U
    (d1<                     @   s*  d dl Z d dlZd dlZd dlZd dlZd dlmZ d dlmZm	Z	m
Z
mZmZmZ d dlZd dlZddlmZ ddlmZ z4d dlZejejj eejjjdsed	ZW n ek
r   ed
ZY nX ddddZedddZ d a!da"d,e#ej$e%e#ee	e#ef  eej$ ee% ee# ee	e#ef  dd
ddZ&de%e%e#de	e#eee'ee'df e
e' f  f e
d dddZ(ej$e
d e'e%ej$dddZ)d-e#ee%ef eee%ef  e#e#eej$ej$e	e#ef f d!d"d#Z*ded$d%d&Z+de
e' d$d'd(Z,d.e#e#ee
e' ee% f d)d*d+Z-dS )/    N)Fraction)AnyDictListOptionalTupleUnion   )_log_api_usage_once   )
_video_opt	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.
)returnc                   C   s   t ttrtd S N
isinstanceav	Exception r   r   8/tmp/pip-unpacked-wheel-vx7f76es/torchvision/io/video.py_check_av_available(   s    
r   c                   C   s   t tt S r   r   r   r   r   r   _av_available-   s    r   
   libx264)
filenamevideo_arrayfpsvideo_codecoptionsaudio_array	audio_fpsaudio_codecaudio_optionsr   c	                 C   s  t j st j stt t  t j|t jd	 }t
|trJt|}tj| ddv}	|	j||d}
|jd |
_|jd |
_|dkrdnd	|
_|pi |
_|d
k	rtddddddddddd
}|	j||d}|pi |_|jd }|dkrdnd}|	jjd jj}t|| }t |	 |}tjj|||d}||_||D ]}|	 | qH| D ]}|	 | qb|D ]8}tj!j|d	d}d|_"|
|D ]}|	 | qqx|
 D ]}|	 | qW 5 Q R X d
S )a|  
    Writes a 4d tensor in [T, H, W, C] format in a video file

    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
        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
    dtypew)mode)rater	   r   Z
libx264rgbZyuv420pZrgb24Nz<f8z<f4z<i2z<i4u1)
ZdblZdblpZfltZfltpZs16Zs16pZs32Zs32pu8Zu8pr   ZstereoZmono)formatZlayout)r*   NONE)#torchjitis_scripting
is_tracingr
   write_videor   	as_tensoruint8numpyr   floatnproundr   openZ
add_streamshapewidthheightZpix_fmtr   streamsaudior*   namer$   ZastypeZ
AudioFrameZfrom_ndarrayZsample_rateencodeZmux
VideoFramer   )r   r   r   r   r   r   r    r!   r"   	containerstreamZaudio_format_dtypesZa_streamZnum_channelsZaudio_layoutZaudio_sample_fmtZformat_dtypeframepacketimgr   r   r   r0   6   sX    





r0   zav.container.Containerzav.stream.Stream.zav.frame.Frame)r@   start_offset
end_offsetpts_unitrA   stream_namer   c                    s.  t d7 a t t td kr t  |dkrfttd|j   tdkrptt	 d|j   n
t
d i d}d}|jdkr|jj}|rd|kr|d}	||	d  }
td	|
}|d krtd
|
}|d k	r|ddk}}t|d d}|rt|| d}z| j|dd|d W n tjk
r>   g  Y S X d}zXt| jf |D ]B\}}||j< |j krV|r||k r|d7 }qV qqVW n tjk
r   Y nX  fddtD }tdkr*dkr*kr*fddD }t|dkr*t|}|d|  |S )Nr   secinfzBThe pts_unit 'pts' gives wrong results. Please use pts_unit 'sec'.T   videos   DivXs   DivX(\d+)Build(\d+)(\w)s   DivX(\d+)b(\d+)(\w)      pr   F)Z	any_frameZbackwardrA   c                    s2   g | ]*}| j   kr" krn q| qS r   pts.0irF   framesrE   r   r   
<listcomp>   s
        z%_read_from_stream.<locals>.<listcomp>c                    s   g | ]}| k r|qS r   r   rQ   )rE   r   r   rV      s      )_CALLED_TIMES_GC_COLLECTION_INTERVALgcZcollectintmathfloor	time_baser4   ceilwarningswarntypecodec_context	extradatafindresearchgroupmaxseekr   AVError	enumeratedecoderP   sortedleninsert)r@   rE   rF   rG   rA   rH   Zshould_buffermax_buffer_sizerc   posdoZseek_offsetZbuffer_countZ_idxrB   resultZpreceding_framesZfirst_frame_ptsr   rT   r   _read_from_stream   s^    	




"ru   )aframesaudio_frames	ref_startref_endr   c           
      C   s|   |d j |d j  }}| jd }|| d | }d}|}	||k rPt|| | }||krht|| | }	| d d ||	f S )Nr   r   )rP   r8   rZ   )
rv   rw   rx   ry   startendZtotal_aframesZstep_per_aframeZs_idxZe_idxr   r   r   _align_audio_frames   s    
r}   rP   THWC)r   	start_ptsend_ptsrG   output_formatr   c              	   C   s|  t j st j stt | }|dkr<td| dddlm	} t
j| sbtd|  | dkr|t| |||S t  |dkrtd	}||k rtd
| d| i }g }g }tj}	ztj| dd}
|
jjr|
jjd j}	|
jjr4t|
||||
jjd ddi}|
jjd j}|dk	r4t||d< |
jjrnt|
||||
jjd ddi}|
jjd j|d< W 5 Q R X W n tjk
r   Y nX dd |D }dd |D }|rt t|}nt j dt j!d}|rHt"|d}t |}|dkr8t#t$%|d|	  }|td	kr8t#t$&|d|	  }t'||||}nt j dt j(d}|dkrr|)dddd}|||fS )a  
    Reads a video from a file, returning both the video frames as well as
    the audio frames

    Args:
        filename (str): path to the video file
        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)
    )r~   TCHWz5output_format should be either 'THWC' or 'TCHW', got .r   get_video_backendzFile not found: pyavNrJ   z7end_pts should be larger than start_pts, got start_pts=z and end_pts=ignoreZmetadata_errorsrL   	video_fpsr<   r    c                 S   s   g | ]}|   qS r   )Zto_rgb
to_ndarrayrR   rB   r   r   r   rV   D  s     zread_video.<locals>.<listcomp>c                 S   s   g | ]}|  qS r   )r   r   r   r   r   rV   E  s     )r   r   r   rM   r#   r   rI   )r   r   r   rM   r	   )*r,   r-   r.   r/   r
   
read_videoupper
ValueErrortorchvisionr   ospathexistsRuntimeErrorr   Z_read_videor   r4   Zdefault_timebaser   r7   r;   r<   r]   rL   ru   average_rater'   rj   r1   r5   stackemptyr2   ZconcatenaterZ   r[   r\   r^   r}   Zfloat32Zpermute)r   r   r   rG   r   r   infoZvideo_framesrw   Zaudio_timebaser@   r   Zvframes_listZaframes_listZvframesrv   r   r   r   r      s    





 


r   )r@   r   c                 C   s*   | j d jj}|d krdS d|kr&dS dS )Nr   Fs   LavcT)r;   rb   rc   )r@   rc   r   r   r   !_can_read_timestamps_from_packets^  s    r   c                 C   s8   t | rdd | jddD S dd | jddD S d S )Nc                 S   s   g | ]}|j d k	r|j qS r   rO   rR   xr   r   r   rV   j  s     
 z,_decode_video_timestamps.<locals>.<listcomp>r   )rL   c                 S   s   g | ]}|j d k	r|j qS r   rO   r   r   r   r   rV   l  s     
 )r   Zdemuxrl   )r@   r   r   r   _decode_video_timestampsg  s    r   )r   rG   r   c           	   
      s2  t j st j stt ddlm} | dkr>t	| |S t
  d}g }zvtj| dd^}|jjr|jjd }|j zt|}W n& tjk
r   td|   Y nX t|j}W 5 Q R X W nB tjk
r } z d|  d	| }t|t W 5 d}~X Y nX |  |d
kr* fdd|D }||fS )a   
    List the video frames timestamps.

    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

    r   r   r   Nr   r   z Failed decoding frames for file zFailed to open container for z; Caught error: rI   c                    s   g | ]}|  qS r   r   r   Zvideo_time_baser   r   rV     s     z)read_video_timestamps.<locals>.<listcomp>)r,   r-   r.   r/   r
   read_video_timestampsr   r   r   Z_read_video_timestampsr   r   r7   r;   rL   r]   r   rj   r_   r`   r4   r   RuntimeWarningsort)	r   rG   r   r   rP   r@   Zvideo_streamemsgr   r   r   r   o  s2    

r   )r   NNNNN)r   NrP   r~   )rP   ).rY   r[   r   re   r_   Z	fractionsr   typingr   r   r   r   r   r   r3   r5   r,   utilsr
    r   r   loggingZ	set_levelERRORhasattrrL   rB   r?   ImportErrorr   boolr   rW   rX   strZTensorr4   r0   rZ   ru   r}   r   r   r   r   r   r   r   r   <module>   s    	
	      Y$Q       
q	