o V hG@sddlZddlZddlZddlZddlZddlmZddlmZm Z m Z m Z m Z m Z ddlZddlZddlmZddlmZddlmZz,ddlZejejjeejjjd s]ed Zzej Z Wn e!ynej"Z YnwWn ey}ed ZYnwd=d dZ#d e$fddZ%da&da'      d>de(dej)de*de(de e e(efde ej)de e*de e(de e e(efd dfddZ+ddd e*d!e*d"e(d#d$d%e e(e e e,e e,d&fe e,ffd e d'fd(d)Z-d*ej)d+e d'd,e,d-e*d ej)f d.d/Z.  0 1d?de(d2e e*efd3e e e*efd"e(d4e(d e ej)ej)e e(efff d5d6Z/ddd e$fd7d8Z0ddd e e,fd9d:Z1d@de(d"e(d e e e,e e*ffd;d<Z2dS)AN)Fraction)AnyDictListOptionalTupleUnion)_log_api_usage_once) _video_opt) _raise_video_deprecation_warning pict_typea+Your version of PyAV is too old for the necessary video operations in torchvision. If you are on Python 3.5, you will have to build from source (the conda-forge packages are not up-to-date). See https://github.com/mikeboers/PyAV#installation for instructions on how to install PyAV on your system. zPyAV is not installed, and is necessary for the video operations in torchvision. See https://github.com/mikeboers/PyAV#installation for instructions on how to install PyAV on your system. returncCstttrtdSN isinstanceav ExceptionrrH/var/www/vscode/kcb/lib/python3.10/site-packages/torchvision/io/video.py_check_av_available,s rcCs ttt Srrrrrr _av_available1s r libx264filename video_arrayfps video_codecoptions audio_array audio_fps audio_codec audio_optionsc  Csttjstjsttttj|tj dj dd}t |t r*t |}tj|dd} | j||d} |jd| _|jd| _|d krKd nd | _|pQi| _|d urd d ddddddddd } | j||d} |pni| _|jd} | dkr{dnd}| jjdjj}t | |}t|j dd|}tjj|||d}||_| |D]}| !|q| D]}| !|q|D]0}tj"j|d d}zd|_#Wnt$yddl%m&}|j'|_#Ynw| |D]}| !|qq| D]}| !|qWd d S1swYd 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)rater r libx264rgbyuv420prgb24Nzs2z%_read_from_stream..csg|]}|kr|qSrrrt)rfrrryr) _CALLED_TIMES_GC_COLLECTION_INTERVALgccollectintmathfloor time_baserFceilwarningswarntype codec_context extradatafindresearchgroupmaxseek FFmpegError enumeratedecoderssortedleninsert)r[rfrgrhr\ri should_buffermax_buffer_sizerposdo seek_offset buffer_count_idxrcresultpreceding_framesfirst_frame_ptsrrwr_read_from_streamsh              "raframes audio_frames ref_startref_endc Cs||dj|dj}}|jd}||d|}d}|} ||kr(t|||}||kr4t|||} |dd|| fS)Nrr )rsrKr) rrrrstartend total_aframesstep_per_aframes_idxe_idxrrr_align_audio_framess rrsTHWC start_ptsend_pts output_formatc Csttjstjstt|}|dvr!td|dddl m }|dkrEt j |s9td|t||||\}}}nt|durPtd }||kr^td |d |i}g} g} tj} zdtj|d d S} | jjr{| jjdj} | jjrt| |||| jjdddi} | jjdj} | durt| |d<| jjrt| |||| jjdddi} | jjdj|d<Wdn1swYWn tyYnwdd| D}dd| D}|rtt |}ntj!dtj"d}|r.t#|d}t|}|dkr&t$t%&|d| }|td kr&t$t%'|d| }t(|| ||}ntj!dtj)d}|dkrC|*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: Nrkz7end_pts should be larger than start_pts, got start_pts=z and end_pts=ignoremetadata_errorsrm video_fpsrPr!cSsg|]}|qSr)to_rgb to_ndarrayrurcrrrryqrzzread_video..cSsg|]}|qSr)rrrrrryr)rr r rnr$r rj)r rrrnr )+r r>r?r@rAr read_videoupper ValueError torchvisionrospathexists RuntimeErrorr _read_videorrFdefault_timebaserrIrOrPrrmr average_rater)rrCrGstackemptyrD concatenaterrrrrfloat32permute)rrrrhrrvframesrinfo video_framesraudio_timebaser[r vframes_list aframes_listrrrrs             rcCs*|jdjj}|dur dSd|vrdSdS)NrFsLavcT)rOrr)r[rrrr!_can_read_timestamps_from_packetss rcCs4t|rdd|jddDSdd|jddDS)NcSg|] }|jdur|jqSrrrruxrrrryz,_decode_video_timestamps..r)rmcSrrrrrrrrryr)rdemuxr)r[rrr_decode_video_timestampssrc sDttjstjsttddlm}|dkr"t ||St d}g}zCt j |dd2}|jjr\|jjd}|jzt|}WntyVtd|Ynwt|j}Wdn1sfwYWn ty}zd|d |}t|tWYd}~nd}~ww||d krfd d |D}||fS) a[DEPREACTED] List the video frames timestamps. .. warning:: DEPRECATED: All the video decoding and encoding capabilities of torchvision are deprecated from version 0.22 and will be removed in version 0.24. We recommend that you migrate to `TorchCodec `__, where we'll consolidate the future decoding/encoding capabilities of PyTorch Note that the function decodes the whole video frame-by-frame. Args: filename (str): path to the video file pts_unit (str, optional): unit in which timestamp values will be returned either 'pts' or 'sec'. Defaults to 'pts'. Returns: pts (List[int] if pts_unit = 'pts', List[Fraction] if pts_unit = 'sec'): presentation timestamps for each one of the frames in the video. video_fps (float, optional): the frame rate for the video rrrNrrz Failed decoding frames for file zFailed to open container for z; Caught error: rjcsg|]}|qSrrrvideo_time_baserrryrz)read_video_timestamps..)r r>r?r@rAr read_video_timestampsrrr _read_video_timestampsrrrIrOrmrrrrrrFrRuntimeWarningsort) rrhrrrsr[ video_streamemsgrrrrs@        r)rN)rNNNNN)rNrsrrr)3r}rrrr fractionsrtypingrrrrrrrErGr>utilsr r _video_deprecation_warningr rlogging set_levelERRORhasattrrmrcrX ImportErrorrAttributeErrorAVErrorrboolrr{r|strTensorrFrBrrrrrrrrrrrs            y$ P   y ,