fix: prevent docker images grep from failing build with set -e

The last line of the build script uses grep to display built images,
but with set -e, grep returning no matches causes the entire script
to fail. This is a classic set -e footgun.

The grep pattern was also broken: docker images shows REPOSITORY and
TAG as separate columns (no ':' between them), so 'xiaoxia-saas.*:'
never matched. Changed to two-stage grep and added || true safety net.

The build and push were actually succeeding - only the final display
line was causing the failure.

.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
@@ -44,9 +46,15 @@ OSS_ACCESS_KEY_SECRET=your-access-key-secret
 OSS_BUCKET_NAME=xiaoxia-autocut
 
 # ==================== CosyVoice 语音合成配置 ====================
+# 注意：base_url 只需写到 /api/v1，具体路径由代码拼接
+# 模型: cosyvoice-v3-flash (推荐，支持系统音色，性价比高)
+#       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/services/aigc/text2audio
-COSYVOICE_MODEL=cosyvoice-v1
-COSYVOICE_VOICE=longxiaochun
+COSYVOICE_BASE_URL=https://dashscope.aliyuncs.com/api/v1
+COSYVOICE_MODEL=cosyvoice-v3-flash
+COSYVOICE_VOICE=longxiaochun_v3
 COSYVOICE_SAMPLE_RATE=22050
 COSYVOICE_FORMAT=mp3

.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
@@ -42,10 +41,15 @@ OSS_DIRECT_UPLOAD_MAX_MB=2000
 OSS_DIRECT_UPLOAD_EXPIRE_SECONDS=900
 
 # ==================== CosyVoice 语音合成（必须配置）====================
+# 注意：base_url 只需写到 /api/v1，具体路径由代码拼接
+# 模型: cosyvoice-v3-flash (推荐，支持系统音色，性价比高)
+#       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)
 COSYVOICE_API_KEY=CHANGE_ME_COSYVOICE_API_KEY
-COSYVOICE_BASE_URL=https://dashscope.aliyuncs.com/api/v1/services/aigc/text2audio
-COSYVOICE_MODEL=cosyvoice-v1
-COSYVOICE_VOICE=longxiaochun
+COSYVOICE_BASE_URL=https://dashscope.aliyuncs.com/api/v1
+COSYVOICE_MODEL=cosyvoice-v3-flash
+COSYVOICE_VOICE=longxiaochun_v3
 COSYVOICE_SAMPLE_RATE=22050
 COSYVOICE_FORMAT=mp3
 

.flake8

@@ -2,6 +2,7 @@
 max-line-length = 120
 exclude =
     .git,
+    .cache,
     __pycache__,
     .venv,
     venv,

.gitea/workflows/auto-merge.yml

@@ -1,65 +0,0 @@
-name: Auto Merge PRs
-
-on:
-  schedule:
-    - cron: '0 */6 * * *'
-  workflow_dispatch:
-
-jobs:
-  auto-merge:
-    runs-on: saas
-    timeout-minutes: 10
-    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']}/{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:
-              top_prefix = tar.getmembers()[0].name.split('/', 1)[0] + '/'
-              for member in tar.getmembers():
-                  name = member.name
-                  if name == top_prefix[:-1]:
-                      continue
-                  if name.startswith(top_prefix):
-                      member.name = name[len(top_prefix):]
-                  if member.name:
-                      tar.extract(member, '.')
-          PY
-
-      - name: Auto merge develop PRs
-        run: |
-          bash scripts/auto_merge_prs.sh develop
-
-      - name: Auto merge main PRs (release only)
-        run: |
-          bash scripts/auto_merge_prs.sh main

.gitea/workflows/ci-cd.yml

@@ -19,10 +19,14 @@ on:
 permissions:
   contents: read
 
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
 jobs:
   validate:
     name: Validate Code Quality And Tests
-    runs-on: ubuntu-22.04
+    runs-on: host
     timeout-minutes: 10
 
     env:
@@ -80,7 +84,7 @@ jobs:
         shell: sh
         run: |
           set -eu
-          python --version
+          python3 --version
           python3 -m pip --version
           echo "CI environment is ready"
 
@@ -158,13 +162,255 @@ jobs:
             python3 scripts/check_migration_safety.py --allow-medium-risk
           fi
 
-      - name: Run unit tests
+  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:
-          USE_IN_MEMORY_DB: "true"
+          GITHUB_TOKEN: ${{ github.token }}
         run: |
           set -eu
-          PYTHONPATH="$PWD/apps/api:$PWD" python3 -m pytest tests/unit -q
+          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
+          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: $(ffmpeg -version | head -1)"
+          else
+            echo "Warning: ffmpeg installation failed or not available, some tests may be skipped"
+          fi
+
+      - name: Install dependencies
+        shell: sh
+        run: |
+          set -eu
+          python3 -m pip install -q -r requirements-base.txt
+          python3 -m pip install -q -r requirements.txt
+          python3 -m pip install -q -r requirements-dev.txt
+          pytest --version
+
+      - name: Run unit tests with coverage
+        shell: sh
+        run: |
+          set -eu
+          PYTHONPATH="$PWD/apps/api:$PWD" python3 -m coverage run \
+            --source=apps/api/app,packages \
+            --omit="*/migrations/*,*/tests/*,*/test_*.py,*/site-packages/*" \
+            --branch \
+            -m pytest tests/unit -q
+          python3 -m coverage report --show-missing
+          python3 -m coverage xml -o coverage.xml
+          python3 -m coverage report --fail-under=60 > /dev/null
+
+      - name: CI failure notification
+        if: failure()
+        shell: sh
+        env:
+          GITEA_TOKEN: ${{ secrets.GITEA_TOKEN }}
+          CI_WEBHOOK_URL: ${{ secrets.CI_WEBHOOK_URL }}
+        run: |
+          set +e
+          FAILED_JOB="Unit Tests" python3 scripts/ci_notify_failure.py
+
+
+  integration-tests:
+    name: Integration Tests
+    runs-on: host
+    timeout-minutes: 20
+    if: always()
+    needs: validate
+
+    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
+        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: Verify CI environment
+        shell: sh
+        run: |
+          set -eu
+          python3 --version
+          python3 -m pip --version
+          echo "CI environment is ready"
+
+      - 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: 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: $(ffmpeg -version | head -1)"
+          else
+            echo "Warning: ffmpeg installation failed or not available, some tests may be skipped"
+          fi
+
+      - name: Start Redis
+        shell: sh
+        run: |
+          set -eu
+          REDIS_CONTAINER="ci-redis-${GITHUB_RUN_ID:-$$}"
+          echo "REDIS_CONTAINER=$REDIS_CONTAINER" >> "$GITHUB_ENV"
+          docker rm -f "$REDIS_CONTAINER" 2>/dev/null || true
+          docker run -d --name "$REDIS_CONTAINER" \
+            -P \
+            --health-cmd "redis-cli ping" \
+            --health-interval 2s \
+            --health-timeout 2s \
+            --health-retries 10 \
+            redis:7-alpine
+          REDIS_PORT=$(docker port "$REDIS_CONTAINER" 6379/tcp | cut -d: -f2)
+          echo "Redis port: $REDIS_PORT"
+          echo "REDIS_URL=redis://127.0.0.1:$REDIS_PORT/0" >> "$GITHUB_ENV"
+          for i in $(seq 1 15); do
+            if docker inspect --format='{{.State.Health.Status}}' "$REDIS_CONTAINER" 2>/dev/null | grep -q healthy; then
+              echo "Redis is ready on port $REDIS_PORT"
+              break
+            fi
+            echo "Waiting for Redis... ($i/15)"
+            sleep 2
+          done
+          docker inspect --format='{{.State.Health.Status}}' "$REDIS_CONTAINER" | grep -q healthy
 
       - name: Start PostgreSQL for integration tests
         shell: sh
@@ -208,8 +454,15 @@ jobs:
         shell: sh
         run: |
           set -eu
-          pip install -q pytest-rerunfailures
-          PYTHONPATH="$PWD/apps/api:$PWD" python3 -m pytest tests/integration -q --timeout=60 -x --reruns 2 --reruns-delay 1 -m "not performance"
+          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/*" \
+            --branch \
+            -m pytest tests/integration -q --timeout=60 -x --reruns 2 --reruns-delay 1 -m "not performance"
+          python3 -m coverage report --show-missing
+          python3 -m coverage xml -o coverage.xml
+          python3 -m coverage report --fail-under=40 > /dev/null  # 集成测试覆盖率门槛较低，核心目标是功能验证
 
       - name: Run API performance baseline tests
         shell: sh
@@ -253,25 +506,44 @@ jobs:
           exit 0
 
 
-      - name: Cleanup PostgreSQL
+      - name: Cleanup PostgreSQL & Redis
         if: always()
         shell: sh
         run: |
           docker rm -f "${PG_CONTAINER:-ci-pg-validate}" 2>/dev/null || true
+          docker rm -f "${REDIS_CONTAINER:-ci-redis-int}" 2>/dev/null || true
           echo "PostgreSQL container cleaned up"
+          echo "Redis container cleaned up"
 
-      - name: Build summary
-        if: github.ref == 'refs/heads/develop' || github.ref == 'refs/heads/main'
+      - name: Coverage summary
+        if: always()
+        shell: sh
+        env:
+          COVERAGE_THRESHOLD: "40"
+        run: |
+          set +e
+          echo "=== 覆盖率汇总 ==="
+          python3 scripts/ci_coverage_summary.py
+      - name: Notify CI failure
+        if: failure()
         shell: sh
         run: |
-          set -eu
-          echo "Build completed successfully!"
-          echo "Branch: ${GITHUB_REF_NAME}"
-          echo "Commit: ${GITHUB_SHA}"
+          set +e
+          echo "=== CI 失败通知 ==="
+          FAILED_JOB="Validate Code Quality And Tests" python3 scripts/ci_notify_failure.py
+
+      - name: Notify CI failure - Integration Tests
+        if: failure()
+        shell: sh
+        run: |
+          set +e
+          echo "=== CI 失败通知 ==="
+          FAILED_JOB="Integration Tests" python3 scripts/ci_notify_failure.py
+
 
   frontend-lint:
     name: Frontend Lint
-    runs-on: ubuntu-22.04
+    runs-on: host
     timeout-minutes: 10
 
     steps:
@@ -370,13 +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
@@ -501,6 +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
@@ -569,6 +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
@@ -634,6 +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
@@ -714,6 +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
@@ -779,6 +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
@@ -847,3 +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
+

.gitea/workflows/test-ssh-secret.yml

@@ -1,69 +0,0 @@
-name: Test SSH Secret
-on:
-  push:
-    branches: [develop]
-    paths:
-      - '.gitea/workflows/test-ssh-secret.yml'
-
-jobs:
-  test-ssh:
-    runs-on: ubuntu-22.04
-    steps:
-      - name: Install SSH client
-        run: |
-          which ssh || (apt-get update && apt-get install -y openssh-client)
-          ssh -V
-
-      - name: Debug environment
-        run: |
-          echo "=== Environment ==="
-          echo "Runner hostname: $(hostname)"
-          echo "Runner IP: $(hostname -i || echo 'unknown')"
-          echo "Current user: $(whoami)"
-          echo "=== Secrets check ==="
-          if [ -n "$STAGING_SSH_HOST" ]; then
-            echo "STAGING_SSH_HOST: [SET] value_length=${#STAGING_SSH_HOST}"
-          else
-            echo "STAGING_SSH_HOST: [EMPTY]"
-          fi
-          if [ -n "$STAGING_SSH_USER" ]; then
-            echo "STAGING_SSH_USER: [SET] value_length=${#STAGING_SSH_USER}"
-          else
-            echo "STAGING_SSH_USER: [EMPTY]"
-          fi
-          if [ -n "$STAGING_SSH_KEY" ]; then
-            echo "STAGING_SSH_KEY: [SET] value_length=${#STAGING_SSH_KEY}"
-          else
-            echo "STAGING_SSH_KEY: [EMPTY]"
-          fi
-        env:
-          STAGING_SSH_HOST: ${{ secrets.STAGING_SSH_HOST }}
-          STAGING_SSH_USER: ${{ secrets.STAGING_SSH_USER }}
-          STAGING_SSH_KEY: ${{ secrets.STAGING_SSH_KEY }}
-
-      - name: Setup SSH key
-        run: |
-          mkdir -p ~/.ssh
-          chmod 700 ~/.ssh
-          echo "$STAGING_SSH_KEY" > ~/.ssh/id_ed25519
-          chmod 600 ~/.ssh/id_ed25519
-          ssh-keygen -y -f ~/.ssh/id_ed25519 > ~/.ssh/id_ed25519.pub 2>/dev/null || echo "No public key generated"
-          echo "=== SSH Key fingerprint ==="
-          ssh-keygen -lf ~/.ssh/id_ed25519 || echo "Key fingerprint failed"
-        env:
-          STAGING_SSH_KEY: ${{ secrets.STAGING_SSH_KEY }}
-
-      - name: Test SSH connection
-        run: |
-          echo "Attempting SSH connection to $STAGING_SSH_HOST..."
-          ssh -i ~/.ssh/id_ed25519 \
-            -o StrictHostKeyChecking=no \
-            -o UserKnownHostsFile=/dev/null \
-            -o ConnectTimeout=10 \
-            -o BatchMode=yes \
-            -v \
-            $STAGING_SSH_USER@$STAGING_SSH_HOST "echo 'SSH_CONNECTION_SUCCESS' && hostname && whoami"
-          echo "=== SSH Test Complete ==="
-        env:
-          STAGING_SSH_HOST: ${{ secrets.STAGING_SSH_HOST }}
-          STAGING_SSH_USER: ${{ secrets.STAGING_SSH_USER }}

.gitea/workflows/tests.yml

@@ -1,163 +0,0 @@
-name: Tests
-
-on:
-  pull_request:
-    branches: [ main ]
-
-jobs:
-  test:
-    runs-on: runtime-builder
-
-    steps:
-    - name: Checkout code
-      shell: sh
-      run: |
-        set -eu
-        python - <<'PY'
-        import io
-        import os
-        import tarfile
-        import time
-        import urllib.error
-        import urllib.request
-
-        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']}"})
-        # Retry up to 5 times with backoff for transient 5xx errors
-        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: Show Python version
-      shell: sh
-      run: |
-        set -eu
-        python --version
-        python -m pip --version
-
-    - name: Install dependencies
-      shell: sh
-      run: |
-        set -eu
-        python -m pip install --upgrade pip -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
-        python -m pip install -r requirements.txt -r requirements-dev.txt -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
-
-    - name: Run unit tests
-      shell: sh
-      run: |
-        set -eu
-        PYTHONPATH="$PWD/apps/api:$PWD" python -m pytest tests/unit -q
-
-    - name: Run integration tests
-      shell: sh
-      run: |
-        set -eu
-        PYTHONPATH="$PWD/apps/api:$PWD" python -m pytest tests/integration -q --timeout=60 -x
-
-  lint:
-    runs-on: runtime-builder
-
-    steps:
-    - name: Checkout code
-      shell: sh
-      run: |
-        set -eu
-        python - <<'PY'
-        import io
-        import os
-        import tarfile
-        import time
-        import urllib.error
-        import urllib.request
-
-        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']}"})
-        # Retry up to 5 times with backoff for transient 5xx errors
-        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 dependencies
-      shell: sh
-      run: |
-        set -eu
-        python -m pip install --upgrade pip -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
-        python -m pip install -r requirements.txt -r requirements-dev.txt -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
-
-    - name: Run Black (check only)
-      shell: sh
-      run: |
-        set -eu
-        python -m black --check alembic apps packages tests scripts
-
-    - name: Run Flake8
-      shell: sh
-      run: |
-        set -eu
-        python -m flake8 apps packages tests --count --statistics

.gitignore

@@ -6,6 +6,7 @@ dist/
 coverage/
 
 # Python / backend
+.cache/
 .venv/
 venv/
 .venv-ci-root/

alembic/versions/035_add_editing_mode_to_templates.py

@@ -0,0 +1,26 @@
+"""Add editing_mode to edit_templates
+
+Revision ID: 035_editing_mode
+Revises: 034_cms_enhance
+Create Date: 2026-07-09
+"""
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision = "035_editing_mode"
+down_revision = "034_cms_enhance"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "edit_templates",
+        sa.Column("editing_mode", sa.String(20), nullable=False, server_default="one_take"),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("edit_templates", "editing_mode")

alembic/versions/036_expand_uuid_fields_to_36.py

@@ -0,0 +1,68 @@
+"""Expand UUID fields from varchar(32) to varchar(36)
+
+All UUID fields across all tables were varchar(32), but standard UUIDs with
+hyphens are 36 characters (e.g. 550e8400-e29b-41d4-a716-446655440000).
+This caused StringDataRightTruncation errors on insert.
+
+Revision ID: 036_expand_uuid_36
+Revises: 035_editing_mode
+Create Date: 2026-07-10
+"""
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision = "036_expand_uuid_36"
+down_revision = "035_editing_mode"
+branch_labels = None
+depends_on = None
+
+
+# ── 表 → 需要扩容的列 ─────────────────────────────────────────────────────────
+
+_TABLES: dict[str, list[str]] = {
+    "projects": ["id", "owner_user_id"],
+    "edit_templates": ["id"],
+    "edit_plans": ["id", "template_id", "source_edit_plan_id", "project_id", "created_by_user_id"],
+    "template_clip_configs": ["id", "template_id"],
+    "edit_plan_clips": ["id", "plan_id", "template_clip_config_id", "asset_id"],
+    "ingest_jobs": ["id", "project_id", "library_id", "result_asset_id"],
+    "classification_jobs": ["id", "project_id", "asset_id"],
+    "generation_tasks": [
+        "id",
+        "project_id",
+        "strategy_id",
+        "asset_library_id",
+        "voice_library_id",
+        "created_by_user_id",
+        "source_edit_plan_id",
+        "batch_id",
+    ],
+    "generated_videos": ["id", "project_id", "generation_task_id", "duplicate_of"],
+    "jobs": ["id", "project_id", "source_id", "created_by_user_id"],
+}
+
+
+def upgrade() -> None:
+    for table, columns in _TABLES.items():
+        for col in columns:
+            op.alter_column(
+                table,
+                col,
+                existing_type=sa.String(32),
+                type_=sa.String(36),
+                existing_nullable=None,
+            )
+
+
+def downgrade() -> None:
+    for table, columns in reversed(list(_TABLES.items())):
+        for col in columns:
+            op.alter_column(
+                table,
+                col,
+                existing_type=sa.String(36),
+                type_=sa.String(32),
+                existing_nullable=None,
+            )

alembic/versions/037_add_logs_to_generation_tasks.py

@@ -0,0 +1,26 @@
+"""Add logs field to generation_tasks
+
+Revision ID: 037_generation_logs
+Revises: 036_expand_uuid_36
+Create Date: 2026-07-10
+"""
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision = "037_generation_logs"
+down_revision = "036_expand_uuid_36"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "generation_tasks",
+        sa.Column("logs", sa.Text(), nullable=False, server_default="[]"),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("generation_tasks", "logs")

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

@@ -6,10 +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）
+
+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。
 """
@@ -21,36 +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.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
-from fastapi import APIRouter, Depends, HTTPException, Query, status
+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()
@@ -203,20 +186,47 @@ class GenerateCoverResponse(BaseModel):
     cover: dict[str, Any] = Field(..., description="封面数据（type / image_url / frame_time 等）")
 
 
+# ── 基于模板生成剪辑计划 Schemas ─────────────────────────────────────────────
+
+
+class GenerateFromTemplateRequest(BaseModel):
+    """基于模板生成剪辑计划请求体"""
+
+    template_id: str = Field(..., description="剪辑模板 ID")
+    asset_ids: List[str] = Field(default_factory=list, description="素材 ID 列表")
+    project_id: str = Field(default="", description="所属项目 ID")
+    name: str = Field(default="", description="计划名称（为空则自动取模板名）")
+
+
+class _PlanClipItem(BaseModel):
+    """片段响应体"""
+
+    id: str
+    clip_type: str
+    order: int
+    asset_id: str
+    text_content: str
+    start_time: float
+    duration: float
+    transition_effect: str
+    status: str
+    config: Optional[dict[str, Any]] = None
+    created_at: datetime
+    updated_at: datetime
+
+    model_config = {"from_attributes": True}
+
+
+class GenerateFromTemplateResponse(BaseModel):
+    """基于模板生成剪辑计划响应体"""
+
+    plan: EditPlanResponse
+    clips: List[_PlanClipItem]
+
+
 # ── 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,
@@ -232,7 +242,7 @@ def _to_response(p: EditPlan) -> EditPlanResponse:
     )
 
 
-# ── Routes ────────────────────────────────────────────────────────────────────
+# ── CRUD Routes ───────────────────────────────────────────────────────────────
 
 
 @router.get("", response_model=EditPlanListResponse)
@@ -270,7 +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(
@@ -312,7 +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)
 
 
@@ -328,7 +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)
@@ -370,7 +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:
@@ -412,7 +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),
@@ -424,7 +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(
@@ -438,643 +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)
-        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,
-    )
+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

@@ -1,279 +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")
-    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")
-    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
-    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,
-        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,
-            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,
-            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

@@ -1,9 +1,18 @@
+import logging
 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.celery_app import celery_app
+from app.core.storage import OSSStorageService, get_storage_service
+from app.core.task_enqueue import (
+    GLOBAL_PENDING_LIMIT,
+    USER_PENDING_LIMIT,
+    GlobalQueueFull,
+    UserPendingLimitExceeded,
+    safe_enqueue_generation_task,
+)
 from app.dependencies import (
     get_asset_library_repository,
     get_asset_repository,
@@ -30,18 +39,10 @@ from packages.application import (
     ListGeneratedVideosByTaskUseCase,
 )
 
+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,
@@ -56,6 +57,7 @@ def _to_generation_task_response(task) -> GenerationTaskResponse:
         source_edit_plan_id=task.source_edit_plan_id or "",
         asset_select_mode=getattr(task, "asset_select_mode", ""),
         batch_id=getattr(task, "batch_id", ""),
+        logs=getattr(task, "logs", "[]"),
         status=task.status,
         progress=task.progress,
         result_count=task.result_count,
@@ -63,7 +65,7 @@ def _to_generation_task_response(task) -> GenerationTaskResponse:
     )
 
 
-def _to_generated_video_response(item) -> GeneratedVideoResponse:
+def _to_generated_video_response(item, download_url: str | None = None) -> GeneratedVideoResponse:
     return GeneratedVideoResponse(
         id=item.id,
         project_id=item.project_id,
@@ -76,6 +78,7 @@ def _to_generated_video_response(item) -> GeneratedVideoResponse:
         width=item.width,
         height=item.height,
         fps=item.fps,
+        download_url=download_url,
     )
 
 
@@ -179,19 +182,37 @@ def create_generation_task(
     asset_library_repository: Any = Depends(get_asset_library_repository),
     asset_repository: Any = Depends(get_asset_repository),
 ) -> BatchGenerationTaskResponse:
-    project_id, asset_library_id = _resolve_project_and_library(
-        request, project_repository, asset_library_repository, asset_repository, authenticated_user
+    logger.info(
+        "[生成任务] 接收请求: user_id=%s, template_id=%s, asset_count=%d, mode=%s, count=%d",
+        authenticated_user.user.id,
+        request.template_id,
+        len(request.asset_ids),
+        request.asset_select_mode,
+        request.count,
     )
 
+    try:
+        project_id, asset_library_id = _resolve_project_and_library(
+            request, project_repository, asset_library_repository, asset_repository, authenticated_user
+        )
+    except HTTPException as e:
+        logger.warning("[生成任务] 校验失败: %s", e.detail)
+        raise
+
     # asset_library 存在性校验（仅在提供了 asset_library_id 时）
     resolved_asset_ids: list[str] = list(request.asset_ids)
     if asset_library_id:
         library = asset_library_repository.get(asset_library_id)
         if library is None or (project_id and library.project_id != project_id):
+            logger.warning("[生成任务] 素材库不存在: library_id=%s", asset_library_id)
             raise HTTPException(status_code=404, detail=f"AssetLibrary {asset_library_id} not found")
 
         assets = asset_repository.find_by_library(asset_library_id)
-        _ensure_library_has_ready_video_assets(assets)
+        try:
+            _ensure_library_has_ready_video_assets(assets)
+        except HTTPException as e:
+            logger.warning("[生成任务] 素材校验失败: %s", e.detail)
+            raise
 
         # 素材库自动匹配：当未显式指定 asset_ids 时，按模式自动选取
         if not resolved_asset_ids:
@@ -204,30 +225,85 @@ def create_generation_task(
     use_case = CreateGenerationTaskUseCase(generation_task_repository)
     count = request.count
     created_tasks = []
+    failed_tasks = []
+    user_id = authenticated_user.user.id
     # 同批次任务共享 batch_id，用于视频查重时批次内比对
     batch_id = uuid.uuid4().hex if count > 1 else ""
 
-    for _ in range(count):
-        task = use_case.execute(
-            CreateGenerationTaskCommand(
-                project_id=project_id,
-                asset_library_id=asset_library_id,
-                strategy_id=request.strategy_id,
-                voice_library_id=request.voice_library_id,
-                template_id=request.template_id,
-                asset_ids=resolved_asset_ids,
-                title_ids=request.title_ids,
-                voice_ids=request.voice_ids,
-                created_by_user_id=authenticated_user.user.id,
-                source_edit_plan_id=request.source_edit_plan_id,
-                asset_select_mode=request.asset_select_mode,
-                batch_id=batch_id,
+    # 预检查：批量提交前先看会不会超限，避免建一半才拒
+    try:
+        user_pending = generation_task_repository.count_pending_by_user(user_id)
+        global_pending = generation_task_repository.count_pending_total()
+        if user_pending + count > USER_PENDING_LIMIT:
+            raise UserPendingLimitExceeded(
+                user_id=user_id, pending_count=user_pending + count, limit=USER_PENDING_LIMIT
             )
-        )
-        celery_app.send_task("worker.generate_video", args=[task.id])
-        created_tasks.append(task)
+        if global_pending + count > GLOBAL_PENDING_LIMIT:
+            raise GlobalQueueFull(pending_count=global_pending + count, limit=GLOBAL_PENDING_LIMIT)
+    except UserPendingLimitExceeded as e:
+        raise HTTPException(
+            status_code=429,
+            detail=f"您的待处理任务过多（当前 {e.pending_count - count}/{e.limit}，本次提交 {count} 个），请等待完成后再提交",
+        ) from e
+    except GlobalQueueFull as e:
+        raise HTTPException(
+            status_code=503,
+            detail="系统繁忙，请稍后再试",
+        ) from e
 
-    items = [_to_generation_task_response(t) for t in created_tasks]
+    try:
+        for _ in range(count):
+            task = use_case.execute(
+                CreateGenerationTaskCommand(
+                    project_id=project_id,
+                    asset_library_id=asset_library_id,
+                    strategy_id=request.strategy_id,
+                    voice_library_id=request.voice_library_id,
+                    template_id=request.template_id,
+                    asset_ids=resolved_asset_ids,
+                    title_ids=request.title_ids,
+                    voice_ids=request.voice_ids,
+                    created_by_user_id=user_id,
+                    source_edit_plan_id=request.source_edit_plan_id,
+                    asset_select_mode=request.asset_select_mode,
+                    batch_id=batch_id,
+                )
+            )
+            try:
+                if safe_enqueue_generation_task(
+                    task,
+                    generation_task_repository,
+                    user_id=user_id,
+                    log_prefix="[生成任务]",
+                    log_task_status=True,
+                ):
+                    created_tasks.append(task)
+                else:
+                    failed_tasks.append(task)
+            except UserPendingLimitExceeded:
+                # 兜底：如果预检查后又并发提交了，在这里也拦住
+                failed_tasks.append(task)
+                if not created_tasks:
+                    raise HTTPException(
+                        status_code=429,
+                        detail="您的待处理任务过多，请等待完成后再提交",
+                    )
+                break
+            except GlobalQueueFull:
+                failed_tasks.append(task)
+                if not created_tasks:
+                    raise HTTPException(
+                        status_code=503,
+                        detail="系统繁忙，请稍后再试",
+                    )
+                break
+    except HTTPException:
+        raise
+    except Exception as e:
+        logger.error("[生成任务] 创建失败: %s", e, exc_info=True)
+        raise HTTPException(status_code=500, detail="创建生成任务失败，请稍后重试或查看任务日志")
+
+    items = [_to_generation_task_response(t) for t in created_tasks + failed_tasks]
     return BatchGenerationTaskResponse(items=items, total=len(items))
 
 
@@ -254,7 +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)
 
 
@@ -265,15 +341,20 @@ def list_generation_results(
     generation_task_repository: Any = Depends(get_generation_task_repository),
     generated_video_repository: Any = Depends(get_generated_video_repository),
     project_repository: Any = Depends(get_project_repository),
+    storage_service: OSSStorageService = Depends(get_storage_service),
 ) -> ListGeneratedVideosResponse:
     task = generation_task_repository.get(task_id)
     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)
-    return ListGeneratedVideosResponse(items=[_to_generated_video_response(item) for item in items])
+    responses = []
+    for item in items:
+        download_url = storage_service.get_download_url(item.file_url, expires_seconds=86400)
+        responses.append(_to_generated_video_response(item, download_url=download_url))
+    return ListGeneratedVideosResponse(items=responses)
 
 
 @router.post("/tasks/{task_id}/retry", response_model=GenerationTaskResponse)
@@ -292,6 +373,21 @@ def retry_generation_task(
     if status_val != "failed":
         raise HTTPException(status_code=409, detail="Only failed tasks can be retried")
 
+    user_id = authenticated_user.user.id
+    # 预检查：创建前判断，>= 上限就拒绝
+    user_pending = generation_task_repository.count_pending_by_user(user_id)
+    global_pending = generation_task_repository.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",
+        )
+
     use_case = CreateGenerationTaskUseCase(generation_task_repository)
     retried = use_case.execute(
         CreateGenerationTaskCommand(
@@ -303,10 +399,28 @@ def retry_generation_task(
             asset_ids=task.asset_ids,
             title_ids=task.title_ids,
             voice_ids=task.voice_ids,
-            created_by_user_id=authenticated_user.user.id,
+            created_by_user_id=user_id,
             source_edit_plan_id=task.source_edit_plan_id or "",
             asset_select_mode=getattr(task, "asset_select_mode", ""),
         )
     )
-    celery_app.send_task("worker.generate_video", args=[retried.id])
+    try:
+        if not safe_enqueue_generation_task(
+            retried,
+            generation_task_repository,
+            user_id=user_id,
+            log_prefix="[生成任务]",
+            log_task_status=True,
+        ):
+            logger.warning("[生成任务] 重试入队失败: task_id=%s", retried.id)
+    except UserPendingLimitExceeded:
+        raise HTTPException(
+            status_code=429,
+            detail="您的待处理任务过多，请等待完成后再提交",
+        ) from None
+    except GlobalQueueFull:
+        raise HTTPException(
+            status_code=503,
+            detail="系统繁忙，请稍后再试",
+        ) from None
     return _to_generation_task_response(retried)

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/task_center.py

@@ -1,7 +1,15 @@
+import logging
 from typing import Any
 
 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,
+    GlobalQueueFull,
+    UserPendingLimitExceeded,
+    safe_enqueue_generation_task,
+)
 from app.dependencies import (
     get_generation_task_repository,
     get_ingest_job_repository,
@@ -22,8 +30,9 @@ from packages.application import (
     SubmitIngestJobUseCase,
 )
 
-router = APIRouter()
+logger = logging.getLogger(__name__)
 
+router = APIRouter()
 
 def _humanize_task_error(error_message: str) -> str:
     raw = (error_message or "").strip()
@@ -139,6 +148,21 @@ def retry_task_by_id(
     if _status_value(task.status) != "failed":
         raise HTTPException(status_code=409, detail="Only failed tasks can be retried")
 
+    user_id = authenticated_user.user.id
+    # 预检查
+    user_pending = generation_task_repository.count_pending_by_user(user_id)
+    global_pending = generation_task_repository.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",
+        )
+
     use_case = CreateGenerationTaskUseCase(generation_task_repository)
     retried = use_case.execute(
         CreateGenerationTaskCommand(
@@ -150,10 +174,24 @@ def retry_task_by_id(
             asset_ids=task.asset_ids,
             title_ids=task.title_ids,
             voice_ids=task.voice_ids,
-            created_by_user_id=authenticated_user.user.id,
+            created_by_user_id=user_id,
         )
     )
-    celery_app.send_task("worker.generate_video", args=[retried.id])
+    try:
+        if not safe_enqueue_generation_task(
+            retried, generation_task_repository, user_id=user_id, log_prefix="[任务中心]"
+        ):
+            logger.warning("[任务中心] 用户级重试入队失败: task_id=%s", retried.id)
+    except UserPendingLimitExceeded:
+        raise HTTPException(
+            status_code=429,
+            detail="您的待处理任务过多，请等待完成后再提交",
+        ) from None
+    except GlobalQueueFull:
+        raise HTTPException(
+            status_code=503,
+            detail="系统繁忙，请稍后再试",
+        ) from None
     return UserTaskResponse(
         id=f"generation:{retried.id}",
         task_type="generation",
@@ -221,6 +259,22 @@ def retry_project_task(
             raise HTTPException(status_code=404, detail="Generation task not found")
         if _status_value(task.status) != "failed":
             raise HTTPException(status_code=409, detail="Only failed tasks can be retried")
+
+        user_id = authenticated_user.user.id
+        # 预检查
+        user_pending = generation_task_repository.count_pending_by_user(user_id)
+        global_pending = generation_task_repository.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",
+            )
+
         use_case = CreateGenerationTaskUseCase(generation_task_repository)
         retried = use_case.execute(
             CreateGenerationTaskCommand(
@@ -232,10 +286,24 @@ def retry_project_task(
                 asset_ids=task.asset_ids,
                 title_ids=task.title_ids,
                 voice_ids=task.voice_ids,
-                created_by_user_id=authenticated_user.user.id,
+                created_by_user_id=user_id,
             )
         )
-        celery_app.send_task("worker.generate_video", args=[retried.id])
+        try:
+            if not safe_enqueue_generation_task(
+                retried, generation_task_repository, user_id=user_id, log_prefix="[任务中心]"
+            ):
+                logger.warning("[任务中心] 项目级重试入队失败: task_id=%s", retried.id)
+        except UserPendingLimitExceeded:
+            raise HTTPException(
+                status_code=429,
+                detail="您的待处理任务过多，请等待完成后再提交",
+            ) from None
+        except GlobalQueueFull:
+            raise HTTPException(
+                status_code=503,
+                detail="系统繁忙，请稍后再试",
+            ) from None
         return _generation_task_to_project_response(retried)
     if task_type == "ingest":
         job = ingest_job_repository.get(source_id)

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

@@ -7,6 +7,7 @@ from typing import Optional
 
 from app.auth import AuthenticatedUser, get_current_user
 from app.dependencies import (
+    get_audio_url_signer,
     get_cosyvoice_service,
     get_db_session,
     get_user_repository,
@@ -56,7 +57,10 @@ def _get_repository(session: Session = Depends(get_db_session)) -> SQLAlchemyTTS
     return SQLAlchemyTTSJobRepository(session)
 
 
-def _to_response(job) -> TTSJobResponse:
+def _to_response(job, sign_url=None) -> TTSJobResponse:
+    output_url = job.output_audio_url
+    if sign_url and output_url:
+        output_url = sign_url(output_url)
     return TTSJobResponse(
         id=job.id,
         user_id=job.user_id,
@@ -66,7 +70,7 @@ def _to_response(job) -> TTSJobResponse:
         project_id=job.project_id,
         voice_clone_profile_id=job.voice_clone_profile_id,
         status=job.status,
-        output_audio_url=job.output_audio_url,
+        output_audio_url=output_url,
         output_audio_key=job.output_audio_key,
         duration=job.duration,
         file_size=job.file_size,
@@ -176,6 +180,7 @@ def list_tts_jobs(
     status_filter: Optional[str] = Query(None, alias="status"),
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     repository: SQLAlchemyTTSJobRepository = Depends(_get_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> ListTTSJobResponse:
     """列出用户的 TTS 合成任务。"""
     user_id = authenticated_user.user.id
@@ -183,7 +188,7 @@ def list_tts_jobs(
     skip = (page - 1) * page_size
     items, total = use_case.execute(user_id, status=status_filter, skip=skip, limit=page_size)
     return ListTTSJobResponse(
-        items=[_to_response(j) for j in items],
+        items=[_to_response(j, sign_url) for j in items],
         total=total,
         page=page,
         page_size=page_size,
@@ -195,6 +200,7 @@ def get_tts_job(
     job_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     repository: SQLAlchemyTTSJobRepository = Depends(_get_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> TTSJobResponse:
     """获取 TTS 任务详情。"""
     user_id = authenticated_user.user.id
@@ -203,7 +209,7 @@ def get_tts_job(
         job = use_case.execute(job_id, user_id)
     except TTSJobNotFoundError:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="TTS job not found")
-    return _to_response(job)
+    return _to_response(job, sign_url)
 
 
 @router.get("/jobs/{job_id}/status", response_model=TTSStatusResponse)
@@ -211,6 +217,7 @@ def get_tts_job_status(
     job_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     repository: SQLAlchemyTTSJobRepository = Depends(_get_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> TTSStatusResponse:
     """查询 TTS 合成状态（用于前端轮询）。"""
     user_id = authenticated_user.user.id
@@ -219,10 +226,13 @@ def get_tts_job_status(
         job = use_case.execute(job_id, user_id)
     except TTSJobNotFoundError:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="TTS job not found")
+    output_url = job.output_audio_url
+    if output_url:
+        output_url = sign_url(output_url)
     return TTSStatusResponse(
         id=job.id,
         status=job.status,
-        output_audio_url=job.output_audio_url,
+        output_audio_url=output_url,
         error_message=job.error_message,
         duration=job.duration,
         retry_count=job.retry_count,
@@ -231,7 +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),
@@ -243,7 +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(
@@ -258,6 +268,7 @@ def save_tts_job_to_library(
     tts_repository: SQLAlchemyTTSJobRepository = Depends(_get_repository),
     voice_library_repository: SQLAlchemyVoiceLibraryRepository = Depends(get_voice_library_repository),
     user_repository: UserRepository = Depends(get_user_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> SaveToLibraryResponse:
     """将已完成的 TTS 合成结果保存到配音库。
 
@@ -328,7 +339,7 @@ def save_tts_job_to_library(
     return SaveToLibraryResponse(
         id=item.id,
         name=item.name,
-        audio_url=item.audio_url,
+        audio_url=sign_url(item.audio_url) if item.audio_url else "",
         duration=item.duration,
         voice_id=item.voice_id,
         voice_name=item.voice_name,
@@ -361,10 +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,
 )
 
@@ -38,6 +37,7 @@ router = APIRouter()
 
 
 def _to_response(profile) -> VoiceCloneProfileResponse:
+    # source_audio_url 是用户传入的原始 URL（可能是外部地址），不做预签名转换
     return VoiceCloneProfileResponse(
         id=profile.id,
         user_id=profile.user_id,
@@ -172,6 +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,
@@ -184,7 +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

@@ -7,8 +7,9 @@ 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_db_session, get_user_repository
+from app.dependencies import get_audio_url_signer, get_db_session, get_user_repository
 from app.schemas.voice import (
     PresetVoiceItemResponse,
     PresetVoiceListResponse,
@@ -50,7 +51,10 @@ def _get_clone_profile_repository(session: Session = Depends(get_db_session)) ->
     return SQLAlchemyVoiceCloneProfileRepository(session)
 
 
-def _to_response(item) -> VoiceLibraryItemResponse:
+def _to_response(item, sign_url=None) -> VoiceLibraryItemResponse:
+    audio = item.audio_url
+    if sign_url and audio:
+        audio = sign_url(audio)
     return VoiceLibraryItemResponse(
         id=item.id,
         user_id=item.user_id,
@@ -59,7 +63,7 @@ def _to_response(item) -> VoiceLibraryItemResponse:
         voice_provider=item.voice_provider,
         voice_id=item.voice_id,
         voice_name=item.voice_name,
-        audio_url=item.audio_url,
+        audio_url=audio,
         duration=item.duration,
         file_size=item.file_size,
         status=item.status,
@@ -70,16 +74,20 @@ def _to_response(item) -> VoiceLibraryItemResponse:
     )
 
 
-def _to_unified_response(item, profile_id_map: dict | None = None) -> UnifiedVoiceItemResponse:
+def _to_unified_response(item, profile_id_map: dict | None = None, sign_url=None) -> UnifiedVoiceItemResponse:
     """将数据库音色转换为统一响应格式。
 
     Args:
         item: VoiceLibraryItem
         profile_id_map: voice_id → profile_id 映射，用于填充 voice_clone_profile_id
+        sign_url: 音频URL预签名函数
     """
     profile_id = None
     if profile_id_map and item.voice_id:
         profile_id = profile_id_map.get(item.voice_id)
+    audio = item.audio_url
+    if sign_url and audio:
+        audio = sign_url(audio)
     return UnifiedVoiceItemResponse(
         id=item.id,
         type="clone",
@@ -89,7 +97,7 @@ def _to_unified_response(item, profile_id_map: dict | None = None) -> UnifiedVoi
         language="zh-CN",
         voice_id=item.voice_id,
         voice_provider=item.voice_provider or "cosyvoice",
-        audio_url=item.audio_url,
+        audio_url=audio,
         duration=item.duration,
         file_size=item.file_size,
         status=item.status,
@@ -118,13 +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）====================
 
 
@@ -140,6 +141,7 @@ def list_voices_unified(
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     voice_repository: SQLAlchemyVoiceLibraryRepository = Depends(_get_voice_repository),
     clone_profile_repository: SQLAlchemyVoiceCloneProfileRepository = Depends(_get_clone_profile_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> UnifiedVoiceListResponse:
     """获取配音列表（预置音色 + 用户克隆音色）。
 
@@ -167,7 +169,7 @@ def list_voices_unified(
         # 批量查询 voice_id → profile_id 映射，填充 voice_clone_profile_id
         voice_ids = [i.voice_id for i in clone_items_raw if i.voice_id]
         profile_id_map = clone_profile_repository.find_profile_ids_by_voice_ids(voice_ids) if voice_ids else {}
-        clone_items = [_to_unified_response(i, profile_id_map) for i in clone_items_raw]
+        clone_items = [_to_unified_response(i, profile_id_map, sign_url) for i in clone_items_raw]
 
     # 组装结果
     if type == "preset":
@@ -224,6 +226,7 @@ def list_voices_legacy(
     limit: int = Query(50, ge=1, le=200),
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     voice_repository: SQLAlchemyVoiceLibraryRepository = Depends(_get_voice_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> ListVoiceLibraryResponse:
     """原有配音列表接口（仅返回用户克隆音色）。
 
@@ -233,7 +236,7 @@ def list_voices_legacy(
     use_case = ListVoiceLibraryUseCase(voice_repository)
     items, total = use_case.execute(user_id, status=status_filter, skip=skip, limit=limit)
     return ListVoiceLibraryResponse(
-        items=[_to_response(i) for i in items],
+        items=[_to_response(i, sign_url) for i in items],
         total=total,
     )
 
@@ -243,13 +246,14 @@ def get_voice(
     voice_id: str,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     voice_repository: SQLAlchemyVoiceLibraryRepository = Depends(_get_voice_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> VoiceLibraryItemResponse:
     user_id = authenticated_user.user.id
     use_case = GetVoiceLibraryUseCase(voice_repository)
     item = use_case.execute(voice_id, user_id)
     if item is None:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Voice not found")
-    return _to_response(item)
+    return _to_response(item, sign_url)
 
 
 @router.post("", response_model=VoiceLibraryItemResponse, status_code=status.HTTP_201_CREATED)
@@ -258,9 +262,10 @@ def create_voice(
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     voice_repository: SQLAlchemyVoiceLibraryRepository = Depends(_get_voice_repository),
     user_repository: UserRepository = Depends(get_user_repository),
+    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,
@@ -283,7 +288,7 @@ def create_voice(
             status_code=status.HTTP_429_TOO_MANY_REQUESTS,
             detail=f"配音库配额已满（{exc.used}/{exc.limit}），请升级套餐",
         )
-    return _to_response(item)
+    return _to_response(item, sign_url)
 
 
 @router.put("/{voice_id}", response_model=VoiceLibraryItemResponse)
@@ -292,6 +297,7 @@ def update_voice(
     request: UpdateVoiceLibraryRequest,
     authenticated_user: AuthenticatedUser = Depends(get_current_user),
     voice_repository: SQLAlchemyVoiceLibraryRepository = Depends(_get_voice_repository),
+    sign_url=Depends(get_audio_url_signer),
 ) -> VoiceLibraryItemResponse:
     user_id = authenticated_user.user.id
     command = UpdateVoiceLibraryCommand(
@@ -313,10 +319,10 @@ def update_voice(
         item = use_case.execute(command)
     except NotFoundError:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Voice not found")
-    return _to_response(item)
+    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),
@@ -327,4 +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
@@ -75,19 +73,42 @@ 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"
+
+    @field_validator("OSS_ACCESS_KEY_ID", mode="before")
+    @classmethod
+    def validate_oss_access_key_id(cls, v):
+        if (v is None or v == "") and os.getenv("APP_ENV", "development") != "development":
+            raise ValueError(
+                "OSS_ACCESS_KEY_ID must be set via environment variable in non-development environments. "
+                "Check the server .env file (e.g. /var/lib/xiaoxia-saas-staging/.env)."
+            )
+        return v or ""
+
+    @field_validator("OSS_ACCESS_KEY_SECRET", mode="before")
+    @classmethod
+    def validate_oss_access_key_secret(cls, v):
+        if (v is None or v == "") and os.getenv("APP_ENV", "development") != "development":
+            raise ValueError(
+                "OSS_ACCESS_KEY_SECRET must be set via environment variable in non-development environments. "
+                "Check the server .env file (e.g. /var/lib/xiaoxia-saas-staging/.env)."
+            )
+        return v or ""
+
     OSS_DIRECT_UPLOAD_MAX_MB: int = Field(
         default=2000,
         validation_alias=AliasChoices("OSS_DIRECT_UPLOAD_MAX_MB", "MAX_UPLOAD_SIZE_MB"),
     )
     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

@@ -1,243 +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:
-                    auth = oss2.Auth(
-                        settings.OSS_ACCESS_KEY_ID,
-                        settings.OSS_ACCESS_KEY_SECRET,
-                    )
-                    self.bucket = oss2.Bucket(
-                        auth,
-                        settings.OSS_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 _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
-            return self.get_url(self._normalize_storage_key(storage_key_or_url))
-
-        storage_key = self._normalize_storage_key(storage_key_or_url)
-        try:
-            return self.bucket.sign_url("GET", storage_key, expires_seconds)
-        except Exception:
-            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()
-    return _storage_service
+__all__ = ["OSSStorageService", "get_storage_service", "get_shared_storage_service"]

apps/api/app/core/task_enqueue.py

@@ -0,0 +1,221 @@
+import logging
+from typing import Any
+
+from app.core.celery_app import celery_app
+
+logger = logging.getLogger(__name__)
+
+# ── 限流阈值常量（全系统统一管理，不要在业务代码里硬编码） ──
+USER_PENDING_LIMIT = 3  # 单用户 pending 上限
+GLOBAL_PENDING_LIMIT = 20  # 全局 pending 上限
+
+
+class UserPendingLimitExceeded(Exception):
+    """用户 pending 任务数超限，返回 429。"""
+
+    def __init__(self, user_id: str, pending_count: int, limit: int):
+        self.user_id = user_id
+        self.pending_count = pending_count
+        self.limit = limit
+        super().__init__(f"用户 {user_id} pending 任务数 {pending_count} 超过上限 {limit}")
+
+
+class GlobalQueueFull(Exception):
+    """全局限流，返回 503。"""
+
+    def __init__(self, pending_count: int, limit: int):
+        self.pending_count = pending_count
+        self.limit = limit
+        super().__init__(f"系统 pending 任务数 {pending_count} 超过上限 {limit}")
+
+
+def check_queue_limits(
+    user_id: str,
+    generation_task_repository: Any,
+    *,
+    user_pending_limit: int = USER_PENDING_LIMIT,
+    global_pending_limit: int = GLOBAL_PENDING_LIMIT,
+) -> None:
+    """检查队列限流（预检查用，任务创建前调用），超限抛对应异常。
+
+    边界语义：>= 上限即拒绝（达到上限就不能再加新任务）。
+
+    Args:
+        user_id: 用户 ID
+        generation_task_repository: 任务仓储
+        user_pending_limit: 单用户 pending 上限，默认 USER_PENDING_LIMIT
+        global_pending_limit: 全局 pending 上限，默认 GLOBAL_PENDING_LIMIT
+
+    Raises:
+        GlobalQueueFull: 全局超限时抛出（优先级更高，先查全局）
+        UserPendingLimitExceeded: 用户超限时抛出
+    """
+    # 先查全局（系统级保护优先级更高）
+    global_pending = generation_task_repository.count_pending_total()
+    if global_pending >= global_pending_limit:
+        logger.warning(
+            "[队列限流] 全局 pending 任务数超限: %d/%d, user_id=%s",
+            global_pending,
+            global_pending_limit,
+            user_id,
+        )
+        raise GlobalQueueFull(pending_count=global_pending, limit=global_pending_limit)
+
+    # 再查用户级
+    if user_id:
+        user_pending = generation_task_repository.count_pending_by_user(user_id)
+        if user_pending >= user_pending_limit:
+            logger.warning(
+                "[队列限流] 用户 pending 任务数超限: user_id=%s, count=%d/%d",
+                user_id,
+                user_pending,
+                user_pending_limit,
+            )
+            raise UserPendingLimitExceeded(user_id=user_id, pending_count=user_pending, limit=user_pending_limit)
+
+
+def _mark_task_failed_safely(
+    task: Any,
+    generation_task_repository: Any,
+    log_prefix: str,
+    reason: str,
+) -> None:
+    """安全地把任务标记为 failed，更新失败只打日志不崩溃。"""
+    try:
+        task.mark_failed(f"任务被限流拒绝: {reason}")
+        generation_task_repository.update(task)
+    except Exception as update_err:
+        logger.error(
+            "%s 限流后更新状态也失败: task_id=%s error=%s",
+            log_prefix,
+            task.id,
+            update_err,
+            exc_info=True,
+        )
+
+
+def safe_enqueue_generation_task(
+    task: Any,
+    generation_task_repository: Any,
+    *,
+    user_id: str = "",
+    log_prefix: str = "[任务队列]",
+    log_task_status: bool = False,
+    user_pending_limit: int = USER_PENDING_LIMIT,
+    global_pending_limit: int = GLOBAL_PENDING_LIMIT,
+) -> bool:
+    """安全入队：入队前限流检查 → 发送 Celery 任务 → 入队后最终校验兜底。
+
+    边界说明：
+    入队前检查用 > 而非 >=。因为调用此函数时 task 已经是 pending 状态并计入 DB，
+    pending 总数包含了当前任务本身。pending > limit 等价于"其他任务数 >= limit"，
+    与预检查的 >= 语义一致（都是达到上限就拒绝新任务）。
+
+    入队后最终校验：发送 Celery 成功后再查一次 DB 计数，处理并发竞态场景
+    （两个请求同时通过入队前检查，后到的那个在这里被兜住）。
+
+    Args:
+        task: 生成任务对象，需有 id 属性和 mark_failed 方法（状态已为 pending）
+        generation_task_repository: 任务仓储，用于更新状态
+        user_id: 用户 ID，传了才做用户级限流检查
+        log_prefix: 日志前缀，便于区分调用来源
+        log_task_status: 成功日志中是否额外打印任务状态
+        user_pending_limit: 单用户 pending 上限，默认 USER_PENDING_LIMIT
+        global_pending_limit: 全局 pending 上限，默认 GLOBAL_PENDING_LIMIT
+
+    Returns:
+        True 表示入队成功，False 表示入队失败（已标记为 failed）
+
+    Raises:
+        GlobalQueueFull: 全局 pending 超限时抛出，任务会被标记为 failed
+        UserPendingLimitExceeded: 用户 pending 超限时抛出，任务会被标记为 failed
+    """
+    # ── 入队前检查：任务已是 pending，用 > 判断（包含当前任务） ──
+
+    # 全局限流检查（始终生效）
+    global_pending = generation_task_repository.count_pending_total()
+    if global_pending > global_pending_limit:
+        logger.warning(
+            "[队列限流] 全局 pending 任务数超限(入队前): %d/%d, user_id=%s",
+            global_pending,
+            global_pending_limit,
+            user_id or "unknown",
+        )
+        exc = GlobalQueueFull(pending_count=global_pending, limit=global_pending_limit)
+        _mark_task_failed_safely(task, generation_task_repository, log_prefix, str(exc))
+        raise exc
+
+    # 用户级限流检查（传了 user_id 才做）
+    if user_id:
+        user_pending = generation_task_repository.count_pending_by_user(user_id)
+        if user_pending > user_pending_limit:
+            logger.warning(
+                "[队列限流] 用户 pending 任务数超限(入队前): user_id=%s, count=%d/%d",
+                user_id,
+                user_pending,
+                user_pending_limit,
+            )
+            exc = UserPendingLimitExceeded(user_id=user_id, pending_count=user_pending, limit=user_pending_limit)
+            _mark_task_failed_safely(task, generation_task_repository, log_prefix, str(exc))
+            raise exc
+
+    # ── 发送 Celery 任务 ──
+    try:
+        celery_app.send_task("worker.generate_video", args=[task.id])
+    except Exception as e:
+        logger.error(
+            "%s 入队失败，标记为失败: task_id=%s error=%s",
+            log_prefix,
+            task.id,
+            e,
+            exc_info=True,
+        )
+        try:
+            task.mark_failed(f"任务入队失败: {e}")
+            generation_task_repository.update(task)
+        except Exception as update_err:
+            logger.error(
+                "%s 入队失败后更新状态也失败: task_id=%s error=%s",
+                log_prefix,
+                task.id,
+                update_err,
+                exc_info=True,
+            )
+        return False
+
+    # ── 入队后最终校验：并发竞态兜底 ──
+    # 发送成功后再查一次，防止两个请求同时通过入队前检查导致超限
+    global_after = generation_task_repository.count_pending_total()
+    user_after = generation_task_repository.count_pending_by_user(user_id) if user_id else 0
+
+    global_over = global_after > global_pending_limit
+    user_over = bool(user_id and user_after > user_pending_limit)
+
+    if global_over or user_over:
+        if global_over:
+            reason = f"全局 pending 超限(入队后): {global_after}/{global_pending_limit}"
+            exc: Exception = GlobalQueueFull(pending_count=global_after, limit=global_pending_limit)
+        else:
+            reason = f"用户 pending 超限(入队后): {user_after}/{user_pending_limit}"
+            exc = UserPendingLimitExceeded(user_id=user_id, pending_count=user_after, limit=user_pending_limit)
+
+        logger.warning(
+            "[队列限流] %s, task_id=%s, user_id=%s — 回滚状态为 failed",
+            reason,
+            task.id,
+            user_id or "unknown",
+        )
+        _mark_task_failed_safely(task, generation_task_repository, log_prefix, reason)
+        raise exc
+
+    # 入队成功日志
+    if log_task_status:
+        logger.info(
+            "%s 入队成功: task_id=%s, status=%s",
+            log_prefix,
+            task.id,
+            task.status,
+        )
+    else:
+        logger.info("%s 入队成功: task_id=%s", log_prefix, task.id)
+    return True

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)
 
@@ -201,7 +189,37 @@ def get_voice_clone_profile_repository(
 
 
 def get_cosyvoice_service():
-    """Provide the CosyVoice service instance."""
+    """Provide the CosyVoice service instance.
+
+    注入 OSS 音频URL预签名函数，确保私有bucket下的参考音频
+    能被 CosyVoice 服务器下载。
+    """
+    from app.core.storage import get_storage_service
+
     from packages.application.cosyvoice_service import CosyVoiceService
 
-    return CosyVoiceService()
+    storage = get_storage_service()
+
+    def _sign_audio_url(url: str) -> str:
+        """对音频URL做预签名，私有bucket下 CosyVoice 服务器才能下载."""
+        return storage.get_download_url(url, expires_seconds=86400)
+
+    return CosyVoiceService(audio_url_signer=_sign_audio_url)
+
+
+def get_audio_url_signer():
+    """提供音频URL预签名函数（24小时有效期）。
+
+    用于所有 API 返回给前端的音频 URL，确保私有 OSS bucket 下可正常访问。
+    空 URL、非 OSS URL 直接原样返回；签名失败时回退到原始 URL。
+    """
+    from app.core.storage import get_storage_service
+
+    storage = get_storage_service()
+
+    def sign_audio_url(url: str) -> str:
+        if not url:
+            return url
+        return storage.get_download_url(url, expires_seconds=86400)
+
+    return sign_audio_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/generation_task.py

@@ -1,4 +1,6 @@
-from pydantic import BaseModel, Field, model_validator
+import json
+
+from pydantic import BaseModel, Field, field_validator, model_validator
 
 
 class CreateGenerationTaskRequest(BaseModel):
@@ -62,6 +64,21 @@ class GenerationTaskResponse(BaseModel):
     progress: float
     result_count: int
     error_message: str
+    logs: list[dict] = Field(default_factory=list)
+
+    @field_validator("logs", mode="before")
+    @classmethod
+    def _parse_logs(cls, v: object) -> list[dict]:
+        """将 JSON 字符串解析为 list[dict]。"""
+        if isinstance(v, str):
+            try:
+                parsed = json.loads(v)
+                return parsed if isinstance(parsed, list) else []
+            except (json.JSONDecodeError, TypeError):
+                return []
+        if isinstance(v, list):
+            return v
+        return []
 
 
 class BatchGenerationTaskResponse(BaseModel):

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/__init__.py

@@ -4,6 +4,7 @@ from .auto_clip_service import AutoClipService
 from .edit_plan_service import EditPlanService
 from .edit_template_service import EditTemplateService
 from .job_service import JobService
+from .plan_generator_service import PlanGeneratorService
 from .video_compose_service import VideoComposeService
 
 __all__ = [
@@ -11,5 +12,6 @@ __all__ = [
     "EditPlanService",
     "EditTemplateService",
     "JobService",
+    "PlanGeneratorService",
     "VideoComposeService",
 ]

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/edit_template_service.py

@@ -100,6 +100,7 @@ class EditTemplateService:
         *,
         description: str = "",
         template_type: str = "default",
+        editing_mode: str = "one_take",
         config: Optional[dict[str, Any]] = None,
         preview_url: str = "",
         sort_weight: int = 0,
@@ -124,6 +125,7 @@ class EditTemplateService:
             name=clean_name,
             description=description,
             template_type=template_type,
+            editing_mode=editing_mode,
             config=config,
             preview_url=preview_url,
             sort_weight=sort_weight,
@@ -139,6 +141,7 @@ class EditTemplateService:
         name: Optional[str] = None,
         description: Optional[str] = None,
         template_type: Optional[str] = None,
+        editing_mode: Optional[str] = None,
         config: Optional[dict[str, Any]] = None,
         preview_url: Optional[str] = None,
         sort_weight: Optional[int] = None,
@@ -165,6 +168,7 @@ class EditTemplateService:
             name=new_name,
             description=description.strip() if description is not None else existing.description,
             template_type=template_type.strip() if template_type is not None else existing.template_type,
+            editing_mode=editing_mode.strip() if editing_mode is not None else existing.editing_mode,
             config=config if config is not None else existing.config,
             preview_url=preview_url.strip() if preview_url is not None else existing.preview_url,
             sort_weight=sort_weight if sort_weight is not None else existing.sort_weight,

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

@@ -0,0 +1,391 @@
+"""PlanGeneratorService — 基于模板+素材自动生成剪辑计划.
+
+核心职责:
+- 根据 EditTemplate 的 editing_mode 和 TemplateClipConfig 列表，
+  自动生成 EditPlan + EditPlanClip 列表
+- 四种模式素材分配策略:
+  - ONE_TAKE: 素材顺序分配给 main 类型 clips
+  - PIP: 第1个素材→main(全屏背景)，其余→overlay clips
+  - VOICE_OVER: 素材→main clips (B-roll)，标记需要配音叠加
+  - VOICE_PIP: 第1个→background, 第2个→corner_voice, 其余→b_roll
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any, List
+
+from sqlalchemy.orm import Session
+
+from packages.adapters.sqlalchemy_impl import (
+    SQLAlchemyEditPlanClipRepository,
+    SQLAlchemyEditPlanRepository,
+)
+from packages.domain.config_schemas import normalize_plan_config
+from packages.domain.edit_plan import EditPlan
+from packages.domain.edit_plan_clip import EditPlanClip
+from packages.domain.edit_template import EditTemplate
+from packages.domain.editing_mode import EditingMode
+from packages.domain.template_clip_config import ClipType, TemplateClipConfig
+
+logger = logging.getLogger(__name__)
+
+
+# ── 默认片段时长（秒） ────────────────────────────────────────────────────────
+_DEFAULT_CLIP_DURATION = 5.0
+_DEFAULT_INTRO_DURATION = 3.0
+_DEFAULT_OUTRO_DURATION = 3.0
+
+
+class PlanGeneratorService:
+    """剪辑计划生成器
+
+    基于模板 + 素材，自动生成 EditPlan 及 EditPlanClip 列表。
+    """
+
+    def __init__(self, db: Session) -> None:
+        self._plan_repo = SQLAlchemyEditPlanRepository(db)
+        self._clip_repo = SQLAlchemyEditPlanClipRepository(db)
+
+    # ── 公开接口 ─────────────────────────────────────────────────────────────
+
+    def generate_from_template(
+        self,
+        template: EditTemplate,
+        clip_configs: List[TemplateClipConfig],
+        asset_ids: List[str],
+        *,
+        project_id: str = "",
+        created_by_user_id: str = "",
+        name: str = "",
+    ) -> dict[str, Any]:
+        """基于模板+素材生成剪辑计划
+
+        Args:
+            template: 剪辑模板实体
+            clip_configs: 模板片段配置列表（可为空，自动生成默认结构）
+            asset_ids: 素材 ID 列表
+            project_id: 所属项目 ID
+            created_by_user_id: 创建者用户 ID
+            name: 计划名称（为空则自动取模板名）
+
+        Returns:
+            dict: {"plan": EditPlan, "clips": List[EditPlanClip]}
+        """
+        editing_mode = template.editing_mode or EditingMode.ONE_TAKE.value
+        plan_name = name.strip() or f"{template.name} - 剪辑计划"
+
+        # 1. 构建 plan config（继承模板的 title/subtitle/bgm，记录 editing_mode）
+        plan_config = self._build_plan_config(template, editing_mode)
+
+        # 2. 创建 EditPlan
+        plan = EditPlan.create(
+            template_id=template.id,
+            name=plan_name,
+            config=plan_config,
+            total_duration=0.0,
+            project_id=project_id,
+            created_by_user_id=created_by_user_id,
+        )
+        plan = self._plan_repo.create(plan)
+        logger.info(
+            "生成剪辑计划: plan_id=%s template=%s mode=%s assets=%d",
+            plan.id,
+            template.id,
+            editing_mode,
+            len(asset_ids),
+        )
+
+        # 3. 生成片段列表
+        if clip_configs:
+            clips = self._create_clips_from_configs(plan.id, clip_configs)
+        else:
+            clips = self._generate_default_clips(plan.id, editing_mode, len(asset_ids))
+
+        # 4. 按 editing_mode 分配素材
+        if asset_ids:
+            self._distribute_assets(clips, asset_ids, editing_mode)
+
+        # 5. 持久化所有 clips 并计算总时长
+        created_clips: List[EditPlanClip] = []
+        total_duration = 0.0
+        for clip in clips:
+            saved = self._clip_repo.create(clip)
+            created_clips.append(saved)
+            total_duration += saved.duration
+
+        # 6. 更新 plan 的 total_duration
+        plan.total_duration = total_duration
+        plan = self._plan_repo.update(plan)
+
+        # 7. 流转到 editing 状态
+        try:
+            plan.start_editing()
+            plan = self._plan_repo.update(plan)
+        except ValueError as exc:
+            logger.warning("计划状态流转失败: plan_id=%s error=%s", plan.id, exc)
+
+        logger.info(
+            "剪辑计划生成完成: plan_id=%s clips=%d duration=%.1f",
+            plan.id,
+            len(created_clips),
+            total_duration,
+        )
+
+        return {"plan": plan, "clips": created_clips}
+
+    # ── 内部方法 ─────────────────────────────────────────────────────────────
+
+    def _build_plan_config(
+        self,
+        template: EditTemplate,
+        editing_mode: str,
+    ) -> dict[str, Any]:
+        """从模板配置构建 plan config"""
+        template_config = template.config or {}
+        plan_config: dict[str, Any] = {
+            "editing_mode": editing_mode,
+        }
+        # 继承模板的 cover/title/subtitle/bgm 配置
+        for key in ("cover", "title", "subtitle", "bgm"):
+            if key in template_config:
+                plan_config[key] = template_config[key]
+
+        return normalize_plan_config(plan_config)
+
+    def _create_clips_from_configs(
+        self,
+        plan_id: str,
+        clip_configs: List[TemplateClipConfig],
+    ) -> List[EditPlanClip]:
+        """从 TemplateClipConfig 列表创建 EditPlanClip 列表（未持久化）"""
+        clips: List[EditPlanClip] = []
+        # 按 order 排序
+        sorted_configs = sorted(clip_configs, key=lambda c: c.order)
+
+        for cfg in sorted_configs:
+            # 计算时长：取 min_duration 和 max_duration 的中间值
+            if cfg.min_duration > 0 and cfg.max_duration > 0:
+                duration = (cfg.min_duration + cfg.max_duration) / 2
+            elif cfg.min_duration > 0:
+                duration = cfg.min_duration
+            elif cfg.max_duration > 0:
+                duration = cfg.max_duration
+            else:
+                duration = _DEFAULT_CLIP_DURATION
+
+            # clip_type 可能是枚举或字符串
+            clip_type = cfg.clip_type.value if hasattr(cfg.clip_type, "value") else cfg.clip_type
+
+            # transition_effect 可能是枚举或字符串
+            transition = (
+                cfg.transition_effect.value if hasattr(cfg.transition_effect, "value") else cfg.transition_effect
+            )
+
+            clip = EditPlanClip.create(
+                plan_id=plan_id,
+                clip_type=clip_type,
+                order=cfg.order,
+                template_clip_config_id=cfg.id,
+                text_content=getattr(cfg, "text_template", "") or "",
+                duration=duration,
+                transition_effect=transition or "cut",
+            )
+            clips.append(clip)
+
+        return clips
+
+    def _generate_default_clips(
+        self,
+        plan_id: str,
+        editing_mode: str,
+        asset_count: int,
+    ) -> List[EditPlanClip]:
+        """无 clip_configs 时，根据 editing_mode 生成默认 clip 结构
+
+        - ONE_TAKE: N 个 main clips（N = asset_count，至少1个）
+        - PIP: 1 个 main + (N-1) 个 overlay（N = asset_count）
+        - VOICE_OVER: N 个 main clips + 标记需要配音
+        - VOICE_PIP: 1 个 background + 1 个 corner_voice + (N-2) 个 b_roll
+        """
+        n = max(asset_count, 1)
+        clips: List[EditPlanClip] = []
+        order = 0
+
+        if editing_mode == EditingMode.PIP.value:
+            # 1 个 main（全屏背景）
+            clips.append(
+                EditPlanClip.create(
+                    plan_id=plan_id,
+                    clip_type=ClipType.MAIN.value,
+                    order=order,
+                    duration=_DEFAULT_CLIP_DURATION,
+                )
+            )
+            order += 1
+            # 剩余为 overlay
+            for i in range(1, n):
+                clips.append(
+                    EditPlanClip.create(
+                        plan_id=plan_id,
+                        clip_type="overlay",
+                        order=order,
+                        duration=_DEFAULT_CLIP_DURATION,
+                    )
+                )
+                order += 1
+
+        elif editing_mode == EditingMode.VOICE_OVER.value:
+            # N 个 main clips（B-roll）
+            for i in range(n):
+                clips.append(
+                    EditPlanClip.create(
+                        plan_id=plan_id,
+                        clip_type=ClipType.MAIN.value,
+                        order=order,
+                        duration=_DEFAULT_CLIP_DURATION,
+                        config={"role": "b_roll"},
+                    )
+                )
+                order += 1
+
+        elif editing_mode == EditingMode.VOICE_PIP.value:
+            # 1 个 background
+            clips.append(
+                EditPlanClip.create(
+                    plan_id=plan_id,
+                    clip_type="background",
+                    order=order,
+                    duration=_DEFAULT_CLIP_DURATION,
+                )
+            )
+            order += 1
+            # 1 个 corner_voice
+            clips.append(
+                EditPlanClip.create(
+                    plan_id=plan_id,
+                    clip_type="corner_voice",
+                    order=order,
+                    duration=_DEFAULT_CLIP_DURATION,
+                )
+            )
+            order += 1
+            # 剩余为 b_roll
+            for i in range(2, n):
+                clips.append(
+                    EditPlanClip.create(
+                        plan_id=plan_id,
+                        clip_type="b_roll",
+                        order=order,
+                        duration=_DEFAULT_CLIP_DURATION,
+                    )
+                )
+                order += 1
+
+        else:
+            # ONE_TAKE: N 个 main clips
+            for i in range(n):
+                clips.append(
+                    EditPlanClip.create(
+                        plan_id=plan_id,
+                        clip_type=ClipType.MAIN.value,
+                        order=order,
+                        duration=_DEFAULT_CLIP_DURATION,
+                    )
+                )
+                order += 1
+
+        return clips
+
+    def _distribute_assets(
+        self,
+        clips: List[EditPlanClip],
+        asset_ids: List[str],
+        editing_mode: str,
+    ) -> None:
+        """按 editing_mode 将素材分配到 clips（就地修改，未持久化）
+
+        分配策略:
+        - ONE_TAKE: 素材按顺序依次分配给 main 类型 clips
+        - PIP: 第1个素材→main(全屏背景)，其余→交替分配给 overlay clips
+        - VOICE_OVER: 素材→main clips (B-roll)
+        - VOICE_PIP: 第1个→background, 第2个→corner_voice, 其余→b_roll
+        """
+        if not asset_ids or not clips:
+            return
+
+        if editing_mode == EditingMode.ONE_TAKE.value:
+            self._distribute_one_take(clips, asset_ids)
+        elif editing_mode == EditingMode.PIP.value:
+            self._distribute_pip(clips, asset_ids)
+        elif editing_mode == EditingMode.VOICE_OVER.value:
+            self._distribute_voice_over(clips, asset_ids)
+        elif editing_mode == EditingMode.VOICE_PIP.value:
+            self._distribute_voice_pip(clips, asset_ids)
+        else:
+            # 未知模式，退化为 one_take
+            self._distribute_one_take(clips, asset_ids)
+
+    def _distribute_one_take(
+        self,
+        clips: List[EditPlanClip],
+        asset_ids: List[str],
+    ) -> None:
+        """ONE_TAKE: 素材按顺序依次分配给 main 类型 clips"""
+        main_clips = [c for c in clips if c.clip_type == ClipType.MAIN.value]
+        for i, clip in enumerate(main_clips):
+            if i < len(asset_ids):
+                clip.assign_asset(asset_ids[i])
+
+    def _distribute_pip(
+        self,
+        clips: List[EditPlanClip],
+        asset_ids: List[str],
+    ) -> None:
+        """PIP: 第1个素材→main(全屏背景)，其余→overlay clips"""
+        # 第1个素材 → main clip
+        main_clips = [c for c in clips if c.clip_type == ClipType.MAIN.value]
+        if main_clips and asset_ids:
+            main_clips[0].assign_asset(asset_ids[0])
+
+        # 其余素材 → overlay clips
+        overlay_clips = [c for c in clips if c.clip_type == "overlay"]
+        remaining = asset_ids[1:]
+        for i, clip in enumerate(overlay_clips):
+            if i < len(remaining):
+                clip.assign_asset(remaining[i])
+
+    def _distribute_voice_over(
+        self,
+        clips: List[EditPlanClip],
+        asset_ids: List[str],
+    ) -> None:
+        """VOICE_OVER: 素材→main clips (B-roll)"""
+        main_clips = [c for c in clips if c.clip_type == ClipType.MAIN.value]
+        for i, clip in enumerate(main_clips):
+            if i < len(asset_ids):
+                clip.assign_asset(asset_ids[i])
+
+    def _distribute_voice_pip(
+        self,
+        clips: List[EditPlanClip],
+        asset_ids: List[str],
+    ) -> None:
+        """VOICE_PIP: 第1个→background, 第2个→corner_voice, 其余→b_roll"""
+        bg_clips = [c for c in clips if c.clip_type == "background"]
+        corner_clips = [c for c in clips if c.clip_type == "corner_voice"]
+        broll_clips = [c for c in clips if c.clip_type == "b_roll"]
+
+        # 第1个素材 → background
+        if bg_clips and len(asset_ids) > 0:
+            bg_clips[0].assign_asset(asset_ids[0])
+
+        # 第2个素材 → corner_voice
+        if corner_clips and len(asset_ids) > 1:
+            corner_clips[0].assign_asset(asset_ids[1])
+
+        # 其余素材 → b_roll
+        remaining = asset_ids[2:]
+        for i, clip in enumerate(broll_clips):
+            if i < len(remaining):
+                clip.assign_asset(remaining[i])

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

@@ -1,7 +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。
 """
 
-from .processor import VideoProcessor, VideoResult
+# 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
 
-__all__ = ["VideoProcessor", "VideoResult"]
+__all__ = [
+    "ffmpeg_utils",
+    "oss_helpers",
+    "dedup_helpers",
+]

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/dedup_helpers.py

@@ -0,0 +1,126 @@
+"""查重辅助函数 — 从 generation.py 提取的 GeneratedVideo 记录 + 查重逻辑.
+
+供 render_edit_plan 和 generate_video 共同复用，
+创建 GeneratedVideo 记录后计算指纹并执行项目级 + 批次内查重。
+"""
+
+from __future__ import annotations
+
+import logging
+from uuid import uuid4
+
+from sqlalchemy.orm import Session
+
+logger = logging.getLogger(__name__)
+
+
+def create_video_record_and_dedup(
+    *,
+    generation_task_id: str,
+    project_id: str,
+    batch_id: str,
+    file_url: str,
+    file_size: int,
+    duration: float,
+    video_path: str,
+    mode: str,
+    session: Session,
+    width: int = 1280,
+    height: int = 720,
+    fps: float = 25.0,
+) -> int:
+    """创建 GeneratedVideo 记录，计算指纹并执行查重（历史 + 批次）。
+
+    Args:
+        generation_task_id: 生成任务 ID
+        project_id: 项目 ID
+        batch_id: 批次 ID（可为空字符串）
+        file_url: 视频文件 URL
+        file_size: 文件大小（字节）
+        duration: 视频时长（秒）
+        video_path: 视频本地路径（用于计算指纹）
+        mode: 剪辑模式名称
+        session: 数据库会话
+        width: 视频宽度
+        height: 视频高度
+        fps: 视频帧率
+
+    Returns:
+        创建的视频记录数量（1 表示成功，0 表示失败）
+    """
+    from video_processing.dedup import VideoDeduplicator
+
+    from packages.adapters.sqlalchemy_impl.generated_video_repository import (
+        SQLAlchemyGeneratedVideoRepository,
+    )
+    from packages.domain import GeneratedVideo
+
+    try:
+        video_id = uuid4().hex
+        generated_video = GeneratedVideo(
+            id=video_id,
+            project_id=project_id,
+            generation_task_id=generation_task_id,
+            name=f"generated-{generation_task_id[:8]}.mp4",
+            file_url=file_url,
+            file_size=file_size,
+            duration=duration,
+            width=width,
+            height=height,
+            fps=fps,
+            status="completed",
+            generation_params={"mode": mode},
+        )
+
+        video_repo = SQLAlchemyGeneratedVideoRepository(session)
+        video_repo.create(generated_video)
+
+        # 计算视频指纹
+        deduplicator = VideoDeduplicator()
+        try:
+            fingerprint = deduplicator.compute_fingerprint(video_path)
+        except Exception as fp_err:
+            logger.warning("Fingerprint computation failed for %s: %s", video_id, fp_err)
+            session.commit()
+            return 1
+
+        generated_video.video_fingerprint = fingerprint.to_dict()
+
+        # (a) 历史成片查重
+        duplicate_result = deduplicator.check_duplicate(fingerprint, project_id, session)
+
+        # (b) 批次内查重（仅当有 batch_id 时）
+        if not duplicate_result and batch_id:
+            duplicate_result = deduplicator.check_batch_duplicate(fingerprint, batch_id, video_id, session)
+
+        if duplicate_result:
+            generated_video.is_duplicate = True
+            generated_video.duplicate_of = duplicate_result["duplicate_of"]
+            logger.info(
+                "Duplicate detected: %s -> %s (reason=%s, similarity=%.3f)",
+                video_id,
+                duplicate_result["duplicate_of"],
+                duplicate_result["reason"],
+                duplicate_result["similarity"],
+            )
+        else:
+            generated_video.is_duplicate = False
+            generated_video.duplicate_of = None
+
+        video_repo.update(generated_video)
+        session.commit()
+        logger.info(
+            "GeneratedVideo record created: %s (task=%s, dup=%s)",
+            video_id,
+            generation_task_id,
+            generated_video.is_duplicate,
+        )
+        return 1
+    except Exception as e:
+        logger.error(
+            "Failed to create video record / dedup for task %s: %s",
+            generation_task_id,
+            e,
+        )
+        session.rollback()
+        return 0

apps/worker/video_processing/editing_modes.py

@@ -1,700 +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 subprocess
-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
-
-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()
-        self._ffmpeg_bin = "ffmpeg"
-        self._ffprobe_bin = "ffprobe"
-
-    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"""
-        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 execution failed: {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"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 = [
-            self._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,
-        ]
-        self._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 = [
-                self._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,
-            ]
-            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/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 = [
-                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/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 = [
-                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/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 = [
-            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/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

@@ -0,0 +1,396 @@
+"""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 UnifiedRenderService、VideoComposeService 等复用。
+"""
+
+from __future__ import annotations
+
+import logging
+import shutil
+import subprocess  # nosec B404
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+# ── 常量 ──────────────────────────────────────────────────────────────────────
+
+FFMPEG_BIN: str = shutil.which("ffmpeg") or "ffmpeg"
+FFPROBE_BIN: str = shutil.which("ffprobe") or "ffprobe"
+
+DEFAULT_OUTPUT_WIDTH = 1280
+DEFAULT_OUTPUT_HEIGHT = 720
+DEFAULT_FPS = 25
+
+# xfade 转场映射：transition_effect 名称 → FFmpeg xfade transition 名称
+# 键同时支持 TransitionEffect 枚举值和字符串名称（向后兼容）
+XFADE_TRANSITION_MAP: dict[str, str] = {
+    "fade": "fade",
+    "slideleft": "slideleft",
+    "slide_left": "slideleft",
+    "slideright": "slideright",
+    "slide_right": "slideright",
+    "slideup": "slideup",
+    "slide_up": "slideup",
+    "slidedown": "slidedown",
+    "slide_down": "slidedown",
+    "dissolve": "dissolve",
+    "wipe": "wipeleft",
+    "wipeleft": "wipeleft",
+}
+
+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 ───────────────────────────────────────────────────────────────
+
+
+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
+
+    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
+            command,
+            check=True,
+            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()
+        logger.error(
+            "FFmpeg 命令失败: exit_code=%d command=%s\nstderr:\n%s",
+            e.returncode,
+            " ".join(str(c) for c in command[:20]),  # 截断过长的命令
+            stderr_text[:5000],  # 截断过长的 stderr
+        )
+        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）。
+
+    失败时返回默认值 5.0 秒。
+    """
+    try:
+        result = subprocess.run(  # nosec B603
+            [
+                FFPROBE_BIN,
+                "-v",
+                "error",
+                "-show_entries",
+                "format=duration",
+                "-of",
+                "default=noprint_wrappers=1:nokey=1",
+                str(local_path),
+            ],
+            check=True,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True,
+        )
+        return round(float(result.stdout.strip()), 3)
+    except Exception:
+        return 5.0
+
+
+def probe_video_info(video_path: str) -> dict[str, Any]:
+    """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,
+            "video_codec": str, "audio_codec": str, "pix_fmt": str,
+            "has_audio": bool,
+        }
+        失败时返回默认值。
+    """
+    try:
+        result = subprocess.run(  # nosec B603
+            [
+                FFPROBE_BIN,
+                "-v",
+                "error",
+                "-show_entries",
+                "stream=width,height,r_frame_rate,duration,codec_name,codec_type,pix_fmt",
+                "-show_entries",
+                "format=duration",
+                "-of",
+                "json",
+                video_path,
+            ],
+            check=True,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True,
+            timeout=15,
+        )
+
+        import json
+
+        info = json.loads(result.stdout)
+        streams = info.get("streams", [])
+        fmt = info.get("format", {})
+
+        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 = 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
+        else:
+            fps = float(fps_str) if fps_str else DEFAULT_FPS
+
+        # 时长
+        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)
+        return {
+            "width": DEFAULT_OUTPUT_WIDTH,
+            "height": DEFAULT_OUTPUT_HEIGHT,
+            "duration": 0.0,
+            "fps": DEFAULT_FPS,
+            "video_codec": "",
+            "audio_codec": "",
+            "pix_fmt": "",
+            "has_audio": True,
+        }
+
+
+def normalize_video(
+    input_path: str,
+    output_path: str,
+    *,
+    width: int = DEFAULT_OUTPUT_WIDTH,
+    height: int = DEFAULT_OUTPUT_HEIGHT,
+    fps: int = DEFAULT_FPS,
+) -> dict[str, Any]:
+    """标准化视频（缩放 + 恒定帧率）。
+
+    使用 scale + pad 保持宽高比，黑边填充到目标分辨率。
+
+    Returns:
+        {"width": int, "height": int, "path": str}
+    """
+    command = [
+        FFMPEG_BIN,
+        "-y",
+        "-i",
+        input_path,
+        "-vf",
+        f"scale={width}:{height}:force_original_aspect_ratio=decrease,"
+        f"pad={width}:{height}:(ow-iw)/2:(oh-ih)/2:black,"
+        f"fps={fps}",
+        "-c:v",
+        "libx264",
+        "-crf",
+        "23",
+        "-preset",
+        "medium",
+        "-c:a",
+        "aac",
+        "-b:a",
+        "128k",
+        "-movflags",
+        "+faststart",
+        output_path,
+    ]
+    run_ffmpeg(command)
+    return {"width": width, "height": height, "path": output_path}
+
+
+# ── xfade / concat 滤镜构建 ──────────────────────────────────────────────────
+
+
+def chain_filters(filters: list[str], output_label: str, *, input_label: str = "0:v") -> str:
+    """将滤镜列表串联为 FFmpeg 滤镜字符串。
+
+    例：chain_filters(["scale=1280:720", "fps=25"], "v0")
+    → "[0:v]scale=1280:720,fps=25[v0]"
+    """
+    filter_body = ",".join(filters)
+    return f"[{input_label}]{filter_body}[{output_label}]"
+
+
+def resolve_xfade_transition(transition_name: str) -> str:
+    """将转场效果名称映射为 FFmpeg xfade transition 名称。
+
+    支持 TransitionEffect 枚举值和字符串名称，未知值回退到 "fade"。
+    """
+    # 兼容 TransitionEffect 枚举（有 .value 属性）
+    if hasattr(transition_name, "value"):
+        transition_name = transition_name.value
+    return XFADE_TRANSITION_MAP.get(transition_name, "fade")
+
+
+def build_xfade_filter_chain(
+    clip_durations: list[float],
+    clip_video_labels: list[str],
+    transitions: list[str],
+    *,
+    transition_duration: float = DEFAULT_TRANSITION_DURATION,
+    output_label: str = "outv",
+) -> tuple[str, float]:
+    """构建 xfade 转场滤镜链。
+
+    对每步 xfade 自动钳制 transition duration，确保
+    ``offset + td ≤ first_input_duration``，避免 FFmpeg exit 234。
+
+    Args:
+        clip_durations: 每个片段的时长（必须与 trim 后的实际时长一致）
+        clip_video_labels: 每个片段的视频流标签（如 "v0", "v1"）
+        transitions: 每个片段对应的转场效果（第一个片段的转场被忽略）
+        transition_duration: 转场时长（秒）
+        output_label: 最终输出标签
+
+    Returns:
+        (filter_string, estimated_total_duration)
+    """
+    n = len(clip_durations)
+    parts: list[str] = []
+
+    if n == 0:
+        return "", 0.0
+
+    if n == 1:
+        parts.append(f"[{clip_video_labels[0]}]copy[{output_label}]")
+        return ";".join(parts), clip_durations[0]
+
+    # xfade 链 — 每步动态钳制 td，防止 offset + td > first_input_duration
+    cumulative = 0.0
+    prev_label = clip_video_labels[0]
+    total_transition = 0.0  # 累计已使用的转场时长
+
+    for i in range(1, n):
+        cumulative += clip_durations[i - 1]
+
+        # 当前 xfade 的第一个输入时长
+        if i == 1:
+            first_input_dur = clip_durations[0]
+        else:
+            first_input_dur = cumulative - total_transition
+
+        # 原始 offset 计算
+        offset = max(0.0, cumulative - transition_duration * i)
+
+        # 安全钳制：offset + td 不能超过第一个输入的时长
+        available = max(0.0, first_input_dur - offset)
+        safe_td = min(transition_duration, available)
+
+        # 同时不能超过剩余总时长
+        remaining = max(0.0, sum(clip_durations) - cumulative)
+        safe_td = min(safe_td, remaining)
+        # 同时不能超过当前第二个输入（单个片段）的时长
+        safe_td = min(safe_td, clip_durations[i])
+        safe_td = max(0.001, safe_td)  # 至少 1ms，避免 td=0
+
+        transition = transitions[i] if i < len(transitions) else "cut"
+        xfade_transition = resolve_xfade_transition(transition)
+
+        if i == n - 1:
+            out_label = output_label
+        else:
+            out_label = f"xf{i}"
+
+        parts.append(
+            f"[{prev_label}][{clip_video_labels[i]}]"
+            f"xfade=transition={xfade_transition}"
+            f":duration={safe_td:.3f}"
+            f":offset={offset:.3f}"
+            f"[{out_label}]"
+        )
+        prev_label = out_label
+        total_transition += safe_td
+
+    # 总时长减去转场重叠部分
+    total_duration = sum(clip_durations) - total_transition
+    return ";".join(parts), max(0.0, total_duration)

apps/worker/video_processing/oss_helpers.py

@@ -0,0 +1,265 @@
+"""OSS 工具函数 — 从 generation.py / edit_plan_generation.py 提取的共享 OSS 操作.
+
+提供 OSS 配置读取、Bucket 创建、素材上传/下载、asset_id → 本地路径解析
+等能力，供 render_edit_plan 和 generate_video 共同复用。
+"""
+
+from __future__ import annotations
+
+import hashlib
+import logging
+import os
+import threading
+from pathlib import Path
+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 ──────────────────────────────────────────────────────────────────
+
+
+def oss_settings() -> tuple[str, str, str, str] | None:
+    """获取 OSS 配置。
+
+    Returns:
+        (access_key_id, access_key_secret, endpoint, bucket_name) 元组，
+        配置缺失时返回 None。
+    """
+    access_key_id = os.getenv("OSS_ACCESS_KEY_ID")
+    access_key_secret = os.getenv("OSS_ACCESS_KEY_SECRET")
+    endpoint = os.getenv("OSS_ENDPOINT")
+    bucket_name = os.getenv("OSS_BUCKET_NAME")
+    if not all([access_key_id, access_key_secret, endpoint, bucket_name]):
+        return None
+    return access_key_id, access_key_secret, endpoint, bucket_name
+
+
+def oss_bucket() -> oss2.Bucket | None:
+    """获取 OSS Bucket 实例。
+
+    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。
+    """
+    settings = oss_settings()
+    if settings is None:
+        return None
+    access_key_id, access_key_secret, endpoint, bucket_name = settings
+    # 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,
+        connect_timeout=OSS_CONNECT_TIMEOUT,
+    )
+
+
+def normalize_storage_key(storage_key_or_url: str) -> str:
+    """标准化存储键 — 如果是完整 URL 则提取 path 部分。
+
+    Examples:
+        "https://bucket.oss-cn-hangzhou.aliyuncs.com/path/to/file.mp4"
+        → "path/to/file.mp4"
+        "path/to/file.mp4" → "path/to/file.mp4"
+    """
+    if storage_key_or_url.startswith(("http://", "https://")):
+        return urlparse(storage_key_or_url).path.lstrip("/")
+    return storage_key_or_url.lstrip("/")
+
+
+# ── 上传 / 下载 ───────────────────────────────────────────────────────────────
+
+
+def download_asset(asset_storage_key: str, local_path: Path) -> bool:
+    """从 OSS 下载素材文件到本地路径。
+
+    Args:
+        asset_storage_key: 素材的存储键（或完整 URL）
+        local_path: 本地保存路径
+
+    Returns:
+        True 表示下载成功，False 表示失败。
+    """
+    bucket = oss_bucket()
+    if bucket is None:
+        return False
+    try:
+        bucket.get_object_to_file(normalize_storage_key(asset_storage_key), str(local_path))
+        return local_path.exists() and local_path.stat().st_size > 0
+    except Exception:
+        logger.exception("下载素材失败: %s", asset_storage_key)
+        return False
+
+
+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
+
+    Returns:
+        公开访问 URL，上传失败或 OSS 未配置时返回 None。
+    """
+    bucket = oss_bucket()
+    if bucket is None:
+        return None
+
+    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
+
+    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）。
+
+    Args:
+        storage_key_or_url: 存储键或完整 URL（URL 会自动提取 path）
+        expires_seconds: 签名有效期（秒）
+
+    Returns:
+        预签名 URL，失败或 OSS 未配置时返回 None。
+    """
+    bucket = oss_bucket()
+    if bucket is None:
+        return None
+    try:
+        storage_key = normalize_storage_key(storage_key_or_url)
+        signed = bucket.sign_url("GET", storage_key, expires_seconds)
+        logger.info("生成预签名URL: key=%s url_prefix=%s", storage_key[:80], signed[:60])
+        return signed
+    except Exception:
+        logger.exception("生成预签名URL失败: %s", storage_key_or_url[:80])
+        return None
+
+
+# ── Asset 解析 ────────────────────────────────────────────────────────────────
+
+
+def resolve_asset_path(asset_id: str, work_dir: Path) -> Path | None:
+    """从 asset_id 解析到本地文件路径。
+
+    策略（按优先级）：
+    1. 如果 asset_id 是本地绝对路径（/var/storage/...）→ 直接返回
+    2. 如果 work_dir 下已有缓存文件 → 返回缓存路径
+    3. 从 OSS 下载到 work_dir/{hash}.mp4 → 返回下载路径
+    4. 下载失败 → 返回 None
+
+    缓存策略：以 asset_id 的 SHA256 前 16 位为文件名，避免重复下载。
+    """
+    # 1. 本地绝对路径
+    if asset_id.startswith("/") and os.path.exists(asset_id):
+        return Path(asset_id)
+
+    # 2. 缓存命中
+    cache_hash = hashlib.sha256(asset_id.encode()).hexdigest()[:16]
+    cached_path = work_dir / f"{cache_hash}.mp4"
+    if cached_path.exists() and cached_path.stat().st_size > 0:
+        return cached_path
+
+    # 3. 从 OSS 下载
+    if download_asset(asset_id, cached_path):
+        return cached_path
+
+    return None
+
+
+def resolve_asset_ids_to_paths(
+    asset_ids: list[str],
+    work_dir: Path,
+) -> dict[str, Path]:
+    """批量解析 asset_id → 本地路径。
+
+    Args:
+        asset_ids: 素材 ID 列表
+        work_dir: 工作目录
+
+    Returns:
+        {asset_id: local_path} 映射，仅包含成功解析的条目。
+    """
+    result: dict[str, Path] = {}
+    for aid in asset_ids:
+        local_path = resolve_asset_path(aid, work_dir)
+        if local_path:
+            result[aid] = local_path
+    return result

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

@@ -0,0 +1,986 @@
+"""统一渲染引擎 — 输入 EditPlan + EditPlanClips，按时间线+图层渲染视频.
+
+核心原则（灵应）：渲染引擎是统一的，不判断模式，只按 clip_type/config.role
+分组为图层再合成。
+
+图层分组：
+  main (无 config.role)     → main (z=0)
+  main + config.role=b_roll → broll (z=0，与 main 同层替换)
+  overlay                   → overlay (z=1，画中画叠加)
+  background                → background (z=0，全屏底图)
+  corner_voice              → corner_voice (z=1，右上角小窗)
+  b_roll                    → broll (z=0)
+  intro / outro             → main (z=0，按 order 排在首/尾)
+
+合成流程：
+  1. 每个 clip 先 trim + scale + setpts 预处理
+  2. 同层 clips 按 order 用 xfade 串联
+  3. overlay/corner_voice 层 overlay 到主层
+  4. 如有独立音频轨，amix 混入
+"""
+
+from __future__ import annotations
+
+import logging
+import subprocess
+import time
+from dataclasses import dataclass, field
+from pathlib import Path
+from typing import Any
+
+from video_processing.ffmpeg_utils import (
+    DEFAULT_FPS,
+    DEFAULT_OUTPUT_HEIGHT,
+    DEFAULT_OUTPUT_WIDTH,
+    DEFAULT_TRANSITION_DURATION,
+    FFMPEG_BIN,
+    build_xfade_filter_chain,
+    probe_duration,
+    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__)
+
+
+# ── 数据结构 ──────────────────────────────────────────────────────────────────
+
+
+@dataclass
+class ResolvedClip:
+    """已解析到本地路径的片段。"""
+
+    clip_id: str
+    asset_id: str
+    local_path: Path
+    clip_type: str
+    order: int
+    start_time: float = 0.0
+    duration: float = 0.0  # 0 表示使用素材完整时长
+    transition_effect: str = "cut"
+    config: dict[str, Any] = field(default_factory=dict)
+
+    # 运行时填充
+    actual_duration: float = 0.0  # 素材实际时长（probe 后填充）
+
+
+@dataclass
+class RenderLayer:
+    """渲染图层。"""
+
+    role: str  # "main" | "overlay" | "pip" | "background" | "corner_voice" | "broll" | "audio"
+    clips: list[ResolvedClip] = field(default_factory=list)
+    z_index: int = 0
+    opacity: float = 1.0
+    position: tuple[int, int] | None = None  # (x, y) 偏移，None 表示全屏
+
+
+@dataclass
+class RenderResult:
+    """渲染结果。"""
+
+    output_path: Path
+    duration: float
+    file_size: int
+    width: int
+    height: int
+
+
+# ── clip_type → layer role 映射 ──────────────────────────────────────────────
+
+
+def _resolve_layer_role(clip_type: str, config: dict[str, Any]) -> str:
+    """根据 clip_type 和 config.role 确定图层角色。
+
+    映射规则：
+      intro / outro         → "main"（按 order 排在首/尾）
+      overlay               → "overlay"（画中画叠加，z=1）
+      corner_voice          → "corner_voice"（右上角小窗，z=1）
+      background            → "background"（全屏底图，z=0）
+      b_roll                → "broll"（z=0）
+      main + config.role=b_roll → "broll"
+      main (default)        → "main"
+    """
+    role = config.get("role", "")
+
+    if clip_type in ("intro", "outro"):
+        return "main"
+    if clip_type == "overlay":
+        return "overlay"
+    if clip_type == "corner_voice":
+        return "corner_voice"
+    if clip_type == "background":
+        return "background"
+    if clip_type == "b_roll":
+        return "broll"
+    # main type
+    if role == "b_roll":
+        return "broll"
+    if role == "audio":
+        return "audio"
+    return "main"
+
+
+# ── 图层默认 z_index ─────────────────────────────────────────────────────────
+
+_LAYER_Z_INDEX: dict[str, int] = {
+    "background": -1,
+    "broll": 0,
+    "main": 0,
+    "overlay": 1,
+    "corner_voice": 1,
+    "audio": 2,
+}
+
+# 图层默认 PiP 位置（相对输出画布的偏移）
+_PIP_SCALE = 0.25  # PiP 占主画面的比例
+
+
+# ── 统一渲染引擎 ─────────────────────────────────────────────────────────────
+
+
+class UnifiedRenderService:
+    """统一渲染引擎。
+
+    输入 EditPlan + EditPlanClips + 素材路径映射，按时间线+图层执行渲染。
+    """
+
+    def __init__(
+        self,
+        plan: Any,  # EditPlan
+        clips: list[Any],  # list[EditPlanClip]
+        asset_path_map: dict[str, Path],  # asset_id → local_path
+        work_dir: Path,
+        *,
+        output_width: int = DEFAULT_OUTPUT_WIDTH,
+        output_height: int = DEFAULT_OUTPUT_HEIGHT,
+        output_fps: int = DEFAULT_FPS,
+        transition_duration: float = DEFAULT_TRANSITION_DURATION,
+    ):
+        self.plan = plan
+        self.clips = clips
+        self.asset_path_map = asset_path_map
+        self.work_dir = work_dir
+        self.output_width = output_width
+        self.output_height = output_height
+        self.output_fps = output_fps
+        self.transition_duration = transition_duration
+
+    def render(self) -> 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:
+            raise ValueError("没有可渲染的片段（所有片段素材缺失或下载失败）")
+
+        # 2. 分组为 RenderLayers
+        layers = self._group_clips_into_layers(resolved)
+
+        # 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"
+
+        is_pass_through = self._can_use_pass_through(layers)
+        pass_through_has_audio = False
+        used_stream_copy = False
+
+        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,
+            file_size=file_size,
+            width=width,
+            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]:
+        """将 EditPlanClip 列表解析为 ResolvedClip 列表。
+
+        跳过 asset_id 为空或在 asset_path_map 中找不到的片段。
+        """
+        resolved: list[ResolvedClip] = []
+        for clip in self.clips:
+            asset_id = clip.asset_id
+            if not asset_id:
+                logger.warning("片段无素材: clip_id=%s", clip.id)
+                continue
+
+            local_path = self.asset_path_map.get(asset_id)
+            if local_path is None or not local_path.exists():
+                logger.warning("素材不存在: clip_id=%s asset_id=%s", clip.id, asset_id)
+                continue
+
+            # 探测实际时长
+            try:
+                actual_duration = probe_duration(local_path)
+            except Exception:
+                actual_duration = clip.duration or 5.0
+
+            rc = ResolvedClip(
+                clip_id=clip.id,
+                asset_id=asset_id,
+                local_path=local_path,
+                clip_type=clip.clip_type,
+                order=clip.order,
+                start_time=clip.start_time,
+                duration=clip.duration,
+                transition_effect=clip.transition_effect or "cut",
+                config=clip.config or {},
+                actual_duration=actual_duration,
+            )
+            resolved.append(rc)
+
+        # 按 order 排序
+        resolved.sort(key=lambda c: c.order)
+        return resolved
+
+    def _group_clips_into_layers(self, resolved_clips: list[ResolvedClip]) -> list[RenderLayer]:
+        """将 ResolvedClips 分组为 RenderLayers。
+
+        分组规则见 _resolve_layer_role 函数文档。
+        """
+        layer_map: dict[str, RenderLayer] = {}
+
+        for clip in resolved_clips:
+            role = _resolve_layer_role(clip.clip_type, clip.config)
+            if role not in layer_map:
+                z = _LAYER_Z_INDEX.get(role, 0)
+                layer_map[role] = RenderLayer(role=role, z_index=z)
+            layer_map[role].clips.append(clip)
+
+        # 每个 layer 内的 clips 按 order 排序
+        for layer in layer_map.values():
+            layer.clips.sort(key=lambda c: c.order)
+
+        # 计算 PiP 位置
+        pip_width = int(self.output_width * _PIP_SCALE)
+        margin = 20  # 边距
+
+        if "overlay" in layer_map:
+            layer_map["overlay"].position = (
+                self.output_width - pip_width - margin,
+                margin,
+            )
+        if "corner_voice" in layer_map:
+            layer_map["corner_voice"].position = (
+                self.output_width - pip_width - margin,
+                margin,
+            )
+
+        # 按 z_index 排序返回
+        layers = sorted(layer_map.values(), key=lambda lyr: lyr.z_index)
+        return layers
+
+    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
+        """
+        if not layers:
+            raise ValueError("没有可渲染的图层")
+
+        # 收集所有 clips（按图层顺序，同层按 order）
+        all_clips: list[ResolvedClip] = []
+        for layer in layers:
+            all_clips.extend(layer.clips)
+
+        # 构建输入参数
+        input_args: list[str] = []
+        clip_to_input_idx: dict[str, int] = {}
+        for i, clip in enumerate(all_clips):
+            input_args.extend(["-i", str(clip.local_path)])
+            clip_to_input_idx[clip.clip_id] = i
+
+        filter_parts: list[str] = []
+
+        # Step 1: 预处理每个 clip — scale + setpts
+        # 为每个 clip 生成预处理后的标签 [v0], [v1], ...
+        preprocessed_labels: list[str] = []
+        for i, clip in enumerate(all_clips):
+            label = f"v{i}"
+            role = _resolve_layer_role(clip.clip_type, clip.config)
+
+            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
+            effective_duration = UnifiedRenderService._clip_effective_duration(clip)
+
+            if effective_duration > 0:
+                filters.append(f"trim=duration={effective_duration}")
+                filters.append("setpts=PTS-STARTPTS")
+
+            # scale
+            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}")
+            elif role == "background":
+                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}")
+            else:
+                # 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=increase"
+                )
+                filters.append(f"crop={self.output_width}:{self.output_height}")
+
+            filters.append("setpts=PTS-STARTPTS")
+            filters.append(f"fps={self.output_fps}")
+
+            filter_str = f"[{i}:v]{','.join(filters)}[{label}]"
+            filter_parts.append(filter_str)
+            preprocessed_labels.append(label)
+
+        # Step 2: 同层 clips 用 xfade 串联
+        layer_output_labels: dict[str, str] = {}
+        for layer in layers:
+            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 = [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:
+                # 单 clip 层，直接使用预处理标签
+                layer_output_labels[layer.role] = layer_labels[0]
+            else:
+                # 多 clip 层，用 xfade 串联
+                out_label = f"{layer.role}_merged"
+                xfade_filter, _ = build_xfade_filter_chain(
+                    clip_durations=layer_durations,
+                    clip_video_labels=layer_labels,
+                    transitions=layer_transitions,
+                    transition_duration=self.transition_duration,
+                    output_label=out_label,
+                )
+                if xfade_filter:
+                    filter_parts.append(xfade_filter)
+                layer_output_labels[layer.role] = out_label
+
+        # Step 3: 合成各层
+        # 找到主层 — background 优先作为底图，其次 broll / main
+        final_video_label = None
+
+        if "background" in layer_output_labels:
+            final_video_label = layer_output_labels["background"]
+            # b_roll / main 叠加到 background 上
+            for role in ("broll", "main"):
+                if role in layer_output_labels:
+                    base_label = layer_output_labels[role]
+                    combined_label = f"combined_{role}"
+                    filter_parts.append(
+                        f"[{final_video_label}][{base_label}]" f"overlay=(W-w)/2:(H-h)/2[{combined_label}]"
+                    )
+                    final_video_label = combined_label
+        else:
+            # 无 background 时，取 broll 或 main 作为基础
+            for role in ("broll", "main"):
+                if role in layer_output_labels:
+                    final_video_label = layer_output_labels[role]
+                    break
+
+        if final_video_label is None:
+            # 没有任何主层，使用第一个层
+            final_video_label = layer_output_labels[layers[0].role]
+
+        # 叠加 overlay 层
+        for layer in layers:
+            if layer.role in ("overlay", "corner_voice"):
+                if layer.role not in layer_output_labels:
+                    continue
+                overlay_label = layer_output_labels[layer.role]
+                x, y = layer.position or (
+                    self.output_width - int(self.output_width * _PIP_SCALE) - 20,
+                    20,
+                )
+                combined_label = f"combined_{layer.role}"
+                filter_parts.append(f"[{final_video_label}][{overlay_label}]" f"overlay={x}:{y}[{combined_label}]")
+                final_video_label = combined_label
+
+        # 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
+
+    def _execute_ffmpeg(
+        self,
+        filter_complex: str,
+        input_args: list[str],
+        output_path: Path,
+    ) -> None:
+        """执行 FFmpeg 渲染命令。
+
+        失败时记录完整 filter_complex 以便排查（如 exit code 183）。
+        """
+        command = [
+            FFMPEG_BIN,
+            "-y",
+            *input_args,
+            "-filter_complex",
+            filter_complex,
+            "-map",
+            "[final_video]",
+            "-c:v",
+            "libx264",
+            "-crf",
+            "23",
+            "-preset",
+            "medium",
+            "-pix_fmt",
+            "yuv420p",
+            "-movflags",
+            "+faststart",
+            str(output_path),
+        ]
+
+        logger.info(
+            "执行渲染: plan_id=%s inputs=%d output=%s",
+            self.plan.id,
+            input_args.count("-i"),
+            output_path,
+        )
+        try:
+            run_ffmpeg(command)
+        except subprocess.CalledProcessError as e:
+            # 额外记录 filter_complex，方便排查滤镜链构建问题
+            logger.error(
+                "渲染失败: plan_id=%s exit_code=%d\nfilter_complex:\n%s",
+                self.plan.id,
+                e.returncode,
+                filter_complex[:5000],
+            )
+            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.
+
+        Returns:
+            (duration, file_size, width, height)
+        """
+        info = probe_video_info(str(output_path))
+        file_size = output_path.stat().st_size if output_path.exists() else 0
+        return (
+            info["duration"],
+            file_size,
+            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,
 )