S^(h{dZddlZddlmZmZmZmZmZmZm Z erddl m Z ddl Z ddlmZmZmZddlmZmZmZddlmZmZmZmZmZmZmZmZmZm Z m!Z!m"Z"m#Z#m$Z$dd l%m&Z&m'Z'm(Z(m)Z)m*Z*erddl+Z+e(rddl,Z,e)jZe.Z/ dd e j`d e e1ee1fd e2d e1dee e3efdee1e1ff dZ4GddeZ5dgZ6y)zImage processor class for DPT.N) TYPE_CHECKINGDictIterableListOptionalTupleUnion)DepthEstimatorOutput)BaseImageProcessor BatchFeature get_size_dict)padresizeto_channel_dimension_format)IMAGENET_STANDARD_MEANIMAGENET_STANDARD_STDChannelDimension ImageInputPILImageResamplingget_image_sizeinfer_channel_dimension_formatis_scaled_imageis_torch_availableis_torch_tensormake_list_of_imagesto_numpy_array valid_imagesvalidate_preprocess_arguments) TensorTypefilter_out_non_signature_kwargsis_vision_availableloggingrequires_backends input_image output_sizekeep_aspect_ratiomultipleinput_data_formatreturncdd}t|tr||fn|}t||\}}|\}} ||z } | |z } |r"td| z td| z kr| } n| } || |z|} || |z|} | | fS)Nct||z |z}| ||kDrtj||z |z}||krtj||z |z}|SN)roundmathfloorceil)valr(min_valmax_valxs j/var/www/skyplay_api_hub/venv/lib/python3.12/site-packages/transformers/models/dpt/image_processing_dpt.pyconstrain_to_multiple_ofz>get_resize_output_image_size..constrain_to_multiple_ofFs[ #. !H ,  1w; 3>*X5A w; #.)H4A)r()rN) isinstanceintrabs)r%r&r'r(r)r7 input_height input_width output_height output_width scale_height scale_width new_height new_widths r6get_resize_output_image_sizerE?s 1;;0L; ,R]K .{*DZ&/&;AV ( 0r8image data_formatr)c t|}d|vsd|vrtd|jt||d|df|||} t |f| |||d|S)aE Resize an image to target size `(size["height"], size["width"])`. If `keep_aspect_ratio` is `True`, the image is resized to the largest possible size such that the aspect ratio is preserved. If `ensure_multiple_of` is set, the image is resized to a size that is a multiple of this value. Args: image (`np.ndarray`): Image to resize. size (`Dict[str, int]`): Target size of the output image. keep_aspect_ratio (`bool`, *optional*, defaults to `False`): If `True`, the image is resized to the largest possible size such that the aspect ratio is preserved. ensure_multiple_of (`int`, *optional*, defaults to 1): The image is resized to a size that is a multiple of this value. resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BICUBIC`): Defines the resampling filter to use if resizing the image. Otherwise, the image is resized to size specified in `size`. resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BICUBIC`): Resampling filter to use when resiizing the image. data_format (`str` or `ChannelDimension`, *optional*): The channel dimension format of the image. If not provided, it will be the same as the input image. input_data_format (`str` or `ChannelDimension`, *optional*): The channel dimension format of the input image. If not provided, it will be inferred. rVrWzDThe size dictionary must contain the keys 'height' and 'width'. Got )r&r'r(r))rJrKr_r))r ValueErrorkeysrEr) r[r^rJr'rLrKr_r)r\r&s r6rzDPTImageProcessor.resizesFT" 4 7$#6cdhdmdmdocpqr r2 hg7/'/    #/     r8cd}| t|}t||\}}|||\}} |||\} } t||| f| | ff|S)a) Center pad an image to be a multiple of `multiple`. Args: image (`np.ndarray`): Image to pad. size_divisor (`int`): The width and height of the image will be padded to a multiple of this number. data_format (`ChannelDimension` or `str`, *optional*, defaults to `ChannelDimension.FIRST`): The channel dimension format for the output image. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - Unset: Use the channel dimension format of the input image. input_data_format (`ChannelDimension` or `str`, *optional*): The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - `"none"` or `ChannelDimension.NONE`: image in (height, width) format. c^tj||z |z}||z }|dz}||z }||fS)N)r/r1)rJrSnew_sizepad_size pad_size_leftpad_size_rights r6_get_padz-DPTImageProcessor.pad_image.._get_pad sAyy !45 DH$H$MM% 5N .0 0r8)r_)rrr) r[r^rSr_r)rjrVrWrhri pad_size_toppad_size_bottoms r6 pad_imagezDPTImageProcessor.pad_imagesl8 1  $ >u E &u.?@ (0(F% ~(0 (E% o5M>:\?<[\juvvr8labelcFt|}d||dk(<|dz }d||dk(<|S)Nrr9)r)r[rns r6 reduce_labelzDPTImageProcessor.reduce_labels6u%eqj !esl r8c|r|j|}|r|j||||||}|r|j|| |}| r|j|| | |}| r|j |||}|S)N)r^rJrKr'rLr))r^scaler))r^meanstdr))r^rSr))rrrrescale normalizerm)r[r^rTrIrJrKr'rLrMrNrOrPrQrRrSr)s r6 _preprocesszDPTImageProcessor._preprocess&s$ %%e,E KK!"3#5"3  E LLuNVgLhE NNZYbsNtE NN\]nNoE r8ct|}|r t|rtjd| t |}|j |d|||||||| | | | | |}|t |||}|S)zPreprocesses a single image.zIt looks like you are trying to rescale already rescaled images. If the input images have pixel values between 0 and 1, set `do_rescale=False` to avoid rescaling them again.F)rTrIrJrKr'rLrMrNrOrPrQrRrSr))input_channel_dim)rrlogger warning_oncerryr)r[r^rIrJrKr'rLrMrNrOrPrQrRrSr_r)s r6_preprocess_imagez#DPTImageProcessor._preprocess_imagePs(u% /%0   s   $ >u E   "/1!)%!%/! "  "/{VghE r8segmentation_mapc 2t|}|jdk(r|d}d} tj}nd} | t |d}|j |||||||dd| }| rt j|d }|jt j}|S) z'Preprocesses a single segmentation map.re)N.TFr9) num_channels) r^rTrIrJrKr'rLrOrMr)r)axis) rndimrFIRSTrrynpsqueezeastypeint64) r[rrIrJrKr'rLrTr)added_dimensions r6_preprocess_segmentation_mapz.DPTImageProcessor._preprocess_segmentation_maps**:;  A %/ : "O 0 6 6 #O ($BCSbc$d!++"-/1/,   !zz*:C +22288<r8c (t||fd|i|S)Nsegmentation_maps)rY__call__)r[imagesrr\r]s r6rzDPTImageProcessor.__call__s wV:KVvVVr8rrreturn_tensorsc,||n |j}||n |j}t|}||n |j}||n |j}||n |j }||n |j }| | n |j} | | n |j} | | n |j} | | n |j} | | n |j} ||n |j}||n |j}t|}| t|d}t|s t!dt#|| | | | | |||| |Dcgc]"}|j%|||| | ||||| | | |||$}}d|i}|+|Dcgc]}|j'||||||||}}||d<t)|| Scc}wcc}w) a Preprocess an image or batch of images. Args: images (`ImageInput`): Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If passing in images with pixel values between 0 and 1, set `do_rescale=False`. segmentation_maps (`ImageInput`, *optional*): Segmentation map to preprocess. do_resize (`bool`, *optional*, defaults to `self.do_resize`): Whether to resize the image. size (`Dict[str, int]`, *optional*, defaults to `self.size`): Size of the image after reszing. If `keep_aspect_ratio` is `True`, the image is resized to the largest possible size such that the aspect ratio is preserved. If `ensure_multiple_of` is set, the image is resized to a size that is a multiple of this value. keep_aspect_ratio (`bool`, *optional*, defaults to `self.keep_aspect_ratio`): Whether to keep the aspect ratio of the image. If False, the image will be resized to (size, size). If True, the image will be resized to keep the aspect ratio and the size will be the maximum possible. ensure_multiple_of (`int`, *optional*, defaults to `self.ensure_multiple_of`): Ensure that the image size is a multiple of this value. resample (`int`, *optional*, defaults to `self.resample`): Resampling filter to use if resizing the image. This can be one of the enum `PILImageResampling`, Only has an effect if `do_resize` is set to `True`. do_rescale (`bool`, *optional*, defaults to `self.do_rescale`): Whether to rescale the image values between [0 - 1]. rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`): Rescale factor to rescale the image by if `do_rescale` is set to `True`. do_normalize (`bool`, *optional*, defaults to `self.do_normalize`): Whether to normalize the image. image_mean (`float` or `List[float]`, *optional*, defaults to `self.image_mean`): Image mean. image_std (`float` or `List[float]`, *optional*, defaults to `self.image_std`): Image standard deviation. do_reduce_labels (`bool`, *optional*, defaults to `self.do_reduce_labels`): Whether or not to reduce all label values of segmentation maps by 1. Usually used for datasets where 0 is used for background, and background itself is not included in all classes of a dataset (e.g. ADE20k). The background label will be replaced by 255. return_tensors (`str` or `TensorType`, *optional*): The type of tensors to return. Can be one of: - Unset: Return a list of `np.ndarray`. - `TensorType.TENSORFLOW` or `'tf'`: Return a batch of type `tf.Tensor`. - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`. - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`. - `TensorType.JAX` or `'jax'`: Return a batch of type `jax.numpy.ndarray`. data_format (`ChannelDimension` or `str`, *optional*, defaults to `ChannelDimension.FIRST`): The channel dimension format for the output image. Can be one of: - `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `ChannelDimension.LAST`: image in (height, width, num_channels) format. input_data_format (`ChannelDimension` or `str`, *optional*): The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - `"none"` or `ChannelDimension.NONE`: image in (height, width) format. re)expected_ndimszkInvalid image type. Must be of type PIL.Image.Image, numpy.ndarray, torch.Tensor, tf.Tensor or jax.ndarray.) rMrNrOrPrQrRsize_divisibilityrIrJrK)r^rIrMrOrRrJrKr'rLrNrPrQrSr_r)rH)rrTrIrJrKr'rLr)labels)data tensor_type)rIrJrr'rLrKrMrNrOrPrQrRrSrTrrrarr~rr )r[rrrIrJr'rLrKrMrNrOrPrQrRrSrTrr_r)imgrrs r6 preprocesszDPTImageProcessor.preprocesss2Z"+!6IDNN 'tTYYT"1B1N-TXTjTj3E3Q/W[WnWn'38#-#9Zt +9+E4K^K^'3'?|TEVEV #-#9Zt !*!6IDNN !-4;;'3'?|TEVEV /?/K+QUQfQf$V,  ( 34EVW X F#:  &!)%!* >% $#  " "#%)!"3#5-%#)'"3 #   *'  (): !%11%5%5'%&7'9&72  !  !/DN>BBQ 0 !s %'F  F target_sizesc&|j}|t|t|k7r tdt|r|j }g}t t|D]k}t jjj||jd||dd}|djd}|j|m|S|jd}t |jdDcgc]}|| }}|Scc}w)a4 Converts the output of [`DPTForSemanticSegmentation`] into semantic segmentation maps. Only supports PyTorch. Args: outputs ([`DPTForSemanticSegmentation`]): Raw outputs of the model. target_sizes (`List[Tuple]` of length `batch_size`, *optional*): List of tuples corresponding to the requested final size (height, width) of each prediction. If unset, predictions will not be resized. Returns: semantic_segmentation: `List[torch.Tensor]` of length `batch_size`, where each item is a semantic segmentation map of shape (height, width) corresponding to the target_sizes entry (if `target_sizes` is specified). Each entry of each `torch.Tensor` correspond to a semantic class id. zTMake sure that you pass in as many target sizes as the batch dimension of the logitsr)dimbilinearFrJmode align_cornersr9)logitslenrarnumpyrangetorchnn functional interpolate unsqueezeargmaxappendshape) r[outputsrrsemantic_segmentationidxresized_logits semantic_mapis r6"post_process_semantic_segmentationz4DPTImageProcessor.post_process_semantic_segmentationPs("  #6{c,// j|,+113 $& !S[) ;!&!4!4!@!@3K))a)0|C7Hzin"A" .a077A7> %,,\:  ;%$%+MMaM$8 !GLMbMhMhijMkGl$m!%:1%=$m !$m$$%ns> Drr ct|d|j}|"t|t|k7r tdg}|dgt|zn|}t ||D]s\}}|Yt j jj|jdjd|ddj}|jd |iu|S) a Converts the raw output of [`DepthEstimatorOutput`] into final depth predictions and depth PIL images. Only supports PyTorch. Args: outputs ([`DepthEstimatorOutput`]): Raw outputs of the model. target_sizes (`TensorType` or `List[Tuple[int, int]]`, *optional*): Tensor of shape `(batch_size, 2)` or list of tuples (`Tuple[int, int]`) containing the target size (height, width) of each image in the batch. If left to None, predictions will not be resized. Returns: `List[Dict[str, TensorType]]`: A list of dictionaries of tensors representing the processed depth predictions. rNz]Make sure that you pass in as many target sizes as the batch dimension of the predicted depthrr9bicubicFrpredicted_depth) r$rrraziprrrrrrr)r[rrrresultsdepth target_sizes r6post_process_depth_estimationz/DPTImageProcessor.post_process_depth_estimation{s( $(!11  $3+?3|CT+To 8D8LvO 44R^ "%o|"D 7 E;&++77OOA&003+Iej8') NN-u5 6  7r8)NN)NNNNNNNNNNNNNN)NNNNNNNr-)&__name__ __module__ __qualname____doc__model_input_namesrBICUBICboolrstrr;r floatrrrZrndarrayrrarrayrmrrrryr~rrr!rr PILImagerrrr __classcell__)r]s@r6rGrGis8)V((#'9'A'A"'"#,3!:>9=&*!& 1 138n 1% 1  1  1 1c5j) 1 1U5$u+#567 1E%e"456 1 1sm 1 1 ! 1L#("#'9'A'A>BDH5 zz5 38n5  5  5 % 5 eC)9$9:;5 $E#/?*?$@A5  5 v?CDH +wxx+w+weC)9$9:; +w $E#/?*?$@A +w\*,0$(#'+,0,0%)*.'+:>9=!%&*DH!((#4.(D> ( 38n ( % ($D>(%SM(TN(!(tn(U5$u+#567(E%e"456((sm( $E#/?*?$@A!(Z%)#'+,0,0%)*.'+:>9=!%&*>BDH!11D>138n 1 % 1 $D> 1%SM1TN1!1tn1U5$u+#5671E%e"45611sm1eC)9$9:;1 $E#/?*?$@A!1" #1l%)#'+,0,0+/DH' $' D>' 38n ' % ' $D> ' %SM' #4.' $E#/?*?$@A' TW %&37$(",0,0'+%)*.'+:>9=!%&*+/;?(8(>(>DH'ZCZC$J/ZCD> ZC sm ZC $D> ZC%SMZC%ZCTNZC!ZCtnZCU5$u+#567ZCE%e"456ZCZCsmZC #4.!ZC"!sJ!78#ZC$&%ZC&$E#/?*?$@A'ZC( )ZC'ZCz)%U )%\RV'''uZeCHo1F%LMN' d3 ?# $ 'r8rGr-)7rr/typingrrrrrrr modeling_outputsr rrimage_processing_utilsr r rimage_transformsrrr image_utilsrrrrrrrrrrrrrrutilsr r!r"r#r$rr get_loggerrr|rr;rrrErG__all__rXr8r6rs% NNN8UUHH    H %AE '#'#sHSM)*'#'# '#  c+;&; <= '#  38_ '#Ty*yx  r8