a h/Q@sddlZddlZddlmZddlmZmZddlZddlm Z ddl m Z ze dd Z Wne efyvd Z Yn0e ZeddZGd d d ZGd ddZeeefddddZejejejejejejedddZejejeeefejdddZdd ddddded dddef eeeeeeeeeefeeeeeeefeeejejefdddZeeeeeeefdddZeedd d!Zd1ejeeeeeeeeefeeeeeeeefeeeejejfd"d#d$Z ejeeeeeefd%d&d'Z!ejed%d(d)Z"d2eeeefeeeefeeejeje#eeffd+d,d-Z$d3eeeeeeeefeefd.d/d0Z%dS)4N)Fraction)OptionalUnion) _load_library) _raise_video_deprecation_warning video_readerTFc@s0eZdZeedZddgZeeddddZdS)Timebase numerator denominatorr r N)r r returncCs||_||_dS)Nr )selfr r rG/var/www/auris/lib/python3.9/site-packages/torchvision/io/_video_opt.py__init__szTimebase.__init__)__name__ __module__ __qualname__int__annotations__ __slots__rrrrrr s  r c @s8eZdZeeeeeeeedZgdZddddZdS) VideoMetaData) has_videovideo_timebasevideo_duration video_fps has_audioaudio_timebaseaudio_durationaudio_sample_rateN)rcCs@d|_tdd|_d|_d|_d|_tdd|_d|_d|_dS)NFrrg) rr rrrrrr r!)rrrrr;s  zVideoMetaData.__init__) rrrboolr floatrrrrrrrr%s  r) pts_rangercCs@|d|dkrdkrr?r@rArBrrCrDrEr3rresultvframes _vframe_ptsr'r(r)r1r2r*r+r,inforrr_read_video_from_file}s:@  rM)r;rc Cs~tjj|dddddddddddddddddd}|\ }}}}}}}} } } t|||| | | } |}|}||| fS)z Decode all video- and audio frames in the video. Only pts (presentation timestamp) is returned. The actual frame pixel data is not copied. Thus, it is much faster than read_video(...) rrr4)rFrGr rHr0numpytolist) r;rI_vframes vframe_ptsr'r(r)_aframesr2r*r+r,rLrrr _read_video_timestamps_from_files4  rSc Cs:ttjj|}|\}}}}}}t||||||}|S)zO Probe a video file and return VideoMetaData with info about the video )rrFrGr Zprobe_video_from_filer0) r;rIr'r(r)r*r+r,rLrrr_probe_video_from_files rT) video_datar<r=r>r?r@rArBvideo_timebase_numeratorvideo_timebase_denominatorrCrDrEr3audio_timebase_numeratoraudio_timebase_denominatorrcCstt|t| t|tjsht.tjdddtj|tj d}Wdn1s^0Ytj j ||d||||||d|d|| | | | | d| d||}|\ }}}}}}}}}}| dkrt||| }||fS)a Reads a video from memory, returning both the video frames as the audio frames This function is torchscriptable. Args: video_data (data type could be 1) torch.Tensor, dtype=torch.int8 or 2) python bytes): compressed video content stored in either 1) torch.Tensor 2) python bytes seek_frame_margin (double, optional): seeking frame in the stream is imprecise. Thus, when video_start_pts is specified, we seek the pts earlier by seek_frame_margin seconds read_video_stream (int, optional): whether read video stream. If yes, set to 1. Otherwise, 0 video_width/video_height/video_min_dimension/video_max_dimension (int): together decide the size of decoded frames: - When video_width = 0, video_height = 0, video_min_dimension = 0, and video_max_dimension = 0, keep the original frame resolution - When video_width = 0, video_height = 0, video_min_dimension != 0, and video_max_dimension = 0, keep the aspect ratio and resize the frame so that shorter edge size is video_min_dimension - When video_width = 0, video_height = 0, video_min_dimension = 0, and video_max_dimension != 0, keep the aspect ratio and resize the frame so that longer edge size is video_max_dimension - When video_width = 0, video_height = 0, video_min_dimension != 0, and video_max_dimension != 0, resize the frame so that shorter edge size is video_min_dimension, and longer edge size is video_max_dimension. The aspect ratio may not be preserved - When video_width = 0, video_height != 0, video_min_dimension = 0, and video_max_dimension = 0, keep the aspect ratio and resize the frame so that frame video_height is $video_height - When video_width != 0, video_height == 0, video_min_dimension = 0, and video_max_dimension = 0, keep the aspect ratio and resize the frame so that frame video_width is $video_width - When video_width != 0, video_height != 0, video_min_dimension = 0, and video_max_dimension = 0, resize the frame so that frame video_width and video_height are set to $video_width and $video_height, respectively video_pts_range (list(int), optional): the start and end presentation timestamp of video stream video_timebase_numerator / video_timebase_denominator (float, optional): a rational number which denotes timebase in video stream read_audio_stream (int, optional): whether read audio stream. If yes, set to 1. Otherwise, 0 audio_samples (int, optional): audio sampling rate audio_channels (int optional): audio audio_channels audio_pts_range (list(int), optional): the start and end presentation timestamp of audio stream audio_timebase_numerator / audio_timebase_denominator (float, optional): a rational number which denotes time base in audio stream Returns: vframes (Tensor[T, H, W, C]): the `T` video frames aframes (Tensor[L, K]): the audio frames, where `L` is the number of points and `K` is the number of channels ignore The given buffer is not writablemessageZdtypeNrr)rr& isinstancerFTensorwarningscatch_warningsfilterwarnings frombufferuint8rGr read_video_from_memoryr-r8)rUr<r=r>r?r@rArBrVrWrCrDrEr3rXrYrIrJrKr'r(r)r1r2r*r+r,rrr_read_video_from_memory s@E  .  rg)rUrc Cst|tjsRt.tjdddtj|tjd}Wdn1sH0Ytjj |dddddddddddddddddd}t |\ }}}}}}}} } } t |||| | | } | }| }||| fS) z Decode all frames in the video. Only pts (presentation timestamp) is returned. The actual frame pixel data is not copied. Thus, read_video_timestamps(...) is much faster than read_video(...) rZr[r\r^Nrrr4)r_rFr`rarbrcrdrerGr rfrr0rNrO) rUrIrPrQr'r(r)rRr2r*r+r,rLrrr"_read_video_timestamps_from_memoryzs>  .  rhc Cstt|tjsXt.tjdddtj|tjd}Wdn1sN0Ytj j |}|\}}}}}}t ||||||}|S)zy Probe a video in memory and return VideoMetaData with info about the video This function is torchscriptable rZr[r\r^N) rr_rFr`rarbrcrdrerGr Zprobe_video_from_memoryr0) rUrIr'r(r)r*r+r,rLrrr_probe_video_from_memorys  .ripts)r; start_ptsend_ptspts_unitrc stdurtddkr(tdt|}|j}|j}fdd}d}t} |rrt|j j |j j } || }d} t} |rt|j j |j j } || } t |d|| d| | d\} } }i}|r|j|d <|r|j|d <| | |fS) NinfrjmThe pts_unit 'pts' gives wrong results and will be removed in a follow-up version. Please use pts_unit 'sec'.cs`}}dkrHttd|}|tdkrHttd|}|tdkrXd}||fS)Nsecrrnr4)rmathfloorr#ceil)Z time_baseZ start_offset end_offsetrlrmrkrrget_ptss  z_read_video..get_ptsr:T)r=rBrrCr3rrZ audio_fps)rr#rawarnrTrrdefault_timebaserrr r rrMrr!)r;rkrlrmrLrrrvrBrr3rrJr1_inforrur _read_videosH    rz)r;rmrcsjt|dkrtdt|\}}}|dkrRt|jj|jjfdd|D}|jr^|j nd}||fS)Nrjrorpcsg|] }|qSrr).0xZvideo_time_baserr z*_read_video_timestamps..) rrarwrSrrr r rr)r;rmrj_rLrrr}r_read_video_timestampssr)r9rrrrrr:rrrrrr:rr)rNrj)rj)&rqraZ fractionsrtypingrrrF extensionrZ_video_deprecation_warningrZ_HAS_CPU_VIDEO_DECODER ImportErrorOSErrorZ_HAS_VIDEO_OPTrxr rtuplerr&r`r0r8strr#r"rMlistrSrTrgrhridictrzrrrrrs      !  !    a"#    o ,   ?