feat: 新增仅视频流的视频拼接功能

- 在VideoUtils中新增ffmpeg_concat_medias_video_only函数，仅处理视频流忽略音频
- 新增concat_medias_video_only.py模块，提供Modal函数封装
- 在ffmpeg路由中新增/concat-video-only接口
- 解决音频流缺失导致的"Stream specifier ':a' matches no streams"错误
- 适用于不需要音频的视频拼接场景

README.md

@@ -3,7 +3,7 @@
 ## 部署
 
 ```bash
-modal deploy -m cluster.app 
+modal deploy -m cluster.app --env dev
 ```
 
 ## 本地验证fastapi部署

pyproject.toml

@@ -6,7 +6,7 @@ authors = [
 ]
 description = "Bowong modal云函数以及对于的FastAPI接口"
 readme = "src/BowongModalFunctions/readme.md"
-requires-python = ">=3.11"
+requires-python = ">=3.10"
 dependencies = [
     "fastapi[standard]>=0.115.12",
     "backoff>=2.2.1",

src/BowongModalFunctions/router/ffmpeg.py

@@ -94,6 +94,22 @@ async def concat_media(body: FFMPEGConcatRequest,
     return ModalTaskResponse(success=True, taskId=fn_call.object_id)
 
 
+@router.post("/concat-video-only",
+             summary="发起合并任务(仅视频流)",
+             description="依据AI分析的结果发起合并任务，仅处理视频流，忽略音频",
+             )
+async def concat_media_video_only(body: FFMPEGConcatRequest,
+                                  headers: Annotated[SentryTransactionHeader, Header()]) -> ModalTaskResponse:
+    medias = body.medias
+    fn = modal.Function.from_name(config.modal_app_name, "ffmpeg_concat_medias_video_only",
+                                  environment_name=config.modal_environment)
+    sentry_trace = None
+    if headers.x_trace_id and headers.x_baggage:
+        sentry_trace = SentryTransactionInfo(x_trace_id=headers.x_trace_id, x_baggage=headers.x_baggage)
+    fn_call = fn.spawn(medias, sentry_trace, webhook=body.webhook)
+    return ModalTaskResponse(success=True, taskId=fn_call.object_id)
+
+
 @router.post("/extract-audio", summary="发起音频提取任务", )
 async def extract_audio(body: FFMPEGExtractAudioRequest,
                         headers: Annotated[SentryTransactionHeader, Header()]) -> ModalTaskResponse:

src/BowongModalFunctions/utils/VideoUtils.py

@@ -913,6 +913,63 @@ class VideoUtils:
         image_metadata = VideoUtils.ffprobe_media_metadata(output_path)
         return output_path, image_metadata
 
+    @staticmethod
+    async def ffmpeg_concat_medias_video_only(media_paths: List[str],
+                                              target_width: int = 1080,
+                                              target_height: int = 1920,
+                                              output_path: Optional[str] = None) -> Tuple[str, VideoMetadata]:
+        """
+        将待处理的视频合并为一个视频（仅处理视频流，忽略音频）
+        :param media_paths: 待合并的多个视频文件路径
+        :param target_width: 输出的视频分辨率宽
+        :param target_height: 输出的视频分辨率高
+        :param output_path: 指定输出视频路径
+        :return: 最终合并结果路径，最终合并结果时长
+        """
+
+        total_videos = len(media_paths)
+        if total_videos == 0:
+            raise ValueError("没有可以合并的视频源")
+        if not output_path:
+            output_path = FileUtils.file_path_extend(media_paths[0], "concat_video_only")
+        os.makedirs(os.path.dirname(output_path), exist_ok=True)
+        ffmpeg_cmd = VideoUtils.async_ffmpeg_init()
+        filter_complex = []
+        for input_path in media_paths:
+            ffmpeg_cmd.input(input_path)
+        # 2. 统一所有视频的格式、分辨率和帧率（仅处理视频流）
+        for i in range(total_videos):
+            filter_complex.append(
+                # 先缩放到统一分辨率，然后设置帧率和格式
+                f"[{i}:v]scale={target_width}:{target_height}:force_original_aspect_ratio=decrease,"
+                f"pad={target_width}:{target_height}:(ow-iw)/2:(oh-ih)/2,"
+                f"setsar=1:1,"  # 强制设置SAR
+                f"fps=30,format=yuv420p[v{i}]"
+            )
+        
+        # 3. 准备处理后的视频流的连接字符串（仅视频流）
+        video_streams = "".join(f"[v{i}]" for i in range(total_videos))
+
+        # 4. 使用concat过滤器合并视频（仅视频流）
+        filter_complex.append(
+            f"{video_streams}concat=n={total_videos}:v=1:a=0[vconcated]"
+        )
+
+        ffmpeg_cmd.output(
+            output_path,
+            {
+                "filter_complex": ";".join(filter_complex),
+                "map": "[vconcated]",  # 仅映射视频流
+                "vcodec": "libx264",
+                "crf": 16,
+                "r": 30,
+                # 不处理音频编码相关参数
+            },
+        )
+        await ffmpeg_cmd.execute()
+        video_metadata = VideoUtils.ffprobe_media_metadata(output_path)
+        return output_path, video_metadata
+
     @staticmethod
     async def ffmpeg_stream_record_as_hls(stream_url: str,
                                           segments_output_dir: str,

src/cluster/ffmpeg_apps/concat_medias_video_only.py

@@ -0,0 +1,52 @@
+import modal
+
+from ..ffmpeg_app import ffmpeg_worker_image, app, config, s3_mount, local_copy_to_s3, output_path_prefix
+
+with ffmpeg_worker_image.imports():
+    from BowongModalFunctions.models.requests.models import MediaSourcesRequest
+    from BowongModalFunctions.models.ffmpeg_tasks.models import SentryTransactionInfo, WebhookNotify, FFMPEGResult
+    from BowongModalFunctions.utils.PathUtils import FileUtils
+    from BowongModalFunctions.utils.SentryUtils import SentryUtils
+    from BowongModalFunctions.utils.VideoUtils import VideoUtils
+    import sentry_sdk
+    from typing import Optional, Tuple
+    from modal import current_function_call_id
+
+
+    @app.function(
+        timeout=1800,
+        cloud="aws",
+        # region='ap-northeast',
+        max_containers=config.ffmpeg_worker_concurrency,
+        volumes={
+            s3_mount: modal.CloudBucketMount(
+                bucket_name=config.S3_bucket_name,
+                secret=modal.Secret.from_name("aws-s3-secret",
+                                              environment_name=config.modal_environment),
+            ),
+        }, )
+    @modal.concurrent(max_inputs=1)
+    async def ffmpeg_concat_medias_video_only(medias: MediaSourcesRequest,
+                                              sentry_trace: Optional[SentryTransactionInfo] = None,
+                                              webhook: Optional[WebhookNotify] = None) -> Tuple[
+        FFMPEGResult, Optional[SentryTransactionInfo]]:
+        fn_id = current_function_call_id()
+
+        @SentryUtils.sentry_tracker(name="视频合并任务(仅视频流)", op="ffmpeg.concat_video_only", fn_id=fn_id,
+                                    sentry_trace_id=sentry_trace.x_trace_id if sentry_trace else None,
+                                    sentry_baggage=sentry_trace.x_baggage if sentry_trace else None)
+        @SentryUtils.webhook_handler(webhook=webhook, func_id=fn_id)
+        async def ffmpeg_process(media_sources: MediaSourcesRequest, output_filepath: str) -> FFMPEGResult:
+            input_videos = [f"{s3_mount}/{media_source.cache_filepath}" for media_source in media_sources.inputs]
+            local_output_path, metadata = await VideoUtils.ffmpeg_concat_medias_video_only(media_paths=input_videos,
+                                                                                            output_path=output_filepath)
+            s3_outputs = local_copy_to_s3([local_output_path])
+            return FFMPEGResult(urn=s3_outputs[0], metadata=metadata,
+                                content_length=FileUtils.get_file_size(local_output_path), )
+
+        output_path = f"{output_path_prefix}/{config.modal_environment}/concat_video_only/outputs/{fn_id}/output.mp4"
+        result = await ffmpeg_process(media_sources=medias, output_filepath=output_path)
+        if not sentry_trace:
+            sentry_trace = SentryTransactionInfo(x_trace_id=sentry_sdk.get_traceparent(),
+                                                 x_baggage=sentry_sdk.get_baggage())
+        return result, sentry_trace