%`i/Q%ddlZddlZddlmZddlmZmZddlZddlm Z ddl m Z  e dd Z n#e ef$rd Z YnwxYwe ZeddZGd d ZGd dZdeeefddfdZdejdejdejdejdejdejdefdZdejdejdeeefdejfdZdd ddddded dddef ded ed!ed"ed#ed$ed%ed&eeefd'ed(ed)ed*edeeefd+edeejejeffd,Zdedeeeeeeffd-Zdedefd.Z d=d/ejd ed!ed"ed#ed$ed%ed&eeefd0ed1ed(ed)ed*edeeefd2ed3edeejejff"d4Z d/ejdeeeeeeffd5Z!d/ejdefd6Z" d>ded8eeefd9eeeefd:edeejeje#eefff d;Z$ d?ded:edeeeeeefeeffd<Z%dS)@N)Fraction)OptionalUnion) _load_library) _raise_video_deprecation_warning video_readerTFc4eZdZeedZddgZdededdfdZdS)Timebase numerator denominatorrrreturnNc"||_||_dS)Nr )selfrrs m/home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/torchvision/io/_video_opt.py__init__zTimebase.__init__s #&)__name__ __module__ __qualname__int__annotations__ __slots__rrrr r s[$'<+>'?'?Yq\EVEVEXEXAYAYZZQ<$$&&y|/@/@/B/B)C)CC ??  q !DN"' (8(8"9"9H"DD  zz||atyy{{++1&s9Q<+<+<+>+>'?'?Yq\EVEVEXEXAYAYZZQ<$$&&y|/@/@/B/B)C)CC ??  q !DN"' (8(8"9"9H"DD a!&|'8'8':':!;!; Kraframes aframe_ptsaudio_pts_rangecn|d|d}}|d}t||z dzt|z }d}|}||dkrt|d|z |z }|ddkr'||dkrt|d|z |z }|||ddfS)Nrr)sizer)r) r:r;r<startend num_samplesstep_per_aframes_idxe_idxs r_align_audio_framesrFnsA 23E,,q//KC%K!O,,u[/A/AAO E E q!!!_Q'%/?BCCqRC/!*<$<$<_Q'#-@AA 5;> ""r?rr>filenameseek_frame_marginread_video_stream video_width video_heightvideo_min_dimensionvideo_max_dimensionvideo_pts_ranger read_audio_stream audio_samplesaudio_channelsr$ctt|t| tjj||d||||||d|d|j|j| | | | d| d| j| j}|\ }}}}}}}}}}t||||||}| dkrt||| }|||fS)ab Reads a video from a file, returning both the video frames and the audio frames Args: filename (str): path to the video file 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 $video_height, respectively video_pts_range (list(int), optional): the start and end presentation timestamp of video stream video_timebase (Fraction, optional): a Fraction rational number which denotes timebase in video stream read_audio_stream (int, optional): whether read audio stream. If yes, set to 1. Otherwise, 0 audio_samples (int, optional): audio sampling rate audio_channels (int optional): audio channels audio_pts_range (list(int), optional): the start and end presentation timestamp of audio stream audio_timebase (Fraction, optional): a Fraction rational number which denotes time base in audio stream Returns vframes (Tensor[T, H, W, C]): the `T` video frames aframes (Tensor[L, K]): the audio frames, where `L` is the number of points and `K` is the number of audio_channels info (Dict): metadata for the video and audio. Can contain the fields video_fps (float) and audio_fps (int) rr) r r-torchopsr read_video_from_filerrr9r5rF)rIrJrKrLrMrNrOrPr rQrRrSr<r$resultvframes _vframe_ptsr.r/r0r:r;r1r2r3infos r_read_video_from_filer\}s@%&&&/"""/""" Y # 8 8  " "'F*qwmG[)T9gz9Vbdm iy)\9 U UD}}%gz?KK GT !!rcRtjj|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(...) rrr>)rUrVr rWr9numpytolist) rIrX_vframes vframe_ptsr.r/r0_aframesr;r1r2r3r[s r _read_video_timestamps_from_filercs Y # 8 8                  'F*rxnHj)T9h IWcen iy)\9 U UD!!##**,,J!!##**,,J z4 ''rcttjj|}|\}}}}}}t ||||||}|S)zO Probe a video file and return VideoMetaData with info about the video )r rUrVr probe_video_from_filer9) rIrXr.r/r0r1r2r3r[s r_probe_video_from_filerfsV%&&& Y # 9 9( C CFEKBItY < iy)\9 U UD Kr video_datavideo_timebase_numeratorvideo_timebase_denominatoraudio_timebase_numeratoraudio_timebase_denominatorc`tt|t| t|tjsat j5t jddtj|tj }dddn #1swxYwYtj 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 writablemessagedtypeNrr)r r- isinstancerUTensorwarningscatch_warningsfilterwarnings frombufferuint8rVr read_video_from_memoryr5rF)rgrJrKrLrMrNrOrPrhrirQrRrSr<rjrkrXrYrZr.r/r0r:r;r1r2r3s r_read_video_from_memoryr{ sJ%&&&/"""/""" j%, / /I  $ & & I I  #H6X Y Y Y Y)*EKHHHJ I I I I I I I I I I I I I I I Y # : :  " "'F,qwmG[)T9gz9Vbdm}}%gz?KK G s7BB!$B!cdt|tjsatj5tjddtj|tj}dddn #1swxYwYtjj |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(...) rmrnrorqNrrr>)rsrUrtrurvrwrxryrVr rzr r9r^r_) rgrXr`rar.r/r0rbr;r1r2r3r[s r"_read_video_timestamps_from_memoryr}zs j%, / /I  $ & & I I  #H6X Y Y Y Y)*EKHHHJ I I I I I I I I I I I I I I IY # : :                  'F*%&&&qwnHj)T9h IWcen iy)\9 U UD!!##**,,J!!##**,,J z4 ''s7A11A58A5ctt|tjsat j5t jddtj|tj}dddn #1swxYwYtj j |}|\}}}}}}t||||||}|S)zy Probe a video in memory and return VideoMetaData with info about the video This function is torchscriptable rmrnrorqN) r rsrUrtrurvrwrxryrVr probe_video_from_memoryr9) rgrXr.r/r0r1r2r3r[s r_probe_video_from_memoryrs%&&& j%, / /I  $ & & I I  #H6X Y Y Y Y)*EKHHHJ I I I I I I I I I I I I I I IY # ; ;J G GFEKBItY < iy)\9 U UD Ks7A??BBpts start_ptsend_ptspts_unitc ttddkrtjdt |}|j}|j}fd}d}t} |r/t|j j |j j } || }d} t} |r/t|j j |j j } || } t|d|| d| | \} } }i}|r |j|d<|r |j|d <| | |fS) NinfrmThe pts_unit 'pts' gives wrong results and will be removed in a follow-up version. Please use pts_unit 'sec'.c }}dkrattjd|z z}|tdkr'ttjd|z z}|tdkrd}||fS)Nsecrrr>)rmathfloorr)ceil) time_base start_offset end_offsetrrrs rget_ptsz_read_video..get_ptss  u  tz)q9}*EFFGGLU5\\)) 7a)m+D!E!EFF u % %JZ''rrHT)rKrPr rQr<r$r" audio_fps)r r)ruwarnrfrr#default_timebaserr rrr$r\r"r&)rIrrrr[rr#rrPr r<r$rYr:_infos ``` r _read_videorsn %&&&,,5  >   "( + +DII ( ( ( ( ( ( (O%N2!$"5"?ATA`aa!'.11O%N2!$"5"?ATA`aa!'.112'%'%GWd E,!^k4!3k GU ""rct|dkrtjdt|\}}}|dkr2t |jj|jjfd|D}|jr|j nd}||fS)Nrrrcg|]}|zSrr).0xvideo_time_bases r z*_read_video_timestamps..s000qq?"000r) r rurrcrr rrrr")rIrr_r[r"rs @r_read_video_timestampsrs%&&&5  >   4H==LCD5"4#6#@$BUBabb0000C000"&.:dI  >r)rGrrrrrrHrrrrrrHrr)rNr)r)&rru fractionsrtypingrrrU extensionr_video_deprecation_warningr _HAS_CPU_VIDEO_DECODER ImportErrorOSError_HAS_VIDEO_OPTrr rtuplerr-rtr9rFstrr)r(r\listrcrfr{r}rdictrrrrrrs """""""" %%%%%%HHHHHH#M.!!!!W###"#(8Aq>> ' ' ' ' ' ' ' '%%%%%%%%B U38_     | ,||  ,  | @ # \ #',| #FKCQTHo # \ # # # #" $"  './"'./^"^"^"^"^" ^"  ^"  ^"^"38_^"^"^"^"^"38_^"^" 5<} 45^"^"^"^"B (s (uT#YS S`=`7a ( ( ( (FS] $  '.$%&''.$%&'!kk kkk k  k  kk38_k"k!$kkkk38_k"k !$!k" 5< %&#kkkk\(( (( 49d3i ./((((((((V *)*04 ;#;#;#UH_%;#eE8O, -;# ;#  5<tCJ'7 78 ;#;#;#;#~$)  5cDN* +Xe_ <=s 6 AA