优化google/upload接口的速度

* 优化google/upload接口的速度
* 切割视频接口新增输出质量选项

---------

Merge request URL: https://g-ldyi2063.coding.net/p/dev/d/modalDeploy/git/merge/4790?initial=true
Co-authored-by: shuohigh@gmail.com

src/BowongModalFunctions/config.py

@@ -22,4 +22,6 @@ class WorkerConfig(BaseSettings):
     comfyui_s3_input: Optional[str] = Field(default="comfyui-input", description="ComfyUI input S3文件夹名")
     comfyui_s3_output: Optional[str] = Field(default="comfyui-output", description="ComfyUI output S3文件夹名")
 
-    modal_config: Any = SettingsConfigDict()
+    modal_config: Any = SettingsConfigDict(json_schema_extra={
+        "description": "可通过本地环境变量加载对应Field, 不区分大小写, Modal创建App的Image时可通过dotenv加载指定.env文件写入到Docker Image的系统变量",
+    })

src/BowongModalFunctions/models/web_model.py

@@ -7,7 +7,7 @@ from pydantic import BaseModel, Field, field_validator, ConfigDict, HttpUrl, com
 from .ffmpeg_worker_model import FFMpegSliceSegment
 from .media_model import MediaSource, MediaSources, MediaProtocol
 from ..config import WorkerConfig
-from ..utils.VideoUtils import VideoMetadata
+from ..utils.VideoUtils import VideoMetadata, FFMPEGSliceOptions
 
 config = WorkerConfig()
 
@@ -124,6 +124,7 @@ class FFMPEGConvertStreamResponse(BaseFFMPEGTaskStatusResponse):
 class FFMPEGSliceRequest(BaseFFMPEGTaskRequest):
     media: MediaSource = Field(description="待切割的媒体源")
     markers: List[FFMpegSliceSegment] = Field(description="按照时间顺序排序过的切割标记数组")
+    options: FFMPEGSliceOptions = Field(default_factory=FFMPEGSliceOptions, description="切割质量选项")
 
     @field_validator('media', mode='before')
     @classmethod
@@ -428,14 +429,15 @@ class FFMPEGStreamRecordRequest(BaseFFMPEGTaskRequest):
     stream_source: str = Field(description="直播源地址")
     segment_duration: int = Field(default=5, description="hls片段时长(秒)")
     recording_timeout: int = Field(default=300, description="hls流无内容后等待的时长(秒)")
-    monitor_timeout: int = Field(default=36000, description="录制监控最大时长(秒), 默认为10小时, 不可大于12小时", le=43200)
+    monitor_timeout: int = Field(default=36000, description="录制监控最大时长(秒), 默认为10小时, 不可大于12小时",
+                                 le=43200)
 
 
 class GeminiRequest(BaseFFMPEGTaskRequest):
     media_hls_url: MediaSource = Field(default="", description="视频流录制HLS地址 hls://格式 需录制超过20分钟")
     product_list: List[Union[str, dict]] = Field(description="商品名列表(时间倒序)"),
     start_time: str = Field(default="00:00:00.000", description="开始时间(hls)")
-    end_time:str = Field(default="00:20:00.000", description="结束时间(hls)")
+    end_time: str = Field(default="00:20:00.000", description="结束时间(hls)")
 
     @field_validator('media_hls_url', mode='before')
     @classmethod
@@ -451,9 +453,11 @@ class GeminiRequest(BaseFFMPEGTaskRequest):
         else:
             raise pydantic.ValidationError("media格式读取失败")
 
+
 class GeminiResultResponse(BaseFFMPEGTaskStatusResponse):
     result: str = Field(default="", description="推理出的json")
 
+
 class LiveProduct(BaseModel):
     title: str = Field(default="", description="商品标题")
     leaf_category: str = Field(default="", description="商品分类")
@@ -462,19 +466,23 @@ class LiveProduct(BaseModel):
     cover: str = Field(default="", description="商品封面图链接")
     detail_url: str = Field(default="", description="商品详情链接")
 
+
 class LiveProductCaches(BaseModel):
     room_id: str = Field(default="", description="直播间room_id/Room room_id")
     author_id: str = Field(default="", description="作者id/Author id")
     update_time: str = Field(default=0, description="商品列表更新时间")
     count: int = Field(default=0, description="缓存直播间商品数量")
-    product_list: List[LiveProduct] = Field(default=None,description="缓存直播间商品列表")
+    product_list: List[LiveProduct] = Field(default=None, description="缓存直播间商品列表")
+
 
 class MonitorLiveRoomProductRequest(BaseModel):
     cookie: str = Field(default="YOUR_COOKIE", description="用户网页版抖音Cookie/Your web version of Douyin Cookie")
     room_id: str = Field(default="", description="直播间room_id/Room room_id")
     author_id: str = Field(default="", description="作者id/Author id")
 
+
 class LiveRoomProductCachesResponse(BaseModel):
-    status: int = Field(default=None, description="缓存状态/0-正常返回 1-直播已结束 2-IP风控 3-请求Tikhub API错误 4-内部错误")
+    status: int = Field(default=None,
+                        description="缓存状态/0-正常返回 1-直播已结束 2-IP风控 3-请求Tikhub API错误 4-内部错误")
     message: str = Field(default="", description="错误信息")
-    cache_json: str = Field(default="", description="缓存内容/Json文本")
+    cache_json: str = Field(default="", description="缓存内容/Json文本")

src/BowongModalFunctions/router/ffmpeg.py

@@ -1,11 +1,10 @@
 from typing import Annotated, cast
 
 import modal
-from fastapi import APIRouter, Depends, Header, Body
+from fastapi import APIRouter, Depends, Header
-from fastapi.responses import Response, JSONResponse
+from fastapi.responses import Response
 from loguru import logger
 from starlette import status
-import m3u8
 
 from ..config import WorkerConfig
 from ..middleware.authorization import verify_token
@@ -51,8 +50,7 @@ sentry_header_schema = {
         "description": "",
         "headers": sentry_header_schema
     },
-},
+}, )
-            )
 async def get_task_status(task_id: str, response: Response) -> BaseFFMPEGTaskStatusResponse:
     logger.info(f"Get task [{task_id}]")
     task_info = await ModalUtils.get_modal_task_status(task_id)
@@ -72,10 +70,10 @@ async def slice_media(request: FFMPEGSliceRequest,
     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(media=request.media, markers=request.markers, sentry_trace=sentry_trace, webhook=request.webhook)
+    fn_call = fn.spawn(media=request.media, markers=request.markers, options=request.options,
+                       sentry_trace=sentry_trace, webhook=request.webhook)
     return ModalTaskResponse(success=True, taskId=fn_call.object_id)
 
-
 @router.post("/convert", summary="发起直播流转换任务", )
 async def convert_media(request: FFMPEGConvertStreamRequest,
                         headers: Annotated[SentryTransactionHeader, Header()]) -> ModalTaskResponse:

src/BowongModalFunctions/router/google.py

@@ -1,6 +1,6 @@
 import json
 import os
-from typing import Annotated, Optional, cast
+from typing import Annotated, Optional, cast, List
 
 import modal
 import sentry_sdk
@@ -25,6 +25,7 @@ router = APIRouter(prefix="/google", tags=["Google"])
 class GoogleAPIKeyHeaders(BaseModel):
     x_google_api_key: Optional[str] = Field(description="Google API Key")
 
+
 class BundleHeaders(BaseModel):
     x_google_api_key: Optional[str] = Field(description="Google API Key")
     x_trace_id: str = Field(description="Sentry Transaction ID", default=None)
@@ -41,8 +42,12 @@ async def upload_file_multipart(file: UploadFile,
         raise HTTPException(status_code=status.HTTP_401_UNAUTHORIZED, detail="Missing Google API Key")
     content_length = file.size
     content_type = file.content_type
-    if content_type not in ['video/mp4', 'video/mpeg', 'video/mov', 'video/avi', 'video/x-flv', 'video/mpg',
+    # 允许上传的视频
-                            'video/webm', 'video/wmv', 'video/3gpp']:
+    google_content_type_allow_list = ['video/mp4', 'video/mpeg', 'video/mov', 'video/avi', 'video/x-flv', 'video/mpg',
+                                      'video/webm', 'video/wmv', 'video/3gpp', ]
+    # 允许上传的图片
+    google_content_type_allow_list.extend(["image/png", "image/jpeg", "image/webp", "image/heic", "image/heif"])
+    if content_type not in google_content_type_allow_list:
         raise HTTPException(status_code=status.HTTP_415_UNSUPPORTED_MEDIA_TYPE)
     logger.info(f"Uploading name = {file.filename}, size = {content_length}, type = {content_type} to google file")
     with httpx.Client(timeout=1800) as client:
@@ -63,7 +68,7 @@ async def upload_file_multipart(file: UploadFile,
 
         upload_url = pre_upload_response.headers.get("X-Goog-Upload-Url")
 
-        upload_response = client.post(url=upload_url, content=file.file.read(), headers={
+        upload_response = client.post(url=upload_url, content=file.file, headers={
             "X-Goog-Upload-Offset": "0",
             "X-Goog-Upload-Command": "upload, finalize",
             "Content-Type": content_type
@@ -97,6 +102,7 @@ async def delete_file(filename: str, headers: Annotated[GoogleAPIKeyHeaders, Hea
         response.raise_for_status()
     return JSONResponse(content=response.json(), status_code=response.status_code)
 
+
 @router.delete('/delete_all', summary="删除所有已上传的文件/第一页")
 async def delete_all(headers: Annotated[GoogleAPIKeyHeaders, Header()]):
     google_api_key = headers.x_google_api_key or os.environ.get("GOOGLE_API_KEY")
@@ -116,6 +122,7 @@ async def delete_all(headers: Annotated[GoogleAPIKeyHeaders, Header()]):
                 return JSONResponse(content={"msg": f"删除文件{filename}失败"}, status_code=resp.status_code)
     return JSONResponse(content=response.json(), status_code=response.status_code)
 
+
 @router.get('/list', summary="列出已上传的文件")
 async def list_files(headers: Annotated[GoogleAPIKeyHeaders, Header()]):
     google_api_key = headers.x_google_api_key or os.environ.get("GOOGLE_API_KEY")
@@ -128,20 +135,25 @@ async def list_files(headers: Annotated[GoogleAPIKeyHeaders, Header()]):
         return JSONResponse(content={}, status_code=response.status_code)
     return JSONResponse(content=response.json() if len(response.text) > 0 else "", status_code=response.status_code)
 
+
 @router.post('/inference_gemini', summary="使用Gemini推理hls视频流指定时间段的打点情况")
 async def inference_gemini(
-                    data:GeminiRequest,
+        data: GeminiRequest,
-                    headers: Annotated[BundleHeaders, Header()],
+        headers: Annotated[BundleHeaders, Header()],
-                    ) -> ModalTaskResponse:
+) -> ModalTaskResponse:
     google_api_key = headers.x_google_api_key or os.environ.get("GOOGLE_API_KEY")
     if not google_api_key:
         raise HTTPException(status_code=status.HTTP_401_UNAUTHORIZED, detail="Missing Google API Key")
-    fn = modal.Function.from_name(config.modal_app_name,"video_hls_slice_inference", environment_name=config.modal_environment)
+    fn = modal.Function.from_name(config.modal_app_name, "video_hls_slice_inference",
+                                  environment_name=config.modal_environment)
     fn_call = fn.spawn(data.media_hls_url, google_api_key, data.product_list, data.start_time, data.end_time,
-                     SentryTransactionInfo(x_trace_id=sentry_sdk.get_traceparent(), x_baggage=sentry_sdk.get_baggage())
+                       SentryTransactionInfo(x_trace_id=sentry_sdk.get_traceparent(),
-                     if headers.x_trace_id is None else SentryTransactionInfo(x_trace_id=headers.x_trace_id, x_baggage=headers.x_baggage))
+                                             x_baggage=sentry_sdk.get_baggage())
+                       if headers.x_trace_id is None else SentryTransactionInfo(x_trace_id=headers.x_trace_id,
+                                                                                x_baggage=headers.x_baggage))
     return ModalTaskResponse(success=True, taskId=fn_call.object_id)
 
+
 @router.get("/inference_gemini/{task_id}", summary="查询Gemini推理hls视频流指定时间段的打点任务")
 async def gemini_status(task_id: str, response: Response) -> GeminiResultResponse:
     task_info = await ModalUtils.get_modal_task_status(task_id)
@@ -150,10 +162,11 @@ async def gemini_status(task_id: str, response: Response) -> GeminiResultRespons
         response.headers["x-baggage"] = task_info.transaction.x_baggage
     try:
         return GeminiResultResponse(taskId=task_id, status=task_info.status, code=cast(int, task_info.error_code.value),
-                                    error=task_info.error_reason, result=json.dumps(task_info.results, ensure_ascii=False)
+                                    error=task_info.error_reason,
+                                    result=json.dumps(task_info.results, ensure_ascii=False)
                                     if task_info.results is not None else "")
     except Exception as e:
         logger.exception(f"获取Gemini状态发生错误 {e}")
         return GeminiResultResponse(taskId=task_id, status=task_info.status, code=cast(int, task_info.error_code.value),
                                     error=task_info.error_reason + f"获取Gemini状态发生错误 {e}"
-                                    if task_info.error_reason else f"获取Gemini状态发生错误 {e}", result="")
+                                    if task_info.error_reason else f"获取Gemini状态发生错误 {e}", result="")

src/BowongModalFunctions/utils/VideoUtils.py

@@ -1,5 +1,4 @@
 import asyncio
-import re
 import shutil
 import tempfile
 from datetime import datetime, timedelta
@@ -7,15 +6,12 @@ from datetime import datetime, timedelta
 import aiofiles
 import aiohttp
 from typing import Union, List, Tuple, Optional, Dict, Any
-from typing import Union, List, Tuple, Optional, Dict, Any
 
 import m3u8
 import numpy as np
 import json, os
 import math
-import retry
 from aiohttp import ClientTimeout
-from urllib3 import Retry
 
 from .TimeUtils import TimeDelta
 from pydantic import BaseModel, ConfigDict, computed_field, Field, field_validator
@@ -39,7 +35,25 @@ from pedalboard.io import AudioFile
 
 from loguru import logger
 from .PathUtils import FileUtils
-from BowongModalFunctions.models.ffmpeg_worker_model import FFMpegSliceSegment
+from ..models.ffmpeg_worker_model import FFMpegSliceSegment
+
+
+class FFMPEGSliceOptions(BaseModel):
+    limit_size: Optional[int] = Field(default=None, description="不超过指定文件(字节)大小, 默认为空不限制输出大小")
+    crf: int = Field(default=16, description="输出视频的质量")
+    fps: int = Field(default=30, description="输出视频的FPS")
+
+    @computed_field(description="解析为字符表达式的文件大小")
+    @property
+    def pretty_limit_size(self) -> Optional[str]:
+        if self.limit_size is not None:
+            for unit in ['B', 'KB', 'MB', 'GB', 'TB']:
+                if self.limit_size < 1024.0:
+                    return f"{self.limit_size:.2f} {unit}"
+                self.limit_size /= 1024.0
+            return f"{self.limit_size:.2f} PB"
+        else:
+            return None
 
 
 class MediaStream(BaseModel):
@@ -408,11 +422,13 @@ class VideoUtils:
 
     @staticmethod
     async def ffmpeg_slice_media(media_path: str, media_markers: List[FFMpegSliceSegment],
+                                 options: FFMPEGSliceOptions,
                                  output_path: Optional[str] = None) -> List[Tuple[str, VideoMetadata]]:
         """
         使用本地视频文件按时间段切割出分段视频
         :param media_path: 本地视频路径
         :param media_markers: 分段起始结束时间标记
+        :param options: 输出切割质量选项
         :param output_path: 最终输出文件路径, 片段会根据指定路径附加_1.mp4， _2.mp4等片段编号
         :return: 输出片段的本地路径
         """
@@ -449,17 +465,20 @@ class VideoUtils:
                     raise ValueError(
                         f"第{i}个切割点结束点{marker.end.total_seconds()}s超出视频时长[0-{metadata.format.duration}s]范围")
             segment_output_path = FileUtils.file_path_extend(output_path, str(i))
-            ffmpeg_cmd.output(segment_output_path,
+            ffmpeg_options = {
-                              map=[f"[cut{i}]", f"[acut{i}]"],
+                "map": [f"[cut{i}]", f"[acut{i}]"],
-                              reset_timestamps="1",
+                "reset_timestamps": "1",
-                              sc_threshold="0",
+                "sc_threshold": "0",
-                              g="1",
+                "g": "1",
-                              force_key_frames="expr:gte(t,n_forced*1)",
+                "force_key_frames": "expr:gte(t,n_forced*1)",
-                              vcodec="libx264",
+                "vcodec": "libx264",
-                              acodec="aac",
+                "acodec": "aac",
-                              crf=16,
+                "crf": options.crf,
-                              r=30,
+                "r": options.fps
-                              )
+            }
+            if options.limit_size:
+                ffmpeg_options["fs"] = options.pretty_limit_size
+            ffmpeg_cmd.output(segment_output_path, options=ffmpeg_options)
             temp_outputs.append(segment_output_path)
 
         await ffmpeg_cmd.execute()

src/cluster/ffmpeg_app.py

@@ -1,13 +1,5 @@
-import asyncio
-from pathlib import Path
-
-import botocore
-import httpx
 import modal
 from dotenv import dotenv_values
-from watchdog.events import DirMovedEvent
-
-from BowongModalFunctions.utils.ModalUtils import ModalUtils
 
 ffmpeg_worker_image = (
     modal.Image.debian_slim(python_version="3.11")
@@ -28,15 +20,18 @@ with ffmpeg_worker_image.imports():
     from typing import List, Optional, Tuple, Dict, Any, Union, Set
     from loguru import logger
     from modal import current_function_call_id
-    from ffmpeg.asyncio import FFmpeg
     from BowongModalFunctions.utils.SentryUtils import SentryUtils
     from BowongModalFunctions.utils.PathUtils import FileUtils
     from BowongModalFunctions.utils.VideoUtils import VideoUtils
+    from BowongModalFunctions.utils.ModalUtils import ModalUtils
     from BowongModalFunctions.models.ffmpeg_worker_model import FFMpegSliceSegment
     from BowongModalFunctions.models.media_model import MediaSources, MediaSource, MediaProtocol
     from BowongModalFunctions.models.web_model import SentryTransactionInfo, WebhookNotify, FFMPEGResult, \
-        WebhookMethodEnum, BaseFFMPEGTaskStatusResponse, TaskStatus
+        WebhookMethodEnum, BaseFFMPEGTaskStatusResponse, TaskStatus, FFMPEGSliceOptions
     from BowongModalFunctions.config import WorkerConfig
+    import httpx
+    from pathlib import Path
+    import asyncio
 
     config = WorkerConfig()
 
@@ -124,6 +119,7 @@ with ffmpeg_worker_image.imports():
     @modal.concurrent(max_inputs=1)
     async def ffmpeg_slice_media(media: MediaSource,
                                  markers: List[FFMpegSliceSegment],
+                                 options: FFMPEGSliceOptions,
                                  sentry_trace: Optional[SentryTransactionInfo] = None,
                                  webhook: Optional[WebhookNotify] = None) -> Tuple[
         List[FFMPEGResult], Optional[SentryTransactionInfo]]:
@@ -134,13 +130,15 @@ with ffmpeg_worker_image.imports():
                                     sentry_baggage=sentry_trace.x_baggage if sentry_trace else None)
         @SentryUtils.webhook_handler(webhook=webhook, func_id=fn_id)
         async def ffmpeg_slice_process(media_source: MediaSource, media_markers: List[FFMpegSliceSegment],
-                                       fn_id: str) -> List[FFMPEGResult]:
+                                       fn_id: str, options: Optional[FFMPEGSliceOptions] = None,
+                                       ) -> List[FFMPEGResult]:
             cache_filepath = f"{s3_mount}/{media_source.cache_filepath}"
             logger.info(f"从{media_source.urn}切割")
             for i, marker in enumerate(media_markers):
                 logger.info(f"[{i}] {marker.start.toFormatStr()} --> {marker.end.toFormatStr()}")
             segments = await VideoUtils.ffmpeg_slice_media(media_path=cache_filepath,
                                                            media_markers=media_markers,
+                                                           options=options,
                                                            output_path=f"{output_path_prefix}/{config.modal_environment}/slice/outputs/{fn_id}/output.mp4")
             return [FFMPEGResult(urn=local_copy_to_s3([segment[0]])[0], metadata=segment[1],
                                  content_length=FileUtils.get_file_size(segment[0])) for segment in segments]
@@ -149,9 +147,9 @@ with ffmpeg_worker_image.imports():
                                     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_hls_slice_process(media_source: MediaSource,
+        async def ffmpeg_hls_slice_process(media_source: MediaSource, media_markers: List[FFMpegSliceSegment],
-                                           media_markers: List[FFMpegSliceSegment],
+                                           fn_id: str,
-                                           fn_id: str) -> List[FFMPEGResult]:
+                                           options: Optional[FFMPEGSliceOptions] = None, ) -> List[FFMPEGResult]:
             hls_m3u8_url = media_source.path
             segments = await VideoUtils.ffmpeg_slice_stream_media_multithread(media_path=hls_m3u8_url,
                                                                               media_markers=media_markers,
@@ -594,7 +592,6 @@ with ffmpeg_worker_image.imports():
                                              )
             self.s3_root_output_dir = s3_mount_output_dir.replace(s3_mount, "")
 
-
         def on_created(self, event: FileCreatedEvent) -> None:
             logger.info(f"[created] {event.src_path}")
             if event.src_path.endswith(".ts"):