_raise(plan_id)
-    # 项目鉴权
-    if plan.project_id:
-        _check_project_access(plan.project_id, current_user.user.id, project_repository)
-
-    clips = svc.list_clips(plan_id=plan_id, skip=0, limit=200)
-    # 按 order 排序
-    clips.sort(key=lambda c: c.order)
-
-    scenes: List[TimelineSceneResponse] = []
-    current_time = 0.0
-
-    for clip in clips:
-        start = current_time
-        end = start + clip.duration
-        color = _CLIP_TYPE_COLORS.get(clip.clip_type, _DEFAULT_COLOR)
-        scene_label = _clip_type_to_scene_label(clip.clip_type, clip.text_content)
-
-        scenes.append(
-            TimelineSceneResponse(
-                scene=scene_label,
-                time=f"{_format_time(start)} - {_format_time(end)}",
-                duration=clip.duration,
-                color=color,
-                clip_id=clip.id,
-                clip_type=clip.clip_type,
-            )
-        )
-        current_time = end
-
-    total_duration = sum(s.duration for s in scenes) or plan.total_duration
-
-    return TimelineResponse(
-        plan_id=plan_id,
-        total_duration=total_duration,
-        scenes=scenes,
-    )
-
-
-# ── 基于模板生成剪辑计划 ─────────────────────────────────────────────────────
-
-
-@router.post(
-    "/generate-from-template",
-    response_model=GenerateFromTemplateResponse,
-    status_code=status.HTTP_201_CREATED,
-)
-def generate_from_template(
-    body: GenerateFromTemplateRequest,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-    project_repository: Any = Depends(get_project_repository),
-) -> GenerateFromTemplateResponse:
-    """基于模板 + 素材自动生成剪辑计划
-
-    流程:
-    1. 获取模板及其片段配置
-    2. 调用 PlanGeneratorService 生成 EditPlan + EditPlanClips
-    3. 返回完整的计划和片段列表
-    """
-    from app.services import EditTemplateService
-
-    # 项目鉴权
-    if body.project_id:
-        _check_project_access(body.project_id, current_user.user.id, project_repository)
-
-    template_svc = EditTemplateService(db)
-
-    # 获取模板
-    try:
-        template = template_svc.get_template_or_raise(body.template_id)
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(exc),
-        )
-
-    # 获取模板片段配置
-    clip_configs = template_svc.list_clip_configs(body.template_id, skip=0, limit=200)
-
-    # 调用 PlanGeneratorService 生成计划
-    generator = PlanGeneratorService(db)
-    result = generator.generate_from_template(
-        template=template,
-        clip_configs=clip_configs,
-        asset_ids=body.asset_ids,
-        project_id=body.project_id,
-        created_by_user_id=current_user.user.id,
-        name=body.name,
-    )
-
-    plan = result["plan"]
-    clips = result["clips"]
-
-    logger.info(
-        "基于模板生成剪辑计划: plan_id=%s template_id=%s clips=%d by user=%s",
-        plan.id,
-        body.template_id,
-        len(clips),
-        current_user.user.id,
-    )
-
-    return GenerateFromTemplateResponse(
-        plan=_to_response(plan),
-        clips=[
-            _PlanClipItem(
-                id=c.id,
-                clip_type=c.clip_type,
-                order=c.order,
-                asset_id=c.asset_id,
-                text_content=c.text_content,
-                start_time=c.start_time,
-                duration=c.duration,
-                transition_effect=c.transition_effect,
-                status=c.status.value if hasattr(c.status, "value") else c.status,
-                config=c.config,
-                created_at=c.created_at,
-                updated_at=c.updated_at,
-            )
-            for c in clips
-        ],
-    )
+router.include_router(generation_router)
+router.include_router(ai_router)
+router.include_router(timeline_router)

apps/api/app/api/routes/edit_plans_ai.py

@@ -0,0 +1,199 @@
+"""剪辑计划 AI 推荐 & 封面生成 API 端点。
+
+从 edit_plans.py 拆分，包含：
+- POST /{plan_id}/ai-recommend    AI 推荐片段方案
+- POST /{plan_id}/generate-cover  AI 生成封面
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any
+
+from app.api.routes._helpers import check_project_access
+from app.api.routes.edit_plans import (
+    AIRecommendClipItem,
+    AIRecommendRequest,
+    AIRecommendResponse,
+    GenerateCoverRequest,
+    GenerateCoverResponse,
+)
+from app.auth import AuthenticatedUser, get_current_user
+from app.dependencies import get_db_session, get_project_repository
+from app.services import EditPlanService
+from fastapi import APIRouter, Depends, HTTPException, status
+from sqlalchemy.orm import Session
+
+from packages.domain.config_schemas import normalize_plan_config
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter()
+
+
+@router.post(
+    "/{plan_id}/ai-recommend",
+    response_model=AIRecommendResponse,
+)
+def ai_recommend_clips(
+    plan_id: str,
+    body: AIRecommendRequest,
+    db: Session = Depends(get_db_session),
+    current_user: AuthenticatedUser = Depends(get_current_user),
+    project_repository: Any = Depends(get_project_repository),
+) -> AIRecommendResponse:
+    """AI 推荐片段方案
+
+    调用 AI 服务分析素材，自动生成片段编排方案并写入剪辑计划。
+
+    流程：
+    1. 验证计划存在且状态为 draft/editing
+    2. 调用 AI 推荐服务（当前为 stub，后续接入真实 AI）
+    3. 清除计划现有片段，按推荐方案重新创建
+    4. 更新计划 config（cover/title/subtitle/bgm）和 total_duration
+    5. 返回推荐方案详情
+    """
+    svc = EditPlanService(db)
+
+    try:
+        plan = svc.get_plan_or_raise(plan_id)
+    except ValueError as exc:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=str(exc))
+
+    if plan.project_id:
+        check_project_access(plan.project_id, current_user.user.id, project_repository)
+
+    plan_status = plan.status.value if hasattr(plan.status, "value") else plan.status
+    if plan_status not in ("draft", "editing"):
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail="当前计划状态不支持AI推荐，请先创建或编辑计划后再试",
+        )
+
+    from apps.worker.worker_app.tasks.ai_tasks import run_ai_recommend
+
+    result = run_ai_recommend(
+        plan_id=plan_id,
+        template_id=plan.template_id,
+        asset_ids=body.asset_ids,
+        editing_mode=body.editing_mode,
+        target_duration=body.target_duration,
+    )
+
+    # 事务保护：清除 → 重建 → 更新 必须在同一逻辑事务中
+    try:
+        svc.delete_all_clips(plan_id)
+
+        for clip_data in result["clips"]:
+            svc.create_clip(
+                plan_id=plan_id,
+                clip_type=clip_data["clip_type"],
+                order=clip_data["order"],
+                text_content=clip_data.get("text_content", ""),
+                duration=clip_data["duration"],
+                transition_effect=clip_data.get("transition_effect", "cut"),
+                asset_id=clip_data.get("asset_id", ""),
+                start_time=clip_data.get("start_time", 0.0),
+                config=clip_data.get("config", {}),
+            )
+
+        normalized_config = normalize_plan_config(result.get("config", {}))
+        svc.update_plan(
+            plan_id,
+            config=normalized_config,
+            total_duration=result["total_duration"],
+        )
+    except Exception:
+        logger.exception("AI 推荐写入失败，plan_id=%s 数据可能不一致", plan_id)
+        try:
+            db.rollback()
+        except Exception as rollback_err:
+            logger.error(
+                "AI 推荐回滚失败，数据库会话可能处于不一致状态: plan_id=%s error=%s",
+                plan_id,
+                rollback_err,
+            )
+        raise HTTPException(
+            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            detail="AI推荐结果保存失败，请稍后重试",
+        )
+
+    logger.info(
+        "AI 推荐片段方案: plan_id=%s clips=%d duration=%.1f by user=%s",
+        plan_id,
+        len(result["clips"]),
+        result["total_duration"],
+        current_user.user.id,
+    )
+
+    return AIRecommendResponse(
+        plan_id=plan_id,
+        clips=[
+            AIRecommendClipItem(
+                clip_type=c["clip_type"],
+                order=c["order"],
+                text_content=c.get("text_content", ""),
+                duration=c["duration"],
+                transition_effect=c.get("transition_effect", "cut"),
+                asset_id=c.get("asset_id", ""),
+                start_time=c.get("start_time", 0.0),
+                config=c.get("config", {}),
+            )
+            for c in result["clips"]
+        ],
+        config=normalized_config,
+        total_duration=result["total_duration"],
+        confidence=result["confidence"],
+    )
+
+
+@router.post(
+    "/{plan_id}/generate-cover",
+    response_model=GenerateCoverResponse,
+)
+def generate_cover(
+    plan_id: str,
+    body: GenerateCoverRequest,
+    db: Session = Depends(get_db_session),
+    current_user: AuthenticatedUser = Depends(get_current_user),
+    project_repository: Any = Depends(get_project_repository),
+) -> GenerateCoverResponse:
+    """AI 生成封面
+
+    调用 AI 服务从视频中选帧或生成封面图，并更新计划 config.cover。
+    """
+    svc = EditPlanService(db)
+
+    try:
+        plan = svc.get_plan_or_raise(plan_id)
+    except ValueError as exc:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=str(exc))
+
+    if plan.project_id:
+        check_project_access(plan.project_id, current_user.user.id, project_repository)
+
+    from apps.worker.worker_app.tasks.ai_tasks import run_generate_cover
+
+    cover_data = run_generate_cover(
+        plan_id=plan_id,
+        asset_ids=body.asset_ids,
+        cover_type=body.cover_type,
+        frame_time=body.frame_time,
+    )
+
+    current_config = dict(plan.config)
+    current_config["cover"] = cover_data
+    normalized = normalize_plan_config(current_config)
+    svc.update_plan_config(plan_id, {"cover": normalized["cover"]})
+
+    logger.info(
+        "AI 封面生成: plan_id=%s type=%s by user=%s",
+        plan_id,
+        body.cover_type,
+        current_user.user.id,
+    )
+
+    return GenerateCoverResponse(
+        plan_id=plan_id,
+        cover=cover_data,
+    )

apps/api/app/api/routes/edit_plans_generation.py

@@ -0,0 +1,383 @@
+"""剪辑计划生成相关 API 端点。
+
+从 edit_plans.py 拆分，包含：
+- POST /{plan_id}/generate          触发剪辑渲染生成
+- GET  /{plan_id}/generation-status 查询生成进度
+- GET  /{plan_id}/generations       查询关联的生成记录
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any
+
+from app.api.routes._helpers import check_project_access
+from app.api.routes.edit_plans import (
+    ClipStatusItem,
+    EditPlanGenerateResponse,
+    EditPlanGenerationsResponse,
+    EditPlanGenerationStatusResponse,
+)
+from app.auth import AuthenticatedUser, get_current_user
+from app.core.celery_app import celery_app
+from app.core.task_enqueue import GLOBAL_PENDING_LIMIT, USER_PENDING_LIMIT
+from app.dependencies import get_asset_library_repository, get_asset_repository, get_db_session, get_project_repository
+from app.services import EditPlanService
+from fastapi import APIRouter, Depends, HTTPException, status
+from sqlalchemy.orm import Session
+
+from packages.adapters.sqlalchemy_impl.generation_task_repository import (
+    SQLAlchemyGenerationTaskRepository,
+)
+from packages.adapters.sqlalchemy_impl.template_clip_config_repository import (
+    SQLAlchemyTemplateClipConfigRepository,
+)
+from packages.adapters.sqlalchemy_impl.template_repository import (
+    SQLAlchemyTemplateRepository,
+)
+from packages.application.generation_tasks import (
+    CreateGenerationTaskCommand,
+    CreateGenerationTaskUseCase,
+)
+from packages.domain.edit_plan import EditPlanStatus
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter()
+
+
+def _auto_fallback_draft_to_editing(svc: EditPlanService, plan_id: str, plan_check) -> None:
+    """自动兜底 1: draft → editing"""
+    if plan_check.status == EditPlanStatus.DRAFT:
+        logger.info("自动兜底: plan=%s draft→editing", plan_id)
+        svc.transition_status(plan_id, EditPlanStatus.EDITING)
+
+
+def _auto_fallback_copy_template_clips(svc: EditPlanService, plan_id: str, plan_check, db: Session) -> None:
+    """自动兜底 2: 无片段 + 有 template_id → 从模板复制片段配置"""
+    existing_clips = svc.count_clips(plan_id)
+    if existing_clips == 0 and plan_check.template_id:
+        logger.info(
+            "自动兜底: plan=%s 无片段，从模板 %s 复制片段配置",
+            plan_id,
+            plan_check.template_id,
+        )
+        clip_config_repo = SQLAlchemyTemplateClipConfigRepository(db)
+        configs = clip_config_repo.list_by_template(plan_check.template_id)
+        if configs:
+            for cfg in configs:
+                svc.create_clip(
+                    plan_id=plan_id,
+                    clip_type=cfg.clip_type.value if hasattr(cfg.clip_type, "value") else cfg.clip_type,
+                    order=cfg.order,
+                    template_clip_config_id=cfg.id,
+                    duration=cfg.default_duration,
+                    transition_effect=(
+                        cfg.transition_effect.value
+                        if hasattr(cfg.transition_effect, "value")
+                        else cfg.transition_effect
+                    ),
+                )
+            logger.info("自动兜底: plan=%s 从新模型 template_clip_configs 复制了 %d 个片段", plan_id, len(configs))
+        else:
+            tpl_repo = SQLAlchemyTemplateRepository(db)
+            segments = tpl_repo.list_segments(plan_check.template_id)
+            for seg in segments:
+                avg_duration = (seg.duration_min + seg.duration_max) / 2
+                svc.create_clip(
+                    plan_id=plan_id,
+                    clip_type="main",
+                    order=seg.segment_order,
+                    duration=avg_duration,
+                    config={
+                        "material_type": seg.material_type or "",
+                        "template_segment_id": seg.id,
+                    },
+                )
+            logger.info("自动兜底: plan=%s 从旧模型 template_segments 复制了 %d 个片段", plan_id, len(segments))
+
+
+def _auto_fallback_assign_assets(
+    svc: EditPlanService,
+    plan_id: str,
+    plan_check,
+) -> list:
+    """自动兜底 3: 为没有素材的片段分配素材。返回剩余无素材片段列表。"""
+    all_clips = svc.list_clips(plan_id)
+    clips_without_asset = [c for c in all_clips if not c.asset_id]
+    config_asset_ids = (plan_check.config or {}).get("asset_ids", [])
+
+    if clips_without_asset and config_asset_ids:
+        logger.info(
+            "自动兜底3: plan=%s 为 %d 个无素材片段分配 %d 个指定素材",
+            plan_id,
+            len(clips_without_asset),
+            len(config_asset_ids),
+        )
+        for i, clip in enumerate(clips_without_asset):
+            asset_idx = i % len(config_asset_ids)
+            svc.assign_asset(clip.id, config_asset_ids[asset_idx])
+        logger.info("自动兜底3: plan=%s 素材分配完成", plan_id)
+        clips_without_asset = []
+
+    return clips_without_asset
+
+
+def _auto_fallback_auto_material_mode(
+    svc: EditPlanService,
+    plan_id: str,
+    plan_check,
+    clips_without_asset: list,
+    asset_library_repo: Any,
+    asset_repo: Any,
+) -> None:
+    """自动兜底 4: 自动素材模式 → 从项目默认视频素材库选取"""
+    if not clips_without_asset:
+        return
+    material_mode = (plan_check.config or {}).get("material_mode", "manual")
+    if material_mode != "auto" or not plan_check.project_id:
+        return
+
+    import random
+
+    logger.info(
+        "自动兜底4: plan=%s 自动素材模式，从项目素材库选取素材 (%d 个片段需要)",
+        plan_id,
+        len(clips_without_asset),
+    )
+    libs = asset_library_repo.find_by_project(plan_check.project_id)
+    video_lib = None
+    for lib in libs:
+        lib_kind = lib.kind.value if hasattr(lib.kind, "value") else lib.kind
+        if lib_kind == "video":
+            video_lib = lib
+            break
+
+    if video_lib:
+        assets = asset_repo.find_by_library(video_lib.id)
+        ready_videos = [
+            a
+            for a in assets
+            if (a.status.value if hasattr(a.status, "value") else a.status) == "ready"
+            and a.mime_type
+            and a.mime_type.startswith("video")
+        ]
+        if ready_videos:
+            random.shuffle(ready_videos)
+            for i, clip in enumerate(clips_without_asset):
+                asset = ready_videos[i % len(ready_videos)]
+                svc.assign_asset(clip.id, asset.id)
+            logger.info(
+                "自动兜底4: plan=%s 从素材库 %s 分配了 %d 个素材给 %d 个片段",
+                plan_id,
+                video_lib.name,
+                len(ready_videos),
+                len(clips_without_asset),
+            )
+        else:
+            logger.warning("自动兜底4: plan=%s 素材库无可用视频素材", plan_id)
+    else:
+        logger.warning("自动兜底4: plan=%s 项目无视频素材库", plan_id)
+
+
+def _check_queue_limits(gen_task_repo, user_id: str) -> None:
+    """队列限流预检查"""
+    try:
+        has_count = hasattr(gen_task_repo, "count_pending_by_user") and hasattr(
+            gen_task_repo, "count_pending_total"
+        )
+        if has_count:
+            user_pending = gen_task_repo.count_pending_by_user(user_id)
+            global_pending = gen_task_repo.count_pending_total()
+            if user_pending >= USER_PENDING_LIMIT:
+                raise HTTPException(
+                    status_code=429,
+                    detail=f"您的待处理任务过多（当前 {user_pending}/{USER_PENDING_LIMIT}），请等待完成后再提交",
+                )
+            if global_pending >= GLOBAL_PENDING_LIMIT:
+                raise HTTPException(
+                    status_code=503,
+                    detail="系统繁忙，请稍后再试",
+                )
+    except HTTPException:
+        raise
+    except Exception as e:
+        logger.warning("[队列限流] 剪辑计划限流检查失败，跳过: %s", e)
+
+
+@router.post("/{plan_id}/generate", response_model=EditPlanGenerateResponse)
+def generate_plan(
+    plan_id: str,
+    db: Session = Depends(get_db_session),
+    current_user: AuthenticatedUser = Depends(get_current_user),
+    project_repository: Any = Depends(get_project_repository),
+    asset_library_repo: Any = Depends(get_asset_library_repository),
+    asset_repo: Any = Depends(get_asset_repository),
+) -> EditPlanGenerateResponse:
+    """触发剪辑计划渲染生成
+
+    前置条件：计划状态必须为 editing，且至少有一个片段。
+    流程：
+    1. 验证计划状态为 editing
+    2. 将 pending 片段标记为 ready
+    3. 创建 GenerationTask
+    4. 调度 Celery 任务 worker.render_edit_plan
+    5. 将计划状态流转为 rendering
+    """
+    svc = EditPlanService(db)
+    plan_check = svc.get_plan(plan_id)
+    if plan_check is None:
+        raise HTTPException(status_code=404, detail=f"剪辑计划不存在: {plan_id}")
+    if plan_check.project_id:
+        check_project_access(plan_check.project_id, current_user.user.id, project_repository)
+
+    # 自动兜底流程
+    _auto_fallback_draft_to_editing(svc, plan_id, plan_check)
+    _auto_fallback_copy_template_clips(svc, plan_id, plan_check, db)
+    clips_without_asset = _auto_fallback_assign_assets(svc, plan_id, plan_check)
+    _auto_fallback_auto_material_mode(svc, plan_id, plan_check, clips_without_asset, asset_library_repo, asset_repo)
+
+    # 检查是否可生成
+    try:
+        can_gen, reason = svc.can_generate(plan_id)
+    except ValueError as exc:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=str(exc))
+    if not can_gen:
+        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=reason)
+
+    # 核心生成流程
+    try:
+        clip_count = svc.mark_clips_ready(plan_id)
+
+        gen_task_repo = SQLAlchemyGenerationTaskRepository(db)
+        user_id = current_user.user.id
+        _check_queue_limits(gen_task_repo, user_id)
+
+        gen_task_use_case = CreateGenerationTaskUseCase(gen_task_repo)
+        plan = svc.get_plan_or_raise(plan_id)
+        gen_task = gen_task_use_case.execute(
+            CreateGenerationTaskCommand(
+                project_id="",
+                template_id=plan.template_id,
+                created_by_user_id=current_user.user.id,
+                source_edit_plan_id=plan_id,
+            )
+        )
+
+        svc.update_plan_config(plan_id, {"generation_task_id": gen_task.id})
+        svc.transition_status(plan_id, EditPlanStatus.RENDERING)
+        celery_app.send_task("worker.render_edit_plan", args=[plan_id])
+
+        updated_plan = svc.get_plan_or_raise(plan_id)
+
+        logger.info(
+            "触发剪辑计划生成: plan_id=%s gen_task_id=%s clips=%d by user=%s",
+            plan_id,
+            gen_task.id,
+            clip_count,
+            current_user.user.id,
+        )
+
+        return EditPlanGenerateResponse(
+            plan_id=plan_id,
+            plan_status=updated_plan.status.value if hasattr(updated_plan.status, "value") else updated_plan.status,
+            generation_task_id=gen_task.id,
+            clip_count=clip_count,
+        )
+    except HTTPException:
+        raise
+    except Exception:
+        logger.exception("触发剪辑计划生成失败: plan_id=%s", plan_id)
+        try:
+            svc.transition_status(plan_id, EditPlanStatus.FAILED)
+        except Exception:
+            logger.warning("标记计划失败状态时异常: plan_id=%s", plan_id)
+        raise HTTPException(
+            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            detail="生成失败，请稍后重试",
+        )
+
+
+@router.get(
+    "/{plan_id}/generation-status",
+    response_model=EditPlanGenerationStatusResponse,
+)
+def get_generation_status(
+    plan_id: str,
+    db: Session = Depends(get_db_session),
+    current_user: AuthenticatedUser = Depends(get_current_user),
+    project_repository: Any = Depends(get_project_repository),
+) -> EditPlanGenerationStatusResponse:
+    """查询剪辑计划生成进度"""
+    svc = EditPlanService(db)
+    try:
+        gen_status = svc.get_generation_status(plan_id)
+    except ValueError as exc:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=str(exc))
+
+    plan = gen_status["plan"]
+    if plan.project_id:
+        check_project_access(plan.project_id, current_user.user.id, project_repository)
+    clips = gen_status["clips"]
+
+    clip_items = [
+        ClipStatusItem(
+            clip_id=c.id,
+            clip_type=c.clip_type,
+            order=c.order,
+            status=c.status.value if hasattr(c.status, "value") else c.status,
+            asset_id=c.asset_id or "",
+            text_content=c.text_content or "",
+            duration=c.duration,
+        )
+        for c in clips
+    ]
+
+    return EditPlanGenerationStatusResponse(
+        plan_id=plan_id,
+        plan_status=plan.status.value if hasattr(plan.status, "value") else plan.status,
+        generation_task_id=gen_status["generation_task_id"],
+        clips=clip_items,
+    )
+
+
+@router.get(
+    "/{plan_id}/generations",
+    response_model=EditPlanGenerationsResponse,
+)
+def list_plan_generations(
+    plan_id: str,
+    db: Session = Depends(get_db_session),
+    current_user: AuthenticatedUser = Depends(get_current_user),
+    project_repository: Any = Depends(get_project_repository),
+) -> EditPlanGenerationsResponse:
+    """查询剪辑计划关联的所有生成记录"""
+    svc = EditPlanService(db)
+    plan = svc.get_plan_or_raise(plan_id)
+    if plan.project_id:
+        check_project_access(plan.project_id, current_user.user.id, project_repository)
+
+    from app.schemas.generation_task import GenerationTaskResponse
+
+    gen_task_repo = SQLAlchemyGenerationTaskRepository(db)
+    tasks = gen_task_repo.list_by_source_edit_plan(plan_id)
+    items = [
+        GenerationTaskResponse(
+            id=t.id,
+            project_id=t.project_id,
+            asset_library_id=t.asset_library_id,
+            strategy_id=t.strategy_id,
+            voice_library_id=t.voice_library_id,
+            template_id=t.template_id,
+            asset_ids=t.asset_ids,
+            title_ids=t.title_ids,
+            voice_ids=t.voice_ids,
+            source_edit_plan_id=t.source_edit_plan_id or "",
+            status=t.status.value if hasattr(t.status, "value") else t.status,
+            progress=t.progress,
+            result_count=t.result_count,
+            error_message=t.error_message,
+        )
+        for t in tasks
+    ]
+    return EditPlanGenerationsResponse(items=items, total=len(items))

apps/api/app/api/routes/edit_plans_timeline.py

@@ -0,0 +1,220 @@
+"""剪辑计划时间线 & 模板生成 API 端点。
+
+从 edit_plans.py 拆分，包含：
+- GET  /{plan_id}/timeline            时间线场景数据
+- POST /generate-from-template        基于模板+素材自动生成剪辑计划
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any, List
+
+from app.api.routes._helpers import check_project_access
+from app.api.routes.edit_plans import (
+    GenerateFromTemplateRequest,
+    GenerateFromTemplateResponse,
+    _PlanClipItem,
+    _to_response,
+)
+from app.auth import AuthenticatedUser, get_current_user
+from app.dependencies import get_db_session, get_project_repository
+from app.services import EditPlanService, PlanGeneratorService
+from fastapi import APIRouter, Depends, HTTPException, Query, status
+from pydantic import BaseModel, Field
+from sqlalchemy.orm import Session
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter()
+
+
+# ── Timeline Schemas ──────────────────────────────────────────────────────────
+
+
+class TimelineSceneResponse(BaseModel):
+    """时间线场景"""
+
+    scene: str = Field(..., description="场景描述")
+    time: str = Field(..., description='时间范围，如 "0:00 - 0:05"')
+    duration: float = Field(..., ge=0, description="时长（秒）")
+    color: str = Field(..., description="展示颜色")
+    clip_id: str = Field(default="", description="关联的片段 ID")
+    clip_type: str = Field(default="", description="片段类型")
+
+
+class TimelineResponse(BaseModel):
+    """时间线响应"""
+
+    plan_id: str
+    total_duration: float
+    scenes: List[TimelineSceneResponse]
+
+
+# clip_type → 颜色映射
+_CLIP_TYPE_COLORS = {
+    "intro": "#6366f1",
+    "title": "#6366f1",
+    "product": "#818cf8",
+    "showcase": "#10b981",
+    "scene": "#10b981",
+    "subtitle": "#f59e0b",
+    "text": "#f59e0b",
+    "cta": "#ef4444",
+    "outro": "#ef4444",
+    "voiceover": "#8b5cf6",
+    "transition": "#64748b",
+}
+
+_DEFAULT_COLOR = "#6366f1"
+
+
+def _format_time(seconds: float) -> str:
+    """将秒数格式化为 M:SS"""
+    m = int(seconds) // 60
+    s = int(seconds) % 60
+    return f"{m}:{s:02d}"
+
+
+def _clip_type_to_scene_label(clip_type: str, text_content: str) -> str:
+    """根据 clip_type 和 text_content 生成场景描述"""
+    type_labels = {
+        "intro": "开场",
+        "title": "标题",
+        "product": "产品展示",
+        "showcase": "场景展示",
+        "scene": "场景",
+        "subtitle": "字幕",
+        "text": "文字",
+        "cta": "结尾 CTA",
+        "outro": "结尾",
+        "voiceover": "配音",
+        "transition": "转场",
+    }
+    label = type_labels.get(clip_type, clip_type or "片段")
+    if text_content:
+        short = text_content[:20].strip()
+        if short:
+            return f"{label} - {short}"
+    return label
+
+
+# ── Routes ────────────────────────────────────────────────────────────────────
+
+
+@router.get(
+    "/{plan_id}/timeline",
+    response_model=TimelineResponse,
+)
+def get_plan_timeline(
+    plan_id: str,
+    db: Session = Depends(get_db_session),
+    current_user: AuthenticatedUser = Depends(get_current_user),
+    project_repository: Any = Depends(get_project_repository),
+) -> TimelineResponse:
+    """获取剪辑计划的时间线场景数据"""
+    svc = EditPlanService(db)
+    plan = svc.get_plan_or_raise(plan_id)
+    if plan.project_id:
+        check_project_access(plan.project_id, current_user.user.id, project_repository)
+
+    clips = svc.list_clips(plan_id=plan_id, skip=0, limit=200)
+    clips.sort(key=lambda c: c.order)
+
+    scenes: List[TimelineSceneResponse] = []
+    current_time = 0.0
+
+    for clip in clips:
+        start = current_time
+        end = start + clip.duration
+        color = _CLIP_TYPE_COLORS.get(clip.clip_type, _DEFAULT_COLOR)
+        scene_label = _clip_type_to_scene_label(clip.clip_type, clip.text_content)
+
+        scenes.append(
+            TimelineSceneResponse(
+                scene=scene_label,
+                time=f"{_format_time(start)} - {_format_time(end)}",
+                duration=clip.duration,
+                color=color,
+                clip_id=clip.id,
+                clip_type=clip.clip_type,
+            )
+        )
+        current_time = end
+
+    total_duration = sum(s.duration for s in scenes) or plan.total_duration
+
+    return TimelineResponse(
+        plan_id=plan_id,
+        total_duration=total_duration,
+        scenes=scenes,
+    )
+
+
+@router.post(
+    "/generate-from-template",
+    response_model=GenerateFromTemplateResponse,
+    status_code=status.HTTP_201_CREATED,
+)
+def generate_from_template(
+    body: GenerateFromTemplateRequest,
+    db: Session = Depends(get_db_session),
+    current_user: AuthenticatedUser = Depends(get_current_user),
+    project_repository: Any = Depends(get_project_repository),
+) -> GenerateFromTemplateResponse:
+    """基于模板 + 素材自动生成剪辑计划"""
+    from app.services import EditTemplateService
+
+    if body.project_id:
+        check_project_access(body.project_id, current_user.user.id, project_repository)
+
+    template_svc = EditTemplateService(db)
+
+    try:
+        template = template_svc.get_template_or_raise(body.template_id)
+    except ValueError as exc:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=str(exc))
+
+    clip_configs = template_svc.list_clip_configs(body.template_id, skip=0, limit=200)
+
+    generator = PlanGeneratorService(db)
+    result = generator.generate_from_template(
+        template=template,
+        clip_configs=clip_configs,
+        asset_ids=body.asset_ids,
+        project_id=body.project_id,
+        created_by_user_id=current_user.user.id,
+        name=body.name,
+    )
+
+    plan = result["plan"]
+    clips = result["clips"]
+
+    logger.info(
+        "基于模板生成剪辑计划: plan_id=%s template_id=%s clips=%d by user=%s",
+        plan.id,
+        body.template_id,
+        len(clips),
+        current_user.user.id,
+    )
+
+    return GenerateFromTemplateResponse(
+        plan=_to_response(plan),
+        clips=[
+            _PlanClipItem(
+                id=c.id,
+                clip_type=c.clip_type,
+                order=c.order,
+                asset_id=c.asset_id,
+                text_content=c.text_content,
+                start_time=c.start_time,
+                duration=c.duration,
+                transition_effect=c.transition_effect,
+                status=c.status.value if hasattr(c.status, "value") else c.status,
+                config=c.config,
+                created_at=c.created_at,
+                updated_at=c.updated_at,
+            )
+            for c in clips
+        ],
+    )

apps/api/app/api/routes/edit_templates.py

@@ -1,289 +0,0 @@
-"""模板管理 API — Phase 8 模板编排引擎.
-
-RESTful CRUD for EditTemplate:
-- GET  /api/v1/edit-templates      列表（分页 + 类型筛选）
-- GET  /api/v1/edit-templates/{id} 详情
-- POST /api/v1/edit-templates      创建（管理员）
-- PUT  /api/v1/edit-templates/{id} 更新
-- DELETE /api/v1/edit-templates/{id} 删除（软删除 → inactive）
-
-业务逻辑委托给 EditTemplateService 服务层。
-"""
-
-from __future__ import annotations
-
-import logging
-from datetime import datetime
-from typing import Any, List, Optional
-
-from app.auth import AuthenticatedUser, get_current_user
-from app.dependencies import get_db_session
-from app.services import EditTemplateService
-from fastapi import APIRouter, Depends, HTTPException, Query, status
-from fastapi.responses import Response
-from pydantic import BaseModel, Field
-from sqlalchemy.orm import Session
-
-from packages.domain.config_schemas import normalize_template_config
-from packages.domain.edit_template import EditTemplate, EditTemplateStatus
-
-logger = logging.getLogger(__name__)
-
-router = APIRouter()
-
-
-# ── Pydantic Schemas ─────────────────────────────────────────────────────────
-
-
-class EditTemplateCreateRequest(BaseModel):
-    """创建模板请求体"""
-
-    name: str = Field(..., min_length=1, max_length=200, description="模板名称")
-    description: str = Field(default="", max_length=2000, description="模板描述")
-    template_type: str = Field(default="default", max_length=50, description="模板类型")
-    editing_mode: str = Field(
-        default="one_take", max_length=20, description="剪辑模式: one_take/pip/voice_over/voice_pip"
-    )
-    config: dict[str, Any] = Field(default_factory=dict, description="模板配置 (JSON)")
-    preview_url: str = Field(default="", max_length=500, description="预览地址")
-    sort_weight: int = Field(default=0, ge=0, le=9999, description="排序权重")
-
-
-class EditTemplateUpdateRequest(BaseModel):
-    """更新模板请求体"""
-
-    name: Optional[str] = Field(default=None, min_length=1, max_length=200, description="模板名称")
-    description: Optional[str] = Field(default=None, max_length=2000, description="模板描述")
-    template_type: Optional[str] = Field(default=None, max_length=50, description="模板类型")
-    editing_mode: Optional[str] = Field(
-        default=None, max_length=20, description="剪辑模式: one_take/pip/voice_over/voice_pip"
-    )
-    config: Optional[dict[str, Any]] = Field(default=None, description="模板配置 (JSON)")
-    preview_url: Optional[str] = Field(default=None, max_length=500, description="预览地址")
-    sort_weight: Optional[int] = Field(default=None, ge=0, le=9999, description="排序权重")
-    status: Optional[str] = Field(default=None, description="状态: active / inactive")
-
-
-class EditTemplateResponse(BaseModel):
-    """模板响应体"""
-
-    id: str
-    name: str
-    description: str
-    template_type: str
-    editing_mode: str
-    config: dict[str, Any]
-    preview_url: str
-    sort_weight: int
-    status: str
-    created_at: datetime
-    updated_at: datetime
-
-    model_config = {"from_attributes": True}
-
-
-class EditTemplateListResponse(BaseModel):
-    """模板列表响应体"""
-
-    items: List[EditTemplateResponse]
-    total: int
-    page: int
-    page_size: int
-
-
-# ── Helpers ───────────────────────────────────────────────────────────────────
-
-
-def _require_admin(current_user: AuthenticatedUser) -> None:
-    """校验当前用户是否为管理员，非管理员返回 403"""
-    if not getattr(current_user.user, "is_admin", False):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail="仅管理员可执行此操作",
-        )
-
-
-def _to_response(t: EditTemplate) -> EditTemplateResponse:
-    return EditTemplateResponse(
-        id=t.id,
-        name=t.name,
-        description=t.description,
-        template_type=t.template_type,
-        editing_mode=t.editing_mode,
-        config=t.config,
-        preview_url=t.preview_url,
-        sort_weight=t.sort_weight,
-        status=t.status.value if hasattr(t.status, "value") else t.status,
-        created_at=t.created_at,
-        updated_at=t.updated_at,
-    )
-
-
-# ── Routes ────────────────────────────────────────────────────────────────────
-
-
-@router.get("", response_model=EditTemplateListResponse)
-def list_templates(
-    page: int = Query(default=1, ge=1, description="页码"),
-    page_size: int = Query(default=20, ge=1, le=100, description="每页数量"),
-    template_type: Optional[str] = Query(default=None, description="按类型筛选"),
-    status_filter: Optional[str] = Query(
-        default=None,
-        alias="status",
-        description="按状态筛选: active / inactive",
-    ),
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-) -> EditTemplateListResponse:
-    """获取模板列表（支持分页、按类型/状态筛选）"""
-    svc = EditTemplateService(db)
-
-    # 解析状态筛选
-    status_enum: Optional[EditTemplateStatus] = None
-    if status_filter:
-        try:
-            status_enum = EditTemplateStatus(status_filter)
-        except ValueError:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=f"无效的状态值: {status_filter}，可选值: active, inactive",
-            )
-
-    skip = (page - 1) * page_size
-    templates = svc.list_templates(
-        template_type=template_type,
-        status=status_enum,
-        skip=skip,
-        limit=page_size,
-    )
-    total = svc.count_templates(
-        template_type=template_type,
-        status=status_enum,
-    )
-
-    return EditTemplateListResponse(
-        items=[_to_response(t) for t in templates],
-        total=total,
-        page=page,
-        page_size=page_size,
-    )
-
-
-@router.get("/{template_id}", response_model=EditTemplateResponse)
-def get_template(
-    template_id: str,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-) -> EditTemplateResponse:
-    """获取单个模板详情"""
-    svc = EditTemplateService(db)
-    try:
-        template = svc.get_template_or_raise(template_id)
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(exc),
-        )
-    return _to_response(template)
-
-
-@router.post("", response_model=EditTemplateResponse, status_code=status.HTTP_201_CREATED)
-def create_template(
-    body: EditTemplateCreateRequest,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-) -> EditTemplateResponse:
-    """创建模板（管理员）"""
-    _require_admin(current_user)
-    svc = EditTemplateService(db)
-    # 标准化 config，填充 cover/title/subtitle/bgm 默认值
-    normalized_config = normalize_template_config(body.config)
-    try:
-        created = svc.create_template(
-            name=body.name,
-            description=body.description,
-            template_type=body.template_type,
-            editing_mode=body.editing_mode,
-            config=normalized_config,
-            preview_url=body.preview_url,
-            sort_weight=body.sort_weight,
-        )
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=str(exc),
-        )
-    logger.info("创建模板: id=%s name=%s by user=%s", created.id, created.name, current_user.user.id)
-    return _to_response(created)
-
-
-@router.put("/{template_id}", response_model=EditTemplateResponse)
-def update_template(
-    template_id: str,
-    body: EditTemplateUpdateRequest,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-) -> EditTemplateResponse:
-    """更新模板"""
-    _require_admin(current_user)
-    svc = EditTemplateService(db)
-
-    # 解析状态
-    status_enum: Optional[EditTemplateStatus] = None
-    if body.status is not None:
-        try:
-            status_enum = EditTemplateStatus(body.status)
-        except ValueError:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=f"无效的状态值: {body.status}，可选值: active, inactive",
-            )
-
-    # 标准化 config（如果提供了）
-    config_to_update = normalize_template_config(body.config) if body.config is not None else None
-
-    try:
-        result = svc.update_template(
-            template_id,
-            name=body.name,
-            description=body.description,
-            template_type=body.template_type,
-            editing_mode=body.editing_mode,
-            config=config_to_update,
-            preview_url=body.preview_url,
-            sort_weight=body.sort_weight,
-            status=status_enum,
-        )
-    except ValueError as exc:
-        err_msg = str(exc)
-        if "不存在" in err_msg:
-            raise HTTPException(
-                status_code=status.HTTP_404_NOT_FOUND,
-                detail=err_msg,
-            )
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=err_msg,
-        )
-    logger.info("更新模板: id=%s by user=%s", template_id, current_user.user.id)
-    return _to_response(result)
-
-
-@router.delete("/{template_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
-def delete_template(
-    template_id: str,
-    db: Session = Depends(get_db_session),
-    current_user: AuthenticatedUser = Depends(get_current_user),
-) -> Response:
-    """删除模板（软删除 → 设为 inactive）"""
-    _require_admin(current_user)
-    svc = EditTemplateService(db)
-    try:
-        svc.deactivate_template(template_id)
-    except ValueError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=str(exc),
-        )
-    logger.info("删除模板(软删除): id=%s by user=%s", template_id, current_user.user.id)
-    return Response(status_code=204)

apps/api/app/api/routes/feature_flags.py

@@ -0,0 +1,194 @@
+"""Feature Flag 内部管理接口。
+
+通过内部 API Key 鉴权，支持查看和修改 Feature Flag 配置。
+主要用于灰度发布期间的动态开关控制。
+
+API:
+  GET  /api/v1/internal/feature-flags          - 列出所有 flag
+  GET  /api/v1/internal/feature-flags/{name}   - 查看单个 flag
+  PUT  /api/v1/internal/feature-flags/{name}   - 设置 flag 配置
+  DELETE /api/v1/internal/feature-flags/{name} - 删除 flag
+
+鉴权：X-API-Key header，走内部 API Key 验证
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Optional
+
+from app.api.routes.auth import _verify_internal_api_key
+from app.config import settings
+from fastapi import APIRouter, Depends, HTTPException, Query, Response, status
+from pydantic import BaseModel, Field
+
+from packages.adapters.redis.feature_flag_store import (
+    FeatureFlagConfig,
+    RedisFeatureFlagStore,
+)
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter(prefix="/internal/feature-flags", tags=["Internal"])
+
+# 允许管理的 flag 白名单（防止误操作其他系统 flag）
+ALLOWED_FLAGS = {
+    "render_engine",
+}
+
+
+def _get_feature_flag_store() -> RedisFeatureFlagStore:
+    """获取 Feature Flag 存储实例。"""
+    return RedisFeatureFlagStore(redis_url=settings.REDIS_URL)
+
+
+class FeatureFlagUpdateRequest(BaseModel):
+    """Feature Flag 更新请求体。"""
+
+    enabled: bool = Field(..., description="是否启用")
+    percentage: int = Field(0, ge=0, le=100, description="灰度百分比 (0-100)")
+    whitelist: list[str] = Field(default_factory=list, description="白名单列表（如 user_id）")
+
+
+class FeatureFlagResponse(BaseModel):
+    """Feature Flag 响应。"""
+
+    name: str
+    enabled: bool
+    percentage: int
+    whitelist: list[str]
+
+    @classmethod
+    def from_config(cls, config: FeatureFlagConfig) -> "FeatureFlagResponse":
+        return cls(
+            name=config.name,
+            enabled=config.enabled,
+            percentage=config.percentage,
+            whitelist=sorted(config.whitelist),
+        )
+
+
+class FeatureFlagCheckResponse(BaseModel):
+    """Flag 激活检查响应。"""
+
+    name: str
+    active: bool
+    identifier: Optional[str] = None
+
+
+def _validate_flag_name(name: str) -> None:
+    """校验 flag 名称是否在允许列表中。"""
+    if name not in ALLOWED_FLAGS:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=f"Unsupported flag: {name}. Allowed: {sorted(ALLOWED_FLAGS)}",
+        )
+
+
+@router.get("", response_model=list[FeatureFlagResponse])
+async def list_feature_flags(
+    _: bool = Depends(_verify_internal_api_key),
+    store: RedisFeatureFlagStore = Depends(_get_feature_flag_store),
+) -> list[FeatureFlagResponse]:
+    """列出所有 Feature Flag。"""
+    try:
+        flags = store.list_all()
+        # 同时返回预定义的 flag（即使未设置也显示默认值）
+        result = []
+        for name in sorted(ALLOWED_FLAGS):
+            config = flags.get(name) or FeatureFlagConfig(name=name, enabled=False)
+            result.append(FeatureFlagResponse.from_config(config))
+        # 加上已存在但不在白名单中的 flag（只读展示）
+        for name, config in flags.items():
+            if name not in ALLOWED_FLAGS:
+                result.append(FeatureFlagResponse.from_config(config))
+        return sorted(result, key=lambda x: x.name)
+    except Exception as exc:
+        logger.error("Failed to list feature flags: %s", exc)
+        raise HTTPException(status_code=500, detail=f"Failed to list flags: {exc}")
+
+
+@router.get("/{name}", response_model=FeatureFlagResponse)
+async def get_feature_flag(
+    name: str,
+    _: bool = Depends(_verify_internal_api_key),
+    store: RedisFeatureFlagStore = Depends(_get_feature_flag_store),
+) -> FeatureFlagResponse:
+    """获取单个 Feature Flag 配置。"""
+    try:
+        config = store.get(name)
+        return FeatureFlagResponse.from_config(config)
+    except Exception as exc:
+        logger.error("Failed to get feature flag %s: %s", name, exc)
+        raise HTTPException(status_code=500, detail=f"Failed to get flag: {exc}")
+
+
+@router.get("/{name}/check", response_model=FeatureFlagCheckResponse)
+async def check_feature_flag(
+    name: str,
+    identifier: Optional[str] = Query(None, description="标识符，如 user_id"),
+    _: bool = Depends(_verify_internal_api_key),
+    store: RedisFeatureFlagStore = Depends(_get_feature_flag_store),
+) -> FeatureFlagCheckResponse:
+    """检查某个标识符是否命中 Feature Flag。"""
+    try:
+        active = store.is_active(name, identifier=identifier)
+        return FeatureFlagCheckResponse(name=name, active=active, identifier=identifier)
+    except Exception as exc:
+        logger.error("Failed to check feature flag %s: %s", name, exc)
+        raise HTTPException(status_code=500, detail=f"Failed to check flag: {exc}")
+
+
+@router.put("/{name}", response_model=FeatureFlagResponse)
+async def update_feature_flag(
+    name: str,
+    request: FeatureFlagUpdateRequest,
+    _: bool = Depends(_verify_internal_api_key),
+    store: RedisFeatureFlagStore = Depends(_get_feature_flag_store),
+) -> FeatureFlagResponse:
+    """更新 Feature Flag 配置。
+
+    只允许修改 ALLOWED_FLAGS 列表中的 flag。
+    """
+    _validate_flag_name(name)
+
+    try:
+        config = FeatureFlagConfig(
+            name=name,
+            enabled=request.enabled,
+            percentage=request.percentage,
+            whitelist=set(request.whitelist),
+        )
+        store.set(config)
+        logger.info(
+            "Feature flag updated: name=%s enabled=%s percentage=%d whitelist=%d",
+            name,
+            config.enabled,
+            config.percentage,
+            len(config.whitelist),
+        )
+        return FeatureFlagResponse.from_config(config)
+    except Exception as exc:
+        logger.error("Failed to update feature flag %s: %s", name, exc)
+        raise HTTPException(status_code=500, detail=f"Failed to update flag: {exc}")
+
+
+@router.delete("/{name}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
+async def delete_feature_flag(
+    name: str,
+    _: bool = Depends(_verify_internal_api_key),
+    store: RedisFeatureFlagStore = Depends(_get_feature_flag_store),
+) :
+    """删除 Feature Flag。
+
+    只允许删除 ALLOWED_FLAGS 列表中的 flag。
+    """
+    _validate_flag_name(name)
+
+    try:
+        deleted = store.delete(name)
+        logger.info("Feature flag deleted: name=%s deleted=%s", name, deleted)
+        pass
+    except Exception as exc:
+        logger.error("Failed to delete feature flag %s: %s", name, exc)
+        raise HTTPException(status_code=500, detail=f"Failed to delete flag: {exc}")

apps/api/app/api/routes/generated_videos.py

@@ -1,123 +0,0 @@
-from typing import Any
-
-from app.auth import AuthenticatedUser, get_current_user
-from app.core.storage import OSSStorageService, get_storage_service
-from app.dependencies import get_generated_video_repository, get_project_repository
-from app.schemas.generated_video import (
-    GeneratedVideoDownloadUrlResponse,
-    GeneratedVideoResponse,
-    ListGeneratedVideosResponse,
-    UpdateGeneratedVideoReviewRequest,
-)
-from fastapi import APIRouter, Depends, HTTPException, Query
-
-from packages.application import (
-    GetGeneratedVideoDownloadUrlUseCase,
-    GetGeneratedVideoUseCase,
-    ListGeneratedVideosUseCase,
-)
-
-router = APIRouter()
-
-
-def _to_generated_video_response(item, download_url: str | None = None) -> GeneratedVideoResponse:
-    return GeneratedVideoResponse(
-        id=item.id,
-        project_id=item.project_id,
-        generation_task_id=item.generation_task_id,
-        name=item.name,
-        file_url=item.file_url,
-        file_size=item.file_size,
-        duration=item.duration,
-        thumbnail_url=item.thumbnail_url,
-        width=item.width,
-        height=item.height,
-        fps=item.fps,
-        status=item.status,
-        review_status=item.review_status,
-        generation_params=item.generation_params,
-        download_url=download_url,
-    )
-
-
-@router.get("", response_model=ListGeneratedVideosResponse)
-def list_generated_videos(
-    project_id: str | None = Query(None),
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    generated_video_repository: Any = Depends(get_generated_video_repository),
-    project_repository: Any = Depends(get_project_repository),
-    storage_service: OSSStorageService = Depends(get_storage_service),
-) -> ListGeneratedVideosResponse:
-    user_id = authenticated_user.user.id
-    use_case = ListGeneratedVideosUseCase(generated_video_repository)
-
-    if project_id:
-        # If project_id provided, check access and filter by project
-        project = project_repository.find_by_id(project_id)
-        if project is None:
-            raise HTTPException(status_code=404, detail=f"Project {project_id} not found")
-        items = use_case.execute(project_id)
-    else:
-        # If no project_id, list all videos from accessible projects
-        accessible_projects = project_repository.find_accessible_projects(user_id)
-        all_items = []
-        for proj in accessible_projects:
-            all_items.extend(use_case.execute(proj.id))
-        items = all_items
-
-    # Generate download URLs for each video
-    responses = []
-    for item in items:
-        download_url = storage_service.get_download_url(item.file_url)
-        responses.append(_to_generated_video_response(item, download_url=download_url))
-    return ListGeneratedVideosResponse(items=responses)
-
-
-@router.get("/{video_id}", response_model=GeneratedVideoResponse)
-def get_generated_video(
-    video_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    generated_video_repository: Any = Depends(get_generated_video_repository),
-    storage_service: OSSStorageService = Depends(get_storage_service),
-) -> GeneratedVideoResponse:
-    use_case = GetGeneratedVideoUseCase(generated_video_repository)
-    item = use_case.execute(video_id)
-    if item is None:
-        raise HTTPException(status_code=404, detail=f"GeneratedVideo {video_id} not found")
-    download_url = storage_service.get_download_url(item.file_url)
-    return _to_generated_video_response(item, download_url=download_url)
-
-
-@router.patch("/{video_id}/review", response_model=GeneratedVideoResponse)
-def update_generated_video_review_status(
-    video_id: str,
-    request: UpdateGeneratedVideoReviewRequest,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    generated_video_repository: Any = Depends(get_generated_video_repository),
-    storage_service: OSSStorageService = Depends(get_storage_service),
-) -> GeneratedVideoResponse:
-    video = generated_video_repository.get(video_id)
-    if video is None:
-        raise HTTPException(status_code=404, detail=f"GeneratedVideo {video_id} not found")
-    video.review_status = request.review_status
-    updated = generated_video_repository.update(video)
-    download_url = storage_service.get_download_url(updated.file_url)
-    return _to_generated_video_response(updated, download_url=download_url)
-
-
-@router.get("/{video_id}/download-url", response_model=GeneratedVideoDownloadUrlResponse)
-def get_generated_video_download_url(
-    video_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    generated_video_repository: Any = Depends(get_generated_video_repository),
-    storage_service: OSSStorageService = Depends(get_storage_service),
-) -> GeneratedVideoDownloadUrlResponse:
-    video = generated_video_repository.get(video_id)
-    if video is None:
-        raise HTTPException(status_code=404, detail=f"GeneratedVideo {video_id} not found")
-    use_case = GetGeneratedVideoDownloadUrlUseCase(generated_video_repository)
-    file_url = use_case.execute(video_id)
-    if file_url is None:
-        raise HTTPException(status_code=404, detail=f"GeneratedVideo {video_id} not found")
-    download_url = storage_service.get_download_url(file_url)
-    return GeneratedVideoDownloadUrlResponse(video_id=video_id, download_url=download_url)

apps/api/app/api/routes/generation_tasks.py

@@ -3,6 +3,7 @@ import random
 import uuid
 from typing import Any
 
+from app.api.routes._helpers import check_project_access
 from app.auth import AuthenticatedUser, get_current_user
 from app.core.storage import OSSStorageService, get_storage_service
 from app.core.task_enqueue import (
@@ -10,7 +11,6 @@ from app.core.task_enqueue import (
     USER_PENDING_LIMIT,
     GlobalQueueFull,
     UserPendingLimitExceeded,
-    check_queue_limits,
     safe_enqueue_generation_task,
 )
 from app.dependencies import (
@@ -43,15 +43,6 @@ logger = logging.getLogger(__name__)
 
 router = APIRouter()
 
-def _check_project_access(project_id: str, user_id: str, project_repository) -> None:
-    """检查用户是否有项目访问权限"""
-    project = project_repository.find_by_id(project_id)
-    if project is None:
-        raise HTTPException(status_code=404, detail=f"Project {project_id} not found")
-    if not project.can_access(user_id):
-        raise HTTPException(status_code=403, detail="Access denied to project")
-
-
 def _to_generation_task_response(task) -> GenerationTaskResponse:
     return GenerationTaskResponse(
         id=task.id,
@@ -339,7 +330,7 @@ def get_generation_task(
     if task is None:
         raise HTTPException(status_code=404, detail=f"GenerationTask {task_id} not found")
     if task.project_id:
-        _check_project_access(task.project_id, authenticated_user.user.id, project_repository)
+        check_project_access(task.project_id, authenticated_user.user.id, project_repository)
     return _to_generation_task_response(task)
 
 
@@ -356,7 +347,7 @@ def list_generation_results(
     if task is None:
         raise HTTPException(status_code=404, detail=f"GenerationTask {task_id} not found")
     if task.project_id:
-        _check_project_access(task.project_id, authenticated_user.user.id, project_repository)
+        check_project_access(task.project_id, authenticated_user.user.id, project_repository)
     use_case = ListGeneratedVideosByTaskUseCase(generated_video_repository)
     items = use_case.execute(task_id)
     responses = []

apps/api/app/api/routes/internal_render.py

@@ -0,0 +1,120 @@
+"""渲染结果内部下载接口。
+
+通过内部 API Key 鉴权，为灰度对比工具等内部系统提供渲染结果下载能力。
+
+API:
+  GET  /api/v1/internal/render/videos/{video_id}/download-url  - 获取单个视频下载URL
+  GET  /api/v1/internal/render/tasks/{task_id}/videos          - 获取任务下所有视频及下载URL
+
+鉴权：X-API-Key header，走内部 API Key 验证
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any
+
+from app.api.routes.auth import _verify_internal_api_key
+from app.core.storage import OSSStorageService, get_storage_service
+from app.dependencies import get_generated_video_repository
+from fastapi import APIRouter, Depends, HTTPException, Query
+from pydantic import BaseModel
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter(prefix="/internal/render", tags=["Internal"])
+
+
+class InternalRenderVideoItem(BaseModel):
+    """内部渲染视频项。"""
+
+    video_id: str
+    generation_task_id: str
+    project_id: str
+    name: str
+    file_url: str
+    file_size: int | None = None
+    duration: float | None = None
+    width: int | None = None
+    height: int | None = None
+    fps: float | None = None
+    status: str
+    download_url: str
+
+
+class InternalRenderTaskVideosResponse(BaseModel):
+    """任务下所有渲染视频响应。"""
+
+    task_id: str
+    count: int
+    videos: list[InternalRenderVideoItem]
+
+
+class InternalRenderDownloadUrlResponse(BaseModel):
+    """单个视频下载URL响应。"""
+
+    video_id: str
+    download_url: str
+
+
+def _video_to_item(video: Any, download_url: str) -> InternalRenderVideoItem:
+    """将 GeneratedVideo 领域对象转为响应项。"""
+    return InternalRenderVideoItem(
+        video_id=video.id,
+        generation_task_id=video.generation_task_id,
+        project_id=video.project_id,
+        name=video.name,
+        file_url=video.file_url,
+        file_size=getattr(video, "file_size", None),
+        duration=getattr(video, "duration", None),
+        width=getattr(video, "width", None),
+        height=getattr(video, "height", None),
+        fps=getattr(video, "fps", None),
+        status=video.status,
+        download_url=download_url,
+    )
+
+
+@router.get("/videos/{video_id}/download-url", response_model=InternalRenderDownloadUrlResponse)
+def get_render_video_download_url(
+    video_id: str,
+    _: bool = Depends(_verify_internal_api_key),
+    generated_video_repository: Any = Depends(get_generated_video_repository),
+    storage_service: OSSStorageService = Depends(get_storage_service),
+) -> InternalRenderDownloadUrlResponse:
+    """获取单个渲染视频的下载URL（预签名）。"""
+    video = generated_video_repository.get(video_id)
+    if video is None:
+        raise HTTPException(status_code=404, detail=f"GeneratedVideo {video_id} not found")
+
+    download_url = storage_service.get_download_url(video.file_url, expires_seconds=86400)
+    logger.info("内部渲染下载URL生成: video_id=%s", video_id)
+    return InternalRenderDownloadUrlResponse(video_id=video_id, download_url=download_url)
+
+
+@router.get("/tasks/{task_id}/videos", response_model=InternalRenderTaskVideosResponse)
+def get_render_task_videos(
+    task_id: str,
+    status: str | None = Query(None, description="按状态筛选，如 completed/failed"),
+    _: bool = Depends(_verify_internal_api_key),
+    generated_video_repository: Any = Depends(get_generated_video_repository),
+    storage_service: OSSStorageService = Depends(get_storage_service),
+) -> InternalRenderTaskVideosResponse:
+    """获取生成任务下所有渲染视频及下载URL。"""
+    videos = generated_video_repository.list_by_generation_task(task_id)
+
+    # 状态筛选
+    if status:
+        videos = [v for v in videos if v.status == status]
+
+    items = []
+    for video in videos:
+        download_url = storage_service.get_download_url(video.file_url, expires_seconds=86400)
+        items.append(_video_to_item(video, download_url))
+
+    logger.info("内部渲染任务视频查询: task_id=%s count=%d", task_id, len(items))
+    return InternalRenderTaskVideosResponse(
+        task_id=task_id,
+        count=len(items),
+        videos=items,
+    )

apps/api/app/api/routes/jobs.py

@@ -1,332 +0,0 @@
-"""Job API 路由 — Phase 8 任务 2.10.
-
-提供统一异步任务管理 RESTful 接口：
-- POST   /api/v1/jobs                        创建任务
-- GET    /api/v1/jobs/{job_id}                任务详情
-- GET    /api/v1/projects/{project_id}/jobs   项目任务列表
-- GET    /api/v1/projects/{project_id}/jobs/stats  任务统计
-- PUT    /api/v1/jobs/{job_id}/progress       更新进度
-- POST   /api/v1/jobs/{job_id}/complete       标记完成
-- POST   /api/v1/jobs/{job_id}/fail           标记失败
-- POST   /api/v1/jobs/{job_id}/retry          重试任务
-- POST   /api/v1/jobs/{job_id}/cancel         取消任务
-- POST   /api/v1/jobs/{job_id}/submit         提交执行
-"""
-
-from __future__ import annotations
-
-import logging
-from typing import Any
-
-from app.auth import AuthenticatedUser, get_current_user
-from app.core.celery_app import celery_app
-from app.dependencies import get_db_session, get_job_repository, get_project_repository
-from app.schemas.job import (
-    CompleteJobRequest,
-    CreateJobRequest,
-    FailJobRequest,
-    JobResponse,
-    JobStatisticsResponse,
-    ListJobsResponse,
-    UpdateProgressRequest,
-    job_to_response,
-)
-from fastapi import APIRouter, Depends, HTTPException, Query, status
-
-from packages.application.jobs import (
-    CancelJobUseCase,
-    CompleteJobCommand,
-    CompleteJobUseCase,
-    CreateJobCommand,
-    CreateJobUseCase,
-    FailJobCommand,
-    FailJobUseCase,
-    GetJobStatisticsUseCase,
-    GetJobUseCase,
-    ListJobsUseCase,
-    RetryJobUseCase,
-    SubmitJobUseCase,
-    UpdateJobProgressCommand,
-    UpdateJobProgressUseCase,
-)
-from packages.domain.job import JobType
-
-logger = logging.getLogger(__name__)
-
-router = APIRouter()
-
-# 任务类型 → Celery task name 映射
-_JOB_TYPE_TO_CELERY_TASK: dict[str, str] = {
-    JobType.VIDEO_COMPOSE: "worker.compose_video",
-    JobType.RENDER_EDIT_PLAN: "worker.render_edit_plan",
-    JobType.ASSET_INGEST: "worker.ingest_asset",
-    JobType.CLASSIFICATION: "worker.classify_asset",
-    JobType.VOICE_EXTRACTION: "worker.extract_voice",
-    JobType.GENERATION: "worker.generate_video",
-}
-
-
-def _check_project_access(project_id: str, user_id: str, project_repository) -> None:
-    """检查用户是否有项目访问权限。"""
-    project = project_repository.find_by_id(project_id)
-    if project is None:
-        raise HTTPException(status_code=404, detail=f"Project {project_id} not found")
-    if not project.can_access(user_id):
-        raise HTTPException(status_code=403, detail="Access denied to project")
-
-
-# ── 创建任务 ──────────────────────────────────────────────────────────────────
-
-
-@router.post("/jobs", response_model=JobResponse, status_code=status.HTTP_201_CREATED)
-def create_job(
-    request: CreateJobRequest,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-    project_repository: Any = Depends(get_project_repository),
-) -> JobResponse:
-    """创建异步任务。
-
-    创建后任务处于 pending 状态，需要调用 /submit 提交执行。
-    """
-    _check_project_access(request.project_id, authenticated_user.user.id, project_repository)
-
-    # 校验 job_type
-    try:
-        JobType(request.job_type)
-    except ValueError:
-        raise HTTPException(
-            status_code=400,
-            detail=f"不支持的任务类型: {request.job_type}，" f"可选值: {[t.value for t in JobType]}",
-        )
-
-    use_case = CreateJobUseCase(job_repo)
-    job = use_case.execute(
-        CreateJobCommand(
-            project_id=request.project_id,
-            job_type=request.job_type,
-            payload=request.payload,
-            source_id=request.source_id,
-            created_by_user_id=authenticated_user.user.id,
-            max_retries=request.max_retries,
-        )
-    )
-
-    return job_to_response(job)
-
-
-# ── 提交执行 ──────────────────────────────────────────────────────────────────
-
-
-@router.post("/jobs/{job_id}/submit", response_model=JobResponse)
-def submit_job(
-    job_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-) -> JobResponse:
-    """提交任务执行。
-
-    将任务状态从 pending 切换为 running，并 dispatch Celery 异步任务。
-    """
-    # 权限检查：先获取任务并验证权限，再执行状态变更
-    job = job_repo.get(job_id)
-    if job is None:
-        raise HTTPException(status_code=404, detail=f"Job {job_id} not found")
-    if job.created_by_user_id and job.created_by_user_id != authenticated_user.user.id:
-        raise HTTPException(status_code=403, detail="Access denied to this job")
-
-    use_case = SubmitJobUseCase(job_repo)
-
-    try:
-        job = use_case.execute(job_id)
-    except ValueError as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-    # Dispatch Celery 任务
-    celery_task_name = _JOB_TYPE_TO_CELERY_TASK.get(job.job_type.value)
-    if celery_task_name:
-        result = celery_app.send_task(celery_task_name, args=[job.id], kwargs=job.payload)
-        job.celery_task_id = result.id
-        job_repo.update(job)
-        logger.info("已提交 Celery 任务: job_id=%s celery_task_id=%s", job.id, result.id)
-
-    return job_to_response(job)
-
-
-# ── 查询接口 ──────────────────────────────────────────────────────────────────
-
-
-@router.get("/jobs/{job_id}", response_model=JobResponse)
-def get_job(
-    job_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-) -> JobResponse:
-    """获取任务详情。"""
-    use_case = GetJobUseCase(job_repo)
-    job = use_case.execute(job_id)
-    if job is None:
-        raise HTTPException(status_code=404, detail=f"Job {job_id} not found")
-    return job_to_response(job)
-
-
-@router.get("/projects/{project_id}/jobs", response_model=ListJobsResponse)
-def list_project_jobs(
-    project_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-    project_repository: Any = Depends(get_project_repository),
-    job_type: str | None = Query(default=None, description="按任务类型过滤"),
-    status_filter: str | None = Query(default=None, alias="status", description="按状态过滤"),
-    limit: int = Query(default=50, ge=1, le=200),
-    offset: int = Query(default=0, ge=0),
-) -> ListJobsResponse:
-    """获取项目下的任务列表。"""
-    _check_project_access(project_id, authenticated_user.user.id, project_repository)
-
-    use_case = ListJobsUseCase(job_repo)
-    jobs = use_case.execute(
-        project_id=project_id,
-        job_type=job_type,
-        status=status_filter,
-        limit=limit,
-        offset=offset,
-    )
-    items = [job_to_response(j) for j in jobs]
-    return ListJobsResponse(items=items, total=len(items))
-
-
-@router.get("/projects/{project_id}/jobs/stats", response_model=JobStatisticsResponse)
-def get_job_statistics(
-    project_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-    project_repository: Any = Depends(get_project_repository),
-) -> JobStatisticsResponse:
-    """获取项目任务统计摘要。"""
-    _check_project_access(project_id, authenticated_user.user.id, project_repository)
-
-    use_case = GetJobStatisticsUseCase(job_repo)
-    stats = use_case.execute(project_id)
-    return JobStatisticsResponse(**stats)
-
-
-# ── 进度更新 ──────────────────────────────────────────────────────────────────
-
-
-@router.put("/jobs/{job_id}/progress", response_model=JobResponse)
-def update_job_progress(
-    job_id: str,
-    request: UpdateProgressRequest,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-) -> JobResponse:
-    """更新任务进度。"""
-    use_case = UpdateJobProgressUseCase(job_repo)
-
-    try:
-        job = use_case.execute(
-            UpdateJobProgressCommand(
-                job_id=job_id,
-                progress=request.progress,
-                current_stage=request.current_stage,
-            )
-        )
-    except ValueError as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-    return job_to_response(job)
-
-
-# ── 完成 / 失败 ────────────────────────────────────────────────────────────────
-
-
-@router.post("/jobs/{job_id}/complete", response_model=JobResponse)
-def complete_job(
-    job_id: str,
-    request: CompleteJobRequest,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-) -> JobResponse:
-    """标记任务完成。"""
-    use_case = CompleteJobUseCase(job_repo)
-
-    try:
-        job = use_case.execute(CompleteJobCommand(job_id=job_id, result=request.result))
-    except ValueError as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-    return job_to_response(job)
-
-
-@router.post("/jobs/{job_id}/fail", response_model=JobResponse)
-def fail_job(
-    job_id: str,
-    request: FailJobRequest,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-) -> JobResponse:
-    """标记任务失败。"""
-    use_case = FailJobUseCase(job_repo)
-
-    try:
-        job = use_case.execute(FailJobCommand(job_id=job_id, error_message=request.error_message))
-    except ValueError as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-    return job_to_response(job)
-
-
-# ── 重试 / 取消 ────────────────────────────────────────────────────────────────
-
-
-@router.post("/jobs/{job_id}/retry", response_model=JobResponse)
-def retry_job(
-    job_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-) -> JobResponse:
-    """重试失败任务。
-
-    将任务重置为 pending，retry_count + 1，但不自动 dispatch。
-    需要再次调用 /submit 提交执行。
-    """
-    # 权限检查：先获取任务并验证权限，再执行状态变更
-    job = job_repo.get(job_id)
-    if job is None:
-        raise HTTPException(status_code=404, detail=f"Job {job_id} not found")
-    if job.created_by_user_id and job.created_by_user_id != authenticated_user.user.id:
-        raise HTTPException(status_code=403, detail="Access denied to this job")
-
-    use_case = RetryJobUseCase(job_repo)
-
-    try:
-        job = use_case.execute(job_id)
-    except ValueError as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-    return job_to_response(job)
-
-
-@router.post("/jobs/{job_id}/cancel", response_model=JobResponse)
-def cancel_job(
-    job_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    job_repo: Any = Depends(get_job_repository),
-) -> JobResponse:
-    """取消任务。"""
-    # 权限检查：先获取任务并验证权限，再执行状态变更
-    job = job_repo.get(job_id)
-    if job is None:
-        raise HTTPException(status_code=404, detail=f"Job {job_id} not found")
-    if job.created_by_user_id and job.created_by_user_id != authenticated_user.user.id:
-        raise HTTPException(status_code=403, detail="Access denied to this job")
-
-    use_case = CancelJobUseCase(job_repo)
-
-    try:
-        job = use_case.execute(job_id)
-    except ValueError as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-    return job_to_response(job)

apps/api/app/api/routes/projects.py

@@ -7,7 +7,7 @@ from app.schemas.project import (
     ListProjectsResponse,
     ProjectResponse,
 )
-from fastapi import APIRouter, Depends, HTTPException, status
+from fastapi import APIRouter, Depends, HTTPException, Response, status
 
 from packages.application import (
     CreateProjectCommand,
@@ -72,12 +72,12 @@ def create_project(
     return _to_project_response(project)
 
 
-@router.delete("/{project_id}")
+@router.delete("/{project_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_project(
     project_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     project_repository: Any = Depends(get_project_repository),
-):
+) -> dict:
     use_case = DeleteProjectUseCase(project_repository)
     try:
         deleted = use_case.execute(project_id, authenticated_user.user.id)
@@ -88,4 +88,4 @@ def delete_project(
         )
     if not deleted:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Project not found")
-    return {"message": "Project deleted successfully"}
+    return

apps/api/app/api/routes/recipes.py

@@ -1,212 +0,0 @@
-"""Recipe CRUD + use routes."""
-
-from __future__ import annotations
-
-from app.auth import AuthenticatedUser, get_current_user
-from app.dependencies import get_db_session, get_user_repository
-from app.schemas.recipe import (
-    CreateRecipeRequest,
-    ListRecipesResponse,
-    RecipeItemResponse,
-    RecipeResponse,
-    UpdateRecipeRequest,
-    UseRecipeResponse,
-)
-from fastapi import APIRouter, Depends, HTTPException, Query, Response, status
-from sqlalchemy.orm import Session
-
-from packages.adapters.sqlalchemy_impl.recipe_repository import SQLAlchemyRecipeRepository
-from packages.application.recipe.commands import (
-    CreateRecipeCommand,
-    RecipeItemCommand,
-    UpdateRecipeCommand,
-)
-from packages.application.recipe.use_cases import (
-    CreateRecipeUseCase,
-    DeleteRecipeUseCase,
-    FeatureDisabledError,
-    GetRecipeUseCase,
-    ListRecipesUseCase,
-    NotFoundError,
-    UpdateRecipeUseCase,
-    UseRecipeUseCase,
-)
-from packages.ports.user_repository import UserRepository
-
-router = APIRouter()
-
-
-def _get_recipe_repository(session: Session = Depends(get_db_session)) -> SQLAlchemyRecipeRepository:
-    return SQLAlchemyRecipeRepository(session)
-
-
-def _get_user_plan(user_id: str, user_repository: UserRepository) -> str:
-    user = user_repository.find_by_id(user_id)
-    if user is None:
-        return "free"
-    return getattr(user, "subscription_plan", "free") or "free"
-
-
-def _item_to_response(item) -> RecipeItemResponse:
-    return RecipeItemResponse(
-        id=item.id,
-        recipe_id=item.recipe_id,
-        item_type=item.item_type,
-        item_id=item.item_id,
-        position=item.position,
-        metadata=item.metadata_,
-    )
-
-
-def _to_response(recipe) -> RecipeResponse:
-    return RecipeResponse(
-        id=recipe.id,
-        user_id=recipe.user_id,
-        name=recipe.name,
-        description=recipe.description,
-        template_id=recipe.template_id,
-        generation_params=recipe.generation_params,
-        items=[_item_to_response(i) for i in getattr(recipe, "items", [])],
-        is_active=recipe.is_active,
-        metadata=recipe.metadata_,
-        created_at=recipe.created_at,
-        updated_at=recipe.updated_at,
-    )
-
-
-@router.get("", response_model=ListRecipesResponse)
-def list_recipes(
-    skip: int = Query(0, ge=0),
-    limit: int = Query(50, ge=1, le=200),
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    recipe_repository: SQLAlchemyRecipeRepository = Depends(_get_recipe_repository),
-) -> ListRecipesResponse:
-    user_id = authenticated_user.user.id
-    use_case = ListRecipesUseCase(recipe_repository)
-    recipes = use_case.execute(user_id, skip=skip, limit=limit)
-    total = recipe_repository.count_by_user(user_id)
-    return ListRecipesResponse(
-        items=[_to_response(r) for r in recipes],
-        total=total,
-    )
-
-
-@router.get("/{recipe_id}", response_model=RecipeResponse)
-def get_recipe(
-    recipe_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    recipe_repository: SQLAlchemyRecipeRepository = Depends(_get_recipe_repository),
-) -> RecipeResponse:
-    user_id = authenticated_user.user.id
-    use_case = GetRecipeUseCase(recipe_repository)
-    recipe = use_case.execute(recipe_id, user_id)
-    if recipe is None:
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Recipe not found")
-    return _to_response(recipe)
-
-
-@router.post("", response_model=RecipeResponse, status_code=status.HTTP_201_CREATED)
-def create_recipe(
-    request: CreateRecipeRequest,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    recipe_repository: SQLAlchemyRecipeRepository = Depends(_get_recipe_repository),
-) -> RecipeResponse:
-    user_id = authenticated_user.user.id
-    command = CreateRecipeCommand(
-        user_id=user_id,
-        name=request.name,
-        description=request.description,
-        template_id=request.template_id,
-        generation_params=request.generation_params,
-        items=[
-            RecipeItemCommand(
-                item_type=ic.item_type,
-                item_id=ic.item_id,
-                position=ic.position,
-                metadata_=ic.metadata_,
-            )
-            for ic in request.items
-        ],
-        metadata_=request.metadata_,
-    )
-    use_case = CreateRecipeUseCase(recipe_repository)
-    recipe = use_case.execute(command)
-    return _to_response(recipe)
-
-
-@router.patch("/{recipe_id}", response_model=RecipeResponse)
-def update_recipe(
-    recipe_id: str,
-    request: UpdateRecipeRequest,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    recipe_repository: SQLAlchemyRecipeRepository = Depends(_get_recipe_repository),
-) -> RecipeResponse:
-    user_id = authenticated_user.user.id
-    command = UpdateRecipeCommand(
-        recipe_id=recipe_id,
-        user_id=user_id,
-        name=request.name,
-        description=request.description,
-        template_id=request.template_id,
-        generation_params=request.generation_params,
-        items=(
-            [
-                RecipeItemCommand(
-                    item_type=ic.item_type,
-                    item_id=ic.item_id,
-                    position=ic.position,
-                    metadata_=ic.metadata_,
-                )
-                for ic in request.items
-            ]
-            if request.items is not None
-            else None
-        ),
-        metadata_=request.metadata_,
-    )
-    use_case = UpdateRecipeUseCase(recipe_repository)
-    try:
-        recipe = use_case.execute(command)
-    except NotFoundError:
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Recipe not found")
-    return _to_response(recipe)
-
-
-@router.delete("/{recipe_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
-def delete_recipe(
-    recipe_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    recipe_repository: SQLAlchemyRecipeRepository = Depends(_get_recipe_repository),
-) -> Response:
-    user_id = authenticated_user.user.id
-    use_case = DeleteRecipeUseCase(recipe_repository)
-    deleted = use_case.execute(recipe_id, user_id)
-    if not deleted:
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Recipe not found")
-    return Response(status_code=204)
-
-
-@router.post("/{recipe_id}/use", response_model=UseRecipeResponse)
-def use_recipe(
-    recipe_id: str,
-    authenticated_user: AuthenticatedUser = Depends(get_current_user),
-    recipe_repository: SQLAlchemyRecipeRepository = Depends(_get_recipe_repository),
-    user_repository: UserRepository = Depends(get_user_repository),
-) -> UseRecipeResponse:
-    user_id = authenticated_user.user.id
-    plan_name = _get_user_plan(user_id, user_repository)
-    use_case = UseRecipeUseCase(recipe_repository)
-    try:
-        result = use_case.execute(recipe_id, user_id, user_plan=plan_name)
-    except FeatureDisabledError as exc:
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=str(exc),
-        )
-    except NotFoundError:
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Recipe not found")
-
-    return UseRecipeResponse(
-        recipe=_to_response(result.recipe),
-        warnings=[{"item_type": w.item_type, "item_id": w.item_id, "position": w.position} for w in result.warnings],
-    )

apps/api/app/api/routes/subscription.py

@@ -92,7 +92,7 @@ def _build_subscription_info(user: AuthenticatedUser) -> SubscriptionInfo:
 @router.get("/current", response_model=SubscriptionInfo)
 async def get_current_subscription(
     current_user: AuthenticatedUser = Depends(get_current_user),
-):
+) -> SubscriptionInfo:
     """获取当前订阅信息"""
     return _build_subscription_info(current_user)
 
@@ -100,7 +100,7 @@ async def get_current_subscription(
 @router.get("/billing-records", response_model=List[BillingRecord])
 async def get_billing_records(
     current_user: AuthenticatedUser = Depends(get_current_user),
-):
+) -> List[BillingRecord]:
     """获取账单记录列表"""
     from packages.adapters.sqlalchemy_impl.billing_repository import SQLAlchemyBillingRepository
     from packages.adapters.sqlalchemy_impl.session import SessionLocal
@@ -134,7 +134,7 @@ async def change_plan(
     request: ChangePlanRequest,
     current_user: AuthenticatedUser = Depends(get_current_user),
     user_repository: UserRepository = Depends(get_user_repository),
-):
+) -> ChangePlanResponse:
     """变更订阅套餐（升级/降级）"""
     # TODO: 接入支付验证（支付宝/微信支付）
     valid_plans = {"free", "standard", "pro", "enterprise"}
@@ -186,7 +186,7 @@ async def change_plan(
 async def cancel_subscription(
     current_user: AuthenticatedUser = Depends(get_current_user),
     user_repository: UserRepository = Depends(get_user_repository),
-):
+) -> SimpleResponse:
     """取消订阅"""
     user = current_user.user
     if user.subscription_plan == "free":
@@ -212,7 +212,7 @@ async def payment_callback(
     amount: float,
     payment_method: str = "alipay",
     payment_id: str = "",
-):
+) -> dict:
     """支付回调 - 在事务中更新账单和订阅状态
 
     注意：生产环境需要验证支付签名
@@ -232,7 +232,7 @@ async def payment_callback(
 
         # 创建账单记录
         record_id = uuid.uuid4().hex
-        record = repo.create(
+        repo.create(
             {
                 "id": record_id,
                 "user_id": user_id,
@@ -263,7 +263,7 @@ async def payment_callback(
 async def toggle_auto_renew(
     request: ToggleAutoRenewRequest,
     current_user: AuthenticatedUser = Depends(get_current_user),
-):
+) -> SimpleResponse:
     """切换自动续费"""
     # TODO: 实际需要在数据库中存储 auto_renew 字段
     status_text = "已开启自动续费" if request.enabled else "已关闭自动续费"

apps/api/app/api/routes/tags.py

@@ -10,7 +10,7 @@ from app.schemas.tag import (
     ListTagsResponse,
     TagResponse,
 )
-from fastapi import APIRouter, Depends, HTTPException
+from fastapi import APIRouter, Depends, HTTPException, Response
 
 from packages.domain import Tag
 
@@ -52,7 +52,7 @@ def create_tag(
     return TagResponse(id=created.id, name=created.name, created_at=created.created_at)
 
 
-@router.delete("/{tag_id}", status_code=204)
+@router.delete("/{tag_id}", status_code=204, response_class=Response)
 def delete_tag(
     tag_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),

apps/api/app/api/routes/templates.py

@@ -206,7 +206,7 @@ def update_template(
     return _to_response(template)
 
 
-@router.delete("/{template_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
+@router.delete("/{template_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_template(
     template_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -217,7 +217,7 @@ def delete_template(
     deleted = use_case.execute(template_id, user_id)
     if not deleted:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Template not found")
-    return Response(status_code=204)
+    return
 
 
 @router.post("/{template_id}/toggle-favorite", response_model=ToggleFavoriteResponse)
@@ -307,7 +307,7 @@ def create_category(
     )
 
 
-@router.delete("/categories/{category_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
+@router.delete("/categories/{category_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_category(
     category_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -318,4 +318,4 @@ def delete_category(
     deleted = use_case.execute(category_id, user_id)
     if not deleted:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Category not found")
-    return Response(status_code=204)
+    return

apps/api/app/api/routes/titles.py

@@ -4,6 +4,7 @@ from __future__ import annotations
 
 from typing import Optional
 
+from app.api.routes._helpers import get_user_plan
 from app.auth import AuthenticatedUser, get_current_user
 from app.dependencies import get_db_session, get_user_repository
 from app.schemas.title_library import (
@@ -51,13 +52,6 @@ def _to_response(item) -> TitleLibraryItemResponse:
     )
 
 
-def _get_user_plan(user_id: str, user_repository: UserRepository) -> str:
-    user = user_repository.find_by_id(user_id)
-    if user is None:
-        return "free"
-    return getattr(user, "subscription_plan", "free") or "free"
-
-
 @router.get("", response_model=ListTitleLibraryResponse)
 def list_titles(
     category: Optional[str] = Query(None),
@@ -98,7 +92,7 @@ def create_title(
     user_repository: UserRepository = Depends(get_user_repository),
 ) -> TitleLibraryItemResponse:
     user_id = authenticated_user.user.id
-    plan_name = _get_user_plan(user_id, user_repository)
+    plan_name = get_user_plan(user_id, user_repository)
     command = CreateTitleLibraryCommand(
         user_id=user_id,
         name=request.name,
@@ -143,7 +137,7 @@ def update_title(
     return _to_response(item)
 
 
-@router.delete("/{title_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
+@router.delete("/{title_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_title(
     title_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -154,4 +148,4 @@ def delete_title(
     deleted = use_case.execute(title_id, user_id)
     if not deleted:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Title not found")
-    return Response(status_code=204)
+    return

apps/api/app/api/routes/tts.py

@@ -241,7 +241,7 @@ def get_tts_job_status(
     )
 
 
-@router.delete("/jobs/{job_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
+@router.delete("/jobs/{job_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_tts_job(
     job_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -253,7 +253,7 @@ def delete_tts_job(
     deleted = use_case.execute(job_id, user_id)
     if not deleted:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="TTS job not found")
-    return Response(status_code=204)
+    return
 
 
 @router.post(
@@ -372,10 +372,10 @@ async def tts_websocket_stream(
         streaming_service = TTSStreamingService(cosyvoice_service)
         await streaming_service.synthesize_and_stream(websocket, params)
     except WebSocketDisconnect:
-        logger.info("WebSocket 客户端断开连接")
+        logger.info("WebSocket 客户端主动断开连接")
     except Exception as e:
         logger.error(f"WebSocket 流式合成异常: {e}", exc_info=True)
         try:
             await websocket.send_json({"type": "error", "message": f"服务异常: {e}"})
-        except Exception:
-            pass
+        except Exception as send_err:
+            logger.warning("WebSocket 错误消息发送失败（连接可能已断开）: %s", send_err)

apps/api/app/api/routes/upload.py

@@ -1,7 +1,8 @@
 import logging
-from typing import Annotated, Any
+from typing import Any
 from uuid import uuid4
 
+from app.api.routes._helpers import require_project_and_library
 from app.auth import AuthenticatedUser, get_current_user
 from app.config import get_settings
 from app.core.celery_app import celery_app
@@ -17,12 +18,11 @@ from app.schemas.upload import (
     DirectUploadCompleteResponse,
     DirectUploadPrepareRequest,
     DirectUploadPrepareResponse,
-    UploadAssetRequest,
     UploadAssetResponse,
 )
 from fastapi import APIRouter, Depends, File, Form, HTTPException, UploadFile, status
 
-from packages.application import GetProjectUseCase, SubmitIngestJobCommand, SubmitIngestJobUseCase
+from packages.application import SubmitIngestJobCommand, SubmitIngestJobUseCase
 
 logger = logging.getLogger(__name__)
 
@@ -80,21 +80,6 @@ def _validate_mime_type(content_type: str | None) -> str:
     return base_type
 
 
-def _require_project_and_library(
-    project_id: str,
-    library_id: str,
-    project_repository: Any,
-    asset_library_repository: Any,
-) -> None:
-    project = GetProjectUseCase(project_repository).execute(project_id)
-    if project is None:
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Project not found")
-
-    libraries = asset_library_repository.find_by_project(project_id)
-    if not any(item.id == library_id for item in libraries):
-        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Asset library not found")
-
-
 def _submit_ingest_job(
     project_id: str,
     library_id: str,
@@ -135,7 +120,7 @@ async def prepare_direct_upload(
     # P2-5: 服务端验证 MIME 类型
     validated_content_type = _validate_mime_type(request.content_type)
 
-    _require_project_and_library(
+    require_project_and_library(
         request.project_id,
         request.library_id,
         project_repository,
@@ -183,7 +168,7 @@ async def complete_direct_upload(
     storage_service: OSSStorageService = Depends(get_storage_service),
 ) -> DirectUploadCompleteResponse:
     """确认浏览器直传完成并创建导入任务。"""
-    _require_project_and_library(
+    require_project_and_library(
         request.project_id,
         request.library_id,
         project_repository,
@@ -252,7 +237,7 @@ async def upload_asset(
     storage_service: OSSStorageService = Depends(get_storage_service),
 ) -> UploadAssetResponse:
     """上传素材文件并触发导入流水线。"""
-    _require_project_and_library(project_id, library_id, project_repository, asset_library_repository)
+    require_project_and_library(project_id, library_id, project_repository, asset_library_repository)
 
     # ── 素材去重检测：上传前检查同素材库 + 同 file_hash ──
     if file_hash:

apps/api/app/api/routes/voice_clones.py

@@ -28,7 +28,6 @@ from packages.application.voice_clone.use_cases import (
     VoiceCloneNotRetryableError,
 )
 from packages.application.voice_clone.workflow import (
-    VoiceCloneWorkflowError,
     VoiceCloneWorkflowService,
 )
 
@@ -173,6 +172,7 @@ def get_voice_clone_status(
     "/{clone_id}",
     status_code=status.HTTP_204_NO_CONTENT,
     response_model=None,
+    response_class=Response,
 )
 def delete_voice_clone(
     clone_id: str,
@@ -185,7 +185,7 @@ def delete_voice_clone(
     deleted = use_case.execute(clone_id, user_id)
     if not deleted:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Voice clone not found")
-    return Response(status_code=204)
+    return
 
 
 @router.post("/{clone_id}/retry", response_model=VoiceCloneProfileResponse)

apps/api/app/api/routes/voices.py

@@ -7,6 +7,7 @@ from __future__ import annotations
 
 from typing import Literal, Optional
 
+from app.api.routes._helpers import get_user_plan
 from app.auth import AuthenticatedUser, get_current_user
 from app.dependencies import get_audio_url_signer, get_db_session, get_user_repository
 from app.schemas.voice import (
@@ -125,13 +126,6 @@ def _preset_to_unified_response(preset) -> UnifiedVoiceItemResponse:
     )
 
 
-def _get_user_plan(user_id: str, user_repository: UserRepository) -> str:
-    user = user_repository.find_by_id(user_id)
-    if user is None:
-        return "free"
-    return getattr(user, "subscription_plan", "free") or "free"
-
-
 # ==================== 统一配音列表（预置 + 克隆）====================
 
 
@@ -271,7 +265,7 @@ def create_voice(
     sign_url=Depends(get_audio_url_signer),
 ) -> VoiceLibraryItemResponse:
     user_id = authenticated_user.user.id
-    plan_name = _get_user_plan(user_id, user_repository)
+    plan_name = get_user_plan(user_id, user_repository)
     command = CreateVoiceLibraryCommand(
         user_id=user_id,
         name=request.name,
@@ -328,7 +322,7 @@ def update_voice(
     return _to_response(item, sign_url)
 
 
-@router.delete("/{voice_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None)
+@router.delete("/{voice_id}", status_code=status.HTTP_204_NO_CONTENT, response_model=None, response_class=Response)
 def delete_voice(
     voice_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
@@ -339,4 +333,4 @@ def delete_voice(
     deleted = use_case.execute(voice_id, user_id)
     if not deleted:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Voice not found")
-    return Response(status_code=204)
+    return

apps/api/app/config.py

@@ -19,27 +19,25 @@ class Settings(BaseSettings):
     # Container bind address; external expose is controlled by Docker/Nginx.
     API_HOST: str = "0.0.0.0"  # nosec: B104
     API_PORT: int = 8000
-    API_PREFIX: str = "/api/v1"
 
     DATABASE_URL: str = "postgresql+psycopg://postgres:postgres@localhost:5432/xiaoxia_saas"
     DATABASE_POOL_SIZE: int = 20
     DATABASE_MAX_OVERFLOW: int = 10  # 调整为合理值：pool_size(20) + max_overflow(10) = 最大30连接
     DATABASE_POOL_TIMEOUT: int = 30
-    DATABASE_POOL_RECYLE: int = 3600
+    DATABASE_POOL_RECYCLE: int = 3600
     USE_IN_MEMORY_DB: bool = False
     AUTO_CREATE_SCHEMA: bool = False
 
     REDIS_URL: str = "redis://localhost:6379/0"
-    REDIS_MAX_CONNECTION: int = 50
     ENABLE_REDIS_SESSIONS: bool = False
 
     # JWT secret key - MUST be set via environment variable, no default allowed
     JWT_SECRET_KEY: Optional[str] = None
-    # 旧的 JWT secret key（用于密钥轮换期间验证旧 token）
-    # 在密钥轮换时，先设置新密钥，旧密钥保留在此处直到所有旧 token 过期
-    JWT_SECRET_KEY_OLD: Optional[str] = None
-    # 密钥轮换天数（到达此天数后建议更换密钥）
-    SECRET_ROTATION_DAYS: int = 90
+
+    # JWT 算法与过期时间（与 .env.example 对齐）
+    JWT_ALGORITHM: str = "HS256"
+    JWT_ACCESS_TOKEN_EXPIRE_MINUTES: int = 30
+    JWT_REFRESH_TOKEN_EXPIRE_DAYS: int = 30
 
     @field_validator("JWT_SECRET_KEY", mode="before")
     @classmethod
@@ -75,7 +73,7 @@ class Settings(BaseSettings):
     CELERY_RESULT_BACKEND: str = "redis://localhost:6379/1"
 
     # OSS 七牛云相关
-    OSS_ENDPOINT: str = "oss-cn-hangzhou.aliiyuncs.com"
+    OSS_ENDPOINT: str = "oss-cn-hangzhou.aliyuncs.com"
     OSS_ACCESS_KEY_ID: str = ""
     OSS_ACCESS_KEY_SECRET: str = ""
     OSS_BUCKET_NAME: str = "xiaoxia-autocut"
@@ -106,9 +104,11 @@ class Settings(BaseSettings):
     )
     OSS_DIRECT_UPLOAD_EXPIRE_SECONDS: int = 900
 
-    LOG_LEVEL: str = "INFO"
     CORS_ORIGINS_RAW: str = "http://localhost:3000,http://localhost:5173,http://localhost:8000"
 
+    # 渲染引擎选择：legacy=旧VideoComposeService，unified=新UnifiedRenderService
+    RENDER_ENGINE: str = "legacy"
+
     model_config = SettingsConfigDict(
         env_file=".env",
         env_file_encoding="utf-8",

apps/api/app/core/__init__.py

@@ -1 +0,0 @@
-"""Core configuration package."""

apps/api/app/core/storage.py

@@ -1,283 +1,14 @@
-"""阿里云 OSS 存储服务"""
+"""Backward-compatible re-export from shared storage.
 
-import base64
-import datetime as dt
-import hashlib
-import hmac
-import json
-import logging
-import os
-from urllib.parse import urlparse
+All storage logic now lives in ``packages.shared.storage``.
+This module keeps old import paths working so existing code
+does not need to change.
+"""
 
-try:
-    import oss2
-except ImportError:  # pragma: no cover - exercised in minimal local/test environments
-    oss2 = None
-from app.config import get_settings
+from packages.shared.storage import SharedStorageService as OSSStorageService
+from packages.shared.storage import (
+    get_shared_storage_service,
+    get_storage_service,
+)
 
-logger = logging.getLogger(__name__)
-
-
-class OSSStorageService:
-    """阿里云 OSS 存储服务"""
-
-    def __init__(self):
-        settings = get_settings()
-        self.bucket_name = settings.OSS_BUCKET_NAME
-        self.public_url = f"https://{settings.OSS_BUCKET_NAME}.{settings.OSS_ENDPOINT}"
-        self.local_url_prefix = os.getenv("GENERATED_FILES_URL_PREFIX", "/generated-files")
-        self.bucket = None
-
-        has_key_id = bool(settings.OSS_ACCESS_KEY_ID)
-        has_key_secret = bool(settings.OSS_ACCESS_KEY_SECRET)
-
-        if has_key_id and has_key_secret:
-            if oss2 is not None:
-                try:
-                    # P0-2 修复：oss2.Bucket 的 endpoint 必须带 https:// 前缀，
-                    # 否则 sign_url 默认生成 HTTP URL。
-                    bucket_endpoint = settings.OSS_ENDPOINT
-                    if not bucket_endpoint.startswith(("http://", "https://")):
-                        bucket_endpoint = f"https://{bucket_endpoint}"
-                    auth = oss2.Auth(
-                        settings.OSS_ACCESS_KEY_ID,
-                        settings.OSS_ACCESS_KEY_SECRET,
-                    )
-                    self.bucket = oss2.Bucket(
-                        auth,
-                        bucket_endpoint,
-                        settings.OSS_BUCKET_NAME,
-                    )
-                    logger.info(
-                        "OSS initialized: endpoint=%s bucket=%s",
-                        settings.OSS_ENDPOINT,
-                        settings.OSS_BUCKET_NAME,
-                    )
-                except Exception as error:
-                    logger.error("Failed to initialize OSS bucket client: %s", error)
-            else:
-                logger.error("oss2 SDK is not installed — OSS operations will fail")
-        else:
-            missing = []
-            if not has_key_id:
-                missing.append("OSS_ACCESS_KEY_ID")
-            if not has_key_secret:
-                missing.append("OSS_ACCESS_KEY_SECRET")
-            logger.error("OSS credentials not configured — missing: %s", ", ".join(missing))
-
-        self.access_key_id = settings.OSS_ACCESS_KEY_ID
-        self.access_key_secret = settings.OSS_ACCESS_KEY_SECRET
-        self.endpoint = settings.OSS_ENDPOINT
-
-    def diagnose(self) -> None:
-        """启动诊断：输出 OSS 配置状态，帮助排查预签名 URL 问题。"""
-        key_id_display = (
-            f"{self.access_key_id[:4]}...{self.access_key_id[-4:]}" if len(self.access_key_id) > 8 else "(empty)"
-        )
-        logger.info(
-            "[OSS诊断] endpoint=%s bucket_name=%s access_key_id=%s",
-            self.endpoint,
-            self.bucket_name,
-            key_id_display,
-        )
-        if self.bucket is None:
-            logger.error(
-                "[OSS诊断] ❌ bucket=None — 预签名URL不可用！"
-                "原因: OSS_ACCESS_KEY_ID/OSS_ACCESS_KEY_SECRET 未配置或 oss2 未安装。"
-                "请检查服务器 .env 文件（如 /var/lib/xiaoxia-saas-staging/.env）"
-            )
-        else:
-            logger.info("[OSS诊断] ✅ bucket 已配置，预签名URL可用")
-
-    def _is_local_generated_url(self, storage_key_or_url: str) -> bool:
-        parsed = urlparse(storage_key_or_url)
-        path = parsed.path if parsed.scheme else storage_key_or_url
-        return path.startswith(f"{self.local_url_prefix}/")
-
-    def create_direct_upload_post(
-        self,
-        storage_key: str,
-        content_type: str,
-        max_size_bytes: int,
-        expires_seconds: int,
-    ) -> dict[str, object]:
-        """创建浏览器直传 OSS 的 PostObject 表单。"""
-        if not self.access_key_id or not self.access_key_secret:
-            raise RuntimeError("OSS storage is not configured")
-        normalized_key = self._normalize_storage_key(storage_key)
-        if not normalized_key.startswith("uploads/"):
-            raise ValueError("direct upload key must be under uploads/")
-
-        expiration = (dt.datetime.now(dt.timezone.utc) + dt.timedelta(seconds=expires_seconds)).strftime(
-            "%Y-%m-%dT%H:%M:%S.000Z"
-        )
-        policy = {
-            "expiration": expiration,
-            "conditions": [
-                {"bucket": self.bucket_name},
-                {"key": normalized_key},
-                ["content-length-range", 1, max_size_bytes],
-                ["starts-with", "$Content-Type", content_type.split("/", 1)[0] + "/" if "/" in content_type else ""],
-            ],
-        }
-        encoded_policy = base64.b64encode(json.dumps(policy, separators=(",", ":")).encode("utf-8")).decode("ascii")
-        signature = base64.b64encode(
-            hmac.new(self.access_key_secret.encode("utf-8"), encoded_policy.encode("utf-8"), hashlib.sha1).digest()
-        ).decode("ascii")
-
-        return {
-            "url": self.public_url,
-            "method": "POST",
-            "storage_key": normalized_key,
-            "expires_at": expiration,
-            "fields": {
-                "key": normalized_key,
-                "OSSAccessKeyId": self.access_key_id,
-                "policy": encoded_policy,
-                "Signature": signature,
-                "success_action_status": "201",
-                "Content-Type": content_type,
-            },
-        }
-
-    def upload_file(
-        self,
-        file_or_path,
-        storage_key: str,
-        content_type: str = "application/octet-stream",
-    ) -> str:
-        """
-        上传文件到 OSS
-
-        Args:
-            file_or_path: 文件对象或本地文件路径
-            storage_key: 存储键（文件路径）
-            content_type: 内容类型
-
-        Returns:
-            文件公网 URL
-        """
-        if self.bucket is None:
-            raise RuntimeError("OSS storage is not configured")
-
-        try:
-            # 如果是字符串路径，从本地文件上传
-            if isinstance(file_or_path, str):
-                self.bucket.put_object_from_file(storage_key, file_or_path, headers={"Content-Type": content_type})
-            else:
-                # 文件对象
-                file_or_path.seek(0)
-                self.bucket.put_object(storage_key, file_or_path, headers={"Content-Type": content_type})
-
-            return f"{self.public_url}/{storage_key}"
-        except Exception as e:
-            raise Exception(f"Failed to upload file to OSS: {e}")
-
-    def get_url(self, storage_key: str) -> str:
-        """获取文件公网 URL"""
-        return f"{self.public_url}/{storage_key}"
-
-    def get_download_url(self, storage_key_or_url: str, expires_seconds: int = 3600) -> str:
-        """
-        获取文件下载签名 URL（用于私有文件）
-
-        Args:
-            storage_key_or_url: 存储键或完整 URL
-            expires_seconds: 过期时间（秒）
-
-        Returns:
-            签名 URL
-        """
-        if self.bucket is None:
-            if self._is_local_generated_url(storage_key_or_url):
-                return storage_key_or_url
-            logger.warning(
-                "get_download_url: OSS bucket not configured, returning raw URL. " "storage_key_or_url=%s",
-                storage_key_or_url[:200],
-            )
-            return self.get_url(self._normalize_storage_key(storage_key_or_url))
-
-        storage_key = self._normalize_storage_key(storage_key_or_url)
-        try:
-            signed = self.bucket.sign_url("GET", storage_key, expires_seconds)
-            logger.info(
-                "get_download_url: signed URL generated. storage_key=%s url_prefix=%s",
-                storage_key[:80],
-                signed[:60],
-            )
-            return signed
-        except Exception:
-            logger.exception(
-                "get_download_url: sign_url failed, falling back to raw URL. " "storage_key=%s",
-                storage_key[:200],
-            )
-            return self.get_url(storage_key)
-
-    def _normalize_storage_key(self, storage_key_or_url: str) -> str:
-        """从 URL 中提取存储键"""
-        if storage_key_or_url.startswith("http://") or storage_key_or_url.startswith("https://"):
-            parsed = urlparse(storage_key_or_url)
-            # 移除开头的 /
-            return parsed.path.lstrip("/")
-        return storage_key_or_url.lstrip("/")
-
-    def download_file(self, storage_key: str, local_path: str):
-        """
-        从 OSS 下载文件到本地
-
-        Args:
-            storage_key: 存储键
-            local_path: 本地文件路径
-        """
-        if self.bucket is None:
-            raise RuntimeError("OSS storage is not configured")
-
-        try:
-            os.makedirs(os.path.dirname(local_path), exist_ok=True)
-            self.bucket.get_object_to_file(storage_key, local_path)
-        except Exception as e:
-            raise Exception(f"Failed to download file from OSS: {e}")
-
-    def delete_file(self, storage_key: str):
-        """
-        删除 OSS 文件
-
-        Args:
-            storage_key: 存储键
-        """
-        if self.bucket is None:
-            return
-
-        try:
-            self.bucket.delete_object(storage_key)
-        except Exception as error:
-            logger.warning(
-                "Failed to delete file from OSS",
-                extra={"storage_key": storage_key, "error": str(error)},
-            )
-
-    def file_exists(self, storage_key: str) -> bool:
-        """
-        检查文件是否存在
-
-        Args:
-            storage_key: 存储键
-
-        Returns:
-            是否存在
-        """
-        if self.bucket is None:
-            return False
-        return self.bucket.object_exists(storage_key)
-
-
-_storage_service = None
-
-
-def get_storage_service() -> OSSStorageService:
-    """获取存储服务实例（全局单例）"""
-    global _storage_service
-    if _storage_service is None:
-        _storage_service = OSSStorageService()
-        _storage_service.diagnose()
-    return _storage_service
+__all__ = ["OSSStorageService", "get_storage_service", "get_shared_storage_service"]

apps/api/app/dependencies.py

@@ -50,20 +50,8 @@ from packages.adapters.sqlalchemy_impl.voice_clone_profile_repository import (
 from packages.adapters.sqlalchemy_impl.voice_library_repository import (
     SQLAlchemyVoiceLibraryRepository,
 )
-from packages.ports.asset_library_repository import AssetLibraryRepository
-from packages.ports.asset_repository import AssetRepository
-from packages.ports.classification_job_repository import ClassificationJobRepository
-from packages.ports.duplication_repository import DuplicationRecordRepository
-from packages.ports.generated_video_repository import GeneratedVideoRepository
-from packages.ports.generation_task_repository import GenerationTaskRepository
-from packages.ports.ingest_job_repository import IngestJobRepository
-from packages.ports.job_repository import JobRepository
-from packages.ports.project_repository import ProjectRepository
 from packages.ports.tag_repository import TagRepository
-from packages.ports.title_library_repository import TitleLibraryRepository
 from packages.ports.user_repository import UserRepository
-from packages.ports.voice_clone_profile_repository import VoiceCloneProfileRepository
-from packages.ports.voice_library_repository import VoiceLibraryRepository
 
 _engine, _SessionLocal = build_session_factory(settings.DATABASE_URL)
 

apps/api/app/middleware/auth.py

@@ -7,12 +7,16 @@ common auth dependencies for backward compatibility.
 
 from __future__ import annotations
 
+import logging
+
 from app.auth import AuthenticatedUser
 from app.auth import get_current_user as get_authenticated_user
 from app.dependencies import get_user_repository
-from fastapi import Depends
+from fastapi import Depends, HTTPException
 from fastapi.security import HTTPAuthorizationCredentials, HTTPBearer
 
+logger = logging.getLogger(__name__)
+
 from packages.domain.entities import User
 from packages.ports.user_repository import UserRepository
 
@@ -33,6 +37,10 @@ async def get_current_user_optional(
         return None
     try:
         authenticated_user = await get_authenticated_user(credentials, user_repository)
-    except HTTPException:
+    except HTTPException as exc:
+        if exc.status_code >= 500:
+            # 服务端错误不应被静默吞掉，记录日志
+            logger.error("可选认证遇到服务端错误，status=%s", exc.status_code, exc_info=True)
+        # 4xx 认证失败（如 token 无效、用户不存在）属于正常流程，返回 None
         return None
     return authenticated_user.user

apps/api/app/middleware/monitoring.py

@@ -6,7 +6,7 @@ import logging
 import time
 from typing import Callable
 
-from fastapi import Request, Response
+from fastapi import Request
 from starlette.middleware.base import BaseHTTPMiddleware
 
 logger = logging.getLogger(__name__)

apps/api/app/schemas/dashboard.py

@@ -1,32 +0,0 @@
-from datetime import datetime
-
-from pydantic import BaseModel, Field
-
-
-class RecentTaskItem(BaseModel):
-    id: str
-    task_type: str = "generation"
-    status: str
-    current_step: str = ""
-    error_message: str = ""
-    updated_at: datetime | None = None
-
-
-class SubscriptionInfo(BaseModel):
-    """用户订阅信息。"""
-
-    plan: str = "free"
-    is_active: bool = False
-
-
-class DashboardOverviewResponse(BaseModel):
-    """Dashboard 概览数据。"""
-
-    total_assets: int = 0
-    used_storage_bytes: int = 0
-    total_titles: int = 0
-    total_voices: int = 0
-    total_tasks: int = 0
-    total_products: int = 0
-    subscription: SubscriptionInfo = Field(default_factory=SubscriptionInfo)
-    recent_tasks: list[RecentTaskItem] = Field(default_factory=list)

apps/api/app/schemas/job.py

@@ -1,109 +0,0 @@
-"""Job API schemas — Phase 8 任务 2.10."""
-
-from __future__ import annotations
-
-from datetime import datetime
-from typing import Any, Optional
-
-from pydantic import BaseModel, Field
-
-
-class CreateJobRequest(BaseModel):
-    """创建任务请求体。"""
-
-    project_id: str = Field(..., min_length=1, description="项目 ID")
-    job_type: str = Field(
-        ...,
-        description="任务类型: video_compose / render_edit_plan / asset_ingest / classification / voice_extraction / generation",
-    )
-    payload: dict[str, Any] = Field(default_factory=dict, description="任务输入参数")
-    source_id: str = Field(default="", description="关联的业务实体 ID（如 edit_plan_id）")
-    max_retries: int = Field(default=3, ge=0, le=10, description="最大重试次数")
-
-
-class UpdateProgressRequest(BaseModel):
-    """更新任务进度请求体。"""
-
-    progress: float = Field(..., ge=0.0, le=100.0, description="进度百分比")
-    current_stage: str = Field(default="", description="当前阶段描述")
-
-
-class CompleteJobRequest(BaseModel):
-    """完成任务请求体。"""
-
-    result: dict[str, Any] = Field(default_factory=dict, description="任务结果")
-
-
-class FailJobRequest(BaseModel):
-    """标记任务失败请求体。"""
-
-    error_message: str = Field(..., min_length=1, description="错误信息")
-
-
-class JobResponse(BaseModel):
-    """任务响应体。"""
-
-    id: str
-    project_id: str
-    job_type: str
-    status: str
-    progress: float
-    current_stage: str
-    payload: dict[str, Any]
-    result: dict[str, Any]
-    error_message: str
-    retry_count: int
-    max_retries: int
-    celery_task_id: str
-    source_id: str
-    created_by_user_id: str
-    is_retryable: bool
-    started_at: Optional[datetime] = None
-    completed_at: Optional[datetime] = None
-    created_at: datetime
-    updated_at: datetime
-
-    model_config = {"from_attributes": True}
-
-
-class ListJobsResponse(BaseModel):
-    """任务列表响应体。"""
-
-    items: list[JobResponse]
-    total: int
-
-
-class JobStatisticsResponse(BaseModel):
-    """任务统计响应体。"""
-
-    project_id: str
-    total: int
-    pending: int
-    running: int
-    success: int
-    failed: int
-
-
-def job_to_response(job) -> JobResponse:
-    """将 Job 领域对象转换为 API 响应。"""
-    return JobResponse(
-        id=job.id,
-        project_id=job.project_id,
-        job_type=job.job_type.value if hasattr(job.job_type, "value") else str(job.job_type),
-        status=job.status.value if hasattr(job.status, "value") else str(job.status),
-        progress=job.progress,
-        current_stage=job.current_stage,
-        payload=job.payload,
-        result=job.result,
-        error_message=job.error_message,
-        retry_count=job.retry_count,
-        max_retries=job.max_retries,
-        celery_task_id=job.celery_task_id,
-        source_id=job.source_id,
-        created_by_user_id=job.created_by_user_id,
-        is_retryable=job.is_retryable,
-        started_at=job.started_at,
-        completed_at=job.completed_at,
-        created_at=job.created_at,
-        updated_at=job.updated_at,
-    )

apps/api/app/schemas/recipe.py

@@ -1,86 +0,0 @@
-"""Recipe API schemas."""
-
-from __future__ import annotations
-
-from datetime import datetime
-from typing import Any, Dict, List, Optional
-
-from pydantic import BaseModel, Field
-
-# ── Response ──
-
-
-class RecipeItemResponse(BaseModel):
-    id: str
-    recipe_id: str
-    item_type: str
-    item_id: str
-    position: int
-    metadata_: Dict[str, Any] = Field(default_factory=dict, alias="metadata")
-
-    class Config:
-        populate_by_name = True
-
-
-class RecipeResponse(BaseModel):
-    id: str
-    user_id: str
-    name: str
-    description: str = ""
-    template_id: str = ""
-    generation_params: Dict[str, Any] = Field(default_factory=dict)
-    items: List[RecipeItemResponse] = Field(default_factory=list)
-    is_active: bool = True
-    metadata_: Dict[str, Any] = Field(default_factory=dict, alias="metadata")
-    created_at: datetime
-    updated_at: datetime
-
-    class Config:
-        populate_by_name = True
-
-
-class ListRecipesResponse(BaseModel):
-    items: List[RecipeResponse]
-    total: int = 0
-
-
-class UseRecipeResponse(BaseModel):
-    recipe: RecipeResponse
-    warnings: List[Dict[str, Any]] = Field(default_factory=list)
-
-
-# ── Request ──
-
-
-class RecipeItemRequest(BaseModel):
-    item_type: str
-    item_id: str
-    position: int = 0
-    metadata_: Dict[str, Any] = Field(default_factory=dict, alias="metadata")
-
-    class Config:
-        populate_by_name = True
-
-
-class CreateRecipeRequest(BaseModel):
-    name: str
-    description: str = ""
-    template_id: str = ""
-    generation_params: Dict[str, Any] = Field(default_factory=dict)
-    items: List[RecipeItemRequest] = Field(default_factory=list)
-    metadata_: Dict[str, Any] = Field(default_factory=dict, alias="metadata")
-
-    class Config:
-        populate_by_name = True
-
-
-class UpdateRecipeRequest(BaseModel):
-    name: Optional[str] = None
-    description: Optional[str] = None
-    template_id: Optional[str] = None
-    generation_params: Optional[Dict[str, Any]] = None
-    items: Optional[List[RecipeItemRequest]] = None
-    metadata_: Optional[Dict[str, Any]] = Field(default=None, alias="metadata")
-
-    class Config:
-        populate_by_name = True

apps/api/app/schemas/subscription.py

@@ -2,7 +2,7 @@
 
 from __future__ import annotations
 
-from typing import List, Optional
+from typing import Optional
 
 from pydantic import BaseModel, Field
 

apps/api/app/services/auto_clip_service.py

@@ -24,7 +24,7 @@ from packages.adapters.sqlalchemy_impl import (
 )
 from packages.domain.asset import AssetType
 from packages.domain.classification import AssetClassification
-from packages.domain.edit_plan_clip import EditPlanClip, EditPlanClipStatus
+from packages.domain.edit_plan_clip import EditPlanClip
 
 logger = logging.getLogger(__name__)
 

apps/api/app/services/edit_plan_service.py

@@ -18,7 +18,6 @@ from packages.adapters.sqlalchemy_impl import (
 )
 from packages.domain.edit_plan import EditPlan, EditPlanStatus
 from packages.domain.edit_plan_clip import EditPlanClip, EditPlanClipStatus
-from packages.domain.generation_task import GenerationTaskStatus
 
 logger = logging.getLogger(__name__)
 

apps/api/app/services/job_service.py

@@ -13,8 +13,6 @@ from __future__ import annotations
 import logging
 from typing import Any
 
-from sqlalchemy.orm import Session
-
 from packages.application.jobs import (
     CancelJobUseCase,
     CompleteJobCommand,

apps/api/app/services/plan_generator_service.py

@@ -13,7 +13,7 @@
 from __future__ import annotations
 
 import logging
-from typing import Any, List, Optional
+from typing import Any, List
 
 from sqlalchemy.orm import Session
 

apps/api/app/services/video_compose_service.py

@@ -16,7 +16,6 @@ FFmpeg 视频合成编排服务：
 from __future__ import annotations
 
 import logging
-import shutil
 from dataclasses import dataclass, field
 from typing import Any
 
@@ -28,7 +27,7 @@ from packages.adapters.sqlalchemy_impl.edit_plan_clip_repository import (
 from packages.adapters.sqlalchemy_impl.edit_plan_repository import (
     SQLAlchemyEditPlanRepository,
 )
-from packages.domain.edit_plan import EditPlan, EditPlanStatus
+from packages.domain.edit_plan import EditPlanStatus
 from packages.domain.edit_plan_clip import EditPlanClip, EditPlanClipStatus
 from packages.domain.template_clip_config import TransitionEffect
 

apps/web/playwright.config.ts

@@ -26,6 +26,9 @@ export default defineConfig({
           use: {
             ...devices["Desktop Chrome"],
             channel: process.env.E2E_BROWSER_CHANNEL || "msedge",
+            launchOptions: {
+              args: ["--disable-gpu", "--disable-software-rasterizer"],
+            },
           },
         },
         {
@@ -51,6 +54,9 @@ export default defineConfig({
           use: {
             ...devices["Desktop Chrome"],
             channel: process.env.E2E_BROWSER_CHANNEL || "msedge",
+            launchOptions: {
+              args: ["--disable-gpu", "--disable-software-rasterizer"],
+            },
           },
         },
       ],

apps/web/src/api/accounts.ts

@@ -1,161 +0,0 @@
-/**
- * 账号管理 Mock API
- *
- * 模拟多平台账号绑定/解绑操作
- * 支持平台：抖音、快手、小红书、微信视频号
- */
-
-/* ── 类型定义 ───────────────────────────────────────────── */
-
-/** 平台 ID */
-export type PlatformId = "douyin" | "kuaishou" | "xiaohongshu" | "wechat";
-
-/** 账号状态 */
-export type AccountStatus = "active" | "expired" | "limited";
-
-/** 已绑定的账号 */
-export interface Account {
-  id: string;
-  platform_id: PlatformId;
-  name: string;
-  avatar?: string;
-  status: AccountStatus;
-  bound_at: string;
-}
-
-/** 平台信息 */
-export interface Platform {
-  id: PlatformId;
-  name: string;
-  subName: string;
-  icon: string;
-  gradient: string;
-}
-
-/** 绑定账号请求 */
-export interface BindAccountRequest {
-  platform_id: PlatformId;
-  name: string;
-}
-
-/* ── 平台配置 ───────────────────────────────────────────── */
-
-export const PLATFORMS: Platform[] = [
-  {
-    id: "douyin",
-    name: "抖音",
-    subName: "短视频发布平台",
-    icon: "📱",
-    gradient: "linear-gradient(135deg, #fe2c55, #25f4ee)",
-  },
-  {
-    id: "kuaishou",
-    name: "快手",
-    subName: "短视频发布平台",
-    icon: "🎬",
-    gradient: "linear-gradient(135deg, #ff4906, #ffba00)",
-  },
-  {
-    id: "xiaohongshu",
-    name: "小红书",
-    subName: "种草笔记发布平台",
-    icon: "📕",
-    gradient: "linear-gradient(135deg, #ff2442, #ff6b6b)",
-  },
-  {
-    id: "wechat",
-    name: "微信视频号",
-    subName: "视频号发布平台",
-    icon: "💬",
-    gradient: "linear-gradient(135deg, #07c160, #4cd964)",
-  },
-];
-
-/* ── Mock 数据 ───────────────────────────────────────────── */
-
-let MOCK_ACCOUNTS: Account[] = [
-  {
-    id: "acc-001",
-    platform_id: "douyin",
-    name: "小虾官方号",
-    avatar: "🦐",
-    status: "active",
-    bound_at: "2025-12-01T10:00:00Z",
-  },
-  {
-    id: "acc-002",
-    platform_id: "douyin",
-    name: "小虾日常",
-    avatar: "🐟",
-    status: "active",
-    bound_at: "2025-12-15T14:30:00Z",
-  },
-  {
-    id: "acc-003",
-    platform_id: "kuaishou",
-    name: "小虾剪辑",
-    avatar: "🎬",
-    status: "active",
-    bound_at: "2026-01-05T09:00:00Z",
-  },
-  {
-    id: "acc-004",
-    platform_id: "xiaohongshu",
-    name: "小虾种草",
-    avatar: "📕",
-    status: "limited",
-    bound_at: "2026-02-20T16:00:00Z",
-  },
-];
-
-/* ── 模拟延迟 ───────────────────────────────────────────── */
-
-const delay = (ms: number) => new Promise((r) => setTimeout(r, ms));
-
-/* ── API 函数 ───────────────────────────────────────────── */
-
-/** 获取指定平台的账号列表 */
-export async function getAccountsByPlatform(
-  platformId: PlatformId,
-): Promise<Account[]> {
-  await delay(300);
-  return MOCK_ACCOUNTS.filter((a) => a.platform_id === platformId);
-}
-
-/** 获取所有平台的账号总数 */
-export async function getAllAccounts(): Promise<Account[]> {
-  await delay(200);
-  return [...MOCK_ACCOUNTS];
-}
-
-/** 绑定新账号 */
-export async function bindAccount(data: BindAccountRequest): Promise<Account> {
-  await delay(500);
-  const newAccount: Account = {
-    id: `acc-${Date.now()}`,
-    platform_id: data.platform_id,
-    name: data.name,
-    avatar: undefined,
-    status: "active",
-    bound_at: new Date().toISOString(),
-  };
-  MOCK_ACCOUNTS = [...MOCK_ACCOUNTS, newAccount];
-  return newAccount;
-}
-
-/** 解绑账号 */
-export async function unbindAccount(accountId: string): Promise<void> {
-  await delay(400);
-  MOCK_ACCOUNTS = MOCK_ACCOUNTS.filter((a) => a.id !== accountId);
-}
-
-/* ── 状态配置 ───────────────────────────────────────────── */
-
-export const ACCOUNT_STATUS_CONFIG: Record<
-  AccountStatus,
-  { label: string; className: string }
-> = {
-  active: { label: "正常", className: "acc-status--active" },
-  expired: { label: "已过期", className: "acc-status--expired" },
-  limited: { label: "受限", className: "acc-status--limited" },
-};

apps/web/src/api/dashboard.ts

@@ -1,42 +0,0 @@
-/**
- * 仪表盘 API
- * Phase 1 新增：用户仪表盘概览
- */
-import apiClient from "./client";
-
-/** 仪表盘概览数据 */
-export interface DashboardOverview {
-  /** 素材总数 */
-  total_assets: number;
-  /** 已用存储（字节） */
-  used_storage_bytes: number;
-  /** 总标题数 */
-  total_titles: number;
-  /** 总配音数 */
-  total_voices: number;
-  /** 生成任务总数 */
-  total_tasks: number;
-  /** 成品总数 */
-  total_products: number;
-  /** 最近生成任务 */
-  recent_tasks: Array<{
-    id: string;
-    task_type: string;
-    status: string;
-    progress: number;
-    user_message: string;
-    created_at: string;
-  }>;
-  /** 订阅信息 */
-  subscription: {
-    plan: "free" | "pro" | "enterprise";
-    status: "active" | "inactive" | "expired";
-    expires_at?: string;
-  };
-}
-
-/** 获取仪表盘概览数据 */
-export const getDashboardOverview = async (): Promise<DashboardOverview> => {
-  const response = await apiClient.get("/dashboard/overview");
-  return response.data;
-};

apps/web/src/components/business/business.css

@@ -1,365 +0,0 @@
-/* V21 业务组件统一样式 */
-
-/* ==================== 按钮 ==================== */
-.xx-primary-btn {
-  background: var(--gradient-primary) !important;
-  color: var(--text-inverse) !important;
-  border: none !important;
-  border-radius: var(--radius-md) !important;
-  padding: 10px 20px !important;
-  font-weight: var(--font-weight-bold) !important;
-  box-shadow: var(--shadow-primary) !important;
-  transition: var(--transition-all) !important;
-  cursor: pointer;
-  height: auto !important;
-}
-
-.xx-primary-btn:hover {
-  box-shadow: var(--shadow-hover) !important;
-  transform: translateY(-1px);
-}
-
-.xx-ghost-btn {
-  background: transparent !important;
-  color: var(--primary-color) !important;
-  border: 2px solid var(--primary-color) !important;
-  border-radius: var(--radius-md) !important;
-  padding: var(--space-sm) 18px !important;
-  font-weight: var(--font-weight-bold) !important;
-  transition: var(--transition-all) !important;
-  cursor: pointer;
-  height: auto !important;
-}
-
-.xx-ghost-btn:hover {
-  background: var(--primary-soft) !important;
-}
-
-/* ==================== 卡片 ==================== */
-.xx-card {
-  background: var(--bg-elevated);
-  border: 1px solid var(--border-color);
-  border-radius: var(--radius-xl);
-  box-shadow: var(--shadow-card);
-  padding: var(--space-lg);
-  margin-bottom: 20px;
-  transition: all var(--transition-slow);
-}
-
-.xx-card:hover {
-  box-shadow: var(--shadow-md);
-  transform: translateY(-2px);
-}
-
-/* ==================== 页面结构 ==================== */
-.xx-page {
-  max-width: 1200px;
-  margin: 0 auto;
-  padding: var(--space-lg);
-}
-
-.xx-page-head {
-  display: flex;
-  justify-content: space-between;
-  align-items: flex-start;
-  gap: 18px;
-  margin-bottom: 28px;
-}
-
-.xx-page-head h2 {
-  font-size: 26px;
-  font-weight: var(--font-weight-extrabold);
-  color: var(--text-primary);
-  margin: 0 0 var(--space-sm);
-}
-
-.xx-page-head p {
-  font-size: var(--font-size-base);
-  color: var(--text-secondary);
-  margin: 0;
-}
-
-/* ==================== 表格样式 ==================== */
-.xx-table-card {
-  background: var(--bg-elevated);
-  border: 1px solid var(--border-color);
-  border-radius: var(--radius-xl);
-  box-shadow: var(--shadow-card);
-  padding: 20px;
-  overflow: hidden;
-}
-
-/* 表格包装器 */
-.xx-table-wrapper {
-  border-radius: var(--radius-lg);
-  overflow: hidden;
-}
-
-/* ==================== 标签/Tag ==================== */
-.xx-tag {
-  padding: var(--space-xs) 12px;
-  border-radius: var(--radius-xs);
-  font-size: 13px;
-  font-weight: var(--font-weight-medium);
-}
-
-.xx-tag-indigo {
-  background: var(--primary-soft);
-  color: var(--primary-color);
-  border: 1px solid var(--color-primary-200);
-}
-
-.xx-tag-success {
-  background: var(--success-soft);
-  color: var(--color-secondary-500);
-  border: 1px solid var(--success-border);
-}
-
-.xx-tag-warning {
-  background: var(--warning-soft);
-  color: var(--accent-dark);
-  border: 1px solid var(--color-accent-200);
-}
-
-.xx-tag-error {
-  background: var(--error-soft);
-  color: var(--error-color);
-  border: 1px solid var(--error-border);
-}
-
-/* ==================== 搜索栏 ==================== */
-.xx-search-bar {
-  margin-bottom: 20px;
-}
-
-.xx-search-input {
-  width: 100%;
-  padding: 12px 18px;
-  border: 2px solid var(--border-color);
-  border-radius: var(--radius-md);
-  font-size: var(--font-size-base);
-  background: var(--bg-primary);
-  transition: var(--transition-all);
-  outline: none;
-}
-
-.xx-search-input:focus {
-  border-color: var(--primary-color);
-  box-shadow: 0 0 0 4px
-    color-mix(in srgb, var(--primary-color) 10%, transparent);
-}
-
-/* ==================== Modal ==================== */
-.xx-modal .ant-modal-content {
-  border-radius: var(--radius-xl);
-  padding: var(--space-lg);
-}
-
-.xx-modal .ant-modal-header {
-  border-radius: var(--radius-xl) var(--radius-xl) 0 0;
-  padding: 20px var(--space-lg);
-  border-bottom: 1px solid var(--border-color);
-}
-
-.xx-modal .ant-modal-title {
-  font-size: var(--font-size-lg);
-  font-weight: var(--font-weight-bold);
-  color: var(--text-primary);
-}
-
-.xx-modal .ant-modal-footer {
-  border-top: 1px solid var(--border-color);
-  padding: var(--space-md) var(--space-lg);
-}
-
-/* ==================== 空状态 ==================== */
-.xx-empty-state {
-  text-align: center;
-  padding: var(--space-3xl) var(--space-lg);
-  color: var(--text-secondary);
-}
-
-.xx-empty-state-icon {
-  font-size: 48px;
-  margin-bottom: var(--space-md);
-}
-
-/* ==================== 网格布局 ==================== */
-.xx-grid-2 {
-  display: grid;
-  grid-template-columns: repeat(2, 1fr);
-  gap: 20px;
-}
-
-.xx-grid-3 {
-  display: grid;
-  grid-template-columns: repeat(3, 1fr);
-  gap: 20px;
-}
-
-.xx-grid-4 {
-  display: grid;
-  grid-template-columns: repeat(4, 1fr);
-  gap: 20px;
-}
-
-@media (max-width: 768px) {
-  .xx-grid-2,
-  .xx-grid-3,
-  .xx-grid-4 {
-    grid-template-columns: 1fr;
-  }
-}
-
-/* ==================== 配额展示 ==================== */
-.xx-quota-item {
-  padding: 20px;
-  background: var(--bg-primary);
-  border: 1px solid var(--border-color);
-  border-radius: var(--radius-lg);
-  transition: var(--transition-all);
-}
-
-.xx-quota-item:hover {
-  border-color: var(--primary-color);
-  box-shadow: 0 8px 24px
-    color-mix(in srgb, var(--primary-color) 10%, transparent);
-}
-
-/* ==================== 进度条 ==================== */
-.xx-progress {
-  margin-top: 12px;
-}
-
-/* ==================== Ant Design 覆盖样式 ==================== */
-/* Table overrides */
-.ant-table-wrapper .ant-table-thead > tr > th {
-  background: var(--bg-secondary) !important;
-  font-weight: var(--font-weight-bold) !important;
-  color: var(--text-primary) !important;
-  border-bottom: 2px solid var(--border-color) !important;
-  padding: 14px var(--space-md) !important;
-}
-
-.ant-table-wrapper .ant-table-tbody > tr > td {
-  padding: 14px var(--space-md) !important;
-  border-bottom: 1px solid var(--color-gray-100) !important;
-}
-
-.ant-table-wrapper .ant-table-tbody > tr:hover > td {
-  background: var(--color-gray-50) !important;
-}
-
-/* Card overrides */
-.ant-card {
-  border-radius: var(--radius-xl) !important;
-  border: 1px solid var(--border-color) !important;
-}
-
-.ant-card-head {
-  border-bottom: 1px solid var(--border-color) !important;
-  min-height: 52px !important;
-  padding: 0 var(--space-lg) !important;
-}
-
-.ant-card-head-title {
-  font-weight: var(--font-weight-bold) !important;
-  font-size: var(--font-size-md) !important;
-  color: var(--text-primary) !important;
-}
-
-.ant-card-body {
-  padding: 20px var(--space-lg) !important;
-}
-
-/* Modal overrides */
-.ant-modal-content {
-  border-radius: var(--radius-xl) !important;
-  overflow: hidden;
-}
-
-.ant-modal-header {
-  padding: 20px var(--space-lg) !important;
-  background: var(--bg-primary) !important;
-}
-
-.ant-modal-title {
-  font-weight: var(--font-weight-bold) !important;
-  font-size: var(--font-size-lg) !important;
-  color: var(--text-primary) !important;
-}
-
-.ant-modal-body {
-  padding: var(--space-lg) !important;
-}
-
-.ant-modal-footer {
-  padding: var(--space-md) var(--space-lg) !important;
-}
-
-/* Button overrides */
-.ant-btn-primary {
-  background: var(--gradient-primary) !important;
-  border: none !important;
-  border-radius: var(--radius-md) !important;
-  box-shadow: var(--shadow-primary) !important;
-  height: auto !important;
-  padding: 10px 20px !important;
-  font-weight: var(--font-weight-bold) !important;
-}
-
-.ant-btn-primary:hover {
-  background: var(--gradient-primary) !important;
-  box-shadow: var(--shadow-hover) !important;
-  transform: translateY(-1px);
-}
-
-/* Tag overrides */
-.ant-tag {
-  border-radius: var(--radius-xs) !important;
-  padding: var(--space-xs) 12px !important;
-  font-weight: var(--font-weight-medium) !important;
-}
-
-/* Select overrides */
-.ant-select-selector {
-  border-radius: var(--radius-md) !important;
-  border-color: var(--border-color) !important;
-}
-
-.ant-select:not(.ant-select-disabled):hover .ant-select-selector {
-  border-color: var(--primary-color) !important;
-}
-
-.ant-select-focused .ant-select-selector {
-  border-color: var(--primary-color) !important;
-  box-shadow: 0 0 0 3px
-    color-mix(in srgb, var(--primary-color) 10%, transparent) !important;
-}
-
-/* Input overrides */
-.ant-input {
-  border-radius: var(--radius-md) !important;
-  border-color: var(--border-color) !important;
-  padding: 10px 14px !important;
-}
-
-.ant-input:hover {
-  border-color: var(--primary-color) !important;
-}
-
-.ant-input:focus {
-  border-color: var(--primary-color) !important;
-  box-shadow: 0 0 0 3px
-    color-mix(in srgb, var(--primary-color) 10%, transparent) !important;
-}
-
-/* Progress overrides */
-.ant-progress-inner {
-  background: var(--color-gray-100) !important;
-  border-radius: var(--radius-xs) !important;
-}
-
-.ant-progress-bg {
-  border-radius: var(--radius-xs) !important;
-}

apps/web/src/components/modals/CloneVoiceModal.tsx

@@ -1,290 +0,0 @@
-/**
- * CloneVoiceModal — 音色克隆弹窗
- *
- * 三步骤状态：input → uploading → success
- * 支持上传音频文件或直接录制（mock，无真实录音）
- *
- * V21 Design System — 零 antd 直接导入
- */
-import React, { useState, useCallback, useRef } from "react";
-import { Modal, Button } from "@/components/ui";
-import { createVoiceClone, toVoiceClone } from "@/api/voiceClone";
-import type { VoiceClone } from "@/api/voiceClone";
-import { uploadAsset } from "@/api/assets";
-import "./clone-voice-modal.css";
-
-/* ── 类型定义 ───────────────────────────────────────────── */
-
-type ModalStep = "input" | "uploading" | "success";
-
-export interface CloneVoiceModalProps {
-  /** 弹窗是否可见 */
-  open: boolean;
-  /** 关闭弹窗回调 */
-  onClose: () => void;
-  /** 克隆成功回调（返回新创建的音色） */
-  onSuccess?: (voice: VoiceClone) => void;
-}
-
-/* ── 默认音色名称计数器 ─────────────────────────────────── */
-
-let cloneCounter = 1;
-
-const getNextDefaultName = (): string => {
-  const name = `我的声音 ${cloneCounter}`;
-  cloneCounter += 1;
-  return name;
-};
-
-/* ── 组件 ───────────────────────────────────────────────── */
-
-const CloneVoiceModal: React.FC<CloneVoiceModalProps> = ({
-  open,
-  onClose,
-  onSuccess,
-}) => {
-  const [step, setStep] = useState<ModalStep>("input");
-  const [voiceName, setVoiceName] = useState("");
-  const [isRecording, setIsRecording] = useState(false);
-  const [selectedFile, setSelectedFile] = useState<File | null>(null);
-  const [dragActive, setDragActive] = useState(false);
-  const fileInputRef = useRef<HTMLInputElement>(null);
-
-  /** 重置弹窗状态 */
-  const resetState = useCallback(() => {
-    setStep("input");
-    setVoiceName("");
-    setSelectedFile(null);
-    setIsRecording(false);
-    setDragActive(false);
-  }, []);
-
-  /** 关闭弹窗 */
-  const handleClose = useCallback(() => {
-    resetState();
-    onClose();
-  }, [resetState, onClose]);
-
-  /** 上传区域点击 */
-  const handleUploadClick = () => {
-    fileInputRef.current?.click();
-  };
-
-  /** 文件选择 */
-  const handleFileChange = (e: React.ChangeEvent<HTMLInputElement>) => {
-    const file = e.target.files?.[0];
-    if (file) {
-      setSelectedFile(file);
-      // 清除之前的录制状态
-      setIsRecording(false);
-    }
-    // 清空 input 以允许重复选择同一文件
-    e.target.value = "";
-  };
-
-  /** 拖拽事件 */
-  const handleDrag = (e: React.DragEvent) => {
-    e.preventDefault();
-    e.stopPropagation();
-    if (e.type === "dragenter" || e.type === "dragover") {
-      setDragActive(true);
-    } else if (e.type === "dragleave") {
-      setDragActive(false);
-    }
-  };
-
-  const handleDrop = (e: React.DragEvent) => {
-    e.preventDefault();
-    e.stopPropagation();
-    setDragActive(false);
-    const file = e.dataTransfer.files?.[0];
-    if (file) {
-      const ext = file.name.split(".").pop()?.toLowerCase();
-      if (ext === "mp3" || ext === "wav") {
-        setSelectedFile(file);
-        setIsRecording(false);
-      }
-    }
-  };
-
-  /** 录制按钮（mock） */
-  const handleRecord = () => {
-    setIsRecording((prev) => !prev);
-    if (!isRecording) {
-      // 开始录制 — 清除已选文件
-      setSelectedFile(null);
-    }
-  };
-
-  /** 开始克隆 */
-  const handleStartClone = async () => {
-    const name = voiceName.trim() || getNextDefaultName();
-    setStep("uploading");
-
-    try {
-      // 先上传音频文件获取真实 URL
-      let audioUrl: string;
-      if (selectedFile) {
-        const formData = new FormData();
-        formData.append("file", selectedFile);
-        formData.append("kind", "voice");
-        const uploadResult = await uploadAsset(formData);
-        audioUrl = uploadResult.url;
-      } else {
-        // 录制功能暂未实现，提示用户上传
-        setStep("input");
-        return;
-      }
-
-      // 提交克隆请求
-      const result = await createVoiceClone({
-        name,
-        audio_url: audioUrl,
-      });
-
-      setStep("success");
-
-      // 2秒后自动关闭
-      setTimeout(() => {
-        onSuccess?.(toVoiceClone(result));
-        handleClose();
-      }, 2000);
-    } catch {
-      setStep("input");
-    }
-  };
-
-  /** 弹窗打开时初始化默认名称 */
-  const handleAfterOpenChange = (visible: boolean) => {
-    if (visible) {
-      setVoiceName(getNextDefaultName());
-    }
-  };
-
-  const canStart = selectedFile || isRecording;
-
-  return (
-    <Modal
-      open={open}
-      onCancel={handleClose}
-      title="🎤 克隆新音色"
-      width={520}
-      footer={null}
-      destroyOnClose
-      afterOpenChange={handleAfterOpenChange}
-    >
-      {/* ── 输入步骤 ──────────────────────────────────── */}
-      {step === "input" && (
-        <div className="cvm-body">
-          {/* 音色名称 */}
-          <div className="cvm-field">
-            <label className="cvm-label">音色名称</label>
-            <input
-              type="text"
-              className="cvm-input"
-              value={voiceName}
-              onChange={(e) => setVoiceName(e.target.value)}
-              placeholder="输入音色名称"
-            />
-          </div>
-
-          {/* 上传区域 */}
-          <div className="cvm-field">
-            <label className="cvm-label">上传音频</label>
-            <div
-              className={`cvm-upload-zone${dragActive ? " cvm-upload-zone--active" : ""}`}
-              onClick={handleUploadClick}
-              onDragEnter={handleDrag}
-              onDragOver={handleDrag}
-              onDragLeave={handleDrag}
-              onDrop={handleDrop}
-            >
-              <div className="cvm-upload-icon">🎵</div>
-              <p className="cvm-upload-title">
-                {selectedFile ? selectedFile.name : "拖拽音频文件到此处"}
-              </p>
-              <p className="cvm-upload-hint">支持 MP3、WAV 格式</p>
-              <input
-                ref={fileInputRef}
-                type="file"
-                accept=".mp3,.wav,audio/mpeg,audio/wav"
-                style={{ display: "none" }}
-                onChange={handleFileChange}
-              />
-            </div>
-          </div>
-
-          {/* 或分隔 */}
-          <div className="cvm-divider">
-            <div className="cvm-divider-line" />
-            <span className="cvm-divider-text">或</span>
-            <div className="cvm-divider-line" />
-          </div>
-
-          {/* 录制区域 */}
-          <div className="cvm-field">
-            <label className="cvm-label">直接录制</label>
-            <div className="cvm-record-area">
-              <p className="cvm-record-hint">
-                {isRecording
-                  ? "录制中…再次点击停止"
-                  : "点击按钮开始录制你的声音"}
-              </p>
-              <button
-                type="button"
-                className={`cvm-record-btn${isRecording ? " cvm-record-btn--recording" : ""}`}
-                onClick={handleRecord}
-              >
-                🎙️
-              </button>
-            </div>
-          </div>
-
-          {/* 提示 */}
-          <div className="cvm-tip">
-            <span className="cvm-tip-icon">💡</span>
-            <span>
-              建议上传10秒~3分钟的清晰语音，环境安静、语速均匀效果最佳
-            </span>
-          </div>
-
-          {/* 底部按钮 */}
-          <div className="cvm-footer">
-            <Button buttonType="ghost" onClick={handleClose}>
-              取消
-            </Button>
-            <Button
-              buttonType="primary"
-              disabled={!canStart}
-              onClick={handleStartClone}
-            >
-              🎤 开始克隆
-            </Button>
-          </div>
-        </div>
-      )}
-
-      {/* ── 上传中步骤 ────────────────────────────────── */}
-      {step === "uploading" && (
-        <div className="cvm-uploading">
-          <div className="cvm-uploading-spinner" />
-          <p className="cvm-uploading-text">正在克隆你的音色…</p>
-          <p className="cvm-uploading-sub">AI 正在分析你的声音特征，请稍候</p>
-        </div>
-      )}
-
-      {/* ── 成功步骤 ──────────────────────────────────── */}
-      {step === "success" && (
-        <div className="cvm-success">
-          <div className="cvm-success-icon">✅</div>
-          <h3 className="cvm-success-title">克隆已提交</h3>
-          <p className="cvm-success-desc">
-            音色正在生成中，完成后将出现在列表中
-          </p>
-        </div>
-      )}
-    </Modal>
-  );
-};
-
-export default CloneVoiceModal;

apps/web/src/components/modals/clone-voice-modal.css

@@ -1,325 +0,0 @@
-/**
- * CloneVoiceModal — V21 Design System
- *
- * 音色克隆弹窗样式
- * 三步骤状态：input → uploading → success
- */
-
-/* ── 弹窗内容区 ─────────────────────────────────────────── */
-
-.cvm-body {
-  display: flex;
-  flex-direction: column;
-  gap: 20px;
-}
-
-/* ── 表单区 ─────────────────────────────────────────────── */
-
-.cvm-field {
-  display: flex;
-  flex-direction: column;
-  gap: 6px;
-}
-
-.cvm-label {
-  font-size: 13px;
-  font-weight: 600;
-  color: var(--text-secondary, #475467);
-}
-
-.cvm-input {
-  width: 100%;
-  padding: 10px 14px;
-  border: 1px solid var(--line, #e4e7ec);
-  border-radius: var(--radius-sm);
-  background: var(--bg-surface, #fff);
-  color: var(--text-primary, #101828);
-  font-size: 14px;
-  line-height: 1.5;
-  transition:
-    border-color 0.2s,
-    box-shadow 0.2s;
-  outline: none;
-}
-
-.cvm-input:focus {
-  border-color: var(--primary, #6366f1);
-  box-shadow: 0 0 0 3px
-    color-mix(in srgb, var(--primary-color) 12%, transparent);
-}
-
-.cvm-input::placeholder {
-  color: var(--muted, #98a2b3);
-}
-
-/* ── 上传区域 ───────────────────────────────────────────── */
-
-.cvm-upload-zone {
-  border: 2px dashed var(--line, #e4e7ec);
-  border-radius: var(--radius-md);
-  padding: 28px 20px;
-  text-align: center;
-  background: var(--bg-subtle, #f8fafc);
-  cursor: pointer;
-  transition:
-    border-color 0.2s,
-    background 0.2s;
-}
-
-.cvm-upload-zone:hover {
-  border-color: var(--primary, #6366f1);
-  background: color-mix(in srgb, var(--primary-color) 4%, transparent);
-}
-
-.cvm-upload-zone.cvm-upload-zone--active {
-  border-color: var(--primary, #6366f1);
-  background: color-mix(in srgb, var(--primary-color) 6%, transparent);
-}
-
-.cvm-upload-icon {
-  font-size: 36px;
-  margin-bottom: 8px;
-  line-height: 1;
-}
-
-.cvm-upload-title {
-  font-size: 14px;
-  font-weight: 600;
-  color: var(--text-primary, #101828);
-  margin: 0 0 4px;
-}
-
-.cvm-upload-hint {
-  font-size: 13px;
-  color: var(--muted, #98a2b3);
-  margin: 0;
-}
-
-/* ── 或分隔线 ───────────────────────────────────────────── */
-
-.cvm-divider {
-  display: flex;
-  align-items: center;
-  gap: 16px;
-  margin: 4px 0;
-}
-
-.cvm-divider-line {
-  flex: 1;
-  height: 1px;
-  background: var(--line, #e4e7ec);
-}
-
-.cvm-divider-text {
-  font-size: 13px;
-  color: var(--muted, #98a2b3);
-  flex-shrink: 0;
-}
-
-/* ── 录制区域 ───────────────────────────────────────────── */
-
-.cvm-record-area {
-  border: 1px solid var(--line, #e4e7ec);
-  border-radius: var(--radius-md);
-  padding: 24px;
-  text-align: center;
-}
-
-.cvm-record-hint {
-  font-size: 13px;
-  color: var(--muted, #98a2b3);
-  margin: 0 0 14px;
-}
-
-.cvm-record-btn {
-  width: 80px;
-  height: 80px;
-  border-radius: 50%;
-  border: none;
-  cursor: pointer;
-  font-size: 32px;
-  line-height: 1;
-  padding: 0;
-  background: linear-gradient(
-    135deg,
-    var(--error-color, #ef4444),
-    var(--error-dark, #dc2626)
-  );
-  color: var(--text-inverse);
-  box-shadow: 0 4px 14px
-    color-mix(in srgb, var(--error-color, #ef4444) 35%, transparent);
-  transition:
-    transform 0.15s,
-    box-shadow 0.15s;
-  display: inline-flex;
-  align-items: center;
-  justify-content: center;
-}
-
-.cvm-record-btn:hover {
-  transform: scale(1.06);
-  box-shadow: 0 6px 20px
-    color-mix(in srgb, var(--error-color, #ef4444) 45%, transparent);
-}
-
-.cvm-record-btn:active {
-  transform: scale(0.96);
-}
-
-.cvm-record-btn--recording {
-  animation: cvm-pulse 1.2s ease-in-out infinite;
-}
-
-@keyframes cvm-pulse {
-  0%,
-  100% {
-    box-shadow: 0 4px 14px
-      color-mix(in srgb, var(--error-color, #ef4444) 35%, transparent);
-  }
-  50% {
-    box-shadow: 0 4px 28px
-      color-mix(in srgb, var(--error-color, #ef4444) 60%, transparent);
-  }
-}
-
-/* ── 提示条 ─────────────────────────────────────────────── */
-
-.cvm-tip {
-  display: flex;
-  align-items: flex-start;
-  gap: 8px;
-  padding: 12px 16px;
-  background: var(--warning-soft, #fef3c7);
-  border-radius: var(--radius-sm);
-  font-size: 13px;
-  color: var(--warning-color, #92400e);
-  line-height: 1.5;
-}
-
-.cvm-tip-icon {
-  flex-shrink: 0;
-  font-size: 14px;
-  line-height: 1.5;
-}
-
-/* ── 底部按钮 ───────────────────────────────────────────── */
-
-.cvm-footer {
-  display: flex;
-  gap: 12px;
-  margin-top: 4px;
-}
-
-.cvm-footer .xx-btn {
-  flex: 1;
-}
-
-/* ── 上传中状态 ─────────────────────────────────────────── */
-
-.cvm-uploading {
-  display: flex;
-  flex-direction: column;
-  align-items: center;
-  justify-content: center;
-  padding: 48px 20px;
-  gap: 16px;
-}
-
-.cvm-uploading-spinner {
-  width: 48px;
-  height: 48px;
-  border: 3px solid var(--line, #e4e7ec);
-  border-top-color: var(--primary, #6366f1);
-  border-radius: 50%;
-  animation: cvm-spin 0.8s linear infinite;
-}
-
-@keyframes cvm-spin {
-  to {
-    transform: rotate(360deg);
-  }
-}
-
-.cvm-uploading-text {
-  font-size: 15px;
-  font-weight: 500;
-  color: var(--text-primary, #101828);
-  margin: 0;
-}
-
-.cvm-uploading-sub {
-  font-size: 13px;
-  color: var(--muted, #98a2b3);
-  margin: 0;
-}
-
-/* ── 成功状态 ───────────────────────────────────────────── */
-
-.cvm-success {
-  display: flex;
-  flex-direction: column;
-  align-items: center;
-  justify-content: center;
-  padding: 48px 20px;
-  gap: 12px;
-}
-
-.cvm-success-icon {
-  font-size: 56px;
-  line-height: 1;
-}
-
-.cvm-success-title {
-  font-size: 18px;
-  font-weight: 700;
-  color: var(--text-primary, #101828);
-  margin: 0;
-}
-
-.cvm-success-desc {
-  font-size: 14px;
-  color: var(--muted, #98a2b3);
-  margin: 0;
-}
-
-/* ── 响应式 ─────────────────────────────────────────────── */
-
-@media (max-width: 768px) {
-  .cvm-overlay {
-    padding: var(--space-md);
-  }
-
-  .cvm-modal {
-    width: 100%;
-    max-width: 100%;
-    padding: var(--space-lg);
-  }
-}
-
-@media (max-width: 576px) {
-  .cvm-upload-zone {
-    padding: 20px 14px;
-  }
-
-  .cvm-record-btn {
-    width: 64px;
-    height: 64px;
-    font-size: 26px;
-  }
-
-  .cvm-footer {
-    flex-direction: column;
-  }
-}
-
-@media (max-width: 480px) {
-  .cvm-record-btn {
-    width: 60px;
-    height: 60px;
-  }
-
-  .cvm-tip {
-    font-size: 12px;
-    padding: var(--space-sm);
-  }
-}

apps/web/src/components/ui/Form.tsx

@@ -1,37 +0,0 @@
-/**
- * V21 Form 表单
- * 封装 Ant Design Form，应用 V21 设计系统样式
- */
-import React from "react";
-import { Form as AntForm } from "antd";
-import type { FormProps as AntFormProps } from "antd";
-import classNames from "classnames";
-import "./ui.css";
-
-export interface FormProps extends AntFormProps {
-  /** 紧凑模式（减小表单项间距） */
-  compact?: boolean;
-}
-
-const Form = ({ className, compact, children, ...rest }: FormProps) => {
-  const v21Class = classNames(
-    "xx-form",
-    compact && "xx-form-compact",
-    className,
-  );
-  return (
-    <AntForm
-      className={v21Class}
-      {...(rest as Omit<FormProps, "className" | "compact" | "children">)}
-    >
-      {children as React.ReactNode}
-    </AntForm>
-  );
-};
-
-/** 导出 Form 的子组件（保持 antd API 一致） */
-export const FormItem = AntForm.Item;
-export const FormList = AntForm.List;
-export const FormProvider = AntForm.Provider;
-
-export default Form;

apps/web/src/components/ui/Pagination.tsx

@@ -1,25 +0,0 @@
-/**
- * V21 Pagination 分页
- * 封装 Ant Design Pagination，应用 V21 设计系统样式
- */
-import React from "react";
-import { Pagination as AntPagination } from "antd";
-import type { PaginationProps as AntPaginationProps } from "antd";
-import classNames from "classnames";
-import "./ui.css";
-
-export interface PaginationProps extends AntPaginationProps {
-  /** 使用 V21 样式 */
-  v21?: boolean;
-}
-
-const Pagination: React.FC<PaginationProps> = ({
-  className,
-  v21 = true,
-  ...rest
-}) => {
-  const v21Class = classNames(v21 && "xx-pagination", className);
-  return <AntPagination className={v21Class} {...rest} />;
-};
-
-export default Pagination;

apps/web/src/components/ui/Table.tsx

@@ -1,29 +0,0 @@
-/**
- * V21 Table 表格
- * 封装 Ant Design Table，应用 V21 设计系统样式
- */
-import React from "react";
-import { Table as AntTable } from "antd";
-import type { TableProps as AntTableProps } from "antd";
-import classNames from "classnames";
-import "./ui.css";
-
-export interface TableProps<
-  RecordType = unknown,
-> extends AntTableProps<RecordType> {
-  /** 使用 V21 样式 */
-  v21?: boolean;
-}
-
-function Table<RecordType extends object = Record<string, unknown>>({
-  className,
-  v21 = true,
-  ...rest
-}: TableProps<RecordType>) {
-  const v21Class = classNames(v21 && "xx-table", className);
-  return <AntTable<RecordType> className={v21Class} {...rest} />;
-}
-
-export default Table as <RecordType extends object = Record<string, unknown>>(
-  props: TableProps<RecordType> & React.RefAttributes<HTMLDivElement>,
-) => React.ReactElement;

apps/web/src/components/ui/index.ts

@@ -15,9 +15,6 @@ export type { SelectProps } from "./Select";
 export { default as Modal } from "./Modal";
 export type { ModalProps } from "./Modal";
 
-export { default as Table } from "./Table";
-export type { TableProps } from "./Table";
-
 export { default as Card } from "./Card";
 export type { CardProps } from "./Card";
 
@@ -26,9 +23,3 @@ export type { TagProps, TagVariant } from "./Tag";
 
 export { Tooltip, Popover } from "./Tooltip";
 export type { TooltipProps, PopoverProps } from "./Tooltip";
-
-export { default as Form, FormItem, FormList, FormProvider } from "./Form";
-export type { FormProps } from "./Form";
-
-export { default as Pagination } from "./Pagination";
-export type { PaginationProps } from "./Pagination";

apps/web/src/components/ui/ui.css

@@ -258,51 +258,6 @@
   font-size: 24px !important;
 }
 
-/* ============================================================
-   Table 表格
-   ============================================================ */
-.xx-table .ant-table {
-  background: var(--bg-primary) !important;
-  color: var(--text-primary) !important;
-  border-radius: var(--radius-md) !important;
-  overflow: hidden;
-}
-
-.xx-table .ant-table-thead > tr > th {
-  background: var(--bg-secondary) !important;
-  color: var(--text-secondary) !important;
-  font-weight: var(--font-weight-semibold) !important;
-  border-bottom: 1px solid var(--border-color) !important;
-  font-size: var(--font-size-sm) !important;
-  text-transform: uppercase;
-  letter-spacing: var(--letter-spacing-wide);
-}
-
-.xx-table .ant-table-tbody > tr > td {
-  border-bottom: 1px solid var(--border-light) !important;
-  color: var(--text-primary) !important;
-  transition: var(--transition-fast) !important;
-}
-
-.xx-table .ant-table-tbody > tr:hover > td {
-  background: var(--primary-soft) !important;
-}
-
-.xx-table .ant-table-tbody > tr:last-child > td {
-  border-bottom: none !important;
-}
-
-/* 排序图标 */
-.xx-table .ant-table-column-sorter-up.active,
-.xx-table .ant-table-column-sorter-down.active {
-  color: var(--primary-color) !important;
-}
-
-/* 分页 */
-.xx-table .ant-pagination {
-  padding: var(--space-md) 0 !important;
-}
-
 /* ============================================================
    Card 卡片
    ============================================================ */
@@ -429,68 +384,6 @@
   border-bottom: 1px solid var(--border-light) !important;
 }
 
-/* ============================================================
-   Form 表单
-   ============================================================ */
-.xx-form .ant-form-item-label > label {
-  color: var(--text-primary) !important;
-  font-weight: var(--font-weight-medium) !important;
-  font-size: var(--font-size-base) !important;
-}
-
-.xx-form .ant-form-item-explain-error {
-  color: var(--error-color) !important;
-  font-size: var(--font-size-sm) !important;
-}
-
-.xx-form .ant-form-item {
-  margin-bottom: var(--space-lg) !important;
-}
-
-/* 表单项间距紧凑 */
-.xx-form-compact .ant-form-item {
-  margin-bottom: var(--space-md) !important;
-}
-
-/* ============================================================
-   Pagination 分页
-   ============================================================ */
-.xx-pagination .ant-pagination-item {
-  border-radius: var(--radius-xs) !important;
-  border-color: var(--border-color) !important;
-  transition: var(--transition-fast) !important;
-}
-
-.xx-pagination .ant-pagination-item a {
-  color: var(--text-primary) !important;
-}
-
-.xx-pagination .ant-pagination-item:hover {
-  border-color: var(--primary-color) !important;
-}
-
-.xx-pagination .ant-pagination-item:hover a {
-  color: var(--primary-color) !important;
-}
-
-.xx-pagination .ant-pagination-item-active {
-  background: var(--gradient-primary) !important;
-  border-color: transparent !important;
-}
-
-.xx-pagination .ant-pagination-item-active a {
-  color: var(--text-inverse) !important;
-}
-
-.xx-pagination .ant-pagination-prev .ant-pagination-item-link,
-.xx-pagination .ant-pagination-next .ant-pagination-item-link {
-  border-radius: var(--radius-xs) !important;
-  color: var(--text-secondary) !important;
-}
-
-.xx-pagination .ant-pagination-disabled .ant-pagination-item-link {
-  color: var(--text-disabled) !important;
-}
 
 /* ============================================================
    响应式
@@ -517,9 +410,6 @@
     padding: var(--space-md) !important;
   }
 
-  .xx-form .ant-form-item {
-    margin-bottom: var(--space-md) !important;
-  }
 }
 
 @media (max-width: 480px) {
@@ -532,3 +422,23 @@
     padding: 8px 16px !important;
   }
 }
+
+
+/* ── xx-card antd 子元素覆盖样式（从 Admin.css 迁移） ── */
+/* AdminComingSoon 等页面使用 <Card className="xx-card"> 时需要 */
+/* .xx-card 基础样式和 :hover 已在 global.css 中定义（V21 设计系统） */
+
+.xx-card .ant-card-head {
+  border-bottom: 1px solid var(--border-color);
+  padding: 20px 24px;
+}
+
+.xx-card .ant-card-head-title {
+  font-weight: 800;
+  font-size: 17px;
+  color: var(--text-primary);
+}
+
+.xx-card .ant-card-body {
+  padding: 24px;
+}

apps/web/src/config/navigation.tsx

@@ -1,211 +0,0 @@
-/**
- * 统一导航配置
- * Header 和 Sidebar 共用此数据源
- */
-import React from "react";
-import {
-  DashboardOutlined,
-  VideoCameraOutlined,
-  FileOutlined,
-  AudioOutlined,
-  FileTextOutlined,
-  TrophyOutlined,
-  AppstoreOutlined,
-  HistoryOutlined,
-  ControlOutlined,
-  CrownOutlined,
-  ScanOutlined,
-  EditOutlined,
-  FolderOutlined,
-} from "@ant-design/icons";
-
-/** 导航项定义 */
-export interface NavItem {
-  key: string;
-  label: string;
-  path: string;
-  icon: React.ReactNode;
-}
-
-/** 导航分组定义 */
-export interface NavGroup {
-  title: string;
-  items: NavItem[];
-}
-
-/**
- * 扁平导航列表（Header 使用）
- */
-export const NAV_ITEMS: NavItem[] = [
-  {
-    key: "dashboard",
-    label: "概览",
-    path: "/app/dashboard",
-    icon: <DashboardOutlined />,
-  },
-  {
-    key: "assets",
-    label: "素材库",
-    path: "/app/assets",
-    icon: <FileOutlined />,
-  },
-  {
-    key: "titles",
-    label: "标题库",
-    path: "/app/titles",
-    icon: <FileTextOutlined />,
-  },
-  {
-    key: "voices",
-    label: "配音库",
-    path: "/app/voices",
-    icon: <AudioOutlined />,
-  },
-  {
-    key: "voice-clone",
-    label: "我的音色",
-    path: "/app/voice-clone",
-    icon: <AudioOutlined />,
-  },
-  {
-    key: "voice-materials",
-    label: "配音素材库",
-    path: "/app/voice-materials",
-    icon: <AudioOutlined />,
-  },
-  {
-    key: "templates",
-    label: "模板库",
-    path: "/app/templates",
-    icon: <AppstoreOutlined />,
-  },
-  {
-    key: "editing-planner",
-    label: "剪辑编辑器",
-    path: "/app/editing-planner",
-    icon: <EditOutlined />,
-  },
-  {
-    key: "my-templates",
-    label: "我的模板",
-    path: "/app/my-templates",
-    icon: <FolderOutlined />,
-  },
-  {
-    key: "generate",
-    label: "一键生成",
-    path: "/app/generate",
-    icon: <VideoCameraOutlined />,
-  },
-  {
-    key: "history",
-    label: "任务历史",
-    path: "/app/history",
-    icon: <HistoryOutlined />,
-  },
-  {
-    key: "products",
-    label: "成品库",
-    path: "/app/products",
-    icon: <TrophyOutlined />,
-  },
-  {
-    key: "duplication",
-    label: "查重",
-    path: "/app/duplication",
-    icon: <ScanOutlined />,
-  },
-];
-
-/**
- * 分组导航列表（Sidebar 使用）
- */
-export const NAV_GROUPS: NavGroup[] = [
-  {
-    title: "创作工具",
-    items: [
-      {
-        key: "dashboard",
-        label: "首页",
-        path: "/app/dashboard",
-        icon: <DashboardOutlined />,
-      },
-      {
-        key: "generate",
-        label: "一键生成",
-        path: "/app/generate",
-        icon: <VideoCameraOutlined />,
-      },
-    ],
-  },
-  {
-    title: "资源管理",
-    items: [
-      {
-        key: "assets",
-        label: "素材库",
-        path: "/app/assets",
-        icon: <FileOutlined />,
-      },
-      {
-        key: "voices",
-        label: "配音库",
-        path: "/app/voices",
-        icon: <AudioOutlined />,
-      },
-      {
-        key: "voice-clone",
-        label: "我的音色",
-        path: "/app/voice-clone",
-        icon: <AudioOutlined />,
-      },
-      {
-        key: "voice-materials",
-        label: "配音素材库",
-        path: "/app/voice-materials",
-        icon: <AudioOutlined />,
-      },
-      {
-        key: "titles",
-        label: "标题库",
-        path: "/app/titles",
-        icon: <FileTextOutlined />,
-      },
-      {
-        key: "products",
-        label: "成片库",
-        path: "/app/products",
-        icon: <TrophyOutlined />,
-      },
-      {
-        key: "templates",
-        label: "模板库",
-        path: "/app/templates",
-        icon: <AppstoreOutlined />,
-      },
-    ],
-  },
-  {
-    title: "系统",
-    items: [
-      {
-        key: "history",
-        label: "任务历史",
-        path: "/app/history",
-        icon: <HistoryOutlined />,
-      },
-      {
-        key: "admin",
-        label: "控制台",
-        path: "/app/admin",
-        icon: <ControlOutlined />,
-      },
-      {
-        key: "subscription",
-        label: "订阅管理",
-        path: "/app/subscription",
-        icon: <CrownOutlined />,
-      },
-    ],
-  },
-];

apps/web/src/pages/accounts/Accounts.tsx

@@ -2,194 +2,62 @@
  * 账号管理页面 — V21 Design System
  *
  * 展示多平台账号绑定状态（抖音/快手/小红书/微信视频号）
- * 支持绑定/解绑操作
+ * 后端账号管理 API 尚未就绪，当前展示占位状态
  *
  * 零 antd 直接导入，全部使用 CSS 变量
  */
-import React, { useState, useCallback } from "react";
-import { useQueries, useMutation, useQueryClient } from "@tanstack/react-query";
+import React from "react";
 import { Button } from "@/components/ui";
 import PageHead from "@/components/layout/PageHead";
-import {
-  PLATFORMS,
-  getAccountsByPlatform,
-  unbindAccount,
-  bindAccount,
-  ACCOUNT_STATUS_CONFIG,
-  type Platform,
-  type Account,
-  type PlatformId,
-} from "@/api/accounts";
 import "./accounts.css";
 
-/* ── Toast 系统 ─────────────────────────────────────────── */
+/* ── 类型定义 ───────────────────────────────────────────── */
 
-interface Toast {
-  id: number;
-  message: string;
-  type: "success" | "error";
+export type PlatformId = "douyin" | "kuaishou" | "xiaohongshu" | "wechat";
+
+export interface Platform {
+  id: PlatformId;
+  name: string;
+  subName: string;
+  icon: string;
+  gradient: string;
 }
 
-let toastIdCounter = 0;
-
-/* ── 平台卡片组件 ───────────────────────────────────────── */
-
-interface PlatformCardProps {
-  platform: Platform;
-  accounts: Account[];
-  isLoading: boolean;
-  onBind: (platformId: PlatformId) => void;
-  onUnbind: (accountId: string, accountName: string) => void;
-}
-
-const PlatformCard: React.FC<PlatformCardProps> = ({
-  platform,
-  accounts,
-  isLoading,
-  onBind,
-  onUnbind,
-}) => {
-  return (
-    <div className="acc-card">
-      {/* 平台头部 */}
-      <div className="acc-card-header">
-        <div
-          className="acc-card-icon"
-          style={{ background: platform.gradient }}
-        >
-          {platform.icon}
-        </div>
-        <div>
-          <h3 className="acc-card-title">{platform.name}</h3>
-          <p className="acc-card-subtitle">{platform.subName}</p>
-        </div>
-      </div>
-
-      {/* 账号列表 */}
-      <div className="acc-account-list">
-        {isLoading ? (
-          <div className="acc-empty">
-            <p className="acc-empty-text">加载中…</p>
-          </div>
-        ) : accounts.length > 0 ? (
-          accounts.map((account) => {
-            const statusCfg = ACCOUNT_STATUS_CONFIG[account.status];
-            return (
-              <div key={account.id} className="acc-account-row">
-                <div
-                  className="acc-account-avatar"
-                  style={{ background: platform.gradient }}
-                >
-                  {account.avatar || platform.icon}
-                </div>
-                <div className="acc-account-info">
-                  <div className="acc-account-name">{account.name}</div>
-                  <span className={`acc-status-pill ${statusCfg.className}`}>
-                    {statusCfg.label}
-                  </span>
-                </div>
-                <Button
-                  buttonType="ghost"
-                  buttonSize="sm"
-                  onClick={() => onUnbind(account.id, account.name)}
-                >
-                  解绑
-                </Button>
-              </div>
-            );
-          })
-        ) : (
-          <div className="acc-empty">
-            <div className="acc-empty-icon">🔓</div>
-            <p className="acc-empty-text">暂未绑定账号</p>
-          </div>
-        )}
-      </div>
-
-      {/* 绑定按钮 */}
-      <Button buttonType="ghost" onClick={() => onBind(platform.id)}>
-        + 绑定新账号
-      </Button>
-    </div>
-  );
-};
+/** 支持的平台列表 */
+const PLATFORMS: Platform[] = [
+  {
+    id: "douyin",
+    name: "抖音",
+    subName: "短视频发布",
+    icon: "🎵",
+    gradient: "linear-gradient(135deg, #000 0%, #333 100%)",
+  },
+  {
+    id: "kuaishou",
+    name: "快手",
+    subName: "短视频发布",
+    icon: "📹",
+    gradient: "linear-gradient(135deg, #ff6600 0%, #ff9933 100%)",
+  },
+  {
+    id: "xiaohongshu",
+    name: "小红书",
+    subName: "种草笔记 + 视频",
+    icon: "📕",
+    gradient: "linear-gradient(135deg, #fe2c55 0%, #ff6680 100%)",
+  },
+  {
+    id: "wechat",
+    name: "微信视频号",
+    subName: "视频号发布",
+    icon: "💬",
+    gradient: "linear-gradient(135deg, #07c160 0%, #38d97a 100%)",
+  },
+];
 
 /* ── 主页面 ─────────────────────────────────────────────── */
 
 const Accounts: React.FC = () => {
-  const queryClient = useQueryClient();
-  const [toasts, setToasts] = useState<Toast[]>([]);
-
-  /** 显示 toast */
-  const showToast = useCallback((message: string, type: Toast["type"]) => {
-    const id = ++toastIdCounter;
-    setToasts((prev) => [...prev, { id, message, type }]);
-    setTimeout(() => {
-      setToasts((prev) => prev.filter((t) => t.id !== id));
-    }, 3000);
-  }, []);
-
-  /** 查询所有平台的账号 */
-  const _queriesResults = useQueries({
-    queries: PLATFORMS.map((platform) => ({
-      queryKey: ["accounts", platform.id] as const,
-      queryFn: () => getAccountsByPlatform(platform.id),
-    })),
-  });
-  const accountQueries = PLATFORMS.map((platform, i) => ({
-    platform,
-    ..._queriesResults[i],
-  }));
-
-  /** 解绑 mutation */
-  const unbindMutation = useMutation({
-    mutationFn: unbindAccount,
-    onSuccess: () => {
-      queryClient.invalidateQueries({ queryKey: ["accounts"] });
-      showToast("已解绑账号", "success");
-    },
-    onError: () => {
-      showToast("解绑失败", "error");
-    },
-  });
-
-  /** 绑定 mutation（mock） */
-  const bindMutation = useMutation({
-    mutationFn: bindAccount,
-    onSuccess: () => {
-      queryClient.invalidateQueries({ queryKey: ["accounts"] });
-      showToast("账号绑定成功", "success");
-    },
-    onError: () => {
-      showToast("绑定失败", "error");
-    },
-  });
-
-  /** 绑定新账号（mock：直接创建） */
-  const handleBind = (platformId: PlatformId) => {
-    const platform = PLATFORMS.find((p) => p.id === platformId);
-    if (!platform) return;
-
-    const name = window.prompt(`请输入要绑定的${platform.name}账号名称：`);
-    if (name && name.trim()) {
-      bindMutation.mutate({ platform_id: platformId, name: name.trim() });
-    }
-  };
-
-  /** 解绑账号 */
-  const handleUnbind = (accountId: string, accountName: string) => {
-    if (window.confirm(`确定解绑账号「${accountName}」吗？`)) {
-      unbindMutation.mutate(accountId);
-    }
-  };
-
-  /** 统计已绑定账号数 */
-  const totalBound = accountQueries.reduce(
-    (sum, q) => sum + (q.data?.length ?? 0),
-    0,
-  );
-  const totalPlatforms = PLATFORMS.length;
-
   return (
     <div className="acc-page">
       <PageHead
@@ -197,17 +65,34 @@ const Accounts: React.FC = () => {
         description="绑定您的社交平台账号，用于视频一键发布到各平台"
       />
 
-      {/* 平台卡片网格 */}
+      {/* 平台卡片网格 — 占位状态 */}
       <div className="acc-grid">
-        {accountQueries.map(({ platform, data, isLoading }) => (
-          <PlatformCard
-            key={platform.id}
-            platform={platform}
-            accounts={data ?? []}
-            isLoading={isLoading}
-            onBind={handleBind}
-            onUnbind={handleUnbind}
-          />
+        {PLATFORMS.map((platform) => (
+          <div key={platform.id} className="acc-card">
+            <div className="acc-card-header">
+              <div
+                className="acc-card-icon"
+                style={{ background: platform.gradient }}
+              >
+                {platform.icon}
+              </div>
+              <div>
+                <h3 className="acc-card-title">{platform.name}</h3>
+                <p className="acc-card-subtitle">{platform.subName}</p>
+              </div>
+            </div>
+
+            <div className="acc-account-list">
+              <div className="acc-empty">
+                <div className="acc-empty-icon">🔒</div>
+                <p className="acc-empty-text">功能即将上线</p>
+              </div>
+            </div>
+
+            <Button buttonType="ghost" disabled>
+              即将支持绑定
+            </Button>
+          </div>
         ))}
       </div>
 
@@ -215,22 +100,10 @@ const Accounts: React.FC = () => {
       <div className="acc-stats-bar">
         <span className="acc-stats-icon">📊</span>
         <span className="acc-stats-text">
-          已绑定 <span className="acc-stats-highlight">{totalBound}</span>{" "}
-          个账号 / 支持{" "}
-          <span className="acc-stats-highlight">{totalPlatforms}</span> 个平台
+          支持 <span className="acc-stats-highlight">{PLATFORMS.length}</span>{" "}
+          个平台，功能即将上线
         </span>
       </div>
-
-      {/* Toast 提示 */}
-      {toasts.length > 0 && (
-        <div className="vc-toast-container">
-          {toasts.map((t) => (
-            <div key={t.id} className={`vc-toast vc-toast--${t.type}`}>
-              {t.type === "success" ? "✅" : "❌"} {t.message}
-            </div>
-          ))}
-        </div>
-      )}
     </div>
   );
 };

apps/web/src/pages/accounts/accounts.css

@@ -82,77 +82,6 @@
   gap: 10px;
 }
 
-/* ── 账号行 ─────────────────────────────────────────────── */
-
-.acc-account-row {
-  display: flex;
-  align-items: center;
-  gap: 10px;
-  padding: 12px;
-  background: var(--bg-secondary);
-  border-radius: var(--radius-md);
-  transition: var(--transition-fast);
-}
-
-.acc-account-row:hover {
-  background: var(--primary-soft);
-}
-
-.acc-account-avatar {
-  width: 36px;
-  height: 36px;
-  border-radius: 50%;
-  display: grid;
-  place-items: center;
-  color: var(--text-inverse);
-  font-weight: var(--font-weight-bold);
-  font-size: var(--font-size-sm);
-  flex-shrink: 0;
-}
-
-.acc-account-info {
-  flex: 1;
-  min-width: 0;
-}
-
-.acc-account-name {
-  font-size: var(--font-size-base);
-  font-weight: var(--font-weight-semibold);
-  color: var(--text-primary);
-  white-space: nowrap;
-  overflow: hidden;
-  text-overflow: ellipsis;
-}
-
-/* ── 状态标签 ───────────────────────────────────────────── */
-
-.acc-status-pill {
-  display: inline-flex;
-  align-items: center;
-  gap: 4px;
-  padding: 2px 8px;
-  border-radius: var(--radius-full);
-  font-size: var(--font-size-xs);
-  font-weight: var(--font-weight-semibold);
-  line-height: 1.6;
-  margin-top: 2px;
-}
-
-.acc-status--active {
-  background: var(--success-soft);
-  color: var(--secondary-color);
-}
-
-.acc-status--expired {
-  background: var(--error-soft);
-  color: var(--error-color);
-}
-
-.acc-status--limited {
-  background: var(--warning-soft);
-  color: var(--accent-color);
-}
-
 /* ── 空状态 ─────────────────────────────────────────────── */
 
 .acc-empty {
@@ -235,17 +164,6 @@
     font-size: var(--font-size-base);
   }
 
-  .acc-account-row {
-    padding: 10px;
-    gap: 8px;
-  }
-
-  .acc-account-avatar {
-    width: 32px;
-    height: 32px;
-    font-size: var(--font-size-xs);
-  }
-
   .acc-stats-bar {
     padding: var(--space-sm) var(--space-md);
     font-size: var(--font-size-sm);

apps/web/src/pages/admin/Admin.css

@@ -1,482 +1,20 @@
-/* V21 Admin 页面样式 */
+/* Admin 页面样式（Phase 3 精简）
+ *
+ * 原始 477 行 → 精简至仅保留实际使用的 class。
+ * 已迁移至 global.css / ui.css 的样式不再重复定义：
+ *   .xx-page-head       → global.css
+ *   .xx-primary-btn     → global.css
+ *   .xx-tag / .xx-card  → ui.css / global.css
+ *
+ * 以下 class 仅被 AdminComingSoon.tsx 使用。
+ */
 
-/* 页面容器 */
-.dashboard-page,
-.analytics-page,
-.user-management-page,
-.log-viewer-page,
-.system-monitor-page,
 .admin-coming-soon-page {
   padding: 32px;
   max-width: 1400px;
   margin: 0 auto;
 }
 
-/* 页面头部 */
-.xx-page-head {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-xl);
-  box-shadow: 0 24px 70px rgba(15, 23, 42, 0.09);
-  padding: 32px 40px;
-  margin-bottom: 32px;
-  display: flex;
-  justify-content: space-between;
-  align-items: center;
-}
-
-.xx-page-head-content {
-  display: flex;
-  flex-direction: column;
-  gap: 8px;
-}
-
-.xx-page-head h2 {
-  font-size: 28px;
-  font-weight: 900;
-  color: var(--slate, #0f172a);
-  margin: 0;
-  letter-spacing: -0.02em;
-}
-
-.xx-page-head p {
-  font-size: 15px;
-  color: var(--muted, #64748b);
-  margin: 0;
-}
-
-.xx-page-head-actions {
-  display: flex;
-  gap: 12px;
-  align-items: center;
-}
-
-/* 简化页面头部（无操作按钮） */
-.xx-page-head-simple {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-xl);
-  box-shadow: 0 24px 70px rgba(15, 23, 42, 0.09);
-  padding: 32px 40px;
-  margin-bottom: 32px;
-}
-
-.xx-page-head-simple h2 {
-  font-size: 28px;
-  font-weight: 900;
-  color: var(--slate, #0f172a);
-  margin: 0 0 8px;
-  letter-spacing: -0.02em;
-}
-
-.xx-page-head-simple p {
-  font-size: 15px;
-  color: var(--muted, #64748b);
-  margin: 0;
-}
-
-/* V21 卡片 */
-.xx-card {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-lg);
-  box-shadow: 0 10px 30px rgba(15, 23, 42, 0.06);
-  padding: 24px;
-  transition: all 0.3s;
-}
-
-.xx-card:hover {
-  box-shadow: 0 16px 40px rgba(15, 23, 42, 0.08);
-}
-
-.xx-card .ant-card-head {
-  border-bottom: 1px solid rgba(226, 232, 240, 0.8);
-  padding: 20px 24px;
-}
-
-.xx-card .ant-card-head-title {
-  font-weight: 800;
-  font-size: 17px;
-  color: var(--slate, #0f172a);
-}
-
-.xx-card .ant-card-body {
-  padding: 24px;
-}
-
-/* 统计卡片网格 - 4列 */
-.xx-grid-4 {
-  display: grid;
-  grid-template-columns: repeat(4, 1fr);
-  gap: 24px;
-  margin-bottom: 32px;
-}
-
-@media (max-width: 1200px) {
-  .xx-grid-4 {
-    grid-template-columns: repeat(2, 1fr);
-  }
-}
-
-@media (max-width: 768px) {
-  .xx-grid-4 {
-    grid-template-columns: 1fr;
-  }
-}
-
-/* 统计卡片 */
-.xx-stat-card {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-lg);
-  box-shadow: 0 10px 30px rgba(15, 23, 42, 0.06);
-  padding: 24px;
-  transition: all 0.3s;
-}
-
-.xx-stat-card:hover {
-  transform: translateY(-2px);
-  box-shadow: 0 16px 40px rgba(15, 23, 42, 0.08);
-}
-
-.xx-stat-card-header {
-  display: flex;
-  justify-content: space-between;
-  align-items: flex-start;
-  margin-bottom: 16px;
-}
-
-.xx-stat-card-icon {
-  width: 48px;
-  height: 48px;
-  border-radius: var(--radius-md);
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  font-size: 22px;
-}
-
-.xx-stat-card-icon.primary {
-  background: linear-gradient(135deg, #6366f1, #4f46e5);
-  color: white;
-}
-
-.xx-stat-card-icon.success {
-  background: linear-gradient(135deg, #34d399, #10b981);
-  color: white;
-}
-
-.xx-stat-card-icon.warning {
-  background: linear-gradient(135deg, #fbbf24, #f59e0b);
-  color: white;
-}
-
-.xx-stat-card-icon.purple {
-  background: linear-gradient(135deg, #a78bfa, #8b5cf6);
-  color: white;
-}
-
-.xx-stat-card-icon.info {
-  background: linear-gradient(135deg, #60a5fa, #3b82f6);
-  color: white;
-}
-
-.xx-stat-card-icon.orange {
-  background: linear-gradient(135deg, #fb923c, #f97316);
-  color: white;
-}
-
-.xx-stat-card-label {
-  font-size: 14px;
-  color: var(--muted, #64748b);
-  font-weight: 500;
-  margin-bottom: 8px;
-}
-
-.xx-stat-card-value {
-  font-size: 32px;
-  font-weight: 900;
-  color: var(--slate, #0f172a);
-  line-height: 1.2;
-  letter-spacing: -0.02em;
-}
-
-.xx-stat-card-value.primary {
-  color: var(--indigo, #4f46e5);
-}
-
-.xx-stat-card-value.success {
-  color: var(--green, #10b981);
-}
-
-.xx-stat-card-value.warning {
-  color: var(--amber, #f59e0b);
-}
-
-.xx-stat-card-value.purple {
-  color: #8b5cf6;
-}
-
-.xx-stat-card-growth {
-  font-size: 13px;
-  color: var(--green, #10b981);
-  font-weight: 600;
-  margin-top: 8px;
-  display: flex;
-  align-items: center;
-  gap: 4px;
-}
-
-/* 数据表格 */
-.xx-table-wrapper {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-lg);
-  box-shadow: 0 10px 30px rgba(15, 23, 42, 0.06);
-  overflow: hidden;
-}
-
-.xx-table-wrapper .ant-table {
-  background: transparent;
-}
-
-.xx-table-wrapper .ant-table-thead > tr > th {
-  background: rgba(248, 250, 252, 0.8);
-  font-weight: 800;
-  font-size: 13px;
-  color: var(--slate, #0f172a);
-  text-transform: uppercase;
-  letter-spacing: 0.05em;
-  padding: 16px 20px;
-}
-
-.xx-table-wrapper .ant-table-tbody > tr > td {
-  padding: 16px 20px;
-  font-size: 14px;
-}
-
-.xx-table-wrapper .ant-table-tbody > tr:hover > td {
-  background: rgba(79, 70, 229, 0.03);
-}
-
-/* V21 按钮 */
-.xx-primary-btn {
-  background: linear-gradient(135deg, #6366f1, #4f46e5) !important;
-  color: white !important;
-  border: none !important;
-  border-radius: var(--radius-md) !important;
-  font-weight: 700 !important;
-  box-shadow: 0 8px 20px rgba(79, 70, 229, 0.25) !important;
-  transition: all 0.2s !important;
-}
-
-.xx-primary-btn:hover {
-  box-shadow: 0 12px 28px rgba(79, 70, 229, 0.3) !important;
-  transform: translateY(-1px);
-}
-
-.xx-ghost-btn {
-  background: white !important;
-  border: 1px solid rgba(226, 232, 240, 0.95) !important;
-  color: var(--slate, #0f172a) !important;
-  border-radius: var(--radius-md) !important;
-  font-weight: 600 !important;
-  transition: all 0.2s !important;
-}
-
-.xx-ghost-btn:hover {
-  border-color: var(--indigo, #4f46e5) !important;
-  color: var(--indigo, #4f46e5) !important;
-}
-
-/* V21 输入框 */
-.xx-search-input {
-  border-radius: var(--radius-md) !important;
-  border: 1px solid rgba(226, 232, 240, 0.95) !important;
-  padding: 8px 16px !important;
-}
-
-.xx-search-input:hover,
-.xx-search-input:focus {
-  border-color: var(--indigo, #4f46e5) !important;
-  box-shadow: 0 0 0 3px rgba(79, 70, 229, 0.1) !important;
-}
-
-/* V21 Select */
-.xx-select {
-  border-radius: var(--radius-md) !important;
-}
-
-.xx-select:hover,
-.xx-select:focus {
-  border-color: var(--indigo, #4f46e5) !important;
-  box-shadow: 0 0 0 3px rgba(79, 70, 229, 0.1) !important;
-}
-
-/* V21 Tag */
-.xx-tag {
-  border-radius: var(--radius-xs) !important;
-  font-weight: 600 !important;
-  font-size: 12px !important;
-  padding: 4px 10px !important;
-}
-
-.xx-tag.info {
-  background: rgba(59, 130, 246, 0.1) !important;
-  color: #3b82f6 !important;
-  border: 1px solid rgba(59, 130, 246, 0.2) !important;
-}
-
-.xx-tag.success {
-  background: rgba(16, 185, 129, 0.1) !important;
-  color: #10b981 !important;
-  border: 1px solid rgba(16, 185, 129, 0.2) !important;
-}
-
-.xx-tag.warning {
-  background: rgba(245, 158, 11, 0.1) !important;
-  color: #f59e0b !important;
-  border: 1px solid rgba(245, 158, 11, 0.2) !important;
-}
-
-.xx-tag.error {
-  background: rgba(239, 68, 68, 0.1) !important;
-  color: #ef4444 !important;
-  border: 1px solid rgba(239, 68, 68, 0.2) !important;
-}
-
-.xx-tag.debug {
-  background: rgba(100, 116, 139, 0.1) !important;
-  color: #64748b !important;
-  border: 1px solid rgba(100, 116, 139, 0.2) !important;
-}
-
-/* V21 Progress */
-.xx-progress-primary .ant-progress-circle .ant-progress-text {
-  color: var(--indigo, #4f46e5) !important;
-  font-weight: 700 !important;
-}
-
-.xx-progress-success .ant-progress-circle .ant-progress-text {
-  color: var(--green, #10b981) !important;
-  font-weight: 700 !important;
-}
-
-.xx-progress-warning .ant-progress-circle .ant-progress-text {
-  color: var(--amber, #f59e0b) !important;
-  font-weight: 700 !important;
-}
-
-/* 图表容器 */
-.xx-chart-container {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-lg);
-  box-shadow: 0 10px 30px rgba(15, 23, 42, 0.06);
-  padding: 24px;
-}
-
-.xx-chart-container .ant-card-head {
-  border-bottom: 1px solid rgba(226, 232, 240, 0.8);
-  padding: 20px 24px;
-}
-
-.xx-chart-container .ant-card-head-title {
-  font-weight: 800;
-  font-size: 17px;
-  color: var(--slate, #0f172a);
-}
-
-/* 筛选器区域 */
-.xx-filter-bar {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-lg);
-  box-shadow: 0 10px 30px rgba(15, 23, 42, 0.06);
-  padding: 20px 24px;
-  margin-bottom: 24px;
-  display: flex;
-  flex-wrap: wrap;
-  gap: 16px;
-  align-items: center;
-}
-
-/* 资源监控卡片 */
-.xx-resource-card {
-  background: rgba(255, 255, 255, 0.94);
-  border: 1px solid rgba(226, 232, 240, 0.95);
-  border-radius: var(--radius-lg);
-  box-shadow: 0 10px 30px rgba(15, 23, 42, 0.06);
-  padding: 24px;
-  text-align: center;
-}
-
-.xx-resource-card-icon {
-  width: 64px;
-  height: 64px;
-  border-radius: var(--radius-md);
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  margin: 0 auto 16px;
-  font-size: 28px;
-}
-
-.xx-resource-card-label {
-  font-size: 14px;
-  color: var(--muted, #64748b);
-  font-weight: 600;
-  margin-bottom: 16px;
-}
-
-/* 日期选择器 */
-.xx-date-picker {
-  border-radius: var(--radius-md) !important;
-}
-
-/* Drawer */
-.xx-drawer .ant-drawer-header {
-  border-bottom: 1px solid rgba(226, 232, 240, 0.8);
-  padding: 20px 24px;
-}
-
-.xx-drawer .ant-drawer-title {
-  font-weight: 800;
-  font-size: 18px;
-  color: var(--slate, #0f172a);
-}
-
-/* 日志详情 */
-.xx-log-detail {
-  padding: 4px 0;
-}
-
-.xx-log-detail-label {
-  font-size: 13px;
-  font-weight: 700;
-  color: var(--slate, #0f172a);
-  margin-bottom: 4px;
-}
-
-.xx-log-detail-value {
-  font-size: 14px;
-  color: var(--muted, #64748b);
-  background: rgba(248, 250, 252, 0.8);
-  padding: 12px;
-  border-radius: var(--radius-sm);
-  margin-bottom: 16px;
-}
-
-.xx-log-detail-code {
-  font-family: "JetBrains Mono", "Fira Code", monospace;
-  background: rgba(248, 250, 252, 0.8);
-  padding: 12px;
-  border-radius: var(--radius-sm);
-  white-space: pre-wrap;
-  word-break: break-all;
-}
-
-/* Result 页面居中 */
 .xx-result-center {
   display: flex;
   justify-content: center;
@@ -487,31 +25,3 @@
 .xx-result-center .ant-result {
   padding: 48px;
 }
-
-/* 刷新时间显示 */
-.xx-refresh-time {
-  font-size: 13px;
-  color: var(--muted, #64748b);
-  margin-right: 12px;
-}
-
-/* 图表配色覆盖 */
-.recharts-text {
-  fill: #64748b !important;
-  font-size: 12px !important;
-}
-
-.recharts-cartesian-grid-horizontal line,
-.recharts-cartesian-grid-vertical line {
-  stroke: rgba(226, 232, 240, 0.8) !important;
-}
-
-/* 图表 Legend */
-.recharts-legend-wrapper {
-  padding-top: 16px !important;
-}
-
-.recharts-legend-item-text {
-  color: #64748b !important;
-  font-size: 13px !important;
-}

apps/web/src/pages/dashboard/Dashboard.tsx

@@ -1,451 +1,97 @@
 /**
  * 控制台页面 — V21 设计系统
  * KPI 卡片网格 + 快速入口 + 最近任务卡片列表 + 使用统计图表 + 公告
- * 使用 mock 数据，CSS 变量，V21 组件
+ * CSS 变量，V21 组件
  */
 import React from "react";
 import { useNavigate } from "react-router-dom";
-import { Button, Tag } from "@/components/ui";
-import {
-  VideoCameraOutlined,
-  AppstoreOutlined,
-  ThunderboltOutlined,
-  DatabaseOutlined,
-  FileTextOutlined,
-} from "@ant-design/icons";
+import { Button } from "@/components/ui";
+import { DatabaseOutlined } from "@ant-design/icons";
 import "./dashboard.css";
 
-/* ============================================================
- * Mock 数据
- * ============================================================ */
-interface KpiItem {
-  key: string;
-  icon: string;
-  iconGradient: string;
-  value: string;
-  label: string;
-  trend: string;
-  trendDirection: "up" | "down" | "neutral";
-  accent: string;
-}
+/* ── 主组件 ─────────────────────────────────────────────── */
 
-const kpiData: KpiItem[] = [
-  {
-    key: "projects",
-    icon: "video",
-    iconGradient: "linear-gradient(135deg, #6366f1, #4f46e5)",
-    value: "12",
-    label: "项目总数",
-    trend: "↑ 2 本月新增",
-    trendDirection: "up",
-    accent: "#6366f1",
-  },
-  {
-    key: "assets",
-    icon: "appstore",
-    iconGradient: "linear-gradient(135deg, #0ea5e9, #0284c7)",
-    value: "486",
-    label: "素材总数",
-    trend: "↑ 38 本月上传",
-    trendDirection: "up",
-    accent: "#0ea5e9",
-  },
-  {
-    key: "generations",
-    icon: "thunderbolt",
-    iconGradient: "linear-gradient(135deg, #10b981, #059669)",
-    value: "156",
-    label: "本月生成数",
-    trend: "↑ 23% 较上月",
-    trendDirection: "up",
-    accent: "#10b981",
-  },
-  {
-    key: "storage",
-    icon: "database",
-    iconGradient: "linear-gradient(135deg, #f59e0b, #d97706)",
-    value: "2.4GB",
-    label: "存储空间",
-    trend: "已用 24%",
-    trendDirection: "neutral",
-    accent: "#f59e0b",
-  },
-];
-
-interface QuickEntry {
-  id: string;
-  icon: string;
-  iconGradient: string;
-  title: string;
-  description: string;
-  path: string;
-}
-
-const quickEntries: QuickEntry[] = [
-  {
-    id: "titles",
-    icon: "filetext",
-    iconGradient: "linear-gradient(135deg, #6366f1, #4f46e5)",
-    title: "标题库",
-    description: "24条标题 · 5个分类",
-    path: "/app/titles",
-  },
-  {
-    id: "assets",
-    icon: "appstore",
-    iconGradient: "linear-gradient(135deg, #0ea5e9, #0284c7)",
-    title: "素材库",
-    description: "486个素材 · 3个素材库",
-    path: "/app/assets",
-  },
-  {
-    id: "generate",
-    icon: "thunderbolt",
-    iconGradient: "linear-gradient(135deg, #10b981, #059669)",
-    title: "一键生成",
-    description: "开始创作新视频",
-    path: "/app/generate",
-  },
-  {
-    id: "products",
-    icon: "video",
-    iconGradient: "linear-gradient(135deg, #f59e0b, #d97706)",
-    title: "成片库",
-    description: "89个成片 · 3个待复核",
-    path: "/app/products",
-  },
-];
-
-type TaskStatus = "completed" | "processing" | "pending" | "failed";
-
-interface RecentTask {
-  id: string;
-  name: string;
-  type: string;
-  template: string;
-  status: TaskStatus;
-  date: string;
-  duration?: string;
-}
-
-const statusLabel: Record<TaskStatus, string> = {
-  completed: "已完成",
-  processing: "进行中",
-  pending: "排队中",
-  failed: "失败",
-};
-
-const recentTasks: RecentTask[] = [
-  {
-    id: "t-1",
-    name: "产品介绍视频_春季促销",
-    type: "视频生成",
-    template: "商品展示模板",
-    status: "completed",
-    date: "2026-07-01 09:30",
-    duration: "2分18秒",
-  },
-  {
-    id: "t-2",
-    name: "品牌宣传片_终版",
-    type: "视频生成",
-    template: "品牌宣传模板",
-    status: "processing",
-    date: "2026-07-01 10:15",
-  },
-  {
-    id: "t-3",
-    name: "用户评价合集",
-    type: "视频生成",
-    template: "评价展示模板",
-    status: "completed",
-    date: "2026-06-30 16:42",
-    duration: "1分45秒",
-  },
-  {
-    id: "t-4",
-    name: "新品发布预告",
-    type: "视频生成",
-    template: "新品预告模板",
-    status: "pending",
-    date: "2026-06-30 14:20",
-  },
-  {
-    id: "t-5",
-    name: "活动回顾_618大促",
-    type: "视频生成",
-    template: "活动回顾模板",
-    status: "failed",
-    date: "2026-06-29 11:05",
-  },
-];
-
-interface ChartItem {
-  label: string;
-  value: number;
-}
-
-const weeklyData: ChartItem[] = [
-  { label: "周一", value: 18 },
-  { label: "周二", value: 25 },
-  { label: "周三", value: 32 },
-  { label: "周四", value: 28 },
-  { label: "周五", value: 42 },
-  { label: "周六", value: 15 },
-  { label: "周日", value: 8 },
-];
-
-interface Announcement {
-  id: string;
-  tag: "update" | "notice" | "activity";
-  tagLabel: string;
-  title: string;
-  date: string;
-}
-
-const announcements: Announcement[] = [
-  {
-    id: "a-1",
-    tag: "update",
-    tagLabel: "更新",
-    title: "系统已升级至 v2.0，新增批量生成功能",
-    date: "2026-07-01",
-  },
-  {
-    id: "a-2",
-    tag: "activity",
-    tagLabel: "活动",
-    title: "7月创作挑战赛已开启，参与赢积分奖励",
-    date: "2026-06-28",
-  },
-  {
-    id: "a-3",
-    tag: "notice",
-    tagLabel: "公告",
-    title: "7月3日凌晨 2:00-4:00 系统维护通知",
-    date: "2026-06-25",
-  },
-];
-
-/* ============================================================
- * 工具函数
- * ============================================================ */
-const getGreeting = () => {
-  const hour = new Date().getHours();
-  if (hour < 6) return "夜深了";
-  if (hour < 12) return "早上好";
-  if (hour < 14) return "中午好";
-  if (hour < 18) return "下午好";
-  return "晚上好";
-};
-
-const formatDate = () => {
-  const d = new Date();
-  const weekDays = ["日", "一", "二", "三", "四", "五", "六"];
-  return `${d.getFullYear()}年${d.getMonth() + 1}月${d.getDate()}日 星期${weekDays[d.getDay()]}`;
-};
-
-/** 图标名称 → Ant Design 组件映射 */
-const iconMap: Record<string, React.ReactNode> = {
-  video: <VideoCameraOutlined />,
-  appstore: <AppstoreOutlined />,
-  thunderbolt: <ThunderboltOutlined />,
-  database: <DatabaseOutlined />,
-  filetext: <FileTextOutlined />,
-};
-
-/* ============================================================
- * 组件
- * ============================================================ */
 const Dashboard: React.FC = () => {
   const navigate = useNavigate();
 
-  const maxChart = Math.max(...weeklyData.map((d) => d.value));
-
   return (
     <div className="xx-dashboard-page">
-      {/* ── 欢迎头部 ─────────────────────────────────────────── */}
-      <div className="xx-dashboard-welcome">
-        <h2>{getGreeting()}，创作者</h2>
-        <p>{formatDate()} — 欢迎回到小小剪辑控制台</p>
-      </div>
-
-      {/* ── KPI 卡片网格 ─────────────────────────────────────── */}
+      {/* KPI 卡片网格 */}
       <div className="xx-kpi-grid">
-        {kpiData.map((item) => (
-          <div
-            key={item.key}
-            className="xx-kpi-card"
-            style={{ "--kpi-accent": item.accent } as React.CSSProperties}
-          >
-            <div
-              className="xx-kpi-icon"
-              style={{ background: item.iconGradient }}
-            >
-              {iconMap[item.icon] ?? item.icon}
-            </div>
-            <div className="xx-kpi-value">{item.value}</div>
-            <div className="xx-kpi-label">{item.label}</div>
-            <span
-              className={`xx-kpi-trend xx-kpi-trend--${item.trendDirection}`}
-            >
-              {item.trend}
-            </span>
-          </div>
-        ))}
-      </div>
-
-      {/* ── 主内容区：左侧任务+图表 / 右侧公告 ──────────────── */}
-      <div className="xx-dashboard-main">
-        {/* 左列 */}
-        <div className="xx-dashboard-left-col">
-          {/* 最近任务 */}
-          <div className="xx-dashboard-section">
-            <div className="xx-dashboard-section-header">
-              <h3>最近任务</h3>
-              <button onClick={() => navigate("/app/history")}>查看全部</button>
-            </div>
-            <div className="xx-task-list">
-              {recentTasks.map((task) => (
-                <div key={task.id} className="xx-task-item">
-                  <div className="xx-task-info">
-                    <h4>{task.name}</h4>
-                    <span>
-                      {task.type} · 模板：{task.template}
-                    </span>
-                  </div>
-                  <Tag
-                    variant={
-                      task.status === "completed"
-                        ? "success"
-                        : task.status === "processing"
-                          ? "info"
-                          : task.status === "failed"
-                            ? "error"
-                            : "warning"
-                    }
-                  >
-                    {statusLabel[task.status]}
-                  </Tag>
-                  <div className="xx-task-time">
-                    <span>{task.date}</span>
-                    {task.status === "completed"
-                      ? `耗时 ${task.duration}`
-                      : task.status === "processing"
-                        ? "生成中..."
-                        : task.status === "failed"
-                          ? "请重试"
-                          : "等待中"}
-                  </div>
-                  <div className="xx-task-action">
-                    <Button
-                      buttonType="ghost"
-                      buttonSize="sm"
-                      onClick={() => navigate("/app/history")}
-                    >
-                      查看
-                    </Button>
-                  </div>
-                </div>
-              ))}
-            </div>
-          </div>
-
-          {/* 使用统计图表 */}
-          <div className="xx-dashboard-section">
-            <div className="xx-dashboard-section-header">
-              <h3>本周生成趋势</h3>
-              <span className="xx-chart-total">
-                共 {weeklyData.reduce((s, d) => s + d.value, 0)} 次
-              </span>
-            </div>
-            <div className="xx-chart-container">
-              <div className="xx-chart-bars">
-                {weeklyData.map((d, i) => (
-                  <div key={i} className="xx-chart-bar-wrapper">
-                    <div
-                      className="xx-chart-bar"
-                      style={{
-                        height: `${(d.value / maxChart) * 100}%`,
-                      }}
-                    >
-                      <span className="xx-chart-bar-value">{d.value}</span>
-                    </div>
-                  </div>
-                ))}
-              </div>
-              <div className="xx-chart-labels">
-                {weeklyData.map((d, i) => (
-                  <div key={i} className="xx-chart-label">
-                    {d.label}
-                  </div>
-                ))}
-              </div>
-            </div>
-          </div>
-        </div>
-
-        {/* 右列 — 公告 + 存储用量 */}
-        <div className="xx-dashboard-section xx-dashboard-section--start">
-          <div className="xx-dashboard-section-header">
-            <h3>系统公告</h3>
-          </div>
-          <div className="xx-announcement-list">
-            {announcements.map((a) => (
-              <div key={a.id} className="xx-announcement-item">
-                <span
-                  className={`xx-announcement-tag xx-announcement-tag--${a.tag}`}
-                >
-                  {a.tagLabel}
-                </span>
-                <div className="xx-announcement-content">
-                  <h4>{a.title}</h4>
-                  <time>{a.date}</time>
-                </div>
-              </div>
-            ))}
-          </div>
-
-          {/* 存储用量 */}
-          <div className="xx-storage-section">
-            <div className="xx-storage-section-title">存储用量</div>
-            <div className="xx-storage-bar">
-              <div className="xx-storage-bar-track">
-                <div className="xx-storage-bar-fill" style={{ width: "24%" }} />
-              </div>
-              <div className="xx-storage-bar-label">
-                <span>2.4 GB 已用</span>
-                <span>10 GB 总量</span>
-              </div>
-            </div>
-          </div>
+        <div className="xx-dashboard-empty">
+          <p>暂无统计数据</p>
         </div>
       </div>
 
-      {/* ── 快速入口 ─────────────────────────────────────────── */}
+      {/* 快速入口 */}
       <div className="xx-quick-entry-section">
         <div className="xx-quick-entry-header">
           <h3 className="xx-quick-entry-title">快速入口</h3>
         </div>
         <div className="xx-quick-grid">
-          {quickEntries.map((entry) => (
-            <div
-              key={entry.id}
-              className="xx-quick-card"
-              onClick={() => navigate(entry.path)}
-            >
-              <div
-                className="xx-quick-card-icon"
-                style={{ background: entry.iconGradient }}
-              >
-                {iconMap[entry.icon] ?? entry.icon}
-              </div>
-              <h3>{entry.title}</h3>
-              <p>{entry.description}</p>
-            </div>
-          ))}
+          <div className="xx-dashboard-empty">
+            <p>暂无快速入口</p>
+          </div>
         </div>
       </div>
+
+      {/* 最近任务 */}
+      <section className="xx-dashboard-section">
+        <div className="xx-dashboard-section-header">
+          <h3>最近任务</h3>
+          <Button
+            buttonType="ghost"
+            buttonSize="sm"
+            onClick={() => navigate("/app/history")}
+          >
+            查看全部
+          </Button>
+        </div>
+        <div className="xx-task-list">
+          <div className="xx-dashboard-empty">
+            <p>暂无最近任务</p>
+          </div>
+        </div>
+      </section>
+
+      {/* 使用统计 */}
+      <section
+        className="xx-dashboard-section"
+        style={{ marginTop: "var(--space-md)" }}
+      >
+        <div className="xx-dashboard-section-header">
+          <h3>使用统计</h3>
+        </div>
+        <div className="xx-chart-container">
+          <div className="xx-chart-bars">
+            <div className="xx-dashboard-empty" style={{ width: "100%" }}>
+              <DatabaseOutlined style={{ fontSize: 24, marginBottom: 8 }} />
+              <p>暂无统计数据</p>
+            </div>
+          </div>
+        </div>
+      </section>
+
+      {/* 公告 */}
+      <section
+        className="xx-dashboard-section"
+        style={{ marginTop: "var(--space-md)" }}
+      >
+        <div className="xx-dashboard-section-header">
+          <h3>公告</h3>
+        </div>
+        <div className="xx-announcement-list">
+          <div className="xx-announcement-item">
+            <span className="xx-announcement-tag xx-announcement-tag--notice">
+              官方
+            </span>
+            <div className="xx-announcement-content">
+              <h4>欢迎使用小应 SaaS 平台</h4>
+              <time>当前为演示版本，部分功能正在开发中。</time>
+            </div>
+          </div>
+        </div>
+      </section>
     </div>
   );
 };

apps/web/src/pages/dashboard/dashboard.css

@@ -1,6 +1,6 @@
 /**
  * 控制台页面 - V21 设计系统样式
- * KPI 卡片网格 + 快速入口 + 最近任务卡片列表 + 使用统计图表 + 公告
+ * KPI 卡片网格 + 快速入口 + 最近任务 + 使用统计图表 + 公告
  * 统一使用 CSS 变量，支持深色/浅色主题
  */
 @import "../../styles/global.css";
@@ -13,26 +13,6 @@
   padding: var(--space-xl);
 }
 
-/* ============================================================
-   欢迎头部
-   ============================================================ */
-.xx-dashboard-welcome {
-  margin-bottom: var(--space-lg);
-}
-
-.xx-dashboard-welcome h2 {
-  margin: 0 0 var(--space-xs);
-  font-size: var(--font-size-xl);
-  font-weight: var(--font-weight-bold);
-  color: var(--text-primary);
-}
-
-.xx-dashboard-welcome p {
-  margin: 0;
-  font-size: var(--font-size-sm);
-  color: var(--text-secondary);
-}
-
 /* ============================================================
    KPI 卡片网格
    ============================================================ */
@@ -43,76 +23,6 @@
   margin-bottom: var(--space-lg);
 }
 
-.xx-kpi-card {
-  background: linear-gradient(180deg, var(--bg-primary), var(--bg-secondary));
-  border: 1px solid var(--border-color);
-  border-radius: var(--radius-lg);
-  padding: 20px;
-  transition: 0.18s ease;
-  position: relative;
-  overflow: hidden;
-}
-
-.xx-kpi-card:hover {
-  transform: translateY(-2px);
-  box-shadow: var(--shadow-sm);
-}
-
-.xx-kpi-icon {
-  width: 40px;
-  height: 40px;
-  border-radius: var(--radius-md);
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  font-size: var(--font-size-lg);
-  margin-bottom: var(--space-sm);
-}
-
-.xx-kpi-value {
-  font-size: 30px;
-  font-weight: 800;
-  color: var(--text-primary);
-  line-height: 1.2;
-  margin-bottom: var(--space-xs);
-}
-
-.xx-kpi-label {
-  font-size: var(--font-size-sm);
-  color: var(--text-secondary);
-  margin-bottom: var(--space-sm);
-}
-
-.xx-kpi-trend {
-  font-size: 12px;
-  margin-top: 8px;
-  display: inline-flex;
-  align-items: center;
-  gap: 4px;
-}
-
-.xx-kpi-trend--up {
-  color: var(--success-color);
-}
-
-.xx-kpi-trend--down {
-  color: var(--error-color);
-}
-
-.xx-kpi-trend--neutral {
-  color: var(--text-tertiary);
-}
-
-/* ============================================================
-   主内容区两栏布局
-   ============================================================ */
-.xx-dashboard-main {
-  display: grid;
-  grid-template-columns: 1fr 320px;
-  gap: var(--space-md);
-  margin-bottom: var(--space-lg);
-}
-
 /* ============================================================
    区块卡片
    ============================================================ */
@@ -161,7 +71,25 @@
 }
 
 /* ============================================================
-   最近任务卡片列表
+   空状态
+   ============================================================ */
+.xx-dashboard-empty {
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  justify-content: center;
+  padding: var(--space-xl) var(--space-md);
+  text-align: center;
+  color: var(--text-tertiary);
+}
+
+.xx-dashboard-empty p {
+  margin: 0;
+  font-size: var(--font-size-sm);
+}
+
+/* ============================================================
+   最近任务列表
    ============================================================ */
 .xx-task-list {
   display: flex;
@@ -169,86 +97,6 @@
   gap: var(--space-sm);
 }
 
-.xx-task-item {
-  background: var(--bg-primary);
-  border: 1px solid var(--border-color);
-  border-radius: var(--radius-md);
-  padding: var(--space-sm) var(--space-md);
-  display: grid;
-  grid-template-columns: 1fr auto auto auto;
-  gap: var(--space-md);
-  align-items: center;
-  transition: var(--transition-all);
-}
-
-.xx-task-item:hover {
-  border-color: var(--primary-color);
-  box-shadow: var(--shadow-sm);
-}
-
-.xx-task-info h4 {
-  margin: 0 0 var(--space-xs);
-  font-size: var(--font-size-sm);
-  font-weight: 600;
-  color: var(--text-primary);
-}
-
-.xx-task-info span {
-  font-size: var(--font-size-xs);
-  color: var(--text-secondary);
-}
-
-.xx-task-status {
-  display: inline-flex;
-  align-items: center;
-  gap: var(--space-xs);
-  padding: 3px 10px;
-  border-radius: var(--radius-sm);
-  font-size: var(--font-size-xs);
-  font-weight: 500;
-  white-space: nowrap;
-}
-
-.xx-task-status--completed {
-  color: var(--success-color);
-  background: var(--success-soft);
-  border: 1px solid var(--success-border);
-}
-
-.xx-task-status--processing {
-  color: var(--info-color);
-  background: var(--primary-soft);
-  border: 1px solid var(--color-primary-200);
-}
-
-.xx-task-status--pending {
-  color: var(--text-secondary);
-  background: var(--bg-secondary);
-  border: 1px solid var(--border-color);
-}
-
-.xx-task-status--failed {
-  color: var(--error-color);
-  background: var(--error-soft);
-  border: 1px solid var(--error-border);
-}
-
-.xx-task-time {
-  text-align: right;
-  font-size: var(--font-size-xs);
-  color: var(--text-secondary);
-  min-width: 100px;
-}
-
-.xx-task-time span {
-  display: block;
-  margin-bottom: var(--space-xxs);
-}
-
-.xx-task-action {
-  font-size: var(--font-size-xs);
-}
-
 /* ============================================================
    使用统计图表（纯 CSS 柱状图）
    ============================================================ */
@@ -264,69 +112,27 @@
   padding-top: var(--space-sm);
 }
 
-.xx-chart-bar-wrapper {
-  flex: 1;
-  display: flex;
-  flex-direction: column;
-  align-items: center;
-  height: 100%;
-  justify-content: flex-end;
-}
-
-.xx-chart-bar {
-  width: 100%;
-  max-width: 36px;
-  border-radius: var(--radius-sm) var(--radius-sm) 0 0;
-  background: var(--gradient-primary);
-  transition: height 0.6s cubic-bezier(0.34, 1.56, 0.64, 1);
-  position: relative;
-  min-height: 4px;
-  cursor: pointer;
-}
-
-.xx-chart-bar:hover {
-  opacity: 0.85;
-}
-
-.xx-chart-bar:active {
-  opacity: 0.7;
-  transform: scaleY(0.97);
-  transform-origin: bottom;
-}
-
-.xx-chart-bar-value {
-  position: absolute;
-  top: -20px;
-  left: 50%;
-  transform: translateX(-50%);
-  font-size: var(--font-size-xs);
-  font-weight: 600;
-  color: var(--text-primary);
-  white-space: nowrap;
-  opacity: 0;
-  transition: var(--transition-opacity);
-}
-
-.xx-chart-bar:hover .xx-chart-bar-value {
-  opacity: 1;
-}
-
-.xx-chart-labels {
-  display: flex;
-  gap: var(--space-sm);
-  margin-top: var(--space-sm);
-}
-
-.xx-chart-label {
-  flex: 1;
-  text-align: center;
-  font-size: var(--font-size-xs);
-  color: var(--text-secondary);
-}
-
 /* ============================================================
-   快速入口卡片网格
+   快速入口
    ============================================================ */
+.xx-quick-entry-section {
+  margin-bottom: var(--space-md);
+}
+
+.xx-quick-entry-header {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  margin-bottom: var(--space-md);
+}
+
+.xx-quick-entry-title {
+  margin: 0;
+  font-size: var(--font-size-base);
+  font-weight: var(--font-weight-semibold);
+  color: var(--text-primary);
+}
+
 .xx-quick-grid {
   display: grid;
   grid-template-columns: repeat(4, 1fr);
@@ -334,53 +140,6 @@
   margin-bottom: var(--space-lg);
 }
 
-.xx-quick-card {
-  background: var(--bg-primary);
-  border: 1px solid var(--border-color);
-  border-radius: var(--radius-lg);
-  padding: var(--space-lg);
-  cursor: pointer;
-  transition: var(--transition-all);
-  text-align: center;
-}
-
-.xx-quick-card:hover {
-  border-color: var(--primary-color);
-  box-shadow: var(--shadow-md);
-  transform: translateY(-2px);
-}
-
-.xx-quick-card:active {
-  transform: translateY(0) scale(0.98);
-  box-shadow: var(--shadow-sm);
-  transition-duration: 0.1s;
-}
-
-.xx-quick-card-icon {
-  width: 52px;
-  height: 52px;
-  border-radius: var(--radius-md);
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  font-size: var(--font-size-xl);
-  margin: 0 auto var(--space-md);
-  color: var(--text-inverse);
-}
-
-.xx-quick-card h3 {
-  margin: 0 0 var(--space-sm);
-  font-size: var(--font-size-md);
-  font-weight: var(--font-weight-semibold);
-  color: var(--text-primary);
-}
-
-.xx-quick-card p {
-  margin: 0;
-  color: var(--text-secondary);
-  font-size: var(--font-size-sm);
-}
-
 /* ============================================================
    公告区域
    ============================================================ */
@@ -451,90 +210,10 @@
   color: var(--text-secondary);
 }
 
-/* ============================================================
-   存储用量条
-   ============================================================ */
-.xx-storage-bar {
-  margin-top: var(--space-sm);
-}
-
-.xx-storage-bar-track {
-  height: var(--space-sm);
-  background: var(--bg-secondary);
-  border-radius: var(--space-xs);
-  overflow: hidden;
-}
-
-.xx-storage-bar-fill {
-  height: 100%;
-  border-radius: var(--space-xs);
-  background: var(--gradient-primary);
-  transition: width 0.6s ease;
-}
-
-.xx-storage-bar-label {
-  display: flex;
-  justify-content: space-between;
-  font-size: var(--font-size-xs);
-  color: var(--text-secondary);
-  margin-top: var(--space-xs);
-}
-
-/* ============================================================
-   迁移自内联样式的工具类
-   ============================================================ */
-.xx-dashboard-left-col {
-  display: flex;
-  flex-direction: column;
-  gap: var(--space-md);
-}
-
-.xx-chart-total {
-  font-size: var(--font-size-xs);
-  color: var(--text-secondary);
-}
-
-.xx-dashboard-section--start {
-  align-self: start;
-}
-
-.xx-storage-section {
-  margin-top: var(--space-md);
-}
-
-.xx-storage-section-title {
-  font-size: 13px;
-  font-weight: 500;
-  color: var(--text-primary);
-  margin-bottom: 4px;
-}
-
-.xx-quick-entry-section {
-  margin-bottom: var(--space-md);
-}
-
-.xx-quick-entry-header {
-  display: flex;
-  align-items: center;
-  justify-content: space-between;
-  margin-bottom: var(--space-md);
-}
-
-.xx-quick-entry-title {
-  margin: 0;
-  font-size: var(--font-size-base);
-  font-weight: var(--font-weight-semibold);
-  color: var(--text-primary);
-}
-
 /* ============================================================
    响应式
    ============================================================ */
 @media (max-width: 1200px) {
-  .xx-dashboard-main {
-    grid-template-columns: 1fr;
-  }
-
   .xx-kpi-grid {
     grid-template-columns: repeat(2, 1fr);
   }
@@ -557,15 +236,6 @@
     grid-template-columns: 1fr;
   }
 
-  .xx-task-item {
-    grid-template-columns: 1fr;
-    gap: var(--space-sm);
-  }
-
-  .xx-task-time {
-    text-align: left;
-  }
-
   .xx-chart-bars {
     height: 120px;
   }

apps/web/src/pages/generate/GeneratePage.tsx

@@ -69,22 +69,6 @@ const VOICE_GENDER_ICON: Record<string, string> = {
   neutral: "✨",
 };
 
-/* ── 时间线 Mock ──
- * TODO: 后端暂无时间线场景数据 API，当前使用硬编码预览数据。
- *       待后端提供 timeline/scene 接口后替换为真实 API 调用。
- */
-interface TimelineScene {
-  scene: string;
-  time: string;
-  duration: number;
-}
-
-const MOCK_TIMELINE: TimelineScene[] = [
-  { scene: "主讲口播 · 开场钩子", time: "0-8s", duration: 8 },
-  { scene: "产品特写 · B-roll", time: "8-20s", duration: 12 },
-  { scene: "用户反馈 · 结尾", time: "20-30s", duration: 10 },
-];
-
 /* ── 步骤定义 ── */
 const STEPS = [
   { key: 1, label: "选择模板" },
@@ -1758,15 +1742,19 @@ const GeneratePage: React.FC = () => {
           <div className="xx-preview-title">剪辑计划预览</div>
 
           {/* 时间线列表 */}
-          <div className="xx-preview-timeline">
-            {MOCK_TIMELINE.map((item, idx) => (
-              <div key={idx} className="xx-timeline-item">
-                <div className="num">{idx + 1}</div>
-                <span className="scene-name">{item.scene}</span>
-                <span>{item.time}</span>
+          {generated ? (
+            <div className="xx-preview-timeline">
+              <div className="xx-timeline-item">
+                <span className="scene-name">生成完成，可下载或分享视频</span>
               </div>
-            ))}
-          </div>
+            </div>
+          ) : (
+            <div className="xx-preview-timeline">
+              <div className="xx-timeline-item">
+                <span className="scene-name">确认标题后自动生成剪辑计划</span>
+              </div>
+            </div>
+          )}
 
           {/* 生成操作按钮 */}
           <div className="xx-generate-actions">

apps/web/src/pages/generate/generate.css

@@ -612,54 +612,6 @@
   flex: 1;
 }
 
-/* ============================================================
-   按钮（匹配原型 .btn .ghost / .btn .primary）
-   ============================================================ */
-.xx-btn {
-  display: inline-flex;
-  align-items: center;
-  justify-content: center;
-  gap: 6px;
-  height: 42px;
-  padding: 0 20px;
-  border-radius: var(--radius-sm);
-  font-size: 14px;
-  font-weight: 600;
-  cursor: pointer;
-  transition: all 0.15s ease;
-  border: none;
-  outline: none;
-  white-space: nowrap;
-}
-
-.xx-btn:disabled {
-  opacity: 0.5;
-  cursor: not-allowed;
-}
-
-.xx-btn-primary {
-  background: var(--gradient-primary);
-  color: var(--text-inverse);
-  box-shadow: 0 14px 26px rgba(79, 70, 229, 0.22);
-}
-
-.xx-btn-primary:hover:not(:disabled) {
-  transform: translateY(-2px);
-  box-shadow: 0 18px 34px rgba(79, 70, 229, 0.28);
-}
-
-.xx-btn-ghost {
-  background: var(--bg-primary);
-  border: 1px solid var(--border-color);
-  color: var(--text-secondary);
-}
-
-.xx-btn-ghost:hover:not(:disabled) {
-  border-color: var(--info-border);
-  color: var(--primary-dark);
-  background: var(--primary-soft);
-}
-
 /* ============================================================
    右侧预览区 generate-preview
    ============================================================ */

apps/web/src/pages/titles/TitleLibrary.tsx

@@ -36,40 +36,24 @@ type TitleType = "hot" | "normal" | "creative";
 type Industry = "general" | "food" | "tech" | "beauty" | "education" | "travel";
 type Frequency = "all" | "high" | "medium" | "low";
 
-interface CategoryItem {
-  id: string;
-  name: string;
-  count: number;
-}
-
 interface TitleData {
   id: string;
   content: string;
   type: TitleType;
   industry: Industry;
+  category: string;
   usageCount: number;
   isFavorited: boolean;
   createdAt: string;
 }
 
-/* ============================================================
- * Mock 数据
- * ============================================================ */
-const MOCK_CATEGORIES: CategoryItem[] = [
-  { id: "cat-all", name: "全部标题", count: 15 },
-  { id: "cat-1", name: "美食探店", count: 4 },
-  { id: "cat-2", name: "科技数码", count: 3 },
-  { id: "cat-3", name: "生活日常", count: 4 },
-  { id: "cat-4", name: "美妆穿搭", count: 2 },
-  { id: "cat-5", name: "教育学习", count: 2 },
-];
-
 /** 后端 TitleItem → 前端 TitleData 映射 */
 const toTitleData = (item: TitleItem): TitleData => ({
   id: item.id,
   content: item.content,
   type: (item.category as TitleType) || "normal",
   industry: "general",
+  category: item.category || "未分类",
   usageCount: 0,
   isFavorited: false,
   createdAt: item.created_at?.slice(0, 10) || "",
@@ -250,9 +234,8 @@ const TitleCard: React.FC<{
 const TitleLibrary: React.FC = () => {
   const queryClient = useQueryClient();
 
-  /* 分类数据 */
-  const [categories, setCategories] = useState<CategoryItem[]>(MOCK_CATEGORIES);
-  const [activeCatId, setActiveCatId] = useState<string>(MOCK_CATEGORIES[0].id);
+  /* 分类数据 — 从真实标题数据动态派生 */
+  const [activeCatId, setActiveCatId] = useState<string>("cat-all");
 
   /* 标题数据 — 真实 API */
   const { data: apiTitles = [] } = useQuery({
@@ -265,6 +248,23 @@ const TitleLibrary: React.FC = () => {
     [apiTitles],
   );
 
+  /* 从真实标题数据动态派生分类（无需后端分类 API） */
+  const categories = useMemo(() => {
+    const cats = new Map<string, number>();
+    apiTitles.forEach((t) => {
+      const cat = t.category || "未分类";
+      cats.set(cat, (cats.get(cat) || 0) + 1);
+    });
+    return [
+      { id: "cat-all", name: "全部标题", count: apiTitles.length },
+      ...Array.from(cats.entries()).map(([name, count]) => ({
+        id: `cat-${name}`,
+        name,
+        count,
+      })),
+    ];
+  }, [apiTitles]);
+
   /* CRUD mutations */
   const createMutation = useMutation({
     mutationFn: (content: string) => createTitle({ content }),
@@ -301,10 +301,6 @@ const TitleLibrary: React.FC = () => {
   const [editingId, setEditingId] = useState<string | null>(null);
   const [editText, setEditText] = useState("");
 
-  /* 新建分类 */
-  const [createCatModalOpen, setCreateCatModalOpen] = useState(false);
-  const [newCatName, setNewCatName] = useState("");
-
   /* 新建标题 */
   const [createTitleModalOpen, setCreateTitleModalOpen] = useState(false);
   const [newTitleContent, setNewTitleContent] = useState("");
@@ -322,19 +318,11 @@ const TitleLibrary: React.FC = () => {
   const filteredTitles = useMemo(() => {
     let list = titles;
 
-    /* 按分类过滤（"全部标题" 不过滤） */
+    /* 按分类过滤（"全部标题" 不过滤）— 直接匹配后端 category 字段 */
     if (activeCatId !== "cat-all") {
       const catName = activeCategory?.name || "";
-      const catToIndustry: Record<string, Industry> = {
-        美食探店: "food",
-        科技数码: "tech",
-        生活日常: "general",
-        美妆穿搭: "beauty",
-        教育学习: "education",
-      };
-      const mappedIndustry = catToIndustry[catName];
-      if (mappedIndustry) {
-        list = list.filter((t) => t.industry === mappedIndustry);
+      if (catName) {
+        list = list.filter((t) => t.category === catName);
       }
     }
 
@@ -428,33 +416,6 @@ const TitleLibrary: React.FC = () => {
     [deleteMutation],
   );
 
-  /* 新建分类 */
-  const handleCreateCategory = () => {
-    if (!newCatName.trim()) {
-      message.warning("请输入分类名称");
-      return;
-    }
-    const cat: CategoryItem = {
-      id: `cat-${Date.now()}`,
-      name: newCatName.trim(),
-      count: 0,
-    };
-    setCategories((prev) => [...prev, cat]);
-    setActiveCatId(cat.id);
-    setCreateCatModalOpen(false);
-    setNewCatName("");
-    message.success(`分类 "${cat.name}" 创建成功`);
-  };
-
-  /* 删除分类 */
-  const handleDeleteCategory = (id: string) => {
-    setCategories((prev) => prev.filter((c) => c.id !== id));
-    if (activeCatId === id) {
-      setActiveCatId("cat-all");
-    }
-    message.success("分类已删除");
-  };
-
   /* 新建标题 */
   const handleCreateTitle = () => {
     if (!newTitleContent.trim()) {
@@ -541,38 +502,11 @@ const TitleLibrary: React.FC = () => {
                   </h4>
                   <span>{cat.count} 条</span>
                 </div>
-                {cat.id !== "cat-all" && (
-                  <Popconfirm
-                    title={`确定删除分类 "${cat.name}"？`}
-                    onConfirm={(e) => {
-                      e?.stopPropagation();
-                      handleDeleteCategory(cat.id);
-                    }}
-                    onCancel={(e) => e?.stopPropagation()}
-                    okText="删除"
-                    cancelText="取消"
-                  >
-                    <button
-                      className="xx-title-category-delete"
-                      onClick={(e) => e.stopPropagation()}
-                      title="删除分类"
-                    >
-                      <DeleteOutlined />
-                    </button>
-                  </Popconfirm>
-                )}
               </div>
             </div>
           ))}
 
-          {/* 新建分类 */}
-          <div
-            className="xx-title-category-add"
-            onClick={() => setCreateCatModalOpen(true)}
-          >
-            <PlusOutlined />
-            新建分类
-          </div>
+          {/* TODO: 新建分类功能待后端分类 API 就绪后启用 */}
         </div>
 
         {/* ─── 右侧：内容区 ─── */}
@@ -678,35 +612,6 @@ const TitleLibrary: React.FC = () => {
         </div>
       </div>
 
-      {/* ─── 新建分类弹窗 ─── */}
-      <AntModal
-        title="新建分类"
-        open={createCatModalOpen}
-        onCancel={() => setCreateCatModalOpen(false)}
-        onOk={handleCreateCategory}
-        okText="创建"
-        cancelText="取消"
-        destroyOnClose
-      >
-        <div style={{ padding: "8px 0" }}>
-          <div
-            style={{
-              marginBottom: 6,
-              fontSize: "var(--font-size-sm)",
-              color: "var(--text-secondary)",
-            }}
-          >
-            分类名称
-          </div>
-          <Input
-            placeholder="请输入分类名称"
-            value={newCatName}
-            onChange={(e) => setNewCatName(e.target.value)}
-            maxLength={30}
-          />
-        </div>
-      </AntModal>
-
       {/* ─── 新建标题弹窗 ─── */}
       <AntModal
         title="新建标题"

apps/web/src/pages/voice-clone/VoiceClone.tsx

@@ -18,7 +18,7 @@ import {
   CloseCircleOutlined,
 } from "@ant-design/icons";
 import PageHead from "@/components/layout/PageHead";
-import CloneVoiceModal from "@/components/modals/CloneVoiceModal";
+import CloneModal from "@/components/voice/CloneModal";
 import {
   getVoiceClones,
   deleteVoiceClone,
@@ -356,7 +356,7 @@ const VoiceClone: React.FC = () => {
       )}
 
       {/* 克隆音色弹窗 */}
-      <CloneVoiceModal
+      <CloneModal
         open={cloneModalOpen}
         onClose={() => setCloneModalOpen(false)}
         onSuccess={() => {

apps/web/src/router/index.tsx

@@ -170,13 +170,6 @@ export const router = createBrowserRouter([
             Component: m.default,
           })),
       },
-      {
-        path: "my-voices",
-        lazy: () =>
-          import("@/pages/my-voices/MyVoices").then((m) => ({
-            Component: m.default,
-          })),
-      },
       {
         path: "accounts",
         lazy: () =>

apps/worker/video_processing/__init__.py

@@ -1,18 +1,17 @@
 """
 视频处理模块
+
+轻量工具（ffmpeg_utils / oss_helpers / dedup_helpers）顶层直接导出，
+无额外依赖。渲染相关组件（UnifiedRenderService / RenderAdapter /
+VideoProcessor 等）按需从子模块导入，避免 __init__ 阶段引入
+packages / DB 等重依赖。
 """
 
-# 共享工具模块（供 editing_modes / generation / edit_plan_generation 等复用）
+# 共享工具模块（零外部依赖，供 editing_modes / generation / edit_plan_generation 等复用）
 from . import dedup_helpers, ffmpeg_utils, oss_helpers
-from .processor import VideoProcessor, VideoResult
-from .unified_render_service import RenderResult, UnifiedRenderService
 
 __all__ = [
-    "VideoProcessor",
-    "VideoResult",
     "ffmpeg_utils",
     "oss_helpers",
     "dedup_helpers",
-    "UnifiedRenderService",
-    "RenderResult",
 ]

apps/worker/video_processing/dedup.py

@@ -1,10 +1,8 @@
 """Video deduplication module - compute fingerprints and detect duplicates."""
 
 import hashlib
-import json
 import logging
 import os
-import subprocess
 import tempfile
 from dataclasses import dataclass
 from typing import Optional
@@ -93,12 +91,16 @@ class VideoFingerprint:
     resolution: tuple[int, int]
 
     def to_dict(self) -> dict:
+        # 注意：color_histograms 里的值可能是 np.float32（来自 cv2.normalize），
+        # 直接存进 dict 后 SQLAlchemy JSON 序列化会报 "float32 is not JSON serializable"。
+        # 这里统一转成 Python 原生 float。
+        native_histograms = [[float(v) for v in hist] for hist in self.color_histograms]
         return {
             "md5": self.md5,
             "keyframe_phashes": self.keyframe_phashes,
-            "color_histograms": self.color_histograms,
-            "duration": self.duration,
-            "resolution": list(self.resolution),
+            "color_histograms": native_histograms,
+            "duration": float(self.duration),
+            "resolution": [int(self.resolution[0]), int(self.resolution[1])],
         }
 
 
@@ -323,7 +325,6 @@ def check_duplicate_task(self: Task, generated_video_id: str) -> dict:
             raise ValueError(f"Generated video {generated_video_id} not found")
 
         local_path = os.path.join(temp_dir, f"{generated_video_id}.mp4")
-        storage_key = video.file_url.split("/")[-1]
         storage_service.download_file(
             f"projects/{video.project_id}/generated/{generated_video_id}/{generated_video_id}.mp4", local_path
         )

apps/worker/video_processing/editing_modes.py

@@ -1,657 +0,0 @@
-"""
-视频剪辑模式处理器
-支持四种剪辑模式：一镜到底、画中画、口播、口播+画中画
-"""
-
-import logging
-import os
-import sys
-import tempfile
-from dataclasses import dataclass
-
-if sys.version_info >= (3, 11):
-    from enum import StrEnum
-else:
-    from enum import Enum
-
-    class StrEnum(str, Enum):
-        pass
-
-
-from pathlib import Path
-from typing import Optional
-
-from video_processing.ffmpeg_utils import FFMPEG_BIN, probe_video_info, run_ffmpeg
-
-logger = logging.getLogger(__name__)
-
-
-# 从 domain 层导入 EditingMode，避免重复定义
-from packages.domain.editing_mode import EditingMode
-
-
-class PIPPosition(StrEnum):
-    """画中画位置枚举"""
-
-    TOP_LEFT = "top_left"
-    TOP_RIGHT = "top_right"
-    BOTTOM_LEFT = "bottom_left"
-    BOTTOM_RIGHT = "bottom_right"
-
-
-@dataclass
-class EditingModeConfig:
-    """剪辑模式配置"""
-
-    mode: EditingMode
-    output_width: int = 1280
-    output_height: int = 720
-    output_fps: int = 25
-    pip_position: PIPPosition = PIPPosition.TOP_RIGHT
-    pip_scale: float = 0.25  # 画中画占主画面的比例
-    transition_duration: float = 0.5  # 转场时长（秒）
-    output_codec: str = "libx264"
-    output_preset: str = "medium"
-    output_crf: int = 23
-
-
-class EditingModeProcessor:
-    """剪辑模式处理器"""
-
-    def __init__(self, config: EditingModeConfig, work_dir: Optional[str] = None):
-        """
-        初始化剪辑模式处理器
-
-        Args:
-            config: 剪辑模式配置
-            work_dir: 工作目录，默认使用系统临时目录
-        """
-        self.config = config
-        self.work_dir = work_dir or tempfile.gettempdir()
-
-    def process(
-        self,
-        video_paths: list[str],
-        audio_path: Optional[str] = None,
-        output_path: Optional[str] = None,
-    ) -> str:
-        """
-        根据模式处理视频，返回输出文件路径
-
-        Args:
-            video_paths: 视频素材路径列表
-            audio_path: 音频路径（用于口播模式）
-            output_path: 输出文件路径，默认自动生成
-
-        Returns:
-            输出文件路径
-        """
-        if not video_paths:
-            raise ValueError("video_paths cannot be empty")
-
-        self._validate_inputs(video_paths, audio_path)
-
-        if output_path is None:
-            output_path = self._generate_output_path()
-
-        logger.info(f"Processing videos with mode: {self.config.mode}, count: {len(video_paths)}")
-
-        try:
-            if self.config.mode == EditingMode.ONE_TAKE:
-                return self._one_take(video_paths, output_path)
-            elif self.config.mode == EditingMode.PIP:
-                return self._pip(video_paths, output_path)
-            elif self.config.mode == EditingMode.VOICE_OVER:
-                return self._voice_over(video_paths, audio_path, output_path)
-            elif self.config.mode == EditingMode.VOICE_PIP:
-                return self._voice_pip(video_paths, audio_path, output_path)
-            else:
-                raise ValueError(f"Unsupported editing mode: {self.config.mode}")
-        except Exception as e:
-            logger.error(f"Error processing videos: {e}")
-            raise
-
-    def _validate_inputs(self, video_paths: list[str], audio_path: Optional[str]) -> None:
-        """验证输入文件"""
-        for path in video_paths:
-            if not os.path.exists(path):
-                raise FileNotFoundError(f"Video file not found: {path}")
-            if not os.path.getsize(path) > 0:
-                raise ValueError(f"Video file is empty: {path}")
-
-        if audio_path and not os.path.exists(audio_path):
-            raise FileNotFoundError(f"Audio file not found: {audio_path}")
-
-    def _generate_output_path(self) -> str:
-        """生成输出文件路径"""
-        os.makedirs(self.work_dir, exist_ok=True)
-        return os.path.join(self.work_dir, f"output_{self.config.mode}_{os.getpid()}.mp4")
-
-    def _run_ffmpeg(self, command: list[str], capture_output: bool = True) -> tuple:
-        """执行 FFmpeg 命令 — 委托给共享 ffmpeg_utils.run_ffmpeg"""
-        try:
-            return run_ffmpeg(command, capture_output=capture_output)
-        except RuntimeError as e:
-            logger.error(f"FFmpeg error: {e}")
-            raise
-
-    def _get_video_info(self, video_path: str) -> dict:
-        """获取视频信息 — 委托给共享 ffmpeg_utils.probe_video_info，补充 codec/size 字段"""
-        try:
-            info = probe_video_info(video_path)
-            info["codec"] = "unknown"
-            info["size"] = os.path.getsize(video_path) if os.path.exists(video_path) else 0
-            return info
-        except Exception as e:
-            logger.warning(f"Failed to get video info for {video_path}: {e}")
-            return {"width": 0, "height": 0, "fps": 25, "duration": 0, "codec": "unknown", "size": 0}
-
-    def _get_pip_position_offset(
-        self, main_width: int, main_height: int, pip_width: int, pip_height: int
-    ) -> tuple[int, int]:
-        """获取画中画位置偏移量"""
-        margin = 10
-        position_offsets = {
-            PIPPosition.TOP_LEFT: (margin, margin),
-            PIPPosition.TOP_RIGHT: (main_width - pip_width - margin, margin),
-            PIPPosition.BOTTOM_LEFT: (margin, main_height - pip_height - margin),
-            PIPPosition.BOTTOM_RIGHT: (main_width - pip_width - margin, main_height - pip_height - margin),
-        }
-        return position_offsets.get(self.config.pip_position, position_offsets[PIPPosition.TOP_RIGHT])
-
-    def _normalize_video(self, input_path: str, output_path: str) -> dict:
-        """标准化视频格式：先统一帧率，再缩放/填充"""
-        command = [
-            FFMPEG_BIN,
-            "-y",
-            "-i",
-            input_path,
-            "-r",
-            str(self.config.output_fps),  # 先统一帧率
-            "-vf",
-            f"scale={self.config.output_width}:{self.config.output_height}:force_original_aspect_ratio=decrease,pad={self.config.output_width}:{self.config.output_height}:(ow-iw)/2:(oh-ih)/2,setsar=1",
-            "-r",
-            str(self.config.output_fps),
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            "-movflags",
-            "+faststart",
-            "-an",
-            output_path,
-        ]
-        run_ffmpeg(command)
-        return self._get_video_info(output_path)
-
-    def _one_take(self, video_paths: list[str], output_path: str) -> str:
-        """一镜到底模式：顺序拼接视频，添加淡入淡出转场"""
-        if len(video_paths) == 1:
-            return self._normalize_video(video_paths[0], output_path)
-
-        normalized_paths = []
-        for i, path in enumerate(video_paths):
-            normalized = os.path.join(self.work_dir, f"normalized_{i}_{os.getpid()}.mp4")
-            self._normalize_video(path, normalized)
-            normalized_paths.append(normalized)
-
-        durations = [self._get_video_info(p)["duration"] for p in normalized_paths]
-
-        if len(normalized_paths) <= 5:
-            output_path = self._one_take_with_xfade(normalized_paths, durations, output_path)
-        else:
-            output_path = self._one_take_simple_concat(normalized_paths, output_path)
-
-        for p in normalized_paths:
-            try:
-                if p != output_path:
-                    os.remove(p)
-            except Exception as e:
-                logger.warning(f"Operation failed in apps/worker/video_processing/editing_modes.py: {e}", exc_info=True)
-
-        return output_path
-
-    def _one_take_with_xfade(self, normalized_paths: list[str], durations: list[float], output_path: str) -> str:
-        """使用 xfade 滤镜实现转场"""
-        if len(normalized_paths) == 2:
-            transition = self.config.transition_duration
-            offset1 = durations[0] - transition / 2
-
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-i",
-                normalized_paths[0],
-                "-i",
-                normalized_paths[1],
-                "-filter_complex",
-                f"[0:v][1:v]xfade=transition=fade:duration={transition}:offset={offset1}[v]",
-                "-map",
-                "[v]",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                output_path,
-            ]
-            run_ffmpeg(command)
-            return output_path
-        else:
-            return self._one_take_simple_concat(normalized_paths, output_path)
-
-    def _one_take_simple_concat(self, normalized_paths: list[str], output_path: str) -> str:
-        """使用 concat demuxer 简单拼接"""
-        concat_file = os.path.join(self.work_dir, f"concat_list_{os.getpid()}.txt")
-        with open(concat_file, "w") as f:
-            for path in normalized_paths:
-                f.write(f"file '{os.path.abspath(path)}'\n")
-
-        command = [
-            FFMPEG_BIN,
-            "-y",
-            "-f",
-            "concat",
-            "-safe",
-            "0",
-            "-i",
-            concat_file,
-            "-c",
-            "copy",
-            output_path,
-        ]
-        run_ffmpeg(command)
-
-        try:
-            os.remove(concat_file)
-        except Exception as e:
-            logger.warning(f"Operation failed in apps/worker/video_processing/editing_modes.py: {e}", exc_info=True)
-
-        return output_path
-
-    def _pip(self, video_paths: list[str], output_path: str) -> str:
-        """画中画模式：主视频全屏，后续视频叠加在角落"""
-        if not video_paths:
-            raise ValueError("No video paths provided")
-
-        main_video = video_paths[0]
-        main_normalized = os.path.join(self.work_dir, f"main_{os.getpid()}.mp4")
-        main_info = self._normalize_video(main_video, main_normalized)
-
-        if len(video_paths) == 1:
-            os.rename(main_normalized, output_path)
-            return output_path
-
-        pip_width = int(self.config.output_width * self.config.pip_scale)
-        pip_height = int(self.config.output_height * self.config.pip_scale)
-        x_offset, y_offset = self._get_pip_position_offset(
-            self.config.output_width, self.config.output_height, pip_width, pip_height
-        )
-
-        pip_normalized = os.path.join(self.work_dir, f"pip_{os.getpid()}.mp4")
-        pip_info = self._get_video_info(video_paths[1])
-
-        if pip_info["duration"] > main_info["duration"]:
-            temp_pip = os.path.join(self.work_dir, f"pip_temp_{os.getpid()}.mp4")
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-i",
-                video_paths[1],
-                "-t",
-                str(main_info["duration"]),
-                "-vf",
-                f"scale={pip_width}:{pip_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                temp_pip,
-            ]
-            run_ffmpeg(command)
-            pip_normalized_input = temp_pip
-        else:
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-i",
-                video_paths[1],
-                "-vf",
-                f"scale={pip_width}:{pip_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                pip_normalized,
-            ]
-            run_ffmpeg(command)
-            pip_normalized_input = pip_normalized
-
-        if main_info["duration"] > pip_info["duration"]:
-            looped_pip = os.path.join(self.work_dir, f"pip_looped_{os.getpid()}.mp4")
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-stream_loop",
-                "-1",
-                "-i",
-                pip_normalized_input,
-                "-t",
-                str(main_info["duration"]),
-                "-vf",
-                f"scale={pip_width}:{pip_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                looped_pip,
-            ]
-            run_ffmpeg(command)
-            pip_normalized_input = looped_pip
-
-        command = [
-            FFMPEG_BIN,
-            "-y",
-            "-i",
-            main_normalized,
-            "-i",
-            pip_normalized_input,
-            "-filter_complex",
-            f"[0:v][1:v]overlay={x_offset}:{y_offset}[v]",
-            "-map",
-            "[v]",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            output_path,
-        ]
-        run_ffmpeg(command)
-
-        for temp_file in [main_normalized, pip_normalized]:
-            if temp_file and temp_file != output_path:
-                try:
-                    os.remove(temp_file)
-                except Exception as e:
-                    logger.warning(
-                        f"Operation failed in apps/worker/video_processing/editing_modes.py: {e}", exc_info=True
-                    )
-
-        return output_path
-
-    def _voice_over(self, video_paths: list[str], audio_path: Optional[str], output_path: str) -> str:
-        """口播模式：背景画面 + 配音"""
-        if not audio_path:
-            raise ValueError("audio_path is required for VOICE_OVER mode")
-
-        if not video_paths:
-            raise ValueError("No background video provided")
-
-        audio_info = self._get_video_info(audio_path)
-        audio_duration = audio_info["duration"]
-
-        bg_normalized = os.path.join(self.work_dir, f"bg_{os.getpid()}.mp4")
-        bg_info = self._normalize_video(video_paths[0], bg_normalized)
-
-        if bg_info["duration"] < audio_duration:
-            looped_bg = os.path.join(self.work_dir, f"bg_looped_{os.getpid()}.mp4")
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-stream_loop",
-                "-1",
-                "-i",
-                bg_normalized,
-                "-t",
-                str(audio_duration),
-                "-vf",
-                f"scale={self.config.output_width}:{self.config.output_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                looped_bg,
-            ]
-            run_ffmpeg(command)
-            bg_normalized = looped_bg
-        elif bg_info["duration"] > audio_duration:
-            temp_bg = os.path.join(self.work_dir, f"bg_trimmed_{os.getpid()}.mp4")
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-i",
-                bg_normalized,
-                "-t",
-                str(audio_duration),
-                "-c:v",
-                "copy",
-                temp_bg,
-            ]
-            run_ffmpeg(command)
-            bg_normalized = temp_bg
-
-        blurred_bg = os.path.join(self.work_dir, f"bg_blurred_{os.getpid()}.mp4")
-        command = [
-            FFMPEG_BIN,
-            "-y",
-            "-i",
-            bg_normalized,
-            "-vf",
-            f"boxblur=5:5,scale={self.config.output_width}:{self.config.output_height}",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            blurred_bg,
-        ]
-        run_ffmpeg(command)
-
-        command = [
-            FFMPEG_BIN,
-            "-y",
-            "-i",
-            blurred_bg,
-            "-i",
-            audio_path,
-            "-filter_complex",
-            "[0:v]drawbox=x=0:y=0:w=iw:h=ih:color=black@0.3:t=fill[v]",
-            "-map",
-            "[v]",
-            "-map",
-            "1:a",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            "-shortest",
-            output_path,
-        ]
-        run_ffmpeg(command)
-
-        for temp_file in [bg_normalized, blurred_bg]:
-            try:
-                if temp_file != output_path:
-                    os.remove(temp_file)
-            except Exception as e:
-                logger.warning(f"Operation failed in apps/worker/video_processing/editing_modes.py: {e}", exc_info=True)
-
-        return output_path
-
-    def _voice_pip(self, video_paths: list[str], audio_path: Optional[str], output_path: str) -> str:
-        """口播+画中画模式：口播视频在角落，其他视频作为背景"""
-        if not video_paths:
-            raise ValueError("No video paths provided")
-
-        if len(video_paths) == 1:
-            return self._normalize_video(video_paths[0], output_path)
-
-        voice_video = video_paths[0]
-        bg_video = video_paths[1] if len(video_paths) > 1 else video_paths[0]
-
-        voice_normalized = os.path.join(self.work_dir, f"voice_{os.getpid()}.mp4")
-        voice_info = self._normalize_video(voice_video, voice_normalized)
-
-        bg_normalized = os.path.join(self.work_dir, f"bg_{os.getpid()}.mp4")
-        bg_info = self._normalize_video(bg_video, bg_normalized)
-
-        final_duration = min(voice_info["duration"], bg_info["duration"])
-
-        pip_width = int(self.config.output_width * self.config.pip_scale)
-        pip_height = int(self.config.output_height * self.config.pip_scale)
-        x_offset, y_offset = self._get_pip_position_offset(
-            self.config.output_width, self.config.output_height, pip_width, pip_height
-        )
-
-        voice_adjusted = os.path.join(self.work_dir, f"voice_adj_{os.getpid()}.mp4")
-        command = [
-            FFMPEG_BIN,
-            "-y",
-            "-i",
-            voice_normalized,
-            "-t",
-            str(final_duration),
-            "-vf",
-            f"scale={pip_width}:{pip_height}",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            voice_adjusted,
-        ]
-        run_ffmpeg(command)
-
-        bg_adjusted = os.path.join(self.work_dir, f"bg_adj_{os.getpid()}.mp4")
-        command = [
-            FFMPEG_BIN,
-            "-y",
-            "-i",
-            bg_normalized,
-            "-t",
-            str(final_duration),
-            "-c:v",
-            "copy",
-            bg_adjusted,
-        ]
-        run_ffmpeg(command)
-
-        if audio_path:
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-i",
-                bg_adjusted,
-                "-i",
-                voice_adjusted,
-                "-i",
-                audio_path,
-                "-filter_complex",
-                f"[0:v][1:v]overlay={x_offset}:{y_offset}[v]",
-                "-map",
-                "[v]",
-                "-map",
-                "2:a",
-                "-shortest",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                output_path,
-            ]
-        else:
-            command = [
-                FFMPEG_BIN,
-                "-y",
-                "-i",
-                bg_adjusted,
-                "-i",
-                voice_adjusted,
-                "-filter_complex",
-                f"[0:v][1:v]overlay={x_offset}:{y_offset}[v]",
-                "-map",
-                "[v]",
-                "-map",
-                "1:a",
-                "-shortest",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                output_path,
-            ]
-        run_ffmpeg(command)
-
-        for temp_file in [voice_normalized, voice_adjusted, bg_normalized, bg_adjusted]:
-            try:
-                if temp_file != output_path:
-                    os.remove(temp_file)
-            except Exception as e:
-                logger.warning(f"Operation failed in apps/worker/video_processing/editing_modes.py: {e}", exc_info=True)
-
-        return output_path
-
-
-def create_processor(mode: str, work_dir: Optional[str] = None, **kwargs) -> EditingModeProcessor:
-    """便捷工厂函数：创建剪辑模式处理器"""
-    try:
-        editing_mode = EditingMode(mode)
-    except ValueError:
-        raise ValueError(f"Invalid editing mode: {mode}. Valid modes: {[m.value for m in EditingMode]}")
-
-    config = EditingModeConfig(
-        mode=editing_mode,
-        output_width=kwargs.get("output_width", 1280),
-        output_height=kwargs.get("output_height", 720),
-        output_fps=kwargs.get("output_fps", 25),
-        pip_position=PIPPosition(kwargs.get("pip_position", "top_right")),
-        pip_scale=kwargs.get("pip_scale", 0.25),
-        transition_duration=kwargs.get("transition_duration", 0.5),
-    )
-
-    return EditingModeProcessor(config=config, work_dir=work_dir)

apps/worker/video_processing/ffmpeg_utils.py

@@ -1,8 +1,7 @@
-"""FFmpeg 工具函数 — 从 editing_modes.py / video_compose_service.py 提取的共享原语.
+"""FFmpeg 工具函数 — 共享原语.
 
 提供 FFmpeg / FFprobe 调用、视频信息探测、视频标准化、xfade 转场滤镜构建
-等底层能力，供 EditingModeProcessor、VideoComposeService、UnifiedRenderService
-共同复用。
+等底层能力，供 UnifiedRenderService、VideoComposeService 等复用。
 """
 
 from __future__ import annotations
@@ -32,6 +31,10 @@ XFADE_TRANSITION_MAP: dict[str, str] = {
     "slide_left": "slideleft",
     "slideright": "slideright",
     "slide_right": "slideright",
+    "slideup": "slideup",
+    "slide_up": "slideup",
+    "slidedown": "slidedown",
+    "slide_down": "slidedown",
     "dissolve": "dissolve",
     "wipe": "wipeleft",
     "wipeleft": "wipeleft",
@@ -39,6 +42,10 @@ XFADE_TRANSITION_MAP: dict[str, str] = {
 
 DEFAULT_TRANSITION_DURATION = 0.5
 
+# FFmpeg 执行默认超时（秒），防止 FFmpeg hang 住导致 worker 永久阻塞
+# 默认 30 分钟，足够处理大部分短视频渲染；超长视频可单独传参覆盖
+DEFAULT_FFMPEG_TIMEOUT = 1800
+
 
 # ── FFmpeg 执行 ───────────────────────────────────────────────────────────────
 
@@ -47,12 +54,14 @@ def run_ffmpeg(
     command: list[str],
     *,
     capture_output: bool = True,
+    timeout: int | None = DEFAULT_FFMPEG_TIMEOUT,
 ) -> tuple[str, str]:
     """执行 FFmpeg 命令。
 
     Args:
         command: 完整的 ffmpeg 命令列表（含 "ffmpeg" 本身）
         capture_output: 是否捕获 stdout/stderr
+        timeout: 超时时间（秒），默认 1800s（30分钟）；None 表示不设超时（不推荐）
 
     Returns:
         (stdout, stderr) 元组
@@ -60,6 +69,7 @@ def run_ffmpeg(
     Raises:
         subprocess.CalledProcessError: 命令执行失败时抛出，
             异常信息包含完整 stderr 以便排查。
+        subprocess.TimeoutExpired: 超时未完成时抛出，FFmpeg 进程会被 kill。
     """
     try:
         result = subprocess.run(  # nosec B603
@@ -68,8 +78,16 @@ def run_ffmpeg(
             stdout=subprocess.PIPE if capture_output else None,
             stderr=subprocess.PIPE if capture_output else None,
             text=True,
+            timeout=timeout,
         )
         return (result.stdout or "", result.stderr or "")
+    except subprocess.TimeoutExpired:
+        logger.error(
+            "FFmpeg 命令超时 (%ds): command=%s",
+            timeout or -1,
+            " ".join(str(c) for c in command[:20]),
+        )
+        raise
     except subprocess.CalledProcessError as e:
         # 把完整 stderr 打到日志，方便排查 exit code 183 等问题
         stderr_text = (e.stderr or "").strip()
@@ -82,6 +100,41 @@ def run_ffmpeg(
         raise
 
 
+def probe_has_audio(local_path: str | Path) -> bool:
+    """探测文件是否包含音频流。
+
+    Args:
+        local_path: 本地文件路径
+
+    Returns:
+        True 表示有音频流（或探测失败保守返回），False 表示确认无音频流
+    """
+    try:
+        result = subprocess.run(  # nosec B603
+            [
+                FFPROBE_BIN,
+                "-v",
+                "error",
+                "-select_streams",
+                "a:0",
+                "-show_entries",
+                "stream=codec_type",
+                "-of",
+                "default=noprint_wrappers=1:nokey=1",
+                str(local_path),
+            ],
+            check=True,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True,
+            timeout=10,
+        )
+        return result.stdout.strip() == "audio"
+    except Exception:
+        # 探测失败保守返回 True，让 FFmpeg 自己处理（避免误删音频）
+        return True
+
+
 def probe_duration(local_path: str | Path) -> float:
     """用 ffprobe 获取视频时长（秒）。
 
@@ -110,10 +163,14 @@ def probe_duration(local_path: str | Path) -> float:
 
 
 def probe_video_info(video_path: str) -> dict[str, Any]:
-    """获取视频信息（宽、高、时长、fps）。
+    """获取视频信息（宽、高、时长、fps、编码、像素格式）。
 
     Returns:
-        {"width": int, "height": int, "duration": float, "fps": float}
+        {
+            "width": int, "height": int, "duration": float, "fps": float,
+            "video_codec": str, "audio_codec": str, "pix_fmt": str,
+            "has_audio": bool,
+        }
         失败时返回默认值。
     """
     try:
@@ -122,10 +179,8 @@ def probe_video_info(video_path: str) -> dict[str, Any]:
                 FFPROBE_BIN,
                 "-v",
                 "error",
-                "-select_streams",
-                "v:0",
                 "-show_entries",
-                "stream=width,height,r_frame_rate,duration",
+                "stream=width,height,r_frame_rate,duration,codec_name,codec_type,pix_fmt",
                 "-show_entries",
                 "format=duration",
                 "-of",
@@ -136,19 +191,25 @@ def probe_video_info(video_path: str) -> dict[str, Any]:
             stdout=subprocess.PIPE,
             stderr=subprocess.PIPE,
             text=True,
+            timeout=15,
         )
 
         import json
 
         info = json.loads(result.stdout)
-        stream = info.get("streams", [{}])[0]
+        streams = info.get("streams", [])
         fmt = info.get("format", {})
 
-        width = int(stream.get("width", DEFAULT_OUTPUT_WIDTH))
-        height = int(stream.get("height", DEFAULT_OUTPUT_HEIGHT))
+        video_stream = next((s for s in streams if s.get("codec_type") == "video"), {})
+        audio_stream = next((s for s in streams if s.get("codec_type") == "audio"), {})
+
+        width = int(video_stream.get("width", DEFAULT_OUTPUT_WIDTH))
+        height = int(video_stream.get("height", DEFAULT_OUTPUT_HEIGHT))
+        video_codec = video_stream.get("codec_name", "") or ""
+        pix_fmt = video_stream.get("pix_fmt", "") or ""
 
         # 解析帧率
-        fps_str = stream.get("r_frame_rate", "25/1")
+        fps_str = video_stream.get("r_frame_rate", "25/1")
         if "/" in fps_str:
             num, den = fps_str.split("/")
             fps = float(num) / float(den) if float(den) > 0 else DEFAULT_FPS
@@ -156,13 +217,20 @@ def probe_video_info(video_path: str) -> dict[str, Any]:
             fps = float(fps_str) if fps_str else DEFAULT_FPS
 
         # 时长
-        duration = float(fmt.get("duration", 0)) or float(stream.get("duration", 0))
+        duration = float(fmt.get("duration", 0)) or float(video_stream.get("duration", 0))
+
+        has_audio = bool(audio_stream)
+        audio_codec = audio_stream.get("codec_name", "") or ""
 
         return {
             "width": width,
             "height": height,
             "duration": duration,
             "fps": round(fps, 2),
+            "video_codec": video_codec,
+            "audio_codec": audio_codec,
+            "pix_fmt": pix_fmt,
+            "has_audio": has_audio,
         }
     except Exception as e:
         logger.warning("获取视频信息失败: %s, error: %s", video_path, e)
@@ -171,6 +239,10 @@ def probe_video_info(video_path: str) -> dict[str, Any]:
             "height": DEFAULT_OUTPUT_HEIGHT,
             "duration": 0.0,
             "fps": DEFAULT_FPS,
+            "video_codec": "",
+            "audio_codec": "",
+            "pix_fmt": "",
+            "has_audio": True,
         }
 
 

apps/worker/video_processing/oss_helpers.py

@@ -9,14 +9,21 @@ from __future__ import annotations
 import hashlib
 import logging
 import os
+import threading
 from pathlib import Path
-from typing import Optional
 from urllib.parse import urlparse
 
 import oss2
 
 logger = logging.getLogger(__name__)
 
+# OSS 上传配置
+OSS_CONNECT_TIMEOUT = 10  # 连接超时（秒），防止 TCP 握手挂死
+OSS_UPLOAD_TOTAL_TIMEOUT = 300  # 单文件上传总超时（秒），防止网络慢时无限卡住
+OSS_MULTIPART_THRESHOLD = 100 * 1024 * 1024  # 分片上传阈值：100MB 以上走分片
+OSS_PART_SIZE = 8 * 1024 * 1024  # 分片大小：8MB
+OSS_MULTIPART_NUM_THREADS = 3  # 分片上传并发数
+
 
 # ── OSS 配置 ──────────────────────────────────────────────────────────────────
 
@@ -43,6 +50,9 @@ def oss_bucket() -> oss2.Bucket | None:
     P0-2 修复：endpoint 不带 scheme 时自动补 https:// 前缀，
     确保 sign_url 等依赖 scheme 的方法返回 HTTPS URL。
 
+    P0-staging 修复：增加 connect_timeout=10s，防止网络抖动时
+    TCP 握手阶段无限挂死，导致 worker 进程卡死。
+
     Returns:
         oss2.Bucket 实例，配置缺失时返回 None。
     """
@@ -53,7 +63,12 @@ def oss_bucket() -> oss2.Bucket | None:
     # endpoint 无 scheme 时补 https://，与 API 端 storage.py 保持一致
     if not endpoint.startswith(("http://", "https://")):
         endpoint = f"https://{endpoint}"
-    return oss2.Bucket(oss2.Auth(access_key_id, access_key_secret), endpoint, bucket_name)
+    return oss2.Bucket(
+        oss2.Auth(access_key_id, access_key_secret),
+        endpoint,
+        bucket_name,
+        connect_timeout=OSS_CONNECT_TIMEOUT,
+    )
 
 
 def normalize_storage_key(storage_key_or_url: str) -> str:
@@ -96,6 +111,9 @@ def download_asset(asset_storage_key: str, local_path: Path) -> bool:
 def upload_to_oss(local_path: Path, storage_key: str) -> str | None:
     """上传文件到 OSS，返回公开 URL。
 
+    大文件（>100MB）自动走分片上传，降低内存峰值，减少 OOM 风险。
+    上传加总超时保护（默认 300s），防止网络异常时无限挂死。
+
     Args:
         local_path: 本地文件路径
         storage_key: 目标存储键
@@ -106,18 +124,71 @@ def upload_to_oss(local_path: Path, storage_key: str) -> str | None:
     bucket = oss_bucket()
     if bucket is None:
         return None
-    try:
-        bucket.put_object_from_file(storage_key, str(local_path))
-        settings = oss_settings()
-        if settings:
-            _, _, endpoint, bucket_name = settings
-            endpoint_clean = endpoint.replace("https://", "").replace("http://", "")
-            return f"https://{bucket_name}.{endpoint_clean}/{storage_key}"
+
+    result: dict = {"url": None, "error": None, "file_size": 0}
+    done = threading.Event()
+
+    def _do_upload():
+        try:
+            # 尝试获取文件大小，用于分片判断和日志；stat 失败时 fallback 走普通上传
+            try:
+                file_size = local_path.stat().st_size
+                result["file_size"] = file_size
+                use_multipart = file_size >= OSS_MULTIPART_THRESHOLD
+            except OSError:
+                use_multipart = False
+                file_size = 0
+
+            if use_multipart:
+                # 分片上传：降低内存峰值，每片 8MB，3 线程并发
+                logger.info(
+                    "大文件分片上传: storage_key=%s, size=%.1fMB, part_size=%dMB, threads=%d",
+                    storage_key[:80],
+                    file_size / 1024 / 1024,
+                    OSS_PART_SIZE // 1024 // 1024,
+                    OSS_MULTIPART_NUM_THREADS,
+                )
+                oss2.resumable_upload(
+                    bucket,
+                    storage_key,
+                    str(local_path),
+                    multipart_threshold=OSS_MULTIPART_THRESHOLD,
+                    part_size=OSS_PART_SIZE,
+                    num_threads=OSS_MULTIPART_NUM_THREADS,
+                )
+            else:
+                bucket.put_object_from_file(storage_key, str(local_path))
+
+            # 构造返回 URL
+            settings = oss_settings()
+            if settings:
+                _, _, endpoint, bucket_name = settings
+                endpoint_clean = endpoint.replace("https://", "").replace("http://", "")
+                result["url"] = f"https://{bucket_name}.{endpoint_clean}/{storage_key}"
+        except Exception as e:
+            result["error"] = e
+            logger.exception("上传 OSS 失败: %s", storage_key)
+        finally:
+            done.set()
+
+    upload_thread = threading.Thread(target=_do_upload, daemon=True)
+    upload_thread.start()
+    finished = done.wait(timeout=OSS_UPLOAD_TOTAL_TIMEOUT)
+
+    if not finished:
+        logger.error(
+            "OSS 上传超时（%.0fs），强制中止: storage_key=%s, size=%.1fMB",
+            OSS_UPLOAD_TOTAL_TIMEOUT,
+            storage_key[:80],
+            result["file_size"] / 1024 / 1024 if result["file_size"] else 0,
+        )
         return None
-    except Exception:
-        logger.exception("上传 OSS 失败: %s", storage_key)
+
+    if result["error"]:
         return None
 
+    return result["url"]
+
 
 def get_signed_download_url(storage_key_or_url: str, expires_seconds: int = 3600) -> str | None:
     """生成预签名下载 URL（用于私有 bucket 的 URL 校验或临时下载）。

apps/worker/video_processing/processor.py

@@ -2,14 +2,16 @@
 视频处理核心类
 """
 
+import logging
 import os
 import tempfile
 from dataclasses import dataclass
-from pathlib import Path
 from typing import List
 
 import ffmpeg
 
+logger = logging.getLogger(__name__)
+
 
 @dataclass
 class VideoResult:
@@ -131,8 +133,8 @@ class VideoProcessor:
             if concat_file is not None:
                 try:
                     concat_file.close()
-                except Exception:
-                    pass  # 忽略关闭时的错误
+                except OSError as close_err:
+                    logger.warning("临时文件关闭失败: %s", close_err)
 
     def generate_thumbnail(
         self,

apps/worker/video_processing/render_adapter.py

@@ -0,0 +1,299 @@
+"""统一渲染引擎适配层 — Phase 2.
+
+将 EditPlan + EditPlanClips（来自 DB）适配为 UnifiedRenderService 的输入格式，
+封装素材下载、渲染执行、结果上传的完整流程。
+
+职责：
+  1. 从 DB 读取 EditPlan + EditPlanClips
+  2. 下载素材到本地，构建 asset_path_map
+  3. 调用 UnifiedRenderService 执行渲染
+  4. 上传渲染结果到 OSS
+  5. 支持进度回调（对接 JobService）
+"""
+
+from __future__ import annotations
+
+import logging
+import tempfile
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Callable
+
+from sqlalchemy.orm import Session
+from video_processing.oss_helpers import download_asset, upload_to_oss
+from video_processing.unified_render_service import UnifiedRenderService
+
+from packages.adapters.sqlalchemy_impl.edit_plan_clip_repository import SQLAlchemyEditPlanClipRepository
+from packages.adapters.sqlalchemy_impl.edit_plan_repository import SQLAlchemyEditPlanRepository
+from packages.domain.edit_plan import EditPlanStatus
+from packages.domain.edit_plan_clip import EditPlanClip, EditPlanClipStatus
+
+logger = logging.getLogger(__name__)
+
+
+# ── 数据结构 ──────────────────────────────────────────────────────────────────
+
+
+@dataclass
+class RenderAdapterResult:
+    """渲染适配结果。"""
+
+    success: bool
+    output_url: str = ""
+    output_path: Path | None = None
+    duration: float = 0.0
+    file_size: int = 0
+    width: int = 0
+    height: int = 0
+    clip_count: int = 0
+    error_message: str = ""
+
+
+ProgressCallback = Callable[[float, str], None]
+"""进度回调：(progress_0_100, stage_description) → None"""
+
+
+# ── 适配层主体 ────────────────────────────────────────────────────────────────
+
+
+class RenderAdapter:
+    """统一渲染引擎适配层。
+
+    桥接 EditPlan 领域模型与 UnifiedRenderService 图层模型。
+
+    用法::
+
+        adapter = RenderAdapter(db)
+        result = adapter.render_plan(
+            plan_id=plan_id,
+            job_id=job_id,
+            progress_cb=lambda p, s: job_service.update_progress(job_id, p, s),
+        )
+    """
+
+    def __init__(self, db: Session) -> None:
+        self._db = db
+        self._plan_repo = SQLAlchemyEditPlanRepository(db)
+        self._clip_repo = SQLAlchemyEditPlanClipRepository(db)
+
+    # ── 公开方法 ──────────────────────────────────────────────────────────
+
+    def render_plan(
+        self,
+        plan_id: str,
+        *,
+        job_id: str = "",
+        work_dir: Path | None = None,
+        progress_cb: ProgressCallback | None = None,
+    ) -> RenderAdapterResult:
+        """渲染一个 EditPlan。
+
+        完整流程：
+        1. 加载计划与片段
+        2. 下载素材
+        3. 执行统一渲染
+        4. 上传结果
+
+        Args:
+            plan_id: EditPlan ID
+            job_id: 关联的 Job ID（用于结果存储路径）
+            work_dir: 工作目录，不传则使用临时目录
+            progress_cb: 进度回调函数
+
+        Returns:
+            RenderAdapterResult
+        """
+        temp_dir = None
+        try:
+            # 0. 准备工作目录
+            if work_dir is None:
+                temp_dir = tempfile.mkdtemp(prefix="render_")
+                work_dir = Path(temp_dir)
+            work_dir.mkdir(parents=True, exist_ok=True)
+
+            self._report_progress(progress_cb, 5.0, "加载剪辑计划")
+
+            # 1. 加载计划与片段
+            plan = self._plan_repo.get(plan_id)
+            if plan is None:
+                return RenderAdapterResult(
+                    success=False,
+                    error_message=f"剪辑计划不存在: {plan_id}",
+                )
+
+            clips = self._clip_repo.list_by_plan(plan_id, skip=0, limit=10000)
+            ready_clips = [c for c in clips if c.status == EditPlanClipStatus.READY and c.asset_id]
+            ready_clips.sort(key=lambda c: c.order)
+
+            if not ready_clips:
+                return RenderAdapterResult(
+                    success=False,
+                    error_message="没有可渲染的就绪片段",
+                    clip_count=0,
+                )
+
+            logger.info(
+                "开始渲染: plan_id=%s job_id=%s ready_clips=%d engine=unified",
+                plan_id,
+                job_id,
+                len(ready_clips),
+            )
+
+            self._report_progress(progress_cb, 15.0, f"下载素材（{len(ready_clips)} 个）")
+
+            # 2. 下载素材
+            asset_path_map = self._download_assets(ready_clips, work_dir)
+            if not asset_path_map:
+                return RenderAdapterResult(
+                    success=False,
+                    error_message="所有素材下载失败",
+                    clip_count=len(ready_clips),
+                )
+
+            self._report_progress(progress_cb, 40.0, "执行视频渲染")
+
+            # 3. 执行统一渲染
+            render_svc = UnifiedRenderService(
+                plan=plan,
+                clips=ready_clips,
+                asset_path_map=asset_path_map,
+                work_dir=work_dir,
+            )
+            result = render_svc.render()
+
+            self._report_progress(progress_cb, 80.0, "上传渲染结果")
+
+            # 4. 上传结果
+            storage_key = f"rendered/{plan_id}/{job_id or plan_id}.mp4"
+            output_url = upload_to_oss(result.output_path, storage_key)
+
+            self._report_progress(progress_cb, 100.0, "渲染完成")
+
+            logger.info(
+                "[render-adapter] render success: plan_id=%s job_id=%s engine=unified "
+                "duration=%.2fs file_size=%d resolution=%dx%d clip_count=%d",
+                plan_id,
+                job_id,
+                result.duration,
+                result.file_size,
+                result.width,
+                result.height,
+                len(ready_clips),
+            )
+
+            return RenderAdapterResult(
+                success=True,
+                output_url=output_url or "",
+                output_path=result.output_path,
+                duration=result.duration,
+                file_size=result.file_size,
+                width=result.width,
+                height=result.height,
+                clip_count=len(ready_clips),
+            )
+
+        except Exception as exc:
+            logger.exception(
+                "[render-adapter] render failed: plan_id=%s job_id=%s engine=unified error=%s",
+                plan_id,
+                job_id,
+                str(exc)[:200],
+            )
+            return RenderAdapterResult(
+                success=False,
+                error_message=str(exc)[:500],
+            )
+        finally:
+            # 清理临时目录
+            if temp_dir:
+                import shutil
+
+                try:
+                    shutil.rmtree(temp_dir, ignore_errors=True)
+                except Exception as cleanup_err:
+                    logger.warning("临时目录清理失败: path=%s error=%s", temp_dir, cleanup_err)
+
+    def validate_plan(self, plan_id: str) -> tuple[bool, list[str], list[str], int, int]:
+        """校验计划是否可渲染（兼容 VideoComposeService.validate_compose 接口）。
+
+        Returns:
+            (valid, errors, warnings, ready_clip_count, total_clip_count)
+        """
+        errors: list[str] = []
+        warnings: list[str] = []
+
+        plan = self._plan_repo.get(plan_id)
+        if plan is None:
+            return False, [f"剪辑计划不存在: {plan_id}"], [], 0, 0
+
+        if plan.status not in (EditPlanStatus.EDITING, EditPlanStatus.RENDERING):
+            errors.append(f"计划状态不正确，需要 editing 或 rendering，当前: {plan.status}")
+
+        clips = self._clip_repo.list_by_plan(plan_id, skip=0, limit=10000)
+        if not clips:
+            errors.append("计划没有任何片段")
+            return False, errors, warnings, 0, 0
+
+        clips.sort(key=lambda c: c.order)
+
+        ready_count = 0
+        pending_count = 0
+        no_asset_count = 0
+
+        for clip in clips:
+            if clip.status == EditPlanClipStatus.READY:
+                ready_count += 1
+                if not clip.asset_id:
+                    errors.append(f"片段 {clip.id} (order={clip.order}) 没有分配素材")
+                    no_asset_count += 1
+            elif clip.status == EditPlanClipStatus.PENDING:
+                pending_count += 1
+            elif clip.status == EditPlanClipStatus.FAILED:
+                warnings.append(f"片段 {clip.id} (order={clip.order}) 状态为 failed，已跳过")
+
+        if ready_count == 0:
+            errors.append("没有就绪（ready）的片段可以合成")
+
+        if pending_count > 0:
+            warnings.append(f"有 {pending_count} 个片段仍处于 pending 状态")
+
+        return len(errors) == 0, errors, warnings, ready_count, len(clips)
+
+    # ── 内部方法 ──────────────────────────────────────────────────────────
+
+    @staticmethod
+    def _report_progress(progress_cb: ProgressCallback | None, progress: float, stage: str) -> None:
+        """上报进度。"""
+        if progress_cb is not None:
+            try:
+                progress_cb(progress, stage)
+            except Exception:
+                logger.exception("进度回调失败")
+
+    @staticmethod
+    def _download_assets(clips: list[EditPlanClip], work_dir: Path) -> dict[str, Path]:
+        """下载片段素材到本地，返回 asset_id → local_path 映射。
+
+        只保留下载成功的素材。
+        """
+        asset_dir = work_dir / "assets"
+        asset_dir.mkdir(exist_ok=True)
+
+        asset_path_map: dict[str, Path] = {}
+
+        for clip in clips:
+            asset_id = clip.asset_id
+            if not asset_id:
+                continue
+
+            # 生成安全的本地文件名
+            safe_name = f"clip_{clip.order:04d}_{abs(hash(asset_id)) % 100000:05d}.mp4"
+            local_path = asset_dir / safe_name
+
+            if download_asset(asset_id, local_path):
+                asset_path_map[asset_id] = local_path
+                logger.debug("素材下载成功: clip_id=%s asset_id=%s", clip.id, asset_id[:60])
+            else:
+                logger.warning("素材下载失败: clip_id=%s asset_id=%s", clip.id, asset_id[:60])
+
+        return asset_path_map

apps/worker/video_processing/render_audio.py

@@ -0,0 +1,364 @@
+"""音频混音模块 — 从 unified_render_service.py 拆分.
+
+职责：
+- 主图层音频 concat 拼接
+- 独立音频轨 amix 混音
+- 音视频合并（mux）
+
+所有函数接收 RenderContext 获取共享依赖（work_dir、plan_id 等），
+避免直接依赖 UnifiedRenderService 类。
+"""
+
+from __future__ import annotations
+
+import logging
+import subprocess
+from dataclasses import dataclass, field
+from pathlib import Path
+
+# 延迟导入避免循环依赖：unified_render_service 定义 ResolvedClip / RenderLayer，
+# 本模块提供音频函数供 unified_render_service 调用。
+# 使用 from __future__ import annotations + TYPE_CHECKING 解决类型引用。
+from typing import TYPE_CHECKING
+
+from video_processing.ffmpeg_utils import FFMPEG_BIN, probe_has_audio, run_ffmpeg
+
+if TYPE_CHECKING:
+    from video_processing.unified_render_service import RenderLayer, ResolvedClip
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class RenderContext:
+    """渲染上下文 — 提供音频混音所需的共享依赖."""
+
+    work_dir: Path
+    plan_id: str
+    # 音频探测缓存（避免同一 clip 被多次 ffprobe）
+    _audio_cache: dict[str, bool] = field(default_factory=dict)
+
+
+# ── 工具函数 ──────────────────────────────────────────────────────────────────
+
+
+def clip_effective_duration(clip: ResolvedClip) -> float:
+    """计算 clip 的有效时长.
+
+    与 UnifiedRenderService._clip_effective_duration 逻辑一致。
+    """
+    if clip.duration > 0:
+        return min(clip.duration, clip.actual_duration) if clip.actual_duration > 0 else clip.duration
+    return clip.actual_duration if clip.actual_duration > 0 else 0.0
+
+
+def clip_has_audio(ctx: RenderContext, clip: ResolvedClip) -> bool:
+    """探测 clip 是否有音频流（带缓存）.
+
+    避免同一个 clip 被多次 ffprobe 探测。
+    """
+    key = str(clip.local_path)
+    if key not in ctx._audio_cache:
+        ctx._audio_cache[key] = probe_has_audio(clip.local_path)
+    return ctx._audio_cache[key]
+
+
+# ── 音频混音 ──────────────────────────────────────────────────────────────────
+
+
+def mix_audio(
+    ctx: RenderContext,
+    layers: list[RenderLayer],
+    video_duration: float,
+) -> Path | None:
+    """音频后处理混音.
+
+    处理逻辑：
+    1. 主音频源按优先级查找：main > broll（background 不参与主音频，通常是图片无音轨）
+    2. 主图层音频按顺序 concat 拼接
+    3. 独立音频轨（audio role）用 amix 混入
+    4. 输出时长截断到 video_duration
+    5. 无音频流的 clip 会被自动跳过，避免 FFmpeg 引用 [i:a] 失败
+
+    Args:
+        ctx: 渲染上下文
+        layers: 图层列表
+        video_duration: 视频总时长（用于截断音频）
+
+    Returns:
+        混音后的音频文件路径，无音频时返回 None
+    """
+    # 按优先级精确查找主音频图层：main > broll
+    # background 不参与主音频（通常是静态图片，无音轨）
+    layer_map = {layer.role: layer for layer in layers}
+    main_layer = None
+    for role in ("main", "broll"):
+        if role in layer_map and layer_map[role].clips:
+            main_layer = layer_map[role]
+            break
+
+    main_clips: list[ResolvedClip] = main_layer.clips if main_layer else []
+
+    # 没有主视频图层时兜底：检查 overlay/corner_voice 层是否有带音频的素材
+    if not main_clips:
+        for role in ("overlay", "corner_voice"):
+            if role in layer_map and layer_map[role].clips:
+                main_clips = layer_map[role].clips
+                break
+
+    # 收集独立音频轨
+    audio_clips: list[ResolvedClip] = []
+    if "audio" in layer_map:
+        audio_clips = layer_map["audio"].clips
+
+    # ── 防御：过滤掉无音频流的 clip ──
+    main_clips = [c for c in main_clips if clip_has_audio(ctx, c)]
+    audio_clips = [c for c in audio_clips if clip_has_audio(ctx, c)]
+
+    if not main_clips and not audio_clips:
+        return None
+
+    # 构建音频处理命令
+    output_path = ctx.work_dir / f"audio_{ctx.plan_id}.aac"
+
+    # 简单场景：只有主图层 + 无独立音频 → 直接从视频提取音频并拼接
+    if main_clips and not audio_clips:
+        concat_main_audio(ctx, main_clips, output_path, video_duration)
+        return output_path
+
+    # 有独立音频轨 → amix 混音
+    mix_with_independent_audio(ctx, main_clips, audio_clips, output_path, video_duration)
+    return output_path
+
+
+def concat_main_audio(
+    ctx: RenderContext,
+    clips: list[ResolvedClip],
+    output_path: Path,
+    video_duration: float,
+) -> None:
+    """主图层音频 concat 拼接（对齐链路A行为）.
+
+    每个 clip 提取音频 → trim → 按顺序 concat。
+    """
+    if len(clips) == 1:
+        # 单 clip，直接提取音频，截断到 min(clip有效时长, 视频总时长)
+        clip = clips[0]
+        effective_duration = clip_effective_duration(clip)
+        # 最终时长：取 clip 有效时长和视频总时长的较小值
+        # （视频总时长由主图层决定，但单 clip 场景下两者应该一致，仍做保护）
+        final_duration = effective_duration
+        if video_duration > 0 and (final_duration <= 0 or final_duration > video_duration):
+            final_duration = video_duration
+
+        command = [
+            FFMPEG_BIN,
+            "-y",
+            "-i",
+            str(clip.local_path),
+            "-vn",
+            "-acodec",
+            "aac",
+            "-b:a",
+            "128k",
+        ]
+        if final_duration > 0:
+            command.extend(["-t", f"{final_duration:.3f}"])
+        command.append(str(output_path))
+        run_ffmpeg(command)
+        return
+
+    # 多 clip，用 filter_complex concat
+    input_args: list[str] = []
+    filter_parts: list[str] = []
+
+    for i, clip in enumerate(clips):
+        input_args.extend(["-i", str(clip.local_path)])
+        effective_duration = clip_effective_duration(clip)
+        if effective_duration > 0:
+            filter_parts.append(f"[{i}:a]atrim=0:{effective_duration:.3f},asetpts=PTS-STARTPTS[a{i}]")
+        else:
+            filter_parts.append(f"[{i}:a]asetpts=PTS-STARTPTS[a{i}]")
+
+    audio_labels = "".join(f"[a{i}]" for i in range(len(clips)))
+    filter_parts.append(f"{audio_labels}concat=n={len(clips)}:v=0:a=1[outa]")
+
+    # 截断到视频总时长
+    if video_duration > 0:
+        filter_parts.append(f"[outa]atrim=0:{video_duration:.3f}[final_audio]")
+        final_label = "final_audio"
+    else:
+        final_label = "outa"
+
+    filter_complex = ";".join(filter_parts)
+
+    command = [
+        FFMPEG_BIN,
+        "-y",
+        *input_args,
+        "-filter_complex",
+        filter_complex,
+        "-map",
+        f"[{final_label}]",
+        "-acodec",
+        "aac",
+        "-b:a",
+        "128k",
+        str(output_path),
+    ]
+    run_ffmpeg(command)
+
+
+def mix_with_independent_audio(
+    ctx: RenderContext,
+    main_clips: list[ResolvedClip],
+    audio_clips: list[ResolvedClip],
+    output_path: Path,
+    video_duration: float,
+) -> None:
+    """主音频 + 独立音频轨 amix 混音.
+
+    Args:
+        ctx: 渲染上下文
+        main_clips: 主视频 clips（提取音频后 concat）
+        audio_clips: 独立音频轨 clips
+        output_path: 输出路径
+        video_duration: 视频总时长
+    """
+    input_args: list[str] = []
+    filter_parts: list[str] = []
+    mix_labels: list[str] = []
+
+    input_idx = 0
+
+    # 1. 主图层音频 concat
+    if main_clips:
+        for clip in main_clips:
+            input_args.extend(["-i", str(clip.local_path)])
+            effective_duration = clip_effective_duration(clip)
+            if effective_duration > 0:
+                filter_parts.append(
+                    f"[{input_idx}:a]atrim=0:{effective_duration:.3f},asetpts=PTS-STARTPTS[ma{input_idx}]"
+                )
+            else:
+                filter_parts.append(f"[{input_idx}:a]asetpts=PTS-STARTPTS[ma{input_idx}]")
+            input_idx += 1
+
+        if len(main_clips) == 1:
+            mix_labels.append("ma0")
+        else:
+            main_labels = "".join(f"[ma{i}]" for i in range(len(main_clips)))
+            filter_parts.append(f"{main_labels}concat=n={len(main_clips)}:v=0:a=1[main_audio]")
+            mix_labels.append("main_audio")
+
+    # 2. 独立音频轨
+    for j, clip in enumerate(audio_clips):
+        input_args.extend(["-i", str(clip.local_path)])
+        effective_duration = clip_effective_duration(clip)
+        volume = clip.config.get("volume", 1.0) if clip.config else 1.0
+        label = f"ia{j}"
+        filters = []
+        if effective_duration > 0:
+            filters.append(f"atrim=0:{effective_duration:.3f}")
+        filters.append("asetpts=PTS-STARTPTS")
+        if volume != 1.0:
+            filters.append(f"volume={volume}")
+        filter_parts.append(f"[{input_idx}:a]{','.join(filters)}[{label}]")
+        mix_labels.append(label)
+        input_idx += 1
+
+    # 3. amix 混音
+    mix_inputs = "".join(f"[{label}]" for label in mix_labels)
+    n_inputs = len(mix_labels)
+    # normalized=0 保持音量，duration=shortest 取最短
+    filter_parts.append(f"{mix_inputs}amix=inputs={n_inputs}:duration=longest:normalize=0[mixed_audio]")
+
+    # 4. 截断到视频时长
+    if video_duration > 0:
+        filter_parts.append(f"[mixed_audio]atrim=0:{video_duration:.3f}[final_audio]")
+        final_label = "final_audio"
+    else:
+        final_label = "mixed_audio"
+
+    filter_complex = ";".join(filter_parts)
+
+    command = [
+        FFMPEG_BIN,
+        "-y",
+        *input_args,
+        "-filter_complex",
+        filter_complex,
+        "-map",
+        f"[{final_label}]",
+        "-acodec",
+        "aac",
+        "-b:a",
+        "128k",
+        str(output_path),
+    ]
+
+    logger.info(
+        "音频混音: plan_id=%s main_clips=%d audio_clips=%d",
+        ctx.plan_id,
+        len(main_clips),
+        len(audio_clips),
+    )
+    try:
+        run_ffmpeg(command)
+    except subprocess.CalledProcessError as e:
+        logger.error(
+            "音频混音失败: plan_id=%s exit_code=%d\nfilter_complex:\n%s",
+            ctx.plan_id,
+            e.returncode,
+            filter_complex[:3000],
+        )
+        raise
+
+
+def merge_audio_video(
+    ctx: RenderContext,
+    video_path: Path,
+    audio_path: Path,
+    output_path: Path,
+) -> None:
+    """将音频合并到视频中（视频流拷贝，音频直接复用）.
+
+    Args:
+        ctx: 渲染上下文
+        video_path: 无声视频路径
+        audio_path: 音频文件路径
+        output_path: 输出文件路径
+    """
+    command = [
+        FFMPEG_BIN,
+        "-y",
+        "-i",
+        str(video_path),
+        "-i",
+        str(audio_path),
+        "-c:v",
+        "copy",
+        "-c:a",
+        "aac",
+        "-b:a",
+        "128k",
+        "-map",
+        "0:v:0",
+        "-map",
+        "1:a:0",
+        "-shortest",
+        "-movflags",
+        "+faststart",
+        str(output_path),
+    ]
+
+    logger.info("合并音视频: plan_id=%s", ctx.plan_id)
+    try:
+        run_ffmpeg(command)
+    except subprocess.CalledProcessError as e:
+        logger.error(
+            "合并音视频失败: plan_id=%s exit_code=%d",
+            ctx.plan_id,
+            e.returncode,
+        )
+        raise

apps/worker/video_processing/render_engine_resolver.py

@@ -0,0 +1,204 @@
+"""渲染引擎 Feature Flag 解析器。
+
+封装渲染引擎选择逻辑，支持：
+- 环境变量作为默认值（RENDER_ENGINE=legacy/unified）
+- Redis Feature Flag 运行时覆盖（白名单 + 百分比 + 全局开关）
+- 定时刷新，支持热更新不重启 worker
+
+使用方式：
+    resolver = RenderEngineResolver(redis_url="redis://...", default_engine="legacy")
+    engine = resolver.get_engine(user_id="user123")
+    # engine: "legacy" 或 "unified"
+"""
+
+from __future__ import annotations
+
+import logging
+import threading
+from typing import Optional
+
+from packages.adapters.redis.feature_flag_store import (
+    FeatureFlagConfig,
+    FeatureFlagStore,
+    InMemoryFeatureFlagStore,
+    RedisFeatureFlagStore,
+)
+
+logger = logging.getLogger(__name__)
+
+# Feature Flag 名称常量
+FLAG_RENDER_ENGINE = "render_engine"
+
+# 引擎常量
+ENGINE_LEGACY = "legacy"
+ENGINE_UNIFIED = "unified"
+VALID_ENGINES = {ENGINE_LEGACY, ENGINE_UNIFIED}
+
+
+class RenderEngineResolver:
+    """渲染引擎选择器。
+
+    判定逻辑（从高到低）：
+    1. Redis flag 白名单匹配 → unified
+    2. Redis flag 百分比命中 → unified
+    3. Redis flag 全局开启（100%）→ unified
+    4. 环境变量默认值 → legacy / unified
+
+    当 Redis 不可用时，自动降级到环境变量默认值，不影响业务。
+    """
+
+    def __init__(
+        self,
+        default_engine: str = ENGINE_LEGACY,
+        redis_url: Optional[str] = None,
+        refresh_interval: float = 30.0,
+        store: Optional[FeatureFlagStore] = None,
+    ) -> None:
+        """
+        Args:
+            default_engine: 环境变量默认的引擎名（legacy / unified）
+            redis_url: Redis 连接 URL，传 None 时使用内存实现（测试用）
+            refresh_interval: Redis flag 配置刷新间隔（秒）
+            store: 直接传入 store 实例（测试用，优先级高于 redis_url）
+        """
+        self._default_engine = default_engine.lower() if default_engine else ENGINE_LEGACY
+        if self._default_engine not in VALID_ENGINES:
+            logger.warning(
+                "Invalid default engine '%s', fallback to '%s'",
+                self._default_engine,
+                ENGINE_LEGACY,
+            )
+            self._default_engine = ENGINE_LEGACY
+
+        if store is not None:
+            self._store = store
+        elif redis_url:
+            self._store = RedisFeatureFlagStore(redis_url=redis_url)
+        else:
+            self._store = InMemoryFeatureFlagStore()
+            logger.info("No Redis configured, using in-memory feature flag store")
+
+        self._refresh_interval = refresh_interval
+        self._lock = threading.Lock()
+        self._cached_config: Optional[FeatureFlagConfig] = None
+        self._last_refresh: float = 0.0
+
+    def _maybe_refresh(self) -> None:
+        """惰性刷新配置，超过刷新间隔时从存储重新读取。"""
+        import time
+
+        now = time.time()
+        if now - self._last_refresh < self._refresh_interval:
+            return
+
+        try:
+            config = self._store.get(FLAG_RENDER_ENGINE)
+            with self._lock:
+                self._cached_config = config
+                self._last_refresh = now
+        except Exception as exc:
+            logger.warning("Failed to refresh render engine flag: %s", exc)
+            # 刷新失败时保留旧缓存，不中断业务
+            if self._cached_config is None:
+                # 首次就读失败，设一个默认值
+                with self._lock:
+                    self._cached_config = FeatureFlagConfig(name=FLAG_RENDER_ENGINE)
+                    self._last_refresh = now
+
+    def _get_config(self) -> FeatureFlagConfig:
+        """获取当前 flag 配置（带缓存）。"""
+        if self._cached_config is None:
+            self._maybe_refresh()
+        else:
+            self._maybe_refresh()
+        return self._cached_config or FeatureFlagConfig(name=FLAG_RENDER_ENGINE)
+
+    def get_engine(self, user_id: Optional[str] = None) -> str:
+        """获取当前应该使用的渲染引擎。
+
+        Args:
+            user_id: 用户ID，用于白名单匹配和百分比哈希。
+                    传 None 时只看全局开关。
+
+        Returns:
+            "legacy" 或 "unified"
+        """
+        config = self._get_config()
+
+        # 全局关闭 → 用默认值
+        if not config.enabled:
+            return self._default_engine
+
+        # 白名单匹配 / 百分比命中 → unified
+        if config.is_active(user_id):
+            return ENGINE_UNIFIED
+
+        # 未命中灰度 → 用默认值
+        return self._default_engine
+
+    def should_use_unified(self, user_id: Optional[str] = None) -> bool:
+        """便捷方法：是否应该使用统一渲染引擎。"""
+        return self.get_engine(user_id) == ENGINE_UNIFIED
+
+    def force_refresh(self) -> None:
+        """强制立即刷新配置（用于管理接口修改后立即生效）。"""
+        self._last_refresh = 0.0
+        if isinstance(self._store, RedisFeatureFlagStore):
+            self._store.invalidate_cache(FLAG_RENDER_ENGINE)
+        self._maybe_refresh()
+
+    def get_config_snapshot(self) -> dict:
+        """获取当前配置快照（用于管理接口展示）。"""
+        config = self._get_config()
+        return {
+            "flag_name": FLAG_RENDER_ENGINE,
+            "default_engine": self._default_engine,
+            "enabled": config.enabled,
+            "percentage": config.percentage,
+            "whitelist": sorted(config.whitelist),
+            "refresh_interval": self._refresh_interval,
+            "last_refresh": self._last_refresh,
+        }
+
+    def set_flag(self, config: FeatureFlagConfig) -> None:
+        """设置 flag 配置（管理接口用）。"""
+        config.name = FLAG_RENDER_ENGINE
+        self._store.set(config)
+        self.force_refresh()
+
+
+# 全局单例
+_resolver: Optional[RenderEngineResolver] = None
+_resolver_lock = threading.Lock()
+
+
+def get_render_engine_resolver() -> RenderEngineResolver:
+    """获取全局单例（基于 worker 配置）。"""
+    global _resolver
+    if _resolver is not None:
+        return _resolver
+
+    with _resolver_lock:
+        if _resolver is not None:
+            return _resolver
+
+        try:
+            from worker_app.core.config import get_settings
+
+            settings = get_settings()
+            redis_url = getattr(settings, "redis_url", None) or getattr(settings, "broker_url", None)
+            default = getattr(settings, "render_engine", ENGINE_LEGACY)
+            _resolver = RenderEngineResolver(
+                default_engine=default,
+                redis_url=redis_url,
+            )
+            logger.info(
+                "RenderEngineResolver initialized: default=%s, redis=%s",
+                default,
+                bool(redis_url),
+            )
+        except Exception as exc:
+            logger.warning("Failed to init RenderEngineResolver from settings: %s", exc)
+            _resolver = RenderEngineResolver(default_engine=ENGINE_LEGACY)
+
+    return _resolver

apps/worker/video_processing/render_subtitles.py

@@ -0,0 +1,255 @@
+"""ASS 字幕生成模块 — 从 unified_render_service.py 拆分.
+
+职责：
+- 将 title / subtitle 配置转换为 ASS 字幕文件
+- 提供样式计算（颜色、对齐、描边/阴影）
+- 供 UnifiedRenderService._maybe_generate_ass 调用
+"""
+
+from __future__ import annotations
+
+import logging
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+# ── 常量 ──────────────────────────────────────────────────────────────────────
+
+# Title/Subtitle 默认边距（像素）
+TITLE_MARGIN_TOP = 60
+TITLE_MARGIN_BOTTOM = 60
+TITLE_MARGIN_SIDE = 40
+
+
+# ── ASS 字幕工具 ─────────────────────────────────────────────────────────────
+
+
+def _hex_to_ass_color(hex_color: str) -> str:
+    """将 HEX 颜色（#RRGGBB）转换为 ASS &HBBGGRR 格式。"""
+    hex_color = hex_color.lstrip("#")
+    if len(hex_color) != 6:
+        return "&H000000"
+    r, g, b = hex_color[0:2], hex_color[2:4], hex_color[4:6]
+    return f"&H{b.upper()}{g.upper()}{r.upper()}"
+
+
+def _position_to_ass_alignment(position: str) -> int:
+    """将文字位置映射为 ASS \\an 对齐编号。
+
+    ASS 对齐编号（数字小键盘布局）：
+      7 8 9
+      4 5 6
+      1 2 3
+    """
+    mapping = {
+        "top": 8,  # 顶部居中
+        "center": 5,  # 居中
+        "bottom": 2,  # 底部居中
+    }
+    return mapping.get(position, 8)
+
+
+def _build_ass_style(
+    style_name: str,
+    *,
+    font_name: str = "思源黑体",
+    font_size: int = 48,
+    primary_color: str = "&H00FFFFFF",
+    outline_color: str = "&H00000000",
+    outline_width: float = 1.0,
+    shadow_blur: float = 0.0,
+    shadow_offset: tuple[int, int] = (0, 0),
+    bold: bool = False,
+    italic: bool = False,
+    alignment: int = 8,
+    margin_v: int = 60,
+    margin_l: int = 40,
+    margin_r: int = 40,
+) -> str:
+    """构建 ASS Style 行。
+
+    Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour,
+            Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle,
+            BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
+    """
+    bold_val = -1 if bold else 0
+    italic_val = -1 if italic else 0
+
+    # BackColour 用于阴影（BorderStyle=1 时 outline + shadow）
+    back_color = primary_color  # 阴影颜色默认同文字色（带透明度由阴影模糊控制）
+
+    # Shadow 值：ASS 中 Shadow 字段是阴影偏移距离（像素），
+    # 我们用 shadow_offset[1] 作为纵向偏移，模糊由 BorderStyle=3 实现
+    # 简化：BorderStyle=1（outline + drop shadow），Shadow 字段表示阴影深度
+    shadow_depth = shadow_offset[1] if shadow_blur > 0 else 0
+
+    return (
+        f"Style: {style_name},{font_name},{font_size},{primary_color},"
+        f"&H000000FF,{outline_color},{back_color},"
+        f"{bold_val},{italic_val},0,0,100,100,0,0,"
+        f"1,{outline_width},{shadow_depth},{alignment},"
+        f"{margin_l},{margin_r},{margin_v},1"
+    )
+
+
+def _escape_ass_text(text: str) -> str:
+    r"""转义 ASS 文本中的特殊字符。
+
+    ASS 中换行用 \N（硬换行）或 \n（软换行），
+    大括号 {} 用于覆盖样式，需要转义。
+    """
+    # 将实际换行转为 ASS 硬换行
+    text = text.replace("\r\n", "\\N").replace("\n", "\\N").replace("\r", "\\N")
+    # 转义大括号（ASS 用它做样式覆盖标签）
+    text = text.replace("{", "(").replace("}", ")")
+    return text
+
+
+def _format_ass_time(seconds: float) -> str:
+    """将秒数格式化为 ASS 时间格式 H:MM:SS.cc。"""
+    hours = int(seconds // 3600)
+    minutes = int((seconds % 3600) // 60)
+    secs = seconds % 60
+    return f"{hours}:{minutes:02d}:{secs:05.2f}"
+
+
+def generate_ass_subtitles(
+    output_path: Path,
+    *,
+    video_width: int,
+    video_height: int,
+    video_duration: float,
+    title_text: str = "",
+    title_config: dict[str, Any] | None = None,
+    subtitle_text: str = "",
+    subtitle_config: dict[str, Any] | None = None,
+) -> Path:
+    """生成 ASS 字幕文件。
+
+    支持 Title（标题）和 Subtitle（字幕）两种字幕类型，
+    各自可独立配置样式、位置和内容。
+
+    Args:
+        output_path: 输出 ASS 文件路径
+        video_width: 视频宽度（用于 ASS PlayResX）
+        video_height: 视频高度（用于 ASS PlayResY）
+        video_duration: 视频总时长（秒），字幕显示整个时长
+        title_text: 标题文本
+        title_config: 标题样式配置（TitleConfig dict）
+        subtitle_text: 字幕文本
+        subtitle_config: 字幕样式配置（SubtitleConfig dict）
+
+    Returns:
+        生成的 ASS 文件路径
+    """
+    title_config = title_config or {}
+    subtitle_config = subtitle_config or {}
+
+    title_enabled = title_config.get("enabled", True) and bool(title_text.strip())
+    subtitle_enabled = subtitle_config.get("enabled", True) and bool(subtitle_text.strip())
+
+    if not title_enabled and not subtitle_enabled:
+        # 没有字幕，生成空文件（仍返回路径，调用方自行判断是否使用）
+        output_path.write_text("", encoding="utf-8")
+        return output_path
+
+    styles: list[str] = []
+    events: list[str] = []
+
+    # ── Title 样式与事件 ──────────────────────────────────────────────────
+    if title_enabled:
+        title_color = _hex_to_ass_color(title_config.get("color", "#ffffff"))
+        title_stroke = title_config.get("stroke", {}) or {}
+        title_shadow = title_config.get("shadow", {}) or {}
+        stroke_color = _hex_to_ass_color(title_stroke.get("color", "#000000"))
+        stroke_width = float(title_stroke.get("width", 1)) if title_stroke.get("enabled", False) else 0.0
+        shadow_blur = float(title_shadow.get("blur", 4)) if title_shadow.get("enabled", False) else 0.0
+        shadow_offset = (
+            title_shadow.get("offset_x", 2) if title_shadow.get("enabled", False) else 0,
+            title_shadow.get("offset_y", 2) if title_shadow.get("enabled", False) else 0,
+        )
+
+        title_alignment = _position_to_ass_alignment(title_config.get("position", "top"))
+
+        styles.append(
+            _build_ass_style(
+                "TitleStyle",
+                font_name=title_config.get("font", "思源黑体"),
+                font_size=int(title_config.get("size", 48)),
+                primary_color=title_color,
+                outline_color=stroke_color,
+                outline_width=stroke_width,
+                shadow_blur=shadow_blur,
+                shadow_offset=shadow_offset,
+                bold=bool(title_config.get("bold", True)),
+                italic=bool(title_config.get("italic", False)),
+                alignment=title_alignment,
+                margin_v=TITLE_MARGIN_TOP,
+                margin_l=TITLE_MARGIN_SIDE,
+                margin_r=TITLE_MARGIN_SIDE,
+            )
+        )
+
+        # 转义 ASS 特殊字符
+        safe_title_text = _escape_ass_text(title_text)
+
+        events.append(
+            "Dialogue: 0,0:00:00.00," f"{_format_ass_time(video_duration)}," "TitleStyle,,0,0,0,," f"{safe_title_text}"
+        )
+
+    # ── Subtitle 样式与事件 ───────────────────────────────────────────────
+    if subtitle_enabled:
+        sub_color = _hex_to_ass_color(subtitle_config.get("color", "#ffffff"))
+        sub_alignment = _position_to_ass_alignment(subtitle_config.get("position", "bottom"))
+
+        styles.append(
+            _build_ass_style(
+                "SubtitleStyle",
+                font_name=subtitle_config.get("font", "思源黑体"),
+                font_size=int(subtitle_config.get("size", 24)),
+                primary_color=sub_color,
+                outline_color="&H00000000",
+                outline_width=1.0,
+                shadow_blur=0.0,
+                shadow_offset=(0, 0),
+                bold=False,
+                italic=False,
+                alignment=sub_alignment,
+                margin_v=TITLE_MARGIN_BOTTOM,
+                margin_l=TITLE_MARGIN_SIDE,
+                margin_r=TITLE_MARGIN_SIDE,
+            )
+        )
+
+        safe_subtitle_text = _escape_ass_text(subtitle_text)
+
+        events.append(
+            "Dialogue: 0,0:00:00.00,"
+            f"{_format_ass_time(video_duration)},"
+            "SubtitleStyle,,0,0,0,,"
+            f"{safe_subtitle_text}"
+        )
+
+    # ── 组装 ASS 文件 ─────────────────────────────────────────────────────
+    ass_content = f"""[Script Info]
+ScriptType: v4.00+
+PlayResX: {video_width}
+PlayResY: {video_height}
+ScaledBorderAndShadow: yes
+WrapStyle: 2
+Encoding: UTF-8
+
+[V4+ Styles]
+Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding  # noqa: E501
+{chr(10).join(styles)}
+
+[Events]
+Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
+{chr(10).join(events)}
+"""
+
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    output_path.write_text(ass_content, encoding="utf-8")
+    return output_path

apps/worker/video_processing/unified_render_service.py

@@ -22,8 +22,8 @@
 from __future__ import annotations
 
 import logging
-import os
 import subprocess
+import time
 from dataclasses import dataclass, field
 from pathlib import Path
 from typing import Any
@@ -39,6 +39,8 @@ from video_processing.ffmpeg_utils import (
     probe_video_info,
     run_ffmpeg,
 )
+from video_processing.render_audio import RenderContext, merge_audio_video, mix_audio
+from video_processing.render_subtitles import generate_ass_subtitles
 
 logger = logging.getLogger(__name__)
 
@@ -116,6 +118,8 @@ def _resolve_layer_role(clip_type: str, config: dict[str, Any]) -> str:
     # main type
     if role == "b_roll":
         return "broll"
+    if role == "audio":
+        return "audio"
     return "main"
 
 
@@ -165,11 +169,26 @@ class UnifiedRenderService:
         self.transition_duration = transition_duration
 
     def render(self) -> RenderResult:
-        """执行渲染，返回 RenderResult。
+        """执行渲染，返回 RenderResult.
+
+        优化路径：
+        - 单图层单 clip → 直通模式（-vf），性能最优
+        - 其他情况 → 完整 filter_complex 渲染
+
+        字幕渲染流程：
+        1. 视频主渲染（直通或完整链路）
+        2. 如有 title/subtitle，叠加 ASS 字幕
+
+        音频后处理：
+        1. 主图层音频 concat 拼接
+        2. 独立音频轨 amix 混入
+        3. 合并到输出视频
 
         Raises:
             ValueError: 没有可渲染的片段时抛出
         """
+        t_start = time.time()
+
         # 1. 解析 clips → ResolvedClips（跳过无素材的 clip）
         resolved = self._resolve_clips()
         if not resolved:
@@ -178,16 +197,112 @@ class UnifiedRenderService:
         # 2. 分组为 RenderLayers
         layers = self._group_clips_into_layers(resolved)
 
-        # 3. 构建 filter_complex
+        # 3. 计算视频总时长（用于字幕显示时长）
+        video_duration = self._estimate_total_duration(layers)
+
+        # 4. 生成 ASS 字幕文件（如果有 title/subtitle 配置）
+        ass_path = self._maybe_generate_ass(video_duration)
+
+        # 灰度埋点：开始渲染
+        layer_roles = [layer.role for layer in layers]
+        clip_counts = {layer.role: len(layer.clips) for layer in layers}
+        logger.info(
+            "[unified-render] start render: plan_id=%s clip_count=%d layers=%s clip_counts=%s",
+            self.plan.id,
+            len(resolved),
+            layer_roles,
+            clip_counts,
+        )
+
+        # 5. 视频主渲染
+        t_video_start = time.time()
+        video_only_path = self.work_dir / f"rendered_{self.plan.id}_video.mp4"
         output_path = self.work_dir / f"rendered_{self.plan.id}.mp4"
-        filter_complex, input_args = self._build_filter_complex(layers)
 
-        # 4. 执行 FFmpeg
-        self._execute_ffmpeg(filter_complex, input_args, output_path)
+        is_pass_through = self._can_use_pass_through(layers)
+        pass_through_has_audio = False
+        used_stream_copy = False
 
-        # 5. 探测输出
+        if is_pass_through:
+            # 先尝试 stream copy 优化（无重编码，性能提升 10 倍+）
+            # 条件不满足或失败时回退到带滤镜的直通渲染
+            stream_copy_ok = self._try_render_stream_copy(
+                layers, output_path, ass_path=ass_path, video_duration=video_duration
+            )
+            if stream_copy_ok:
+                used_stream_copy = True
+                # stream copy 模式下，直接探测输出是否有音频
+                clip = layers[0].clips[0]
+                info = probe_video_info(str(clip.local_path))
+                pass_through_has_audio = info.get("has_audio", True)
+            else:
+                # 回退到带滤镜的直通渲染
+                pass_through_has_audio = self._render_pass_through(
+                    layers,
+                    output_path,
+                    ass_path=ass_path,
+                    video_duration=video_duration,
+                )
+        else:
+            filter_complex, input_args = self._build_filter_complex(layers, ass_path=ass_path)
+            self._execute_ffmpeg(filter_complex, input_args, video_only_path)
+
+        t_video_end = time.time()
+        video_render_ms = int((t_video_end - t_video_start) * 1000)
+        logger.info(
+            "[unified-render] video render done: plan_id=%s duration_ms=%d pass_through=%s stream_copy=%s",
+            self.plan.id,
+            video_render_ms,
+            is_pass_through,
+            used_stream_copy,
+        )
+
+        # 6. 音频后处理混音（直通场景已合并处理，跳过）
+        t_audio_start = time.time()
+        audio_mix_ms = 0
+        has_audio = False
+
+        if is_pass_through:
+            # 直通场景已在一次调用中完成视频+音频
+            has_audio = pass_through_has_audio
+        else:
+            ctx = RenderContext(work_dir=self.work_dir, plan_id=self.plan.id)
+            audio_path = mix_audio(ctx, layers, video_duration)
+            t_audio_end = time.time()
+            audio_mix_ms = int((t_audio_end - t_audio_start) * 1000)
+            has_audio = audio_path is not None
+            if has_audio:
+                logger.info(
+                    "[unified-render] audio mix done: plan_id=%s duration_ms=%d",
+                    self.plan.id,
+                    audio_mix_ms,
+                )
+                # 7. 合并音视频
+                merge_audio_video(ctx, video_only_path, audio_path, output_path)
+            else:
+                # 无音频，直接用无声视频
+                import shutil
+
+                shutil.copy2(video_only_path, output_path)
+
+        # 8. 探测输出
         duration, file_size, width, height = self._probe_output(output_path)
 
+        t_total = int((time.time() - t_start) * 1000)
+        logger.info(
+            "[unified-render] render done: plan_id=%s total_ms=%d video_ms=%d audio_ms=%d "
+            "output_duration=%.2fs output_size=%d resolution=%dx%d has_audio=%s",
+            self.plan.id,
+            t_total,
+            video_render_ms,
+            audio_mix_ms if has_audio else 0,
+            duration,
+            file_size,
+            width,
+            height,
+            has_audio,
+        )
+
         return RenderResult(
             output_path=output_path,
             duration=duration,
@@ -196,6 +311,382 @@ class UnifiedRenderService:
             height=height,
         )
 
+    def _estimate_total_duration(self, layers: list[RenderLayer]) -> float:
+        """估算视频总时长（用于字幕等需要）。
+
+        取主图层（main/broll/background）的总时长，转场重叠按 transition_duration 估算。
+        """
+        # 找主图层（第一个有视频内容的图层）
+        main_layer = None
+        for role in ("main", "broll", "background"):
+            for layer in layers:
+                if layer.role == role:
+                    main_layer = layer
+                    break
+            if main_layer:
+                break
+
+        if not main_layer or not main_layer.clips:
+            return 0.0
+
+        total = sum(UnifiedRenderService._clip_effective_duration(c) for c in main_layer.clips)
+
+        # 减去转场重叠时间（粗略估算）
+        n_clips = len(main_layer.clips)
+        if n_clips > 1:
+            total -= (n_clips - 1) * self.transition_duration
+
+        return max(0.1, total)
+
+    def _maybe_generate_ass(self, video_duration: float) -> Path | None:
+        """根据 plan.config 生成 ASS 字幕文件。
+
+        Returns:
+            ASS 文件路径，没有字幕时返回 None
+        """
+        config = self.plan.config or {}
+        title_cfg = config.get("title", {}) or {}
+        subtitle_cfg = config.get("subtitle", {}) or {}
+
+        title_enabled = title_cfg.get("enabled", True)
+        subtitle_enabled = subtitle_cfg.get("enabled", True)
+        title_text = title_cfg.get("text", "") or ""
+        subtitle_text = subtitle_cfg.get("text", "") or ""
+
+        has_title = title_enabled and bool(title_text.strip())
+        has_subtitle = subtitle_enabled and bool(subtitle_text.strip())
+
+        if not has_title and not has_subtitle:
+            return None
+
+        ass_path = self.work_dir / f"subtitles_{self.plan.id}.ass"
+
+        generate_ass_subtitles(
+            ass_path,
+            video_width=self.output_width,
+            video_height=self.output_height,
+            video_duration=video_duration,
+            title_text=title_text,
+            title_config=title_cfg,
+            subtitle_text=subtitle_text,
+            subtitle_config=subtitle_cfg,
+        )
+
+        logger.info(
+            "生成字幕: plan_id=%s title=%s subtitle=%s ass=%s",
+            self.plan.id,
+            has_title,
+            has_subtitle,
+            ass_path,
+        )
+        return ass_path
+
+    def _can_use_pass_through(self, layers: list[RenderLayer]) -> bool:
+        """判断是否可以走直通优化路径。
+
+        条件：
+        1. 只有 1 个图层
+        2. 该图层是视频图层（main/broll/background），不是 overlay/corner_voice/audio
+        3. 该图层只有 1 个 clip（无转场需求）
+        """
+        if len(layers) != 1:
+            return False
+        layer = layers[0]
+        if layer.role not in ("main", "broll", "background"):
+            return False
+        if len(layer.clips) != 1:
+            return False
+        return True
+
+    def _can_use_stream_copy(
+        self,
+        clip: ResolvedClip,
+        *,
+        ass_path: Path | None = None,
+        video_duration: float = 0.0,
+    ) -> tuple[bool, str]:
+        """判断是否可以走 stream copy（流拷贝，不重编码）。
+
+        性能提升：10 倍以上（典型场景从 20s → 1-2s）。
+
+        条件：
+        1. 视频编码为 h264（输出目标也是 h264）
+        2. 像素格式为 yuv420p
+        3. 分辨率与输出一致（不需要 scale/crop）
+        4. 帧率与输出一致（误差 < 0.1fps）
+        5. 无字幕叠加（字幕需要滤镜）
+        6. 无 trim 需求（或 trim 后恰好等于原时长）
+        7. 无转场、无特效（单 clip 直通已保证）
+
+        Returns:
+            (是否可以 copy, 原因说明)
+        """
+        # 有字幕 → 需要滤镜 → 不能 copy
+        if ass_path is not None:
+            return False, "有字幕叠加"
+
+        # 探测输入视频参数
+        info = probe_video_info(str(clip.local_path))
+
+        # 编码必须是 h264
+        if info.get("video_codec", "") != "h264":
+            return False, f"视频编码不是h264: {info.get('video_codec', 'unknown')}"
+
+        # 像素格式必须是 yuv420p
+        if info.get("pix_fmt", "") != "yuv420p":
+            return False, f"像素格式不是yuv420p: {info.get('pix_fmt', 'unknown')}"
+
+        # 分辨率必须一致
+        if info.get("width", 0) != self.output_width or info.get("height", 0) != self.output_height:
+            return False, (
+                f"分辨率不匹配: "
+                f"{info.get('width', 0)}x{info.get('height', 0)} "
+                f"vs {self.output_width}x{self.output_height}"
+            )
+
+        # 帧率必须一致（误差 < 0.1fps）
+        fps_diff = abs(info.get("fps", 0) - self.output_fps)
+        if fps_diff > 0.1:
+            return False, f"帧率不匹配: {info.get('fps', 0)} vs {self.output_fps}"
+
+        # 检查是否需要 trim
+        effective_duration = UnifiedRenderService._clip_effective_duration(clip)
+        if effective_duration > 0:
+            # 有 trim 需求但视频时长足够，可用 -ss/-t 实现 copy trim
+            input_duration = info.get("duration", 0)
+            if input_duration <= 0:
+                return False, "无法探测输入时长"
+            # trim 起始点 + 目标时长 <= 输入时长
+            start_time = getattr(clip, "start_time", 0) or 0
+            if start_time + effective_duration > input_duration + 0.1:
+                return False, "trim 超出输入时长"
+
+        # video_duration 截断
+        if video_duration > 0 and effective_duration > 0:
+            final_duration = min(effective_duration, video_duration)
+            if final_duration != effective_duration:
+                # 也需要截断，但 -t 可以 copy 模式下用
+                pass
+
+        return True, "所有条件满足"
+
+    def _try_render_stream_copy(
+        self,
+        layers: list[RenderLayer],
+        output_path: Path,
+        *,
+        ass_path: Path | None = None,
+        video_duration: float = 0.0,
+    ) -> bool:
+        """尝试 stream copy 渲染，成功返回 True，失败返回 False（调用方回退到重编码）。
+
+        stream copy 模式：不重编码，直接拷贝视频/音频流，性能提升 10 倍+。
+        仅用于单 clip 直通场景且满足 copy 条件。
+        """
+        clip = layers[0].clips[0]
+        role = layers[0].role
+
+        # 判断是否满足 copy 条件
+        can_copy, reason = self._can_use_stream_copy(clip, ass_path=ass_path, video_duration=video_duration)
+        if not can_copy:
+            logger.info(
+                "[unified-render] stream_copy 跳过: plan_id=%s reason=%s",
+                self.plan.id,
+                reason,
+            )
+            return False
+
+        # 构建 copy 命令
+        command = [
+            FFMPEG_BIN,
+            "-y",
+        ]
+
+        # trim 支持（-ss 放在 -i 前 = input seeking，速度更快但精度稍差；
+        # 放在 -i 后 = output seeking，精度高但慢）
+        # 这里用 output seeking 保证精度，反正 copy 模式已经很快了
+        start_time = getattr(clip, "start_time", 0) or 0
+        effective_duration = UnifiedRenderService._clip_effective_duration(clip)
+
+        command.extend(["-i", str(clip.local_path)])
+
+        if start_time > 0:
+            command.extend(["-ss", f"{start_time:.3f}"])
+
+        # 计算最终时长
+        final_duration = effective_duration
+        if video_duration > 0 and (final_duration <= 0 or final_duration > video_duration):
+            final_duration = video_duration
+        if final_duration > 0:
+            command.extend(["-t", f"{final_duration:.3f}"])
+
+        # 流拷贝
+        command.extend(
+            [
+                "-c:v",
+                "copy",
+                "-c:a",
+                "copy",
+                "-movflags",
+                "+faststart",
+                str(output_path),
+            ]
+        )
+
+        logger.info(
+            "[unified-render] stream_copy 渲染: plan_id=%s clip=%s role=%s duration=%.2fs",
+            self.plan.id,
+            clip.clip_id,
+            role,
+            final_duration,
+        )
+
+        try:
+            run_ffmpeg(command)
+            # 验证输出文件存在且有大小
+            if output_path.exists() and output_path.stat().st_size > 0:
+                logger.info(
+                    "[unified-render] stream_copy 成功: plan_id=%s size=%d",
+                    self.plan.id,
+                    output_path.stat().st_size,
+                )
+                return True
+            else:
+                logger.warning("[unified-render] stream_copy 输出为空: plan_id=%s", self.plan.id)
+                return False
+        except (subprocess.CalledProcessError, subprocess.TimeoutExpired) as e:
+            logger.warning(
+                "[unified-render] stream_copy 失败，回退到重编码: plan_id=%s error=%s",
+                self.plan.id,
+                str(e)[:200],
+            )
+            # 清理可能的损坏输出文件
+            if output_path.exists():
+                try:
+                    output_path.unlink()
+                except OSError as unlink_err:
+                    logger.warning(
+                        "[unified-render] 损坏输出文件清理失败: path=%s error=%s",
+                        output_path,
+                        unlink_err,
+                    )
+            return False
+
+    def _render_pass_through(
+        self,
+        layers: list[RenderLayer],
+        output_path: Path,
+        *,
+        ass_path: Path | None = None,
+        video_duration: float = 0.0,
+    ) -> bool:
+        """单图层单 clip 直通渲染（使用 -vf 而非 -filter_complex），一次性输出带音频的最终视频。
+
+        性能优化：
+        - 避免 filter_complex 的解析和调度开销，单clip场景性能提升 ~30%
+        - 视频+音频一次FFmpeg调用完成，省去后续音频提取+音视频合并两次调用
+
+        Args:
+            layers: 图层列表（只有1个图层1个clip）
+            output_path: 输出文件路径
+            ass_path: ASS 字幕文件路径，有则叠加字幕
+            video_duration: 视频总时长（用于截断音频，0表示不额外截断）
+
+        Returns:
+            True 表示输出包含音频（近似判断，实际以输出文件为准）
+        """
+        clip = layers[0].clips[0]
+        role = layers[0].role
+
+        # 构建视频滤镜链（与 _build_filter_complex 中预处理逻辑一致）
+        filters: list[str] = []
+
+        # trim
+        effective_duration = UnifiedRenderService._clip_effective_duration(clip)
+
+        if effective_duration > 0:
+            filters.append(f"trim=duration={effective_duration}")
+            filters.append("setpts=PTS-STARTPTS")
+
+        # scale + crop（铺满裁剪）
+        if role in ("overlay", "corner_voice"):
+            pip_w = int(self.output_width * _PIP_SCALE)
+            pip_h = int(self.output_height * _PIP_SCALE)
+            filters.append(f"scale={pip_w}:{pip_h}")
+        else:
+            # main / broll / background: 铺满裁剪
+            filters.append(f"scale={self.output_width}:{self.output_height}" ":force_original_aspect_ratio=increase")
+            filters.append(f"crop={self.output_width}:{self.output_height}")
+
+        filters.append("setpts=PTS-STARTPTS")
+        filters.append(f"fps={self.output_fps}")
+        filters.append("format=yuv420p")
+
+        # 字幕叠加
+        if ass_path is not None:
+            ass_filter_path = str(ass_path).replace("\\", "/").replace(":", "\\:")
+            filters.append(f"subtitles='{ass_filter_path}'")
+
+        vf_str = ",".join(filters)
+
+        # 最终输出时长：取 clip 有效时长和 video_duration 的较小值
+        final_duration = effective_duration
+        if video_duration > 0 and (final_duration <= 0 or final_duration > video_duration):
+            final_duration = video_duration
+
+        command = [
+            FFMPEG_BIN,
+            "-y",
+            "-i",
+            str(clip.local_path),
+            "-vf",
+            vf_str,
+            "-c:v",
+            "libx264",
+            "-crf",
+            "23",
+            "-preset",
+            "medium",
+            "-pix_fmt",
+            "yuv420p",
+            "-movflags",
+            "+faststart",
+        ]
+
+        # 音频处理：background 通常是图片无音频，跳过；其他编码为 aac
+        # background 以外的视频素材，默认带音频
+        has_audio = role != "background"
+        if has_audio:
+            command.extend(["-c:a", "aac", "-b:a", "128k"])
+
+        # 统一截断时长（同时作用于视频和音频）
+        if final_duration > 0:
+            command.extend(["-t", f"{final_duration:.3f}"])
+
+        command.append(str(output_path))
+
+        logger.info(
+            "直通渲染: plan_id=%s clip=%s role=%s duration=%.2fs has_audio=%s",
+            self.plan.id,
+            clip.clip_id,
+            role,
+            effective_duration,
+            has_audio,
+        )
+        try:
+            run_ffmpeg(command)
+        except subprocess.CalledProcessError as e:
+            logger.error(
+                "直通渲染失败: plan_id=%s clip=%s exit_code=%d\nvf=%s",
+                self.plan.id,
+                clip.clip_id,
+                e.returncode,
+                vf_str[:2000],
+            )
+            raise
+
+        return has_audio
+
     # ── 内部方法 ──────────────────────────────────────────────────────────────
 
     def _resolve_clips(self) -> list[ResolvedClip]:
@@ -259,7 +750,6 @@ class UnifiedRenderService:
 
         # 计算 PiP 位置
         pip_width = int(self.output_width * _PIP_SCALE)
-        pip_height = int(self.output_height * _PIP_SCALE)
         margin = 20  # 边距
 
         if "overlay" in layer_map:
@@ -277,9 +767,15 @@ class UnifiedRenderService:
         layers = sorted(layer_map.values(), key=lambda lyr: lyr.z_index)
         return layers
 
-    def _build_filter_complex(self, layers: list[RenderLayer]) -> tuple[str, list[str]]:
+    def _build_filter_complex(
+        self, layers: list[RenderLayer], *, ass_path: Path | None = None
+    ) -> tuple[str, list[str]]:
         """构建 FFmpeg filter_complex 字符串和输入参数列表。
 
+        Args:
+            layers: 图层列表
+            ass_path: ASS 字幕文件路径，有则在最后叠加字幕
+
         Returns:
             (filter_complex_str, input_args_list)
             input_args_list 是 ["-i", path1, "-i", path2, ...] 格式
@@ -311,14 +807,7 @@ class UnifiedRenderService:
             filters: list[str] = []
 
             # trim — 始终将输出截断到有效时长，防止 xfade offset 与实际时长不匹配
-            # 有效时长 = min(指定时长, 实际时长)；若均未设置则跳过
-            effective_duration = 0.0
-            if clip.duration > 0:
-                effective_duration = (
-                    min(clip.duration, clip.actual_duration) if clip.actual_duration > 0 else clip.duration
-                )
-            elif clip.actual_duration > 0:
-                effective_duration = clip.actual_duration
+            effective_duration = UnifiedRenderService._clip_effective_duration(clip)
 
             if effective_duration > 0:
                 filters.append(f"trim=duration={effective_duration}")
@@ -335,11 +824,12 @@ class UnifiedRenderService:
                 )
                 filters.append(f"crop={self.output_width}:{self.output_height}")
             else:
-                # main / broll: scale + pad 保持宽高比
+                # main / broll: 铺满裁剪（scale to cover + center crop）
+                # 对齐链路A编辑器合成行为，与主流短视频平台一致
                 filters.append(
-                    f"scale={self.output_width}:{self.output_height}" ":force_original_aspect_ratio=decrease"
+                    f"scale={self.output_width}:{self.output_height}" ":force_original_aspect_ratio=increase"
                 )
-                filters.append(f"pad={self.output_width}:{self.output_height}" ":(ow-iw)/2:(oh-ih)/2:black")
+                filters.append(f"crop={self.output_width}:{self.output_height}")
 
             filters.append("setpts=PTS-STARTPTS")
             filters.append(f"fps={self.output_fps}")
@@ -354,14 +844,7 @@ class UnifiedRenderService:
             layer_clip_indices = [all_clips.index(c) for c in layer.clips]
             layer_labels = [preprocessed_labels[i] for i in layer_clip_indices]
             # 使用 trim 后的有效时长，与 Step 1 的 trim=duration 保持一致
-            layer_durations = []
-            for i in layer_clip_indices:
-                c = all_clips[i]
-                if c.duration > 0:
-                    eff = min(c.duration, c.actual_duration) if c.actual_duration > 0 else c.duration
-                else:
-                    eff = c.actual_duration if c.actual_duration > 0 else 0.0
-                layer_durations.append(eff)
+            layer_durations = [UnifiedRenderService._clip_effective_duration(all_clips[i]) for i in layer_clip_indices]
             layer_transitions = [all_clips[i].transition_effect for i in layer_clip_indices]
 
             if len(layer_labels) == 1:
@@ -421,7 +904,12 @@ class UnifiedRenderService:
                 filter_parts.append(f"[{final_video_label}][{overlay_label}]" f"overlay={x}:{y}[{combined_label}]")
                 final_video_label = combined_label
 
-        filter_parts.append(f"[{final_video_label}]format=yuv420p[final_video]")
+        # 叠加字幕（如有）+ 最终像素格式
+        if ass_path is not None:
+            ass_filter_path = str(ass_path).replace("\\", "/").replace(":", "\\:")
+            filter_parts.append(f"[{final_video_label}]subtitles='{ass_filter_path}',format=yuv420p[final_video]")
+        else:
+            filter_parts.append(f"[{final_video_label}]format=yuv420p[final_video]")
 
         filter_complex = ";".join(filter_parts)
         return filter_complex, input_args
@@ -476,7 +964,7 @@ class UnifiedRenderService:
             raise
 
     def _probe_output(self, output_path: Path) -> tuple[float, int, int, int]:
-        """探测输出文件的时长、大小、宽高。
+        """探测输出文件的时长、大小、宽高.
 
         Returns:
             (duration, file_size, width, height)
@@ -489,3 +977,10 @@ class UnifiedRenderService:
             info["width"],
             info["height"],
         )
+
+    @staticmethod
+    def _clip_effective_duration(clip: ResolvedClip) -> float:
+        """计算 clip 的有效时长."""
+        if clip.duration > 0:
+            return min(clip.duration, clip.actual_duration) if clip.actual_duration > 0 else clip.duration
+        return clip.actual_duration if clip.actual_duration > 0 else 0.0

apps/worker/video_processing/video_compose_service.py

@@ -1,821 +0,0 @@
-"""
-视频合成服务
-支持多种剪辑模式和转场效果，包含完整的安全校验
-"""
-
-import logging
-import os
-import subprocess
-import tempfile
-from dataclasses import dataclass
-from enum import Enum
-
-try:
-    from enum import StrEnum
-except ImportError:
-
-    class StrEnum(str, Enum):  # type: ignore[no-redef]
-        """Python 3.10 兼容的 StrEnum 回退实现。"""
-
-        pass
-
-
-from pathlib import Path
-from typing import Optional
-
-from packages.domain.editing_mode import EditingMode
-
-logger = logging.getLogger(__name__)
-
-# ========== 安全常量 ==========
-# 允许的输出目录白名单（使用环境变量或系统临时目录，避免硬编码 /tmp）
-_VIDEO_OUTPUT_DIR = os.environ.get("VIDEO_OUTPUT_DIR", os.path.join(tempfile.gettempdir(), "video_output"))
-ALLOWED_OUTPUT_DIRS = [_VIDEO_OUTPUT_DIR, "/var/app/rendered"]
-
-# 允许的输入路径前缀白名单
-ALLOWED_INPUT_PREFIXES = ("s3://", "oss://", "local://", "/var/storage/")
-
-# 允许的转场效果白名单
-ALLOWED_TRANSITIONS = {
-    "fade",
-    "slideleft",
-    "slideright",
-    "dissolve",
-    "wipeleft",
-    "wiperight",
-    "cut",
-    "slideup",
-    "slidedown",
-}
-
-# 转场效果映射
-_XFADE_TRANSITION_MAP = {
-    "fade": "fade",
-    "slideleft": "slideleft",
-    "slideright": "slideright",
-    "dissolve": "dissolve",
-    "wipeleft": "wipeleft",
-    "wiperight": "wiperight",
-    "cut": "cut",
-    "slideup": "slideup",
-    "slidedown": "slidedown",
-}
-
-
-class VideoComposeError(Exception):
-    """视频合成服务异常"""
-
-    pass
-
-
-class PIPPosition(StrEnum):
-    """画中画位置枚举"""
-
-    TOP_LEFT = "top_left"
-    TOP_RIGHT = "top_right"
-    BOTTOM_LEFT = "bottom_left"
-    BOTTOM_RIGHT = "bottom_right"
-
-
-@dataclass
-class Clip:
-    """视频片段"""
-
-    asset_id: str  # 资源ID，对应输入路径
-    start_time: float = 0.0
-    duration: float = 0.0
-    transition: str = "fade"  # 转场效果
-
-
-@dataclass
-class EditingModeConfig:
-    """剪辑模式配置"""
-
-    mode: EditingMode
-    output_width: int = 1280
-    output_height: int = 720
-    output_fps: int = 25
-    pip_position: PIPPosition = PIPPosition.TOP_RIGHT
-    pip_scale: float = 0.25  # 画中画占主画面的比例
-    transition_duration: float = 0.5  # 转场时长（秒）
-    output_codec: str = "libx264"
-    output_preset: str = "medium"
-    output_crf: int = 23
-
-
-class VideoComposeService:
-    """视频合成服务"""
-
-    def __init__(self, config: EditingModeConfig, work_dir: Optional[str] = None):
-        """
-        初始化视频合成服务
-
-        Args:
-            config: 剪辑模式配置
-            work_dir: 工作目录，默认使用系统临时目录
-        """
-        self.config = config
-        self.work_dir = work_dir or tempfile.gettempdir()
-        self._ffmpeg_bin = "ffmpeg"
-        self._ffprobe_bin = "ffprobe"
-
-    def _validate_output_path(self, path: str) -> str:
-        """
-        校验输出路径是否在允许范围内 (P0 修复)
-
-        防止路径穿越攻击，如 /app/config/../../../etc/passwd
-
-        Args:
-            path: 用户提供的输出路径
-
-        Returns:
-            标准化后的绝对路径
-
-        Raises:
-            ValueError: 路径不在允许范围内
-        """
-        abs_path = os.path.abspath(path)
-        for allowed_dir in ALLOWED_OUTPUT_DIRS:
-            allowed_abs = os.path.abspath(allowed_dir)
-            if abs_path.startswith(allowed_abs):
-                return abs_path
-        raise ValueError(f"输出路径不在允许范围内: {path}")
-
-    def _validate_input_path(self, path: str) -> bool:
-        """
-        校验输入路径格式是否合法 (P1-1 修复)
-
-        Args:
-            path: 输入文件路径
-
-        Returns:
-            是否合法
-        """
-        return any(path.startswith(prefix) for prefix in ALLOWED_INPUT_PREFIXES)
-
-    def _validate_transition(self, transition: str) -> str:
-        """
-        校验转场效果是否在白名单内 (P1-2 修复)
-
-        Args:
-            transition: 转场效果名称
-
-        Returns:
-            安全的转场效果名称
-        """
-        if transition not in ALLOWED_TRANSITIONS:
-            logger.warning(f"未知的转场效果 '{transition}'，使用默认 'fade'")
-            return "fade"
-        return transition
-
-    def _get_validated_transition(self, transition: str) -> str:
-        """获取白名单校验后的转场效果名称"""
-        return _XFADE_TRANSITION_MAP.get(self._validate_transition(transition), "fade")
-
-    def compose(self, clips: list[Clip], output_path: Optional[str] = None) -> str:
-        """
-        合成视频
-
-        Args:
-            clips: 视频片段列表，每个片段包含 asset_id 和转场配置
-            output_path: 输出文件路径
-
-        Returns:
-            输出文件路径
-        """
-        if not clips:
-            raise ValueError("clips 不能为空")
-
-        # P1-1: 校验所有输入路径
-        for clip in clips:
-            if not self._validate_input_path(clip.asset_id):
-                raise ValueError(f"不合法的输入路径: {clip.asset_id}")
-
-        # 生成默认输出路径并校验
-        if output_path is None:
-            output_path = self._generate_output_path()
-
-        # P0: 校验输出路径
-        validated_output = self._validate_output_path(output_path)
-
-        logger.info(f"合成视频，片段数: {len(clips)}, 输出: {validated_output}")
-
-        # 获取输入路径列表
-        input_paths = [clip.asset_id for clip in clips]
-
-        try:
-            if self.config.mode == EditingMode.ONE_TAKE:
-                return self._one_take(input_paths, validated_output, clips)
-            elif self.config.mode == EditingMode.PIP:
-                return self._pip(input_paths, validated_output)
-            elif self.config.mode == EditingMode.VOICE_OVER:
-                return self._voice_over(input_paths, validated_output)
-            elif self.config.mode == EditingMode.VOICE_PIP:
-                return self._voice_pip(input_paths, validated_output)
-            else:
-                raise ValueError(f"不支持的剪辑模式: {self.config.mode}")
-        except Exception as e:
-            logger.error(f"视频合成失败: {e}")
-            raise VideoComposeError(f"视频合成失败: {e}") from e
-
-    def _generate_output_path(self) -> str:
-        """生成输出文件路径"""
-        os.makedirs(self.work_dir, exist_ok=True)
-        return os.path.join(self.work_dir, f"output_{self.config.mode}_{os.getpid()}.mp4")
-
-    def _validate_inputs(self, video_paths: list[str], audio_path: Optional[str] = None) -> None:
-        """验证输入文件存在"""
-        for path in video_paths:
-            if not os.path.exists(path):
-                raise FileNotFoundError(f"视频文件不存在: {path}")
-            if not os.path.getsize(path) > 0:
-                raise ValueError(f"视频文件为空: {path}")
-
-        if audio_path and not os.path.exists(audio_path):
-            raise FileNotFoundError(f"音频文件不存在: {audio_path}")
-
-    def _run_ffmpeg(self, command: list[str], capture_output: bool = True) -> tuple:
-        """执行 FFmpeg 命令"""
-        logger.debug(f"Running FFmpeg: {' '.join(command)}")
-        try:
-            result = subprocess.run(
-                command,
-                check=True,
-                stdout=subprocess.PIPE if capture_output else None,
-                stderr=subprocess.PIPE if capture_output else None,
-                text=capture_output,
-            )
-            return result.stdout or "", result.stderr or ""
-        except subprocess.CalledProcessError as e:
-            stderr = e.stderr.decode() if e.stderr else str(e)
-            logger.error(f"FFmpeg error: {stderr}")
-            raise RuntimeError(f"FFmpeg 执行失败: {stderr}") from e
-
-    def _get_video_info(self, video_path: str) -> dict:
-        """获取视频信息"""
-        try:
-            result = subprocess.run(
-                [
-                    self._ffprobe_bin,
-                    "-v",
-                    "error",
-                    "-show_entries",
-                    "stream=width,height,r_frame_rate,duration,codec_name",
-                    "-show_entries",
-                    "format=duration,size",
-                    "-of",
-                    "json",
-                    video_path,
-                ],
-                check=True,
-                stdout=subprocess.PIPE,
-                stderr=subprocess.PIPE,
-                text=True,
-            )
-            import json
-
-            data = json.loads(result.stdout)
-            streams = data.get("streams", [{}])
-            video_stream = next((s for s in streams if s.get("codec_type") == "video"), streams[0] if streams else {})
-            fmt = data.get("format", {})
-
-            fps_str = video_stream.get("r_frame_rate", "25/1")
-            fps_parts = fps_str.split("/")
-            fps = float(fps_parts[0]) / float(fps_parts[1]) if len(fps_parts) == 2 else float(fps_parts[0])
-
-            return {
-                "width": int(video_stream.get("width", 0)),
-                "height": int(video_stream.get("height", 0)),
-                "fps": fps,
-                "duration": float(fmt.get("duration", 0)),
-                "codec": video_stream.get("codec_name", "unknown"),
-                "size": int(fmt.get("size", 0)),
-            }
-        except Exception as e:
-            logger.warning(f"获取视频信息失败 {video_path}: {e}")
-            return {"width": 0, "height": 0, "fps": 25, "duration": 0, "codec": "unknown", "size": 0}
-
-    def _get_pip_position_offset(
-        self, main_width: int, main_height: int, pip_width: int, pip_height: int
-    ) -> tuple[int, int]:
-        """获取画中画位置偏移量"""
-        margin = 10
-        position_offsets = {
-            PIPPosition.TOP_LEFT: (margin, margin),
-            PIPPosition.TOP_RIGHT: (main_width - pip_width - margin, margin),
-            PIPPosition.BOTTOM_LEFT: (margin, main_height - pip_height - margin),
-            PIPPosition.BOTTOM_RIGHT: (main_width - pip_width - margin, main_height - pip_height - margin),
-        }
-        return position_offsets.get(self.config.pip_position, position_offsets[PIPPosition.TOP_RIGHT])
-
-    def _normalize_video(self, input_path: str, output_path: str) -> dict:
-        """标准化视频格式"""
-        command = [
-            self._ffmpeg_bin,
-            "-y",
-            "-i",
-            input_path,
-            "-r",
-            str(self.config.output_fps),
-            "-vf",
-            f"scale={self.config.output_width}:{self.config.output_height}:force_original_aspect_ratio=decrease,pad={self.config.output_width}:{self.config.output_height}:(ow-iw)/2:(oh-ih)/2,setsar=1",
-            "-r",
-            str(self.config.output_fps),
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            "-movflags",
-            "+faststart",
-            "-an",
-            output_path,
-        ]
-        self._run_ffmpeg(command)
-        return self._get_video_info(output_path)
-
-    def _one_take(self, video_paths: list[str], output_path: str, clips: list[Clip]) -> str:
-        """一镜到底模式"""
-        if len(video_paths) == 1:
-            return self._normalize_video(video_paths[0], output_path)
-
-        normalized_paths = []
-        for i, path in enumerate(video_paths):
-            normalized = os.path.join(self.work_dir, f"normalized_{i}_{os.getpid()}.mp4")
-            self._normalize_video(path, normalized)
-            normalized_paths.append(normalized)
-
-        durations = [self._get_video_info(p)["duration"] for p in normalized_paths]
-
-        if len(normalized_paths) <= 5:
-            output_path = self._one_take_with_xfade(normalized_paths, durations, output_path, clips)
-        else:
-            output_path = self._one_take_simple_concat(normalized_paths, output_path)
-
-        for p in normalized_paths:
-            try:
-                if p != output_path:
-                    os.remove(p)
-            except Exception as e:
-                logger.warning(
-                    f"Operation failed in apps/worker/video_processing/video_compose_service.py: {e}", exc_info=True
-                )
-
-        return output_path
-
-    def _one_take_with_xfade(
-        self, normalized_paths: list[str], durations: list[float], output_path: str, clips: list[Clip]
-    ) -> str:
-        """使用 xfade 滤镜实现转场 (P1-2: 转场参数白名单校验)"""
-        if len(normalized_paths) == 2:
-            # 获取当前片段的转场效果并校验白名单
-            transition = "fade"
-            if len(clips) > 1:
-                transition = self._get_validated_transition(clips[1].transition)
-
-            trans_duration = self.config.transition_duration
-            offset1 = durations[0] - trans_duration / 2
-
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-i",
-                normalized_paths[0],
-                "-i",
-                normalized_paths[1],
-                "-filter_complex",
-                f"[0:v][1:v]xfade=transition={transition}:duration={trans_duration}:offset={offset1}[v]",
-                "-map",
-                "[v]",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                output_path,
-            ]
-            self._run_ffmpeg(command)
-            return output_path
-        else:
-            return self._one_take_simple_concat(normalized_paths, output_path)
-
-    def _one_take_simple_concat(self, normalized_paths: list[str], output_path: str) -> str:
-        """使用 concat demuxer 简单拼接"""
-        concat_file = os.path.join(self.work_dir, f"concat_list_{os.getpid()}.txt")
-        with open(concat_file, "w") as f:
-            for path in normalized_paths:
-                f.write(f"file '{os.path.abspath(path)}'\n")
-
-        command = [
-            self._ffmpeg_bin,
-            "-y",
-            "-f",
-            "concat",
-            "-safe",
-            "0",
-            "-i",
-            concat_file,
-            "-c",
-            "copy",
-            output_path,
-        ]
-        self._run_ffmpeg(command)
-
-        try:
-            os.remove(concat_file)
-        except Exception as e:
-            logger.warning(
-                f"Operation failed in apps/worker/video_processing/video_compose_service.py: {e}", exc_info=True
-            )
-
-        return output_path
-
-    def _pip(self, video_paths: list[str], output_path: str) -> str:
-        """画中画模式"""
-        if not video_paths:
-            raise ValueError("No video paths provided")
-
-        main_video = video_paths[0]
-        main_normalized = os.path.join(self.work_dir, f"main_{os.getpid()}.mp4")
-        main_info = self._normalize_video(main_video, main_normalized)
-
-        if len(video_paths) == 1:
-            os.rename(main_normalized, output_path)
-            return output_path
-
-        pip_width = int(self.config.output_width * self.config.pip_scale)
-        pip_height = int(self.config.output_height * self.config.pip_scale)
-        x_offset, y_offset = self._get_pip_position_offset(
-            self.config.output_width, self.config.output_height, pip_width, pip_height
-        )
-
-        pip_normalized = os.path.join(self.work_dir, f"pip_{os.getpid()}.mp4")
-        pip_info = self._get_video_info(video_paths[1])
-
-        if pip_info["duration"] > main_info["duration"]:
-            temp_pip = os.path.join(self.work_dir, f"pip_temp_{os.getpid()}.mp4")
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-i",
-                video_paths[1],
-                "-t",
-                str(main_info["duration"]),
-                "-vf",
-                f"scale={pip_width}:{pip_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                temp_pip,
-            ]
-            self._run_ffmpeg(command)
-            pip_normalized_input = temp_pip
-        else:
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-i",
-                video_paths[1],
-                "-vf",
-                f"scale={pip_width}:{pip_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                pip_normalized,
-            ]
-            self._run_ffmpeg(command)
-            pip_normalized_input = pip_normalized
-
-        if main_info["duration"] > pip_info["duration"]:
-            looped_pip = os.path.join(self.work_dir, f"pip_looped_{os.getpid()}.mp4")
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-stream_loop",
-                "-1",
-                "-i",
-                pip_normalized_input,
-                "-t",
-                str(main_info["duration"]),
-                "-vf",
-                f"scale={pip_width}:{pip_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                looped_pip,
-            ]
-            self._run_ffmpeg(command)
-            pip_normalized_input = looped_pip
-
-        command = [
-            self._ffmpeg_bin,
-            "-y",
-            "-i",
-            main_normalized,
-            "-i",
-            pip_normalized_input,
-            "-filter_complex",
-            f"[0:v][1:v]overlay={x_offset}:{y_offset}[v]",
-            "-map",
-            "[v]",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            output_path,
-        ]
-        self._run_ffmpeg(command)
-
-        for temp_file in [main_normalized, pip_normalized]:
-            if temp_file and temp_file != output_path:
-                try:
-                    os.remove(temp_file)
-                except Exception as e:
-                    logger.warning(
-                        f"Operation failed in apps/worker/video_processing/video_compose_service.py: {e}", exc_info=True
-                    )
-
-        return output_path
-
-    def _voice_over(self, video_paths: list[str], audio_path: str, output_path: str) -> str:
-        """口播模式"""
-        if not audio_path:
-            raise ValueError("audio_path is required for VOICE_OVER mode")
-
-        if not video_paths:
-            raise ValueError("No background video provided")
-
-        audio_info = self._get_video_info(audio_path)
-        audio_duration = audio_info["duration"]
-
-        bg_normalized = os.path.join(self.work_dir, f"bg_{os.getpid()}.mp4")
-        bg_info = self._normalize_video(video_paths[0], bg_normalized)
-
-        if bg_info["duration"] < audio_duration:
-            looped_bg = os.path.join(self.work_dir, f"bg_looped_{os.getpid()}.mp4")
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-stream_loop",
-                "-1",
-                "-i",
-                bg_normalized,
-                "-t",
-                str(audio_duration),
-                "-vf",
-                f"scale={self.config.output_width}:{self.config.output_height}",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                looped_bg,
-            ]
-            self._run_ffmpeg(command)
-            bg_normalized = looped_bg
-        elif bg_info["duration"] > audio_duration:
-            temp_bg = os.path.join(self.work_dir, f"bg_trimmed_{os.getpid()}.mp4")
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-i",
-                bg_normalized,
-                "-t",
-                str(audio_duration),
-                "-c:v",
-                "copy",
-                temp_bg,
-            ]
-            self._run_ffmpeg(command)
-            bg_normalized = temp_bg
-
-        blurred_bg = os.path.join(self.work_dir, f"bg_blurred_{os.getpid()}.mp4")
-        command = [
-            self._ffmpeg_bin,
-            "-y",
-            "-i",
-            bg_normalized,
-            "-vf",
-            f"boxblur=5:5,scale={self.config.output_width}:{self.config.output_height}",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            blurred_bg,
-        ]
-        self._run_ffmpeg(command)
-
-        command = [
-            self._ffmpeg_bin,
-            "-y",
-            "-i",
-            blurred_bg,
-            "-i",
-            audio_path,
-            "-filter_complex",
-            "[0:v]drawbox=x=0:y=0:w=iw:h=ih:color=black@0.3:t=fill[v]",
-            "-map",
-            "[v]",
-            "-map",
-            "1:a",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            "-shortest",
-            output_path,
-        ]
-        self._run_ffmpeg(command)
-
-        for temp_file in [bg_normalized, blurred_bg]:
-            try:
-                if temp_file != output_path:
-                    os.remove(temp_file)
-            except Exception as e:
-                logger.warning(
-                    f"Operation failed in apps/worker/video_processing/video_compose_service.py: {e}", exc_info=True
-                )
-
-        return output_path
-
-    def _voice_pip(self, video_paths: list[str], audio_path: Optional[str], output_path: str) -> str:
-        """口播+画中画模式"""
-        if not video_paths:
-            raise ValueError("No video paths provided")
-
-        if len(video_paths) == 1:
-            return self._normalize_video(video_paths[0], output_path)
-
-        voice_video = video_paths[0]
-        bg_video = video_paths[1] if len(video_paths) > 1 else video_paths[0]
-
-        voice_normalized = os.path.join(self.work_dir, f"voice_{os.getpid()}.mp4")
-        voice_info = self._normalize_video(voice_video, voice_normalized)
-
-        bg_normalized = os.path.join(self.work_dir, f"bg_{os.getpid()}.mp4")
-        bg_info = self._normalize_video(bg_video, bg_normalized)
-
-        final_duration = min(voice_info["duration"], bg_info["duration"])
-
-        pip_width = int(self.config.output_width * self.config.pip_scale)
-        pip_height = int(self.config.output_height * self.config.pip_scale)
-        x_offset, y_offset = self._get_pip_position_offset(
-            self.config.output_width, self.config.output_height, pip_width, pip_height
-        )
-
-        voice_adjusted = os.path.join(self.work_dir, f"voice_adj_{os.getpid()}.mp4")
-        command = [
-            self._ffmpeg_bin,
-            "-y",
-            "-i",
-            voice_normalized,
-            "-t",
-            str(final_duration),
-            "-vf",
-            f"scale={pip_width}:{pip_height}",
-            "-c:v",
-            self.config.output_codec,
-            "-preset",
-            self.config.output_preset,
-            "-crf",
-            str(self.config.output_crf),
-            "-pix_fmt",
-            "yuv420p",
-            voice_adjusted,
-        ]
-        self._run_ffmpeg(command)
-
-        bg_adjusted = os.path.join(self.work_dir, f"bg_adj_{os.getpid()}.mp4")
-        command = [
-            self._ffmpeg_bin,
-            "-y",
-            "-i",
-            bg_normalized,
-            "-t",
-            str(final_duration),
-            "-c:v",
-            "copy",
-            bg_adjusted,
-        ]
-        self._run_ffmpeg(command)
-
-        if audio_path:
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-i",
-                bg_adjusted,
-                "-i",
-                voice_adjusted,
-                "-i",
-                audio_path,
-                "-filter_complex",
-                f"[0:v][1:v]overlay={x_offset}:{y_offset}[v]",
-                "-map",
-                "[v]",
-                "-map",
-                "2:a",
-                "-shortest",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                output_path,
-            ]
-        else:
-            command = [
-                self._ffmpeg_bin,
-                "-y",
-                "-i",
-                bg_adjusted,
-                "-i",
-                voice_adjusted,
-                "-filter_complex",
-                f"[0:v][1:v]overlay={x_offset}:{y_offset}[v]",
-                "-map",
-                "[v]",
-                "-map",
-                "1:a",
-                "-shortest",
-                "-c:v",
-                self.config.output_codec,
-                "-preset",
-                self.config.output_preset,
-                "-crf",
-                str(self.config.output_crf),
-                "-pix_fmt",
-                "yuv420p",
-                output_path,
-            ]
-        self._run_ffmpeg(command)
-
-        for temp_file in [voice_normalized, voice_adjusted, bg_normalized, bg_adjusted]:
-            try:
-                if temp_file != output_path:
-                    os.remove(temp_file)
-            except Exception as e:
-                logger.warning(
-                    f"Operation failed in apps/worker/video_processing/video_compose_service.py: {e}", exc_info=True
-                )
-
-        return output_path
-
-
-def create_compose_service(mode: str, work_dir: Optional[str] = None, **kwargs) -> VideoComposeService:
-    """便捷工厂函数：创建视频合成服务"""
-    try:
-        editing_mode = EditingMode(mode)
-    except ValueError:
-        raise ValueError(f"无效的剪辑模式: {mode}. 有效模式: {[m.value for m in EditingMode]}")
-
-    config = EditingModeConfig(
-        mode=editing_mode,
-        output_width=kwargs.get("output_width", 1280),
-        output_height=kwargs.get("output_height", 720),
-        output_fps=kwargs.get("output_fps", 25),
-        pip_position=PIPPosition(kwargs.get("pip_position", "top_right")),
-        pip_scale=kwargs.get("pip_scale", 0.25),
-        transition_duration=kwargs.get("transition_duration", 0.5),
-    )
-
-    return VideoComposeService(config=config, work_dir=work_dir)

apps/worker/worker_app/core/config.py

@@ -17,6 +17,10 @@ class WorkerSettings(BaseSettings):
     database_pool_recycle: int = 3600
     environment: str = "development"
     auto_create_schema: bool = False
+    redis_url: str = "redis://redis:6379/0"
+
+    # 渲染引擎选择：legacy=旧VideoComposeService，unified=新UnifiedRenderService
+    render_engine: str = "legacy"
 
     model_config = SettingsConfigDict(
         env_file=".env",

apps/worker/worker_app/tasks/asset_analyzer.py

@@ -10,12 +10,10 @@ from __future__ import annotations
 
 import json
 import logging
-import math
 import os
 import subprocess
 import tempfile
 from dataclasses import dataclass, field
-from typing import Any
 
 import numpy as np
 from PIL import Image

apps/worker/worker_app/tasks/classification.py

@@ -1,4 +1,3 @@
-from celery import Task
 from celery.utils.log import get_task_logger
 from worker_app.celery_app import celery_app
 
@@ -9,7 +8,6 @@ from packages.adapters.sqlalchemy_impl.classification_job_repository import (
     SQLAlchemyClassificationJobRepository,
 )
 from packages.domain import (
-    ClassificationJob,
     ClassificationJobStatus,
     ClassificationStatus,
 )

apps/worker/worker_app/tasks/compose_video.py

@@ -5,12 +5,9 @@
 
 from __future__ import annotations
 
-import logging
 import os
-import shutil
 import subprocess
 import tempfile
-from datetime import datetime, timezone
 from pathlib import Path
 
 from celery.utils.log import get_task_logger
@@ -39,6 +36,10 @@ def _get_job_service():
 def compose_video(self, job_id: str, **kwargs):
     """视频合成任务。
 
+    根据 RENDER_ENGINE 配置选择渲染引擎：
+      - legacy: 旧 VideoComposeService（filter_complex 模式）
+      - unified: 新 UnifiedRenderService（图层架构）
+
     Args:
         job_id: JobService 中的任务 ID
         **kwargs: 来自 Job.payload 的额外参数（plan_id, output_path 等）
@@ -56,66 +57,18 @@ def compose_video(self, job_id: str, **kwargs):
             job_service.fail_job(job_id, "Missing plan_id in job payload")
             return {"status": "error", "message": "Missing plan_id"}
 
-        # 标记为 running
-        job_service.update_progress(job_id, progress=10.0, current_stage="初始化合成环境")
+        # 判断使用哪个渲染引擎
+        # 优先级：Redis Feature Flag（白名单 > 百分比） > 环境变量默认
+        from video_processing.render_engine_resolver import get_render_engine_resolver
 
-        # 延迟导入 VideoComposeService
-        from apps.api.app.services.video_compose_service import VideoComposeService
+        resolver = get_render_engine_resolver()
+        user_id = job.created_by_user_id or None
+        engine = resolver.get_engine(user_id=user_id)
 
-        compose_svc = VideoComposeService(db)
-
-        # 校验合成条件
-        job_service.update_progress(job_id, progress=20.0, current_stage="校验合成条件")
-        validation = compose_svc.validate_compose(plan_id)
-        if not validation.valid:
-            error_msg = "; ".join(validation.errors)
-            job_service.fail_job(job_id, f"合成校验失败: {error_msg}")
-            return {"status": "error", "message": error_msg}
-
-        # 构建合成命令
-        job_service.update_progress(job_id, progress=30.0, current_stage="构建 FFmpeg 命令")
-        _output_dir = os.environ.get("VIDEO_OUTPUT_DIR", os.path.join(tempfile.gettempdir(), "video_output"))
-        output_path = os.path.join(_output_dir, f"{job_id}.mp4")
-        compose_cmd = compose_svc.build_compose_command(plan_id, output_path)
-
-        # 执行 FFmpeg
-        job_service.update_progress(job_id, progress=50.0, current_stage="正在执行视频合成")
-        logger.info("Executing FFmpeg for job %s, plan %s", job_id, plan_id)
-
-        try:
-            subprocess.run(
-                compose_cmd.command,
-                check=True,
-                stdout=subprocess.PIPE,
-                stderr=subprocess.PIPE,
-                text=True,
-                timeout=3600,
-            )
-        except subprocess.CalledProcessError as e:
-            job_service.fail_job(job_id, f"FFmpeg 执行失败: {e.stderr[:500]}")
-            raise
-
-        # 上传结果
-        job_service.update_progress(job_id, progress=80.0, current_stage="上传合成结果")
-        storage_key = f"rendered/{plan_id}/{job_id}.mp4"
-
-        from worker_app.tasks.edit_plan_generation import _upload_to_oss
-
-        output_url = _upload_to_oss(Path(output_path), storage_key)
-
-        # 更新 Job 状态为完成
-        result_data = {
-            "plan_id": plan_id,
-            "output_path": output_path,
-            "storage_key": storage_key,
-            "output_url": output_url or "",
-            "estimated_duration": compose_cmd.estimated_duration,
-            "clip_count": len(compose_cmd.clip_chains),
-        }
-        job_service.complete_job(job_id, result=result_data)
-
-        logger.info("视频合成完成: job_id=%s, plan_id=%s", job_id, plan_id)
-        return {"status": "completed", "job_id": job_id, "result": result_data}
+        if engine == "unified":
+            return _compose_with_unified_engine(self, job_service, job, plan_id, db)
+        else:
+            return _compose_with_legacy_engine(self, job_service, job, plan_id, db)
 
     except self.retry_exc as exc:
         logger.warning("视频合成重试中: job_id=%s, exc=%s", job_id, exc)
@@ -129,11 +82,145 @@ def compose_video(self, job_id: str, **kwargs):
         raise self.retry(exc=exc, countdown=60)
     finally:
         db.close()
-        # 清理临时文件
+
+
+def _compose_with_legacy_engine(task, job_service, job, plan_id: str, db) -> dict:
+    """旧引擎渲染路径（VideoComposeService）。"""
+    job_id = job.id
+
+    # 标记为 running
+    job_service.update_progress(job_id, progress=10.0, current_stage="初始化合成环境")
+
+    # 延迟导入 VideoComposeService
+    from apps.api.app.services.video_compose_service import VideoComposeService
+
+    compose_svc = VideoComposeService(db)
+
+    # 校验合成条件
+    job_service.update_progress(job_id, progress=20.0, current_stage="校验合成条件")
+    validation = compose_svc.validate_compose(plan_id)
+    if not validation.valid:
+        error_msg = "; ".join(validation.errors)
+        job_service.fail_job(job_id, f"合成校验失败: {error_msg}")
+        return {"status": "error", "message": error_msg}
+
+    # 构建合成命令
+    job_service.update_progress(job_id, progress=30.0, current_stage="构建 FFmpeg 命令")
+    _output_dir = os.environ.get("VIDEO_OUTPUT_DIR", os.path.join(tempfile.gettempdir(), "video_output"))
+    output_path = os.path.join(_output_dir, f"{job_id}.mp4")
+    compose_cmd = compose_svc.build_compose_command(plan_id, output_path)
+
+    # 执行 FFmpeg
+    job_service.update_progress(job_id, progress=50.0, current_stage="正在执行视频合成")
+    logger.info("Executing FFmpeg for job %s, plan %s", job_id, plan_id)
+
+    try:
+        subprocess.run(
+            compose_cmd.command,
+            check=True,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True,
+            timeout=3600,
+        )
+    except subprocess.CalledProcessError as e:
+        job_service.fail_job(job_id, f"FFmpeg 执行失败: {e.stderr[:500]}")
+        raise
+
+    # 上传结果
+    job_service.update_progress(job_id, progress=80.0, current_stage="上传合成结果")
+    storage_key = f"rendered/{plan_id}/{job_id}.mp4"
+
+    from worker_app.tasks.edit_plan_generation import _upload_to_oss
+
+    output_url = _upload_to_oss(Path(output_path), storage_key)
+
+    # 更新 Job 状态为完成
+    result_data = {
+        "plan_id": plan_id,
+        "output_path": output_path,
+        "storage_key": storage_key,
+        "output_url": output_url or "",
+        "estimated_duration": compose_cmd.estimated_duration,
+        "clip_count": len(compose_cmd.clip_chains),
+        "engine": "legacy",
+    }
+    job_service.complete_job(job_id, result=result_data)
+
+    logger.info("视频合成完成(legacy): job_id=%s, plan_id=%s", job_id, plan_id)
+    return {"status": "completed", "job_id": job_id, "result": result_data}
+
+
+def _compose_with_unified_engine(task, job_service, job, plan_id: str, db) -> dict:
+    """新引擎渲染路径（UnifiedRenderService + RenderAdapter）。"""
+    job_id = job.id
+
+    # 标记为 running
+    job_service.update_progress(job_id, progress=10.0, current_stage="初始化统一渲染引擎")
+
+    from video_processing.render_adapter import RenderAdapter
+
+    adapter = RenderAdapter(db)
+
+    # 校验合成条件
+    job_service.update_progress(job_id, progress=15.0, current_stage="校验合成条件")
+    valid, errors, warnings, ready_count, total_count = adapter.validate_plan(plan_id)
+    if not valid:
+        error_msg = "; ".join(errors)
+        job_service.fail_job(job_id, f"合成校验失败: {error_msg}")
+        return {"status": "error", "message": error_msg}
+
+    # 进度回调
+    def progress_cb(progress: float, stage: str) -> None:
         try:
-            _output_dir = os.environ.get("VIDEO_OUTPUT_DIR", os.path.join(tempfile.gettempdir(), "video_output"))
-            output_path = os.path.join(_output_dir, f"{job_id}.mp4")
-            if Path(output_path).exists():
-                Path(output_path).unlink()
-        except Exception as e:
-            logger.warning(f"Operation failed in apps/worker/worker_app/tasks/compose_video.py: {e}", exc_info=True)
+            job_service.update_progress(job_id, progress=progress, current_stage=stage)
+        except Exception:
+            logger.exception("更新进度失败")
+
+    # 执行渲染
+    job_service.update_progress(job_id, progress=20.0, current_stage="开始渲染")
+    logger.info("统一渲染引擎开始: job_id=%s plan_id=%s", job_id, plan_id)
+
+    result = adapter.render_plan(
+        plan_id=plan_id,
+        job_id=job_id,
+        progress_cb=progress_cb,
+    )
+
+    if not result.success:
+        job_service.fail_job(job_id, f"渲染失败: {result.error_message}")
+        raise RuntimeError(result.error_message)
+
+    # 更新 Job 状态为完成
+    result_data = {
+        "plan_id": plan_id,
+        "output_path": str(result.output_path) if result.output_path else "",
+        "storage_key": f"rendered/{plan_id}/{job_id}.mp4",
+        "output_url": result.output_url,
+        "estimated_duration": result.duration,
+        "clip_count": result.clip_count,
+        "engine": "unified",
+        "width": result.width,
+        "height": result.height,
+        "file_size": result.file_size,
+    }
+    job_service.complete_job(job_id, result=result_data)
+
+    logger.info(
+        "视频合成完成(unified): job_id=%s plan_id=%s duration=%.2fs",
+        job_id,
+        plan_id,
+        result.duration,
+    )
+    return {"status": "completed", "job_id": job_id, "result": result_data}
+
+
+def _cleanup_output(job_id: str) -> None:
+    """清理临时输出文件。"""
+    try:
+        _output_dir = os.environ.get("VIDEO_OUTPUT_DIR", os.path.join(tempfile.gettempdir(), "video_output"))
+        output_path = os.path.join(_output_dir, f"{job_id}.mp4")
+        if Path(output_path).exists():
+            Path(output_path).unlink()
+    except Exception as e:
+        logger.warning(f"清理输出文件失败: {e}", exc_info=True)

apps/worker/worker_app/tasks/edit_plan_generation.py

@@ -1,13 +1,18 @@
-"""剪辑计划渲染任务 — Phase 8 任务 2.05.
+"""剪辑计划渲染任务 — 支持 Feature Flag 灰度.
 
 Celery 任务 worker.render_edit_plan:
   1. 加载 EditPlan + EditPlanClips
-  2. 下载各片段素材
-  3. 使用 UnifiedRenderService 按时间线+图层渲染
+  2. 根据 Feature Flag 选择渲染引擎（legacy / unified）
+  3. 下载各片段素材 + 渲染
   4. 上传渲染结果到 OSS
   5. 创建 GeneratedVideo 记录 + 查重
   6. 更新 EditPlan / EditPlanClip 状态
   7. 更新 GenerationTask 进度
+
+渲染引擎灰度：
+  - 走 Feature Flag (render_engine) 控制
+  - legacy: VideoComposeService + FFmpeg filter_complex
+  - unified: UnifiedRenderService 图层架构
 """
 
 from __future__ import annotations
@@ -16,7 +21,6 @@ import logging
 import tempfile
 from datetime import datetime, timezone
 from pathlib import Path
-from typing import Optional
 
 from worker_app.celery_app import celery_app
 from worker_app.db import SessionLocal
@@ -63,14 +67,268 @@ def _get_repos():
 # ── Celery Task ───────────────────────────────────────────────────────────────
 
 
+def _resolve_render_engine(user_id: str) -> str:
+    """根据 Feature Flag 决定使用哪个渲染引擎。
+
+    Returns:
+        "legacy" 或 "unified"
+    """
+    try:
+        from video_processing.render_engine_resolver import get_render_engine_resolver
+
+        resolver = get_render_engine_resolver()
+        return resolver.get_engine(user_id=user_id)
+    except Exception as exc:
+        logger.warning("获取渲染引擎配置失败，fallback 到 legacy: %s", exc)
+        return "legacy"
+
+
+def _mark_plan_failed(plan_repo, plan_id, gen_task_repo, generation_task_id, error_msg: str):
+    """统一的计划失败标记工具。"""
+    plan = plan_repo.get(plan_id)
+    if plan and plan.status.value == "rendering":
+        plan.mark_failed()
+        plan_repo.update(plan)
+    if generation_task_id:
+        gen_task = gen_task_repo.get(generation_task_id)
+        if gen_task and gen_task.status.value != "failed":
+            gen_task.status = "failed"
+            gen_task.error_message = error_msg
+            gen_task.completed_at = datetime.now(timezone.utc)
+            gen_task_repo.update(gen_task)
+
+
+def _finalize_render_success(
+    plan,
+    plan_repo,
+    clip_repo,
+    gen_task_repo,
+    db,
+    plan_id: str,
+    output_url: str,
+    storage_key: str,
+    duration: float,
+    file_size: int,
+    width: int,
+    height: int,
+    rendered_clip_ids: list[str],
+    failed_clip_ids: list[str],
+    generation_task_id: str,
+    output_path: Path,
+    engine: str,
+) -> dict:
+    """渲染成功后的统一收尾：查重 + 更新状态 + 返回结果。"""
+    # 创建 GeneratedVideo 记录 + 查重
+    project_id = plan.project_id or ""
+    batch_id = plan.config.get("batch_id", "")
+    mode = plan.config.get("mode", "edit_plan")
+    if generation_task_id and project_id:
+        try:
+            create_video_record_and_dedup(
+                generation_task_id=generation_task_id,
+                project_id=project_id,
+                batch_id=batch_id,
+                file_url=output_url or "",
+                file_size=file_size,
+                duration=duration,
+                video_path=str(output_path),
+                mode=mode,
+                session=db,
+                width=width,
+                height=height,
+                fps=OUTPUT_FPS,
+            )
+        except Exception as dedup_err:
+            logger.warning("查重失败（不影响渲染结果）: %s", dedup_err)
+
+    # 更新片段状态为 rendered
+    for clip_id in rendered_clip_ids:
+        clip = clip_repo.get(clip_id)
+        if clip and clip.status.value == "ready":
+            clip.mark_rendered()
+            clip_repo.update(clip)
+
+    # 更新 EditPlan 状态为 completed
+    plan.config["rendered_url"] = output_url or ""
+    plan.config["rendered_storage_key"] = storage_key
+    plan.mark_completed()
+    plan_repo.update(plan)
+
+    # 更新 GenerationTask 状态为 completed
+    if generation_task_id:
+        gen_task = gen_task_repo.get(generation_task_id)
+        if gen_task:
+            gen_task.status = "completed"
+            gen_task.progress = 100.0
+            gen_task.result_count = len(rendered_clip_ids)
+            gen_task.completed_at = datetime.now(timezone.utc)
+            gen_task_repo.update(gen_task)
+
+    logger.info(
+        "剪辑计划渲染完成: plan_id=%s engine=%s rendered=%d failed=%d duration=%.1fs",
+        plan_id,
+        engine,
+        len(rendered_clip_ids),
+        len(failed_clip_ids),
+        duration,
+    )
+
+    return {
+        "status": "completed",
+        "plan_id": plan_id,
+        "rendered_count": len(rendered_clip_ids),
+        "failed_count": len(failed_clip_ids),
+        "output_url": output_url,
+        "duration": duration,
+    }
+
+
+def _render_with_unified(
+    plan,
+    clips,
+    asset_path_map: dict[str, Path],
+    tmpdir_path: Path,
+    rendered_clip_ids: list[str],
+    plan_id: str,
+    generation_task_id: str,
+    plan_repo,
+    clip_repo,
+    gen_task_repo,
+    db,
+) -> dict:
+    """统一渲染引擎路径（UnifiedRenderService 图层架构）。"""
+    render_service = UnifiedRenderService(
+        plan=plan,
+        clips=clips,
+        asset_path_map=asset_path_map,
+        work_dir=tmpdir_path,
+        output_width=OUTPUT_WIDTH,
+        output_height=OUTPUT_HEIGHT,
+        output_fps=int(OUTPUT_FPS),
+    )
+
+    try:
+        render_result = render_service.render()
+    except Exception as render_err:
+        logger.error("渲染失败(unified): %s — %s", plan_id, render_err)
+        _mark_plan_failed(plan_repo, plan_id, gen_task_repo, generation_task_id, f"渲染失败: {render_err}")
+        return {"status": "error", "message": f"渲染失败: {render_err}"}
+
+    output_path = render_result.output_path
+
+    # 上传到 OSS
+    storage_key = f"rendered/{plan_id}/output.mp4"
+    output_url = upload_to_oss(output_path, storage_key)
+
+    failed_clip_ids: list[str] = []
+    return _finalize_render_success(
+        plan=plan,
+        plan_repo=plan_repo,
+        clip_repo=clip_repo,
+        gen_task_repo=gen_task_repo,
+        db=db,
+        plan_id=plan_id,
+        output_url=output_url or "",
+        storage_key=storage_key,
+        duration=render_result.duration,
+        file_size=render_result.file_size,
+        width=render_result.width,
+        height=render_result.height,
+        rendered_clip_ids=rendered_clip_ids,
+        failed_clip_ids=failed_clip_ids,
+        generation_task_id=generation_task_id,
+        output_path=output_path,
+        engine="unified",
+    )
+
+
+def _render_with_legacy(
+    plan,
+    clips,
+    rendered_clip_ids: list[str],
+    failed_clip_ids: list[str],
+    tmpdir_path: Path,
+    plan_id: str,
+    generation_task_id: str,
+    plan_repo,
+    clip_repo,
+    gen_task_repo,
+    db,
+) -> dict:
+    """旧引擎路径（VideoComposeService + FFmpeg filter_complex）。"""
+    import os
+    import subprocess
+
+    from apps.api.app.services.video_compose_service import VideoComposeService
+
+    compose_svc = VideoComposeService(db)
+
+    # 校验合成条件
+    validation = compose_svc.validate_compose(plan_id)
+    if not validation.valid:
+        error_msg = "; ".join(validation.errors)
+        logger.error("合成校验失败(legacy): %s — %s", plan_id, error_msg)
+        _mark_plan_failed(plan_repo, plan_id, gen_task_repo, generation_task_id, f"合成校验失败: {error_msg}")
+        return {"status": "error", "message": error_msg}
+
+    # 构建 FFmpeg 命令
+    output_dir = os.environ.get("VIDEO_OUTPUT_DIR", str(tmpdir_path))
+    output_path = Path(output_dir) / f"{plan_id}.mp4"
+    compose_cmd = compose_svc.build_compose_command(plan_id, str(output_path))
+
+    logger.info("执行 FFmpeg (legacy): plan_id=%s", plan_id)
+    try:
+        subprocess.run(
+            compose_cmd.command,
+            check=True,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True,
+            timeout=3600,
+        )
+    except subprocess.CalledProcessError as e:
+        error_msg = f"FFmpeg 执行失败: {e.stderr[:500]}"
+        logger.error("FFmpeg 执行失败(legacy): %s — %s", plan_id, error_msg)
+        _mark_plan_failed(plan_repo, plan_id, gen_task_repo, generation_task_id, error_msg)
+        return {"status": "error", "message": error_msg}
+
+    # 获取文件大小
+    file_size = output_path.stat().st_size if output_path.exists() else 0
+    duration = compose_cmd.estimated_duration or 0.0
+
+    # 上传到 OSS
+    storage_key = f"rendered/{plan_id}/output.mp4"
+    output_url = upload_to_oss(output_path, storage_key)
+
+    return _finalize_render_success(
+        plan=plan,
+        plan_repo=plan_repo,
+        clip_repo=clip_repo,
+        gen_task_repo=gen_task_repo,
+        db=db,
+        plan_id=plan_id,
+        output_url=output_url or "",
+        storage_key=storage_key,
+        duration=duration,
+        file_size=file_size,
+        width=OUTPUT_WIDTH,
+        height=OUTPUT_HEIGHT,
+        rendered_clip_ids=rendered_clip_ids,
+        failed_clip_ids=failed_clip_ids,
+        generation_task_id=generation_task_id,
+        output_path=output_path,
+        engine="legacy",
+    )
+
+
 @celery_app.task(name="worker.render_edit_plan", bind=True, max_retries=2)
 def render_edit_plan(self, plan_id: str) -> dict:
     """渲染剪辑计划
 
     流程：
     1. 加载 EditPlan + EditPlanClips
-    2. 下载各片段素材到临时目录，构建 asset_path_map
-    3. 使用 UnifiedRenderService 按时间线+图层渲染
+    2. 根据 Feature Flag 选择渲染引擎（legacy / unified）
+    3. 下载素材 + 渲染
     4. 上传渲染结果到 OSS
     5. 创建 GeneratedVideo 记录 + 查重
     6. 更新 EditPlan → completed, EditPlanClips → rendered
@@ -79,6 +337,7 @@ def render_edit_plan(self, plan_id: str) -> dict:
     logger.info("开始渲染剪辑计划: plan_id=%s", plan_id)
 
     generation_task_id = ""
+    engine = "legacy"
 
     for repos in _get_repos():
         plan_repo, clip_repo, gen_task_repo, db = repos
@@ -93,7 +352,12 @@ def render_edit_plan(self, plan_id: str) -> dict:
             # 获取 generation_task_id（提前读取，确保 except 块可用）
             generation_task_id = plan.config.get("generation_task_id", "")
 
-            # 2. 加载片段列表（按 order 排序）
+            # 2. 选择渲染引擎（Feature Flag 灰度控制）
+            user_id = plan.created_by_user_id or ""
+            engine = _resolve_render_engine(user_id)
+            logger.info("剪辑计划渲染引擎: plan_id=%s engine=%s user_id=%s", plan_id, engine, user_id)
+
+            # 3. 加载片段列表（按 order 排序）
             clips = clip_repo.list_by_plan(plan_id, skip=0, limit=10000)
             if not clips:
                 logger.warning("剪辑计划没有片段: %s", plan_id)
@@ -116,6 +380,15 @@ def render_edit_plan(self, plan_id: str) -> dict:
                 rendered_clip_ids: list[str] = []
                 failed_clip_ids: list[str] = []
 
+                # 预先批量查询所有素材的 storage_key（file_url）
+                from packages.adapters.sqlalchemy_impl.models import AssetModel
+
+                clip_asset_ids = [c.asset_id for c in clips if c.asset_id]
+                asset_storage_map: dict[str, str] = {}
+                if clip_asset_ids:
+                    assets = db.query(AssetModel).filter(AssetModel.id.in_(clip_asset_ids)).all()
+                    asset_storage_map = {a.id: a.file_url for a in assets if a.file_url}
+
                 for clip in clips:
                     if not clip.asset_id:
                         # 没有素材的片段跳过，标记为失败
@@ -129,10 +402,22 @@ def render_edit_plan(self, plan_id: str) -> dict:
                         rendered_clip_ids.append(clip.id)
                         continue
 
+                    storage_key = asset_storage_map.get(clip.asset_id)
+                    if not storage_key:
+                        logger.warning(
+                            "片段素材无 storage_key，跳过: clip_id=%s asset_id=%s",
+                            clip.id,
+                            clip.asset_id,
+                        )
+                        clip.mark_failed()
+                        clip_repo.update(clip)
+                        failed_clip_ids.append(clip.id)
+                        continue
+
                     # 下载素材
-                    ext = Path(clip.asset_id).suffix or ".mp4"
+                    ext = Path(storage_key).suffix or ".mp4"
                     local_path = tmpdir_path / f"clip_{clip.order:04d}{ext}"
-                    if download_asset(clip.asset_id, local_path):
+                    if download_asset(storage_key, local_path):
                         asset_path_map[clip.asset_id] = local_path
                         rendered_clip_ids.append(clip.id)
                     else:
@@ -153,100 +438,38 @@ def render_edit_plan(self, plan_id: str) -> dict:
                             gen_task_repo.update(gen_task)
                     return {"status": "error", "message": "所有片段素材下载失败"}
 
-                # 4. 使用 UnifiedRenderService 渲染
-                render_service = UnifiedRenderService(
-                    plan=plan,
-                    clips=clips,
-                    asset_path_map=asset_path_map,
-                    work_dir=tmpdir_path,
-                    output_width=OUTPUT_WIDTH,
-                    output_height=OUTPUT_HEIGHT,
-                    output_fps=int(OUTPUT_FPS),
-                )
+                # 4. 根据引擎选择渲染方式
+                if engine == "unified":
+                    result = _render_with_unified(
+                        plan=plan,
+                        clips=clips,
+                        asset_path_map=asset_path_map,
+                        tmpdir_path=tmpdir_path,
+                        rendered_clip_ids=rendered_clip_ids,
+                        plan_id=plan_id,
+                        generation_task_id=generation_task_id,
+                        plan_repo=plan_repo,
+                        clip_repo=clip_repo,
+                        gen_task_repo=gen_task_repo,
+                        db=db,
+                    )
+                else:
+                    result = _render_with_legacy(
+                        plan=plan,
+                        clips=clips,
+                        rendered_clip_ids=rendered_clip_ids,
+                        failed_clip_ids=failed_clip_ids,
+                        tmpdir_path=tmpdir_path,
+                        plan_id=plan_id,
+                        generation_task_id=generation_task_id,
+                        plan_repo=plan_repo,
+                        clip_repo=clip_repo,
+                        gen_task_repo=gen_task_repo,
+                        db=db,
+                    )
 
-                try:
-                    render_result = render_service.render()
-                except Exception as render_err:
-                    logger.error("渲染失败: %s — %s", plan_id, render_err)
-                    plan.mark_failed()
-                    plan_repo.update(plan)
-                    if generation_task_id:
-                        gen_task = gen_task_repo.get(generation_task_id)
-                        if gen_task:
-                            gen_task.status = "failed"
-                            gen_task.error_message = f"渲染失败: {render_err}"
-                            gen_task.completed_at = datetime.now(timezone.utc)
-                            gen_task_repo.update(gen_task)
-                    return {"status": "error", "message": f"渲染失败: {render_err}"}
-
-                output_path = render_result.output_path
-
-                # 5. 上传到 OSS
-                storage_key = f"rendered/{plan_id}/output.mp4"
-                output_url = upload_to_oss(output_path, storage_key)
-
-                # 6. 创建 GeneratedVideo 记录 + 查重
-                project_id = plan.project_id or ""
-                batch_id = plan.config.get("batch_id", "")
-                mode = plan.config.get("mode", "edit_plan")
-                if generation_task_id and project_id:
-                    try:
-                        create_video_record_and_dedup(
-                            generation_task_id=generation_task_id,
-                            project_id=project_id,
-                            batch_id=batch_id,
-                            file_url=output_url or "",
-                            file_size=render_result.file_size,
-                            duration=render_result.duration,
-                            video_path=str(output_path),
-                            mode=mode,
-                            session=db,
-                            width=render_result.width,
-                            height=render_result.height,
-                            fps=OUTPUT_FPS,
-                        )
-                    except Exception as dedup_err:
-                        logger.warning("查重失败（不影响渲染结果）: %s", dedup_err)
-
-                # 7. 更新片段状态为 rendered
-                for clip_id in rendered_clip_ids:
-                    clip = clip_repo.get(clip_id)
-                    if clip and clip.status.value == "ready":
-                        clip.mark_rendered()
-                        clip_repo.update(clip)
-
-                # 8. 更新 EditPlan 状态为 completed
-                plan.config["rendered_url"] = output_url or ""
-                plan.config["rendered_storage_key"] = storage_key
-                plan.mark_completed()
-                plan_repo.update(plan)
-
-                # 9. 更新 GenerationTask 状态为 completed
-                if generation_task_id:
-                    gen_task = gen_task_repo.get(generation_task_id)
-                    if gen_task:
-                        gen_task.status = "completed"
-                        gen_task.progress = 100.0
-                        gen_task.result_count = len(rendered_clip_ids)
-                        gen_task.completed_at = datetime.now(timezone.utc)
-                        gen_task_repo.update(gen_task)
-
-                logger.info(
-                    "剪辑计划渲染完成: plan_id=%s rendered=%d failed=%d duration=%.1fs",
-                    plan_id,
-                    len(rendered_clip_ids),
-                    len(failed_clip_ids),
-                    render_result.duration,
-                )
-
-                return {
-                    "status": "completed",
-                    "plan_id": plan_id,
-                    "rendered_count": len(rendered_clip_ids),
-                    "failed_count": len(failed_clip_ids),
-                    "output_url": output_url,
-                    "duration": render_result.duration,
-                }
+                result["engine"] = engine
+                return result
 
         except Exception as exc:
             logger.exception("渲染剪辑计划异常: %s", plan_id)

apps/worker/worker_app/tasks/generation.py

@@ -13,15 +13,13 @@
 
 from __future__ import annotations
 
-import json
 import logging
 import os
 import tempfile
 import time
 from dataclasses import dataclass, field
-from datetime import datetime, timezone
 from pathlib import Path
-from typing import Any, Optional
+from typing import Any
 
 from worker_app.celery_app import celery_app
 from worker_app.db import SessionLocal
@@ -69,7 +67,11 @@ def _update_task_status(task_id: str, status_action: str, **kwargs) -> bool:
 
             action(**kwargs)
             repo.update(task)
-            logger.info("GenerationTask 状态更新成功: task_id=%s action=%s", task_id, status_action)
+            logger.info(
+                "GenerationTask 状态更新成功: task_id=%s action=%s",
+                task_id,
+                status_action,
+            )
             return True
         finally:
             session.close()
@@ -113,6 +115,7 @@ from video_processing.oss_helpers import (
     get_signed_download_url,
     upload_to_oss,
 )
+from video_processing.render_engine_resolver import ENGINE_LEGACY, ENGINE_UNIFIED
 from video_processing.unified_render_service import UnifiedRenderService
 
 # ── 虚拟 Plan / Clip（内存中构建，不写数据库） ────────────────────────────────
@@ -124,6 +127,7 @@ class _VirtualPlan:
 
     id: str
     name: str = ""
+    config: dict[str, Any] = field(default_factory=dict)
 
 
 @dataclass
@@ -161,16 +165,17 @@ def _build_plan_and_clips_from_task(
     """
     plan = _VirtualPlan(id=task_id, name=f"Generated-{task_id[:8]}")
 
-    # 为每个下载路径生成合成 asset_id
+    # 为每个下载路径生成合成 asset_id，并预探测素材时长
     asset_path_map: dict[str, Path] = {}
     path_to_asset_id: dict[Path, str] = {}
+    path_duration: dict[Path, float] = {}
     for i, p in enumerate(downloaded_paths):
         asset_id = f"gen_{task_id[:8]}_{i:03d}{p.suffix or '.mp4'}"
         asset_path_map[asset_id] = p
         path_to_asset_id[p] = asset_id
+        path_duration[p] = probe_duration(p)
 
     clips: list[_VirtualClip] = []
-    n = len(downloaded_paths)
 
     if mode == "pip":
         # 1 main + N-1 overlay
@@ -183,6 +188,7 @@ def _build_plan_and_clips_from_task(
                     clip_type=clip_type,
                     order=i,
                     asset_id=path_to_asset_id[p],
+                    duration=path_duration[p],
                 )
             )
     elif mode == "voice_over":
@@ -195,6 +201,7 @@ def _build_plan_and_clips_from_task(
                     clip_type="main",
                     order=i,
                     asset_id=path_to_asset_id[p],
+                    duration=path_duration[p],
                     config={"role": "b_roll"},
                 )
             )
@@ -214,6 +221,7 @@ def _build_plan_and_clips_from_task(
                     clip_type=clip_type,
                     order=i,
                     asset_id=path_to_asset_id[p],
+                    duration=path_duration[p],
                 )
             )
     else:
@@ -226,6 +234,7 @@ def _build_plan_and_clips_from_task(
                     clip_type="main",
                     order=i,
                     asset_id=path_to_asset_id[p],
+                    duration=path_duration[p],
                 )
             )
 
@@ -380,31 +389,37 @@ def _download_library_assets(
 
         session = SessionLocal()
         try:
-            # 构建查询：根据模式选择不同的过滤条件
+            # 构建查询
             query = session.query(AssetModel).filter(
                 AssetModel.status == "ready",
                 AssetModel.file_type.in_(["video", "video/mp4", "video/quicktime"]),
             )
 
-            if asset_library_id:
-                # 素材库模式
-                query = query.filter(AssetModel.asset_library_id == asset_library_id)
+            if asset_ids:
+                # 明确指定了 asset_ids：直接按 ID 查，不预先按 library/project 过滤
+                # 避免项目级素材或跨库素材因为 library_id 不匹配而查不到
+                # 归属安全由后面的归属校验保证
+                query = query.filter(AssetModel.id.in_(asset_ids))
                 logger.info(
-                    "下载素材库视频: asset_library_id=%s asset_ids=%s",
-                    asset_library_id,
-                    asset_ids or "all",
+                    "下载指定素材: asset_ids=%d 个, asset_library_id=%s, project_id=%s",
+                    len(asset_ids),
+                    asset_library_id or "none",
+                    project_id or "none",
                 )
             else:
-                # 项目级模式
-                query = query.filter(AssetModel.project_id == project_id)
-                logger.info(
-                    "下载项目级视频: project_id=%s asset_ids=%s",
-                    project_id,
-                    asset_ids or "all",
-                )
-
-            if asset_ids:
-                query = query.filter(AssetModel.id.in_(asset_ids))
+                # 未指定 asset_ids：按 library 或 project 下载全部 ready 视频
+                if asset_library_id:
+                    query = query.filter(AssetModel.asset_library_id == asset_library_id)
+                    logger.info(
+                        "下载素材库全部视频: asset_library_id=%s",
+                        asset_library_id,
+                    )
+                else:
+                    query = query.filter(AssetModel.project_id == project_id)
+                    logger.info(
+                        "下载项目全部视频: project_id=%s",
+                        project_id,
+                    )
 
             assets = query.order_by(AssetModel.created_at).all()
 
@@ -421,13 +436,15 @@ def _download_library_assets(
                 if missing_ids:
                     raise ValueError(f"素材不存在: asset_ids={sorted(missing_ids)}")
                 for asset in assets:
+                    # 校验素材库归属（只要传了 asset_library_id 就校验）
                     if asset_library_id and asset.asset_library_id != asset_library_id:
                         raise ValueError(
                             f"素材不属于指定素材库: asset_id={asset.id}, "
                             f"expected_asset_library_id={asset_library_id}, "
                             f"actual_asset_library_id={asset.asset_library_id}"
                         )
-                    if not asset_library_id and project_id and asset.project_id != project_id:
+                    # 校验项目归属（只要传了 project_id 就校验）
+                    if project_id and asset.project_id != project_id:
                         raise ValueError(
                             f"素材不属于指定项目: asset_id={asset.id}, "
                             f"expected_project_id={project_id}, "
@@ -445,7 +462,10 @@ def _download_library_assets(
                 if not storage_key:
                     failed_assets.append(f"{asset.name}({asset.id})")
                     logger.warning(
-                        "[task_id=%s] 素材缺少 file_url, 跳过: asset_id=%s name=%s", task_id, asset.id, asset.name
+                        "[task_id=%s] 素材缺少 file_url, 跳过: asset_id=%s name=%s",
+                        task_id,
+                        asset.id,
+                        asset.name,
                     )
                     if gen_task:
                         gen_task.append_log(
@@ -491,7 +511,12 @@ def _download_library_assets(
                         )
                 else:
                     failed_assets.append(f"{asset.name}({asset.id})")
-                    logger.warning("[task_id=%s] Failed to download asset: %s (id=%s)", task_id, asset.name, asset.id)
+                    logger.warning(
+                        "[task_id=%s] Failed to download asset: %s (id=%s)",
+                        task_id,
+                        asset.name,
+                        asset.id,
+                    )
                     if gen_task:
                         gen_task.append_log(
                             "下载素材",
@@ -558,6 +583,380 @@ def _validate_template_exists(template_id: str) -> None:
         session.close()
 
 
+# ── 渲染引擎选择 ─────────────────────────────────────────────────────────────
+
+
+def _resolve_render_engine(user_id: str) -> str:
+    """根据 Feature Flag 决定使用哪个渲染引擎。
+
+    Returns:
+        "legacy" 或 "unified"
+    """
+    try:
+        from video_processing.render_engine_resolver import get_render_engine_resolver
+
+        resolver = get_render_engine_resolver()
+        return resolver.get_engine(user_id=user_id)
+    except Exception as exc:
+        logger.warning("获取渲染引擎配置失败，fallback 到 unified: %s", exc)
+        return ENGINE_UNIFIED
+
+
+# ── 旧引擎渲染（FFmpeg filter_complex） ────────────────────────────────────────
+
+
+def _render_with_legacy_engine(
+    task_id: str,
+    virtual_clips: list[_VirtualClip],
+    asset_path_map: dict[str, Path],
+    work_dir: Path,
+    output_path: Path,
+) -> tuple[float, int]:
+    """旧引擎渲染路径：手动构建 FFmpeg filter_complex 命令。
+
+    说明：generate_video 任务使用虚拟 clips（无 EditPlan 数据库记录），
+    因此无法直接复用 VideoComposeService。这里手动构建等价的 filter_complex
+    命令，与旧引擎行为一致（scale → crop → setpts → trim → setpts，
+    无 fps 归一化，保持原帧率）。
+
+    支持模式：one_take / pip / voice_over / voice_pip
+    - 所有模式统一走 concat 滤镜（与旧引擎多片段逻辑一致）
+
+    Returns:
+        (duration_seconds, file_size_bytes)
+    """
+    import subprocess
+
+    main_clips = [
+        c
+        for c in virtual_clips
+        if c.clip_type in ("main", "b_roll", "background")
+        or (c.clip_type == "main" and c.config.get("role") == "b_roll")
+    ]
+    if not main_clips:
+        main_clips = virtual_clips[:1]
+
+    input_args: list[str] = []
+    video_filters: list[str] = []
+    audio_filters: list[str] = []
+
+    for i, clip in enumerate(main_clips):
+        local_path = asset_path_map.get(clip.asset_id)
+        if not local_path:
+            continue
+        input_args.extend(["-i", str(local_path)])
+
+        duration = clip.duration or 0.0
+
+        # 视频滤镜：scale → crop → setpts → trim → setpts（与旧引擎一致）
+        vf = (
+            f"[{i}:v]"
+            f"scale={OUTPUT_WIDTH}:{OUTPUT_HEIGHT}:force_original_aspect_ratio=increase,"
+            f"crop={OUTPUT_WIDTH}:{OUTPUT_HEIGHT},"
+            f"setpts=PTS-STARTPTS,"
+            f"trim=0:{duration:.3f},"
+            f"setpts=PTS-STARTPTS"
+            f"[v{i}]"
+        )
+        video_filters.append(vf)
+
+        # 音频滤镜：atrim → asetpts
+        af = f"[{i}:a]atrim=0:{duration:.3f},asetpts=PTS-STARTPTS[a{i}]"
+        audio_filters.append(af)
+
+    n = len(main_clips)
+
+    if n == 1:
+        video_label = "[v0]"
+        audio_label = "[a0]"
+    else:
+        # concat 视频
+        v_inputs = "".join(f"[v{i}]" for i in range(n))
+        video_filters.append(f"{v_inputs}concat=n={n}:v=1:a=0[outv]")
+        # concat 音频
+        a_inputs = "".join(f"[a{i}]" for i in range(n))
+        audio_filters.append(f"{a_inputs}concat=n={n}:v=0:a=1[outa]")
+        video_label = "[outv]"
+        audio_label = "[outa]"
+
+    # 组装 filter_complex
+    fc_parts = video_filters + audio_filters
+    filter_complex = ";".join(fc_parts)
+
+    command = [
+        FFMPEG_BIN,
+        "-y",
+        *input_args,
+        "-filter_complex",
+        filter_complex,
+        "-map",
+        video_label,
+        "-map",
+        audio_label,
+        "-c:v",
+        "libx264",
+        "-crf",
+        "23",
+        "-preset",
+        "medium",
+        "-c:a",
+        "aac",
+        "-b:a",
+        "192k",
+        "-movflags",
+        "+faststart",
+        str(output_path),
+    ]
+
+    logger.info("[task_id=%s] [渲染] legacy 引擎 FFmpeg 开始: clips=%d", task_id, n)
+    try:
+        run_ffmpeg(command)
+    except subprocess.CalledProcessError as e:
+        logger.error(
+            "[task_id=%s] [渲染] legacy 引擎 FFmpeg 失败: %s\nfilter_complex: %s",
+            task_id,
+            e,
+            filter_complex[:500],
+        )
+        raise
+
+    file_size = output_path.stat().st_size if output_path.exists() else 0
+    duration = probe_duration(output_path)
+    return duration, file_size
+
+
+# ── generate_video 阶段子函数 ─────────────────────────────────────────────────
+
+
+def _load_task_info(task_id: str) -> dict | None:
+    """从数据库加载 GenerationTask 元数据。
+
+    Returns:
+        包含任务元数据的字典，任务不存在时返回 None。
+    """
+    from packages.adapters.sqlalchemy_impl.generation_task_repository import (
+        SQLAlchemyGenerationTaskRepository,
+    )
+
+    session = SessionLocal()
+    try:
+        task_repo = SQLAlchemyGenerationTaskRepository(session)
+        gen_task = task_repo.get(task_id)
+        if gen_task is None:
+            return None
+
+        return {
+            "project_id": gen_task.project_id,
+            "asset_library_id": gen_task.asset_library_id,
+            "voice_library_id": gen_task.voice_library_id or "",
+            "template_id": getattr(gen_task, "template_id", "") or "",
+            "mode": gen_task.strategy_id or "one_take",
+            "task_asset_ids": list(gen_task.asset_ids or []),
+            "batch_id": getattr(gen_task, "batch_id", "") or "",
+            "user_id": getattr(gen_task, "created_by_user_id", "") or "",
+        }
+    finally:
+        session.close()
+
+
+def _download_all_assets(
+    temp_path: Path,
+    asset_library_id: str,
+    project_id: str,
+    task_asset_ids: list[str],
+    voice_library_id: str,
+    task_id: str,
+) -> tuple[list[Path], str | None]:
+    """下载视频素材和配音素材。
+
+    Returns:
+        (downloaded_videos, audio_path)
+
+    Note: gen_task 不传入下载函数（session 已关闭），
+    主函数在下载前后已有汇总日志。
+    """
+    logger.info("[task_id=%s] [下载素材] 开始下载视频素材", task_id)
+    download_start = time.monotonic()
+    downloaded_videos = _download_library_assets(
+        temp_path,
+        asset_library_id=asset_library_id,
+        project_id=project_id,
+        asset_ids=task_asset_ids or None,
+        task_id=task_id,
+    )
+    download_elapsed = time.monotonic() - download_start
+    logger.info(
+        "[task_id=%s] [下载素材] 完成: 成功=%d个, 耗时=%.1fs",
+        task_id,
+        len(downloaded_videos),
+        download_elapsed,
+    )
+
+    audio_path: str | None = None
+    if voice_library_id:
+        local_audio = temp_path / "voice.mp3"
+        if _download_voice_asset(voice_library_id, local_audio):
+            audio_path = str(local_audio)
+            logger.info("[task_id=%s] [下载配音] 配音下载成功", task_id)
+
+    return downloaded_videos, audio_path
+
+
+def _render_video(
+    task_id: str,
+    downloaded_videos: list[Path],
+    voice_path: str | None,
+    editing_mode,
+    project_id: str,
+    template_id: str,
+    user_id: str,
+    temp_path: Path,
+    output_name: str,
+) -> tuple[Path, float]:
+    """渲染视频（含配音混音）。
+
+    Returns:
+        (output_path, render_duration)
+    """
+    if not downloaded_videos:
+        raise RuntimeError(f"素材下载结果为空: task_id={task_id}")
+
+    # 构建虚拟 plan + clips
+    virtual_plan, virtual_clips, asset_path_map = _build_plan_and_clips_from_task(
+        task_id=task_id,
+        downloaded_paths=downloaded_videos,
+        mode=editing_mode.value,
+    )
+
+    total_duration = sum(c.duration for c in virtual_clips)
+    logger.info(
+        "[task_id=%s] [剪辑计划] 片段数=%d, 总时长=%.1fs",
+        task_id,
+        len(virtual_clips),
+        total_duration,
+    )
+
+    # 选择渲染引擎
+    engine = _resolve_render_engine(user_id) if user_id else ENGINE_UNIFIED
+    logger.info("[task_id=%s] [渲染] 引擎选择: %s (user_id=%s)", task_id, engine, user_id)
+
+    render_start = time.monotonic()
+    render_output_path = temp_path / f"rendered-{task_id}.mp4"
+
+    if engine == ENGINE_LEGACY:
+        render_duration, _ = _render_with_legacy_engine(
+            task_id=task_id,
+            virtual_clips=virtual_clips,
+            asset_path_map=asset_path_map,
+            work_dir=temp_path,
+            output_path=render_output_path,
+        )
+    else:
+        logger.info("[task_id=%s] [渲染] unified 引擎 FFmpeg 渲染开始", task_id)
+        render_service = UnifiedRenderService(
+            plan=virtual_plan,
+            clips=virtual_clips,
+            asset_path_map=asset_path_map,
+            work_dir=temp_path,
+            output_width=OUTPUT_WIDTH,
+            output_height=OUTPUT_HEIGHT,
+            output_fps=int(OUTPUT_FPS),
+        )
+        render_result = render_service.render()
+        render_output_path = render_result.output_path
+        render_duration = render_result.duration
+
+    render_elapsed = time.monotonic() - render_start
+    logger.info(
+        "[task_id=%s] [渲染] %s 引擎完成: 耗时=%.1fs, 时长=%.2fs",
+        task_id,
+        engine,
+        render_elapsed,
+        render_duration,
+    )
+
+    # 配音混音
+    if voice_path:
+        final_path = temp_path / f"final-{task_id}.mp4"
+        try:
+            _mux_audio_track(render_output_path, voice_path, final_path)
+            output_path = final_path
+        except Exception as mux_err:
+            logger.warning("[task_id=%s] [混音] 音频混合失败，使用无音频版本: %s", task_id, mux_err)
+            output_path = render_output_path
+    else:
+        output_path = render_output_path
+
+    return output_path, render_duration
+
+
+def _upload_and_record(
+    task_id: str,
+    output_path: Path,
+    project_id: str,
+    batch_id: str,
+    editing_mode,
+) -> tuple[str, float, int, int]:
+    """上传 OSS、创建视频记录并查重。
+
+    Returns:
+        (file_url, duration, file_size, video_count)
+    """
+    storage_key = f"generated/projects/{project_id}/tasks/{task_id}/{output_path.name}"
+    file_size = output_path.stat().st_size
+
+    # 上传 OSS
+    logger.info("[task_id=%s] [OSS上传] 开始上传: size=%d", task_id, file_size)
+    upload_start = time.monotonic()
+    file_url = upload_to_oss(output_path, storage_key)
+    upload_elapsed = time.monotonic() - upload_start
+    if not file_url:
+        raise RuntimeError(f"OSS 上传失败: task_id={task_id}, storage_key={storage_key}")
+
+    # 校验 URL 可达性（P0-2: 私有 bucket 用预签名 + object_exists 降级）
+    verify_url = get_signed_download_url(file_url, expires_seconds=300) or file_url
+    if not _verify_url_accessible(verify_url):
+        from video_processing.oss_helpers import normalize_storage_key, oss_bucket
+
+        bucket = oss_bucket()
+        key = normalize_storage_key(file_url)
+        if not (bucket and bucket.object_exists(key)):
+            raise RuntimeError(
+                f"OSS 上传后 URL 不可访问且 object_exists 失败: file_url={file_url}, " f"storage_key={storage_key}"
+            )
+        logger.info(
+            "URL 校验失败但 object_exists 确认文件存在，视为上传成功: storage_key=%s",
+            key,
+        )
+
+    logger.info(
+        "[task_id=%s] [OSS上传] 成功: 耗时=%.1fs, file_url=%s",
+        task_id,
+        upload_elapsed,
+        file_url,
+    )
+
+    # 创建 GeneratedVideo 记录 + 查重
+    duration = probe_duration(output_path)
+    dedup_session = SessionLocal()
+    try:
+        video_count = create_video_record_and_dedup(
+            generation_task_id=task_id,
+            project_id=project_id,
+            batch_id=batch_id,
+            file_url=file_url,
+            file_size=file_size,
+            duration=duration,
+            video_path=str(output_path),
+            mode=editing_mode.value,
+            session=dedup_session,
+        )
+    finally:
+        dedup_session.close()
+
+    return file_url, duration, file_size, video_count or 1
+
+
 # ── Celery Task ──────────────────────────────────────────────────────────────
 
 
@@ -580,266 +979,126 @@ def generate_video(self, task_id: str) -> dict:
     Returns:
         生成结果字典
     """
+    from packages.adapters.sqlalchemy_impl.generation_task_repository import (
+        SQLAlchemyGenerationTaskRepository,
+    )
     from packages.domain import EditingMode
 
     logger.info("[task_id=%s] [接收任务] 开始生成视频任务", task_id)
 
-    # 从数据库加载任务信息
-    session = SessionLocal()
+    # ── 1. 加载任务信息 ──────────────────────────────────────────────────────
+    task_info = _load_task_info(task_id)
+    if task_info is None:
+        logger.error("[task_id=%s] [接收任务] 任务不存在", task_id)
+        return {"status": "failed", "error": f"generation task {task_id} not found"}
+
+    project_id = task_info["project_id"]
+    asset_library_id = task_info["asset_library_id"]
+    voice_library_id = task_info["voice_library_id"]
+    template_id = task_info["template_id"]
+    task_asset_ids = task_info["task_asset_ids"]
+    batch_id = task_info["batch_id"]
+    user_id = task_info["user_id"]
+
+    # 加载 gen_task（用于全程进度日志；_flush_logs 使用独立 session 持久化）
+    _session = SessionLocal()
     try:
-        from packages.adapters.sqlalchemy_impl.generation_task_repository import (
-            SQLAlchemyGenerationTaskRepository,
-        )
+        _repo = SQLAlchemyGenerationTaskRepository(_session)
+        gen_task = _repo.get(task_id)
+    finally:
+        _session.close()
 
-        task_repo = SQLAlchemyGenerationTaskRepository(session)
-        gen_task = task_repo.get(task_id)
-        if gen_task is None:
-            logger.error("[task_id=%s] [接收任务] 任务不存在", task_id)
-            return {"status": "failed", "error": f"generation task {task_id} not found"}
-        project_id = gen_task.project_id
-        asset_library_id = gen_task.asset_library_id
-        voice_library_id = gen_task.voice_library_id or ""
-        template_id = getattr(gen_task, "template_id", "") or ""
-        mode = gen_task.strategy_id or "one_take"
-        task_asset_ids = list(gen_task.asset_ids or [])
-        batch_id = getattr(gen_task, "batch_id", "") or ""
-
-        # 记录接收任务日志
+    # 记录接收任务日志
+    if gen_task:
         gen_task.append_log(
             "接收任务",
-            f"模式={mode}, 模板={template_id}, 素材数={len(task_asset_ids)}",
-            mode=mode,
+            f"模式={task_info['mode']}, 模板={template_id}, 素材数={len(task_asset_ids)}",
+            mode=task_info["mode"],
             template_id=template_id,
             asset_count=len(task_asset_ids),
         )
         _flush_logs(task_id, gen_task)
-    finally:
-        session.close()
 
     # 标记任务为 running
     _update_task_status(task_id, "mark_processing")
 
     try:
-        editing_mode = EditingMode(mode)
+        editing_mode = EditingMode(mode) if (mode := task_info["mode"]) else EditingMode.ONE_TAKE
     except ValueError:
         editing_mode = EditingMode.ONE_TAKE
 
     output_name = f"generated-{task_id}.mp4"
-    storage_key = f"generated/projects/{project_id}/tasks/{task_id}/{output_name}"
 
     try:
-        # P1: template_id 存在性校验
         if template_id:
             _validate_template_exists(template_id)
 
-        # P1: asset_ids 归属校验 — 已合并到 _download_library_assets 同一 session（P3-2）
-
         with tempfile.TemporaryDirectory(prefix="xiaoxia-generation-") as temp_dir:
             temp_path = Path(temp_dir)
-            output_path = temp_path / output_name
 
-            # 1. 从素材库/项目下载视频素材
-            logger.info("[task_id=%s] [下载素材] 开始下载视频素材", task_id)
-            download_start = time.monotonic()
-            downloaded_videos = _download_library_assets(
+            # ── 2. 下载素材 ──────────────────────────────────────────────────
+            downloaded_videos, audio_path = _download_all_assets(
                 temp_path,
                 asset_library_id=asset_library_id,
                 project_id=project_id,
-                asset_ids=task_asset_ids or None,
+                task_asset_ids=task_asset_ids,
+                voice_library_id=voice_library_id,
                 task_id=task_id,
-                gen_task=gen_task,
             )
-            download_elapsed = time.monotonic() - download_start
-            logger.info(
-                "[task_id=%s] [下载素材] 完成: 成功=%d个, 耗时=%.1fs",
-                task_id,
-                len(downloaded_videos),
-                download_elapsed,
-            )
-
-            # 重新加载 gen_task 以追加日志（session 已关闭）
-            _session = SessionLocal()
-            try:
-                _repo = SQLAlchemyGenerationTaskRepository(_session)
-                gen_task = _repo.get(task_id)
-            finally:
-                _session.close()
 
             if gen_task:
                 gen_task.append_log(
                     "下载素材",
                     f"成功下载 {len(downloaded_videos)} 个视频素材",
                     count=len(downloaded_videos),
-                    duration=round(download_elapsed, 2),
                 )
                 _flush_logs(task_id, gen_task)
 
-            # 2. 下载配音（如有）
-            audio_path: str | None = None
-            if voice_library_id:
-                local_audio = temp_path / "voice.mp3"
-                if _download_voice_asset(voice_library_id, local_audio):
-                    audio_path = str(local_audio)
-                    logger.info("[task_id=%s] [下载配音] 配音下载成功", task_id)
-
-            # 3. 渲染
-            if not downloaded_videos:
-                # 素材下载为空（不应到达此处，_download_library_assets 已做校验）
-                raise RuntimeError(
-                    f"素材下载结果为空: task_id={task_id}, "
-                    f"asset_library_id={asset_library_id}, project_id={project_id}, "
-                    f"asset_ids={task_asset_ids}"
-                )
-
-            # 构建虚拟 plan + clips + asset_path_map
-            virtual_plan, virtual_clips, asset_path_map = _build_plan_and_clips_from_task(
+            # ── 3. 渲染 + 混音 ───────────────────────────────────────────────
+            output_path, render_duration = _render_video(
                 task_id=task_id,
-                downloaded_paths=downloaded_videos,
-                mode=editing_mode.value,
-            )
-
-            total_duration = sum(c.duration for c in virtual_clips)
-            logger.info(
-                "[task_id=%s] [剪辑计划] 片段数=%d, 总时长=%.1fs",
-                task_id,
-                len(virtual_clips),
-                total_duration,
+                downloaded_videos=downloaded_videos,
+                voice_path=audio_path,
+                editing_mode=editing_mode,
+                project_id=project_id,
+                template_id=template_id,
+                user_id=user_id,
+                temp_path=temp_path,
+                output_name=output_name,
             )
 
             if gen_task:
-                gen_task.append_log(
-                    "剪辑计划",
-                    f"片段数={len(virtual_clips)}, 总时长={total_duration:.1f}s",
-                    segment_count=len(virtual_clips),
-                    total_duration=round(total_duration, 2),
-                )
+                gen_task.append_log("渲染", f"渲染完成, 时长={render_duration:.1f}s")
                 _flush_logs(task_id, gen_task)
 
-            # 使用 UnifiedRenderService 渲染
-            logger.info("[task_id=%s] [渲染] FFmpeg 渲染开始", task_id)
-            render_start = time.monotonic()
-            render_service = UnifiedRenderService(
-                plan=virtual_plan,
-                clips=virtual_clips,
-                asset_path_map=asset_path_map,
-                work_dir=temp_path,
-                output_width=OUTPUT_WIDTH,
-                output_height=OUTPUT_HEIGHT,
-                output_fps=int(OUTPUT_FPS),
-            )
-            render_result = render_service.render()
-            render_elapsed = time.monotonic() - render_start
-            logger.info(
-                "[task_id=%s] [渲染] FFmpeg 渲染完成: 耗时=%.1fs",
-                task_id,
-                render_elapsed,
-            )
-
-            if gen_task:
-                gen_task.append_log(
-                    "渲染",
-                    f"FFmpeg 渲染完成, 耗时={render_elapsed:.1f}s",
-                    duration=round(render_elapsed, 2),
-                )
-                _flush_logs(task_id, gen_task)
-
-            # 4. 如有配音，后处理混音
-            if audio_path:
-                final_path = temp_path / f"final-{task_id}.mp4"
-                try:
-                    _mux_audio_track(render_result.output_path, audio_path, final_path)
-                    # 混音成功，使用混音后的文件
-                    output_path = final_path
-                except Exception as mux_err:
-                    logger.warning("[task_id=%s] [混音] 音频混合失败，使用无音频版本: %s", task_id, mux_err)
-                    output_path = render_result.output_path
-            else:
-                output_path = render_result.output_path
-
-            file_size = output_path.stat().st_size
-            duration = probe_duration(output_path)
-
-            # 5. 上传到 OSS — 失败必须抛异常，不能静默忽略
-            logger.info("[task_id=%s] [OSS上传] 开始上传: size=%d", task_id, file_size)
-            upload_start = time.monotonic()
-            file_url = upload_to_oss(output_path, storage_key)
-            upload_elapsed = time.monotonic() - upload_start
-            if not file_url:
-                # OSS 未配置或上传失败
-                if gen_task:
-                    gen_task.append_log("OSS上传", "上传失败", level="ERROR")
-                    _flush_logs(task_id, gen_task)
-                raise RuntimeError(
-                    f"OSS 上传失败: task_id={task_id}, storage_key={storage_key}, " f"output_path={output_path}"
-                )
-
-            # P0-2 修复：私有 bucket 下裸 URL 永远 403，改用预签名 URL 校验
-            # 先用预签名 URL 校验，失败则降级为检查文件是否存在（object_exists）
-            verify_url = get_signed_download_url(file_url, expires_seconds=300) or file_url
-            if not _verify_url_accessible(verify_url):
-                # 预签名 URL 也访问失败时，退一步用 object_exists 确认上传成功
-                from video_processing.oss_helpers import normalize_storage_key, oss_bucket
-
-                bucket = oss_bucket()
-                key = normalize_storage_key(file_url)
-                if bucket and bucket.object_exists(key):
-                    logger.info("URL 校验失败但 object_exists 确认文件存在，视为上传成功: storage_key=%s", key)
-                    if gen_task:
-                        gen_task.append_log("OSS上传", "URL校验降级: object_exists确认存在", level="WARN")
-                else:
-                    if gen_task:
-                        gen_task.append_log("OSS上传", "上传后URL不可访问", level="ERROR", file_url=file_url)
-                        _flush_logs(task_id, gen_task)
-                    raise RuntimeError(
-                        f"OSS 上传后 URL 不可访问且 object_exists 失败: file_url={file_url}, "
-                        f"storage_key={storage_key}"
-                    )
-
-            logger.info(
-                "[task_id=%s] [OSS上传] 成功: 耗时=%.1fs, file_url=%s",
-                task_id,
-                upload_elapsed,
-                file_url,
+            # ── 4. 上传 OSS + 查重记录 ───────────────────────────────────────
+            file_url, duration, file_size, video_count = _upload_and_record(
+                task_id=task_id,
+                output_path=output_path,
+                project_id=project_id,
+                batch_id=batch_id,
+                editing_mode=editing_mode,
             )
 
             if gen_task:
                 gen_task.append_log(
                     "OSS上传",
-                    f"上传成功, 大小={file_size}, 耗时={upload_elapsed:.1f}s",
+                    f"上传成功, 大小={file_size}",
                     file_size=file_size,
-                    duration=round(upload_elapsed, 2),
                     file_url=file_url,
                 )
                 _flush_logs(task_id, gen_task)
 
-            # 6. 创建 GeneratedVideo 记录 + 查重
-            dedup_session = SessionLocal()
-            try:
-                video_count = create_video_record_and_dedup(
-                    generation_task_id=task_id,
-                    project_id=project_id,
-                    batch_id=batch_id,
-                    file_url=file_url,
-                    file_size=file_size,
-                    duration=duration,
-                    video_path=str(output_path),
-                    mode=editing_mode.value,
-                    session=dedup_session,
-                )
-            finally:
-                dedup_session.close()
+            # ── 5. 标记完成 ──────────────────────────────────────────────────
+            _update_task_status(task_id, "mark_completed", result_count=video_count)
 
-            # 7. 标记任务为 completed
-            _update_task_status(task_id, "mark_completed", result_count=video_count or 1)
-
-            # 记录完成日志
             if gen_task:
                 gen_task.append_log(
                     "任务完成",
                     f"视频生成完成: 时长={duration:.2f}s, 大小={file_size}",
                     duration=round(duration, 2),
                     file_size=file_size,
-                    video_count=video_count or 1,
+                    video_count=video_count,
                 )
                 _flush_logs(task_id, gen_task)
 

apps/worker/worker_app/tasks/ingest.py

@@ -1,9 +1,6 @@
 import subprocess
 from datetime import datetime, timezone
-from typing import Optional
 
-from celery import Celery
-from celery.app.task import Task
 from celery.utils.log import get_task_logger
 from worker_app.celery_app import celery_app
 from worker_app.core.asset_types import infer_mime_type_from_storage_key

apps/worker/worker_app/tasks/voice_extraction.py

@@ -1,14 +1,11 @@
 """Voice extraction tasks - extract voice tracks and background music from videos."""
 
-import json
 import logging
 import os
 import subprocess
 import tempfile
-from typing import Optional
 
 from celery import Task
-from sqlalchemy.orm import Session
 from worker_app.celery_app import celery_app
 from worker_app.db import SessionLocal
 

docs/ci-env-vars.md

@@ -112,7 +112,7 @@
 
 | 变量名 | 用途说明 | 默认值 |
 |--------|---------|--------|
-| `OSS_ENDPOINT` | OSS Endpoint | `oss-cn-hangzhou.aliiyuncs.com` |
+| `OSS_ENDPOINT` | OSS Endpoint | `oss-cn-hangzhou.aliyuncs.com` |
 | `OSS_ACCESS_KEY_ID` | OSS Access Key ID | `""`（空） |
 | `OSS_ACCESS_KEY_SECRET` | OSS Access Key Secret | `""`（空） |
 | `OSS_BUCKET_NAME` | OSS Bucket 名称 | `xiaoxia-autocut` |

packages/__init__.py

@@ -1 +0,0 @@
-"""Packages root."""

packages/adapters/__init__.py

@@ -1 +0,0 @@
-"""Adapters package for external implementations."""

packages/adapters/redis/__init__.py

@@ -1,3 +1,9 @@
+from packages.adapters.redis.feature_flag_store import (
+    FeatureFlagConfig,
+    FeatureFlagStore,
+    InMemoryFeatureFlagStore,
+    RedisFeatureFlagStore,
+)
 from packages.adapters.redis.session_store import (
     NoopSessionStore,
     RedisConfig,
@@ -5,4 +11,13 @@ from packages.adapters.redis.session_store import (
     get_session_store,
 )
 
-__all__ = ["NoopSessionStore", "RedisConfig", "SessionStore", "get_session_store"]
+__all__ = [
+    "FeatureFlagConfig",
+    "FeatureFlagStore",
+    "InMemoryFeatureFlagStore",
+    "NoopSessionStore",
+    "RedisConfig",
+    "RedisFeatureFlagStore",
+    "SessionStore",
+    "get_session_store",
+]

packages/adapters/redis/feature_flag_store.py

@@ -0,0 +1,259 @@
+"""Feature Flag 存储实现。
+
+支持两种后端：
+- RedisFeatureFlagStore：生产环境使用，支持多实例共享、热更新
+- InMemoryFeatureFlagStore：测试/开发环境使用，纯内存
+
+支持的 Flag 类型：
+- 全局开关（enabled: bool）
+- 白名单（whitelist: Set[str]，如 user_id 列表）
+- 百分比切流（percentage: 0-100，基于标识符哈希取模）
+
+判定优先级：白名单 > 百分比 > 全局开关
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import logging
+import threading
+import time
+from abc import ABC, abstractmethod
+from dataclasses import dataclass, field
+from typing import Optional, Set
+
+logger = logging.getLogger(__name__)
+
+# Redis key 前缀
+FEATURE_FLAG_REDIS_PREFIX = "feature_flag:"
+
+
+@dataclass
+class FeatureFlagConfig:
+    """单个 Feature Flag 的配置。"""
+
+    name: str
+    enabled: bool = False
+    percentage: int = 0  # 0-100
+    whitelist: Set[str] = field(default_factory=set)
+
+    def to_dict(self) -> dict:
+        return {
+            "name": self.name,
+            "enabled": self.enabled,
+            "percentage": self.percentage,
+            "whitelist": sorted(self.whitelist),
+        }
+
+    @classmethod
+    def from_dict(cls, data: dict) -> "FeatureFlagConfig":
+        return cls(
+            name=data["name"],
+            enabled=bool(data.get("enabled", False)),
+            percentage=int(data.get("percentage", 0)),
+            whitelist=set(data.get("whitelist", [])),
+        )
+
+    def is_active(self, identifier: Optional[str] = None) -> bool:
+        """判断当前 flag 是否激活。
+
+        判定优先级：
+        1. 全局关闭 → False
+        2. 白名单匹配 → True
+        3. 百分比命中 → True
+        4. 其他 → False
+
+        Args:
+            identifier: 用于白名单匹配和百分比哈希的标识符（如 user_id）。
+                       传 None 时只看全局开关 + 百分比（百分比用随机值）。
+        """
+        if not self.enabled:
+            return False
+
+        # 白名单：精确匹配
+        if identifier and identifier in self.whitelist:
+            return True
+
+        # 百分比：0 直接 False，100 直接 True
+        if self.percentage <= 0:
+            # 没有白名单且百分比为0 → 未启用
+            return False
+        if self.percentage >= 100:
+            return True
+
+        # 基于 identifier 做哈希取模，确保同一用户始终落在同一侧
+        if identifier:
+            hash_val = int(
+                hashlib.md5(f"{self.name}:{identifier}".encode("utf-8")).hexdigest(), 16  # nosec B324
+            )  # nosec B324 - 用于哈希取模做百分比切流，非安全用途
+            return (hash_val % 100) < self.percentage
+
+        # 无 identifier 且百分比在 0-100 之间 → 按比例随机（不保证一致性）
+        import random
+
+        return random.randint(0, 99) < self.percentage
+
+
+class FeatureFlagStore(ABC):
+    """Feature Flag 存储抽象接口。"""
+
+    @abstractmethod
+    def get(self, name: str) -> FeatureFlagConfig:
+        """获取指定 flag 的配置，不存在则返回默认配置（关闭状态）。"""
+        ...
+
+    @abstractmethod
+    def set(self, config: FeatureFlagConfig) -> None:
+        """设置 flag 配置。"""
+        ...
+
+    @abstractmethod
+    def delete(self, name: str) -> bool:
+        """删除 flag，返回是否成功删除。"""
+        ...
+
+    @abstractmethod
+    def list_all(self) -> dict[str, FeatureFlagConfig]:
+        """列出所有 flag。"""
+        ...
+
+    def is_active(self, name: str, identifier: Optional[str] = None) -> bool:
+        """便捷方法：判断 flag 是否激活。"""
+        return self.get(name).is_active(identifier)
+
+
+class InMemoryFeatureFlagStore(FeatureFlagStore):
+    """内存实现，用于测试和本地开发。"""
+
+    def __init__(self) -> None:
+        self._flags: dict[str, FeatureFlagConfig] = {}
+        self._lock = threading.Lock()
+
+    def get(self, name: str) -> FeatureFlagConfig:
+        with self._lock:
+            return self._flags.get(name, FeatureFlagConfig(name=name, enabled=False))
+
+    def set(self, config: FeatureFlagConfig) -> None:
+        with self._lock:
+            self._flags[config.name] = config
+
+    def delete(self, name: str) -> bool:
+        with self._lock:
+            if name in self._flags:
+                del self._flags[name]
+                return True
+            return False
+
+    def list_all(self) -> dict[str, FeatureFlagConfig]:
+        with self._lock:
+            return dict(self._flags)
+
+
+class RedisFeatureFlagStore(FeatureFlagStore):
+    """Redis 实现，支持多实例共享配置。
+
+    每个 flag 存在一个独立的 Redis hash key 中：
+      Key: feature_flag:{name}
+      Fields: enabled, percentage, whitelist(JSON array)
+    """
+
+    def __init__(self, redis_url: str, key_prefix: str = FEATURE_FLAG_REDIS_PREFIX) -> None:
+        import redis as redis_lib
+
+        self._redis = redis_lib.from_url(redis_url, decode_responses=True)
+        self._key_prefix = key_prefix
+        # 本地缓存 + TTL，减少 Redis 调用
+        self._cache: dict[str, tuple[FeatureFlagConfig, float]] = {}
+        self._cache_ttl = 5.0  # 秒，默认5秒本地缓存
+        self._lock = threading.Lock()
+
+    def _redis_key(self, name: str) -> str:
+        return f"{self._key_prefix}{name}"
+
+    def _parse_whitelist(self, raw: Optional[str]) -> Set[str]:
+        if not raw:
+            return set()
+        try:
+            data = json.loads(raw)
+            return set(data) if isinstance(data, list) else set()
+        except (json.JSONDecodeError, TypeError):
+            return set()
+
+    def get(self, name: str) -> FeatureFlagConfig:
+        now = time.time()
+
+        # 先查本地缓存
+        with self._lock:
+            cached = self._cache.get(name)
+            if cached and now - cached[1] < self._cache_ttl:
+                return cached[0]
+
+        # 从 Redis 读取
+        try:
+            key = self._redis_key(name)
+            data = self._redis.hgetall(key)
+            if not data:
+                config = FeatureFlagConfig(name=name, enabled=False)
+            else:
+                config = FeatureFlagConfig(
+                    name=name,
+                    enabled=(data.get("enabled", "0") in ("1", "true", "True")),
+                    percentage=int(data.get("percentage", 0)),
+                    whitelist=self._parse_whitelist(data.get("whitelist")),
+                )
+
+            # 写入本地缓存
+            with self._lock:
+                self._cache[name] = (config, now)
+
+            return config
+        except Exception as exc:
+            logger.warning("Failed to get feature flag %s from Redis: %s", name, exc)
+            # Redis 不可用时返回默认值（关闭），不影响业务
+            return FeatureFlagConfig(name=name, enabled=False)
+
+    def set(self, config: FeatureFlagConfig) -> None:
+        key = self._redis_key(config.name)
+        self._redis.hset(
+            key,
+            mapping={
+                "enabled": "1" if config.enabled else "0",
+                "percentage": str(config.percentage),
+                "whitelist": json.dumps(sorted(config.whitelist), ensure_ascii=False),
+            },
+        )
+        # 失效本地缓存
+        with self._lock:
+            self._cache.pop(config.name, None)
+
+    def delete(self, name: str) -> bool:
+        key = self._redis_key(name)
+        result = self._redis.delete(key)
+        with self._lock:
+            self._cache.pop(name, None)
+        return bool(result)
+
+    def list_all(self) -> dict[str, FeatureFlagConfig]:
+        pattern = f"{self._key_prefix}*"
+        result: dict[str, FeatureFlagConfig] = {}
+        try:
+            cursor = 0
+            while True:
+                cursor, keys = self._redis.scan(cursor=cursor, match=pattern, count=100)
+                for key in keys:
+                    name = key[len(self._key_prefix) :]
+                    result[name] = self.get(name)
+                if cursor == 0:
+                    break
+        except Exception as exc:
+            logger.warning("Failed to list feature flags from Redis: %s", exc)
+        return result
+
+    def invalidate_cache(self, name: Optional[str] = None) -> None:
+        """手动失效本地缓存。"""
+        with self._lock:
+            if name:
+                self._cache.pop(name, None)
+            else:
+                self._cache.clear()

packages/adapters/sqlalchemy_impl/models.py

@@ -1,6 +1,6 @@
 from datetime import datetime, timezone
 
-from sqlalchemy import JSON, Boolean, Column, DateTime, Float, Integer, String, Text, UniqueConstraint, create_engine
+from sqlalchemy import JSON, Boolean, Column, DateTime, Float, Integer, String, Text, UniqueConstraint
 from sqlalchemy.orm import declarative_base
 
 Base = declarative_base()

packages/application/__init__.py

@@ -64,6 +64,7 @@ __all__ = [
     "CreateAssetUseCase",
     "CreateGenerationTaskCommand",
     "CreateGenerationTaskUseCase",
+    "GetGenerationTaskUseCase",
     "CreateJobCommand",
     "CreateJobUseCase",
     "CreateProjectCommand",

packages/application/auth/jwt_handler.py

@@ -12,10 +12,9 @@ JWT 处理器委托层
     payload = jwt_handler.verify_access_token(token)
 """
 
-from datetime import datetime, timedelta
 from typing import Any, Dict, Optional
 
-from packages.application.auth.jwt_service import JWTConfig, JWTService, TokenType
+from packages.application.auth.jwt_service import JWTConfig, JWTService
 
 
 class JWTHandler:

packages/application/auth/jwt_service.py

@@ -208,10 +208,10 @@ def _get_jwt_service():
         kw = dict(secret_key=settings.JWT_SECRET_KEY)
         if hasattr(settings, "JWT_ALGORITHM"):
             kw["algorithm"] = settings.JWT_ALGORITHM
-        if hasattr(settings, "ACCESS_TOKEN_EXPIRE_MINUTES"):
-            kw["access_token_expire_minutes"] = settings.ACCESS_TOKEN_EXPIRE_MINUTES
-        if hasattr(settings, "REFRESH_TOKEN_EXPIRE_DAYS"):
-            kw["refresh_token_expire_days"] = settings.REFRESH_TOKEN_EXPIRE_DAYS
+        if hasattr(settings, "JWT_ACCESS_TOKEN_EXPIRE_MINUTES"):
+            kw["access_token_expire_minutes"] = settings.JWT_ACCESS_TOKEN_EXPIRE_MINUTES
+        if hasattr(settings, "JWT_REFRESH_TOKEN_EXPIRE_DAYS"):
+            kw["refresh_token_expire_days"] = settings.JWT_REFRESH_TOKEN_EXPIRE_DAYS
         _jwt_service_instance = JWTService(JWTConfig(**kw))
     return _jwt_service_instance
 

packages/application/auth/login_use_case.py

@@ -293,7 +293,7 @@ class LogoutUseCase:
         try:
             if request.logout_all_devices:
                 # 删除所有设备的 session
-                count = self.session_store.delete_all_user_sessions(request.user_id)
+                self.session_store.delete_all_user_sessions(request.user_id)
                 return True, None
             else:
                 # 删除当前 session