a hG @sLddlZddlZddlZddlZddlZddlmZddlmZm Z m Z ddl Z ddl Z ddlmZddlmZddlmZzXddlZejejjeejjjd sed Zz ejZWneyejZYn0Wneyed ZYn0dd d dZ e!d ddZ"da#da$d-e%e j&e'e%e e(e%efe e j&e e'e e%e e(e%efdd ddZ)de'e'e%de(e%e e e*e+e*dfe,e*ffe,ddddZ-e j&e,de*e'e j&dddZ.d.e%e e'efe e e'efe%e%e+e j&e j&e(e%effd"d#d$Z/de!d%d&d'Z0de,e*d%d(d)Z1d/e%e%e+e,e*e e'fd*d+d,Z2dS)0N)Fraction)AnyOptionalUnion)_log_api_usage_once) _video_opt) _raise_video_deprecation_warning pict_typea+Your version of PyAV is too old for the necessary video operations in torchvision. If you are on Python 3.5, you will have to build from source (the conda-forge packages are not up-to-date). See https://github.com/mikeboers/PyAV#installation for instructions on how to install PyAV on your system. zPyAV is not installed, and is necessary for the video operations in torchvision. See https://github.com/mikeboers/PyAV#installation for instructions on how to install PyAV on your system. )returncCstttrtdSN isinstanceav ExceptionrrB/var/www/auris/lib/python3.9/site-packages/torchvision/io/video.py_check_av_available,s rcCs ttt Sr rrrrr _av_available1sr libx264) filename video_arrayfps video_codecoptions audio_array audio_fps audio_codec audio_optionsr c  Cs0ttjs"tjs"ttttj|tj dj dd}t |t rXt t|}tj|dd} | j||d} |jd| _|jd| _|d krd nd | _|pi| _|d urd d ddddddddd } | j||d} |pi| _|jd} | dkrdnd}| jjdjj}t| |}t|j dd|}tjj|||d}||_ | !|D]}| "|qZ| !D]}| "|qt|D]f}tj#j|d d}z d|_$Wn(t%yddl&m'}|j(|_$Yn0| !|D]}| "|qܐq| !D]}| "|qWd n1s"0Yd S)a~ [DEPRECATED] Writes a 4d tensor in [T, H, W, C] format in a video file. .. warning:: DEPRECATED: All the video decoding and encoding capabilities of torchvision are deprecated from version 0.22 and will be removed in version 0.24. We recommend that you migrate to `TorchCodec `__, where we'll consolidate the future decoding/encoding capabilities of PyTorch This function relies on PyAV (therefore, ultimately FFmpeg) to encode videos, you can get more fine-grained control by referring to the other options at your disposal within `the FFMpeg wiki `_. Args: filename (str): path where the video will be saved video_array (Tensor[T, H, W, C]): tensor containing the individual frames, as a uint8 tensor in [T, H, W, C] format fps (Number): video frames per second video_codec (str): the name of the video codec, i.e. "libx264", "h264", etc. options (Dict): dictionary containing options to be passed into the PyAV video stream. The list of options is codec-dependent and can all be found from `the FFMpeg wiki `_. audio_array (Tensor[C, N]): tensor containing the audio, where C is the number of channels and N is the number of samples audio_fps (Number): audio sample rate, typically 44100 or 48000 audio_codec (str): the name of the audio codec, i.e. "mp3", "aac", etc. audio_options (Dict): dictionary containing options to be passed into the PyAV audio stream. The list of options is codec-dependent and can all be found from `the FFMpeg wiki `_. Examples:: >>> # Creating libx264 video with CRF 17, for visually lossless footage: >>> >>> from torchvision.io import write_video >>> # 1000 frames of 100x100, 3-channel image. >>> vid = torch.randn(1000, 100, 100, 3, dtype = torch.uint8) >>> write_video("video.mp4", options = {"crf": "17"}) dtypeT)forcew)mode)raterrZ libx264rgbZyuv420pZrgb24Nzz%_read_from_stream..csg|]}|kr|qSrrrS)rGrrrXrY) _CALLED_TIMES_GC_COLLECTION_INTERVALgcZcollectr5mathfloor time_baser4ceilwarningswarntype codec_context extradatafindresearchgroupmaxseek FFmpegError enumeratedecoderRsortedleninsert)rBrGrHrIrCrJZ should_buffermax_buffer_sizereposdoZ seek_offsetZ buffer_countZ_idxrDresultZpreceding_framesZfirst_frame_ptsrrVr_read_from_streams^            "rw)aframes audio_frames ref_startref_endr c Cs||dj|dj}}|jd}||d|}d}|} ||krPt|||}||krht|||} |dd|| fS)Nrr)rRr9r5) rxryrzr{startendZ total_aframesZstep_per_aframeZs_idxZe_idxrrr_align_audio_framess rrRTHWC)r start_ptsend_ptsrI output_formatr c Csttjs"tjs"tt|}|dvrBtd|dddl m }|dkrt j |srtd|t||||\}}}nt|durtd }||krtd |d |i}g} g} tj} ztj|d d } | jjr| jjdj} | jjrDt| |||| jjdddi} | jjdj} | durDt| |d<| jjr~t| |||| jjdddi} | jjdj|d<Wdn1s0YWntyYn0dd| D}dd| D}|rtt |}ntj!dtj"d}|rjt#|d}t|}|dkrZt$t%&|d| }|td krZt$t%'|d| }t(|| ||}ntj!dtj)d}|dkr|*dddd}|||fS)a[DEPRECATED] Reads a video from a file, returning both the video frames and the audio frames .. warning:: DEPRECATED: All the video decoding and encoding capabilities of torchvision are deprecated from version 0.22 and will be removed in version 0.24. We recommend that you migrate to `TorchCodec `__, where we'll consolidate the future decoding/encoding capabilities of PyTorch Args: filename (str): path to the video file. If using the pyav backend, this can be whatever ``av.open`` accepts. start_pts (int if pts_unit = 'pts', float / Fraction if pts_unit = 'sec', optional): The start presentation time of the video end_pts (int if pts_unit = 'pts', float / Fraction if pts_unit = 'sec', optional): The end presentation time pts_unit (str, optional): unit in which start_pts and end_pts values will be interpreted, either 'pts' or 'sec'. Defaults to 'pts'. output_format (str, optional): The format of the output video tensors. Can be either "THWC" (default) or "TCHW". Returns: vframes (Tensor[T, H, W, C] or Tensor[T, C, H, W]): the `T` video frames aframes (Tensor[K, L]): the audio frames, where `K` is the number of channels and `L` is the number of points info (Dict): metadata for the video and audio. Can contain the fields video_fps (float) and audio_fps (int) )rTCHWz5output_format should be either 'THWC' or 'TCHW', got .rget_video_backendpyavzFile not found: NrLz7end_pts should be larger than start_pts, got start_pts=z and end_pts=ignoreZmetadata_errorsrN video_fpsr=rcSsg|]}|qSr)Zto_rgb to_ndarrayrTrDrrrrXqrYzread_video..cSsg|] }|qSr)rrrrrrXrrY)rrrrOr!rrK)rrrrOr)+r r,r-r.r/r read_videoupper ValueError torchvisionrospathexists RuntimeErrorr Z _read_videorr4Zdefault_timebaserr8r<r=r_rNrw average_rater&rlr1r6stackemptyr2Z concatenater5r]r^r`rfloat32Zpermute)rrrrIrrZvframesrxinfoZ video_framesryZaudio_timebaserBrZ vframes_listZ aframes_listrrrrs          6    r)rBr cCs*|jdjj}|durdSd|vr&dSdS)NrFsLavcT)r<rdre)rBrerrr!_can_read_timestamps_from_packetss rcCs8t|rdd|jddDSdd|jddDSdS)NcSsg|]}|jdur|jqSr rQrTxrrrrXrYz,_decode_video_timestamps..r)rNcSsg|]}|jdur|jqSr rQrrrrrXrY)rZdemuxrn)rBrrr_decode_video_timestampssr)rrIr c sHttjs"tjs"ttddlm}|dkrDt ||St d}g}zt j |ddd}|jjr|jjd}|jz t|}Wn"tytd|Yn0t|j}Wdn1s0YWnBty}z(d|d |}t|tWYd}~n d}~00||d kr@fd d |D}||fS) a[DEPREACTED] List the video frames timestamps. .. warning:: DEPRECATED: All the video decoding and encoding capabilities of torchvision are deprecated from version 0.22 and will be removed in version 0.24. We recommend that you migrate to `TorchCodec `__, where we'll consolidate the future decoding/encoding capabilities of PyTorch Note that the function decodes the whole video frame-by-frame. Args: filename (str): path to the video file pts_unit (str, optional): unit in which timestamp values will be returned either 'pts' or 'sec'. Defaults to 'pts'. Returns: pts (List[int] if pts_unit = 'pts', List[Fraction] if pts_unit = 'sec'): presentation timestamps for each one of the frames in the video. video_fps (float, optional): the frame rate for the video rrrNrrz Failed decoding frames for file zFailed to open container for z; Caught error: rKcsg|] }|qSrrrZvideo_time_baserrrXrYz)read_video_timestamps..)r r,r-r.r/rread_video_timestampsrrr Z_read_video_timestampsrrr8r<rNr_rrlrarbr4rRuntimeWarningsort) rrIrrrRrBZ video_streamemsgrrrrs4      ," r)rNNNNN)rNrRr)rR)3r\r]rrgraZ fractionsrtypingrrrr3r6r,utilsrr Z_video_deprecation_warningr rloggingZ set_levelERRORhasattrrNrDr@ ImportErrorrlAttributeErrorZAVErrorrboolrrZr[strZTensorr4dictr0r5tuplelistrwrrrrrrrrrs          z$ Q   y