apps/worker/video_processing/unified_render_service.py

@@ -675,6 +675,27 @@ class UnifiedRenderService:
         config = self.plan.config or {}
         tts_cfg = config.get("tts", {}) or {}
 
+        # 兼容前端顶层字段：voice_id / custom_text / voice_clone_profile_id
+        # 前端一键生成页面传 config.voice_id + config.custom_text，
+        # 统一渲染引擎从 config.tts 读，这里做桥接映射。
+        if not tts_cfg.get("enabled"):
+            top_voice_id = config.get("voice_id", "") or ""
+            top_text = config.get("custom_text", "") or ""
+            if top_voice_id and top_text:
+                tts_cfg = {
+                    "enabled": True,
+                    "voice_id": top_voice_id,
+                    "text": top_text,
+                    "align_mode": "full",
+                    "overlap_mode": "replace",
+                }
+                logger.info(
+                    "[unified-render] 检测到顶层 voice_id+custom_text，桥接到 tts 配置: plan_id=%s voice_id=%s text_len=%d",
+                    self.plan.id,
+                    top_voice_id,
+                    len(top_text),
+                )
+
         tts_config = TtsConfig.parse(tts_cfg)
         if not tts_config.enabled:
             return False

tests/unit/test_unified_render_service.py

@@ -2262,3 +2262,137 @@ class TestConcatNormalizeFourItemsComplete:
         cmd = mock_run.call_args[0][0]
         assert "aac" in cmd
         assert "concat=n=3:v=0:a=1" in " ".join(cmd)
+
+
+class TestVoiceoverTopLevelConfigBridge:
+    """顶层 voice_id + custom_text 桥接到 tts 配置的兼容性测试.
+
+    前端一键生成页面传 config.voice_id + config.custom_text（顶层字段），
+    统一渲染引擎从 config.tts 读取。桥接逻辑确保两条路径都能工作。
+    """
+
+    def test_top_level_voice_id_with_text_triggers_tts(self):
+        """顶层 voice_id + custom_text 能触发 TTS 配音（桥接生效）。"""
+        from unittest.mock import MagicMock
+
+        clips = [_make_clip("c1", "main", order=0, duration=5.0)]
+        asset_paths = {"asset_c1.mp4": Path("/tmp/asset_c1.mp4")}
+        plan = FakePlan(
+            id="plan_tts_001",
+            config={
+                "voice_id": "longxiaoxia_v3",
+                "custom_text": "大家好，欢迎来到我的频道",
+            },
+        )
+        svc = UnifiedRenderService(
+            plan=plan,
+            clips=clips,
+            asset_path_map=asset_paths,
+            work_dir=Path("/tmp/test_tts"),
+        )
+
+        mock_seg = MagicMock()
+        mock_seg.audio_path = Path("/tmp/test_tts/tts/voiceover_full.wav")
+        mock_seg.start_time = 0.0
+        mock_seg.duration = 3.0
+        mock_result = MagicMock()
+        mock_result.success = True
+        mock_result.segments = [mock_seg]
+        mock_result.total_duration = 3.0
+
+        with (
+            _patch_path_exists(),
+            patch("video_processing.unified_render_service.probe_duration", return_value=5.0),
+            patch(
+                "video_processing.tts_engine.TtsEngine.generate_full_voiceover",
+                return_value=mock_result,
+            ),
+        ):
+            resolved = svc._resolve_clips()
+            layers = svc._group_clips_into_layers(resolved)
+            result = svc._maybe_add_voiceover_layer(layers, video_duration=5.0)
+
+        assert result is True, "顶层 voice_id + custom_text 应触发 TTS 配音"
+        # 应有 audio 图层
+        audio_layer = next((layer for layer in layers if layer.role == "audio"), None)
+        assert audio_layer is not None, "应添加 audio 图层"
+        assert len(audio_layer.clips) == 1, "应有 1 个配音片段"
+
+    def test_tts_config_takes_priority(self):
+        """config.tts.enabled 已配置时，以 tts 配置为准，不触发桥接。"""
+        clips = [_make_clip("c1", "main", order=0, duration=5.0)]
+        asset_paths = {"asset_c1.mp4": Path("/tmp/asset_c1.mp4")}
+        # tts.enabled=True 但 text 为空（应失败），顶层有 text
+        plan = FakePlan(
+            id="plan_tts_002",
+            config={
+                "voice_id": "longxiaoxia_v3",
+                "custom_text": "顶层文本不生效",
+                "tts": {
+                    "enabled": True,
+                    "voice_id": "longxiaochun_v3",
+                    "text": "",  # tts 配置里 text 为空
+                },
+            },
+        )
+        svc = UnifiedRenderService(
+            plan=plan,
+            clips=clips,
+            asset_path_map=asset_paths,
+            work_dir=Path("/tmp/test_tts"),
+        )
+
+        with (
+            _patch_path_exists(),
+            patch("video_processing.unified_render_service.probe_duration", return_value=5.0),
+        ):
+            resolved = svc._resolve_clips()
+            layers = svc._group_clips_into_layers(resolved)
+            result = svc._maybe_add_voiceover_layer(layers, video_duration=5.0)
+
+        # tts.enabled=True 但 text 为空 → 生成失败 → 返回 False
+        # 关键是不触发桥接（不会用顶层的 custom_text）
+        assert result is False
+
+    def test_top_level_voice_id_without_text_no_trigger(self):
+        """只有 voice_id 没有 custom_text 不触发 TTS 配音。"""
+        clips = [_make_clip("c1", "main", order=0, duration=5.0)]
+        asset_paths = {"asset_c1.mp4": Path("/tmp/asset_c1.mp4")}
+        plan = FakePlan(
+            id="plan_tts_003",
+            config={"voice_id": "longxiaoxia_v3", "custom_text": ""},
+        )
+        svc = UnifiedRenderService(
+            plan=plan,
+            clips=clips,
+            asset_path_map=asset_paths,
+            work_dir=Path("/tmp/test_tts"),
+        )
+
+        with (
+            _patch_path_exists(),
+            patch("video_processing.unified_render_service.probe_duration", return_value=5.0),
+        ):
+            resolved = svc._resolve_clips()
+            layers = svc._group_clips_into_layers(resolved)
+            result = svc._maybe_add_voiceover_layer(layers, video_duration=5.0)
+
+        assert result is False
+
+    def test_no_voice_config_no_trigger(self):
+        """没有 voice_id 也没有 tts 配置时，不触发配音。"""
+        clips = [_make_clip("c1", "main", order=0, duration=5.0)]
+        asset_paths = {"asset_c1.mp4": Path("/tmp/asset_c1.mp4")}
+        svc = _make_service(clips, asset_paths)
+
+        with (
+            _patch_path_exists(),
+            patch("video_processing.unified_render_service.probe_duration", return_value=5.0),
+        ):
+            resolved = svc._resolve_clips()
+            layers = svc._group_clips_into_layers(resolved)
+            result = svc._maybe_add_voiceover_layer(layers, video_duration=5.0)
+
+        assert result is False
+        # 没有 audio 图层
+        assert not any(layer.role == "audio" for layer in layers)