sg_ddlZddlZddlmZddlmZmZmZmZm Z m Z m Z ddl Z ddlZddlmZddlmZmZmZmZmZmZmZmZmZmZmZmZddlmZm Z m!Z!m"Z"m#Z#m$Z$erddl%Z&ddl'Z&ejPejPe&jRjTejPdk\re&jVjXZ-n e&jVZ-erdd l.m/Z/e-j`e/j`e-jbe/jbe-jde/jde-jfe/jfe-jhe/jhe-jje/jjiZ6er erddl7Z7ejpe9Z:e d e jvd ed ee jved fZGddeZ?GddeZ@eeAe eBeAeeffZCdZDGddeZEdZFdZGdZHdZIde jvdeJfdZKdKdeBdee<fdZLde jvfd ZM dLde jvd!e e eBe eBd"ffde>fd#ZN dLde jvd$e e e>eAfdeBfd%ZOdLde jvd&e>de eBeBffd'ZPd(eeAe ee ffdeJfd)ZQd(eeAe ee ffdeJfd*ZRd+eeeAe ee ffdeJfd,ZSd+eeeAe ee ffdeJfd-ZTdLde eAd fd.e eUdd fd/ZV dLd0e ee eAd fd.e eUde d ed eed ffd1ZW dMd2e eJd3e eUd4e eJd5e e eUeeUfd6e e eUeeUfd7e eJd8e eBd9e eJd:e eeAeBfd;e eJdfd?ZXddddddddddddde>jfd2e eJd3e eUd4e eJd5e e eUeeUfd6e e eUeeUfd7e eJd8e eBd9e eJd:e eeAeBfd;e eJdd@e e eAefdAe e>fdBZZGdCdDZ[dEe?dFe e?d"fd+eeddfdGZ\dHeeAdIeeAfdJZ]y)NN)BytesIO) TYPE_CHECKINGDictIterableListOptionalTupleUnion)version) ExplicitEnum TensorType is_jax_tensoris_numpy_array is_tf_tensoris_torch_availableis_torch_tensoris_torchvision_availableis_vision_availableloggingrequires_backendsto_numpy)IMAGENET_DEFAULT_MEANIMAGENET_DEFAULT_STDIMAGENET_STANDARD_MEANIMAGENET_STANDARD_STDOPENAI_CLIP_MEANOPENAI_CLIP_STDz9.1.0)InterpolationModezPIL.Image.Imagez torch.Tensorz np.ndarrayz np.ndarrrayceZdZdZdZy)ChannelDimensionchannels_first channels_lastN)__name__ __module__ __qualname__FIRSTLASTK/var/www/html/venv/lib/python3.12/site-packages/transformers/image_utils.pyr!r!`s E Dr*r!ceZdZdZdZy)AnnotationFormatcoco_detection coco_panopticN)r$r%r&COCO_DETECTION COCO_PANOPTICr)r*r+r-r-es %N#Mr*r-cdeZdZejj Zej j Zy)AnnotionFormatN)r$r%r&r-r0valuer1r)r*r+r3r3js$%44::N$2288Mr*r3cbtxr$t|tjjSN)r isinstancePILImageimgs r+ is_pil_imager<rs  EZSYY__%EEr*c eZdZdZdZdZdZdZy) ImageTypepillowtorchnumpy tensorflowjaxN)r$r%r&r8TORCHNUMPY TENSORFLOWJAXr)r*r+r>r>vs C E EJ Cr*r>c>t|rtjSt|rtjSt |rtj St|rtjSt|rtjStdt|)NzUnrecognised image type ) r<r>r8rrDrrErrFrrG ValueErrortypeimages r+get_image_typerM~suE}}ueE###U}} /U }= >>r*ct|xs2t|xs%t|xst|xs t |Sr6)r<rrrrr:s r+is_valid_imagerOs8   vs 3 vs7K v|\_O` vdqrudvvr*crt|ttfr|D]}t|ryyt |syy)NFT)r7listtuple valid_imagesrO)imgsr;s r+rSrSs?$u & C$  D ! r*cLt|ttfrt|dSy)NrF)r7rQrRrOr:s r+ is_batchedrVs"#e}%c!f%% r*rLreturnc|jtjk(rytj|dk\xrtj|dkS)zV Checks to see whether the pixel values have already been rescaled to [0, 1]. Frr )dtypenpuint8minmaxrKs r+is_scaled_imager^s> {{bhh 66%=A  4"&&-1"44r*expected_ndimsc Zt|r|St|tjjr|gSt |rU|j |dzk(r t |}|S|j |k(r|g}|Std|dzd|d|j dtdt|d)a Ensure that the input is a list of images. If the input is a single image, it is converted to a list of length 1. If the input is a batch of images, it is converted to a list of images. Args: images (`ImageInput`): Image of images to turn into a list of images. expected_ndims (`int`, *optional*, defaults to 3): Expected number of dimensions for a single input image. If the input image has a different number of dimensions, an error is raised. r z%Invalid image shape. Expected either z or z dimensions, but got z dimensions.ztInvalid image type. Expected either PIL.Image.Image, numpy.ndarray, torch.Tensor, tf.Tensor or jax.ndarray, but got .) rVr7r8r9rOndimrQrIrJ)imagesr_s r+make_list_of_imagesrds& &#))//*xf ;;.1, ,&\F [[N *XF  78J7K4P^O_`KK= .   $V ~Q 0 r*ct|stdt|tr9t |t j j rtj|St|S)NzInvalid image type: ) rOrIrJrr7r8r9rZarrayrr:s r+to_numpy_arrayrgsP # /S {;<<C!Axx} C=r* num_channels.c||nd}t|tr|fn|}|jdk(rd\}}n-|jdk(rd\}}ntd|j|j||vrD|j||vr3t j d|jdtjS|j||vrtjS|j||vrtjStd) a[ Infers the channel dimension format of `image`. Args: image (`np.ndarray`): The image to infer the channel dimension of. num_channels (`int` or `Tuple[int, ...]`, *optional*, defaults to `(1, 3)`): The number of channels of the image. Returns: The channel dimension of the image. r rk)rz(Unsupported number of image dimensions: z4The channel dimension is ambiguous. Got image shape z,. Assuming channels are the first dimension.z(Unable to infer channel dimension format) r7intrbrIshapeloggerwarningr!r'r()rLrh first_dimlast_dims r+infer_channel_dimension_formatrts$0#;<L&0s&CL?L zzQ" 8 q" 8CEJJ<PQQ {{9-%++h2G<2WB5;;-O{ |  %%% Y < /%%% X , .$$$ ? @@r*input_data_formatc| t|}|tjk(r|jdz S|tjk(r|jdz St d|)a Returns the channel dimension axis of the image. Args: image (`np.ndarray`): The image to get the channel dimension axis of. input_data_format (`ChannelDimension` or `str`, *optional*): The channel dimension format of the image. If `None`, will infer the channel dimension from the image. Returns: The channel dimension axis of the image. rkr Unsupported data format: )rtr!r'rbr(rI)rLrus r+get_channel_dimension_axisrxsd :5A,222zzA~ .33 3zzA~ 01B0CD EEr* channel_dimc| t|}|tjk(r|jd|jdfS|tjk(r|jd|jdfSt d|)a Returns the (height, width) dimensions of the image. Args: image (`np.ndarray`): The image to get the dimensions of. channel_dim (`ChannelDimension`, *optional*): Which dimension the channel dimension is in. If `None`, will infer the channel dimension from the image. Returns: A tuple of the image's height and width. rw)rtr!r'ror(rI)rLrys r+get_image_sizer~s{4U; &,,,{{2 B// (-- -{{2 B//4[MBCCr* annotationct|trId|vrEd|vrAt|dttfr(t |ddk(st|ddtryy)Nimage_id annotationsrTFr7dictrQrRlenrs r+"is_valid_annotation_coco_detectionr2s`:t$ * $ Z ' z-04- @  =) *a /:j>WXY>Z\`3a r*ct|trMd|vrId|vrEd|vrAt|dttfr(t |ddk(st|ddtryy)Nr segments_info file_namerTFrrs r+!is_valid_annotation_coco_panopticrAsh:t$ * $ z ) : % z/2T5M B  ?+ , 1Z ?@[\]@^`d5e r*rc&td|DS)Nc32K|]}t|ywr6)r.0anns r+ z3valid_coco_detection_annotations..RsN31#6Nallrs r+ valid_coco_detection_annotationsrQs N+N NNr*c&td|DS)Nc32K|]}t|ywr6)rrs r+rz2valid_coco_panoptic_annotations..VsM#05Mrrrs r+valid_coco_panoptic_annotationsrUs MM MMr*timeoutcttdgt|tr|j ds|j drHt j jttj||j}ntjj|r t j j|}n|j dr|jdd} t!j"|j%}t j jt|}n2t|t j j r|}n t+d t j,j/|}|j1d }|S#t&$r}t)d|d |d }~wwxYw) a3 Loads `image` to a PIL Image. Args: image (`str` or `PIL.Image.Image`): The image to convert to the PIL Image format. timeout (`float`, *optional*): The timeout value in seconds for the URL request. Returns: `PIL.Image.Image`: A PIL Image. visionzhttp://zhttps://rz data:image/,r zIncorrect image source. Must be a valid URL starting with `http://` or `https://`, a valid path to an image file, or a base64 encoded string. Got z. Failed with NzuIncorrect format used for image. Should be an url linking to an image, a base64 string, a local path, or a PIL image.RGB)r load_imager7str startswithr8r9openrrequestsgetcontentospathisfilesplitbase64 decodebytesencode ExceptionrI TypeErrorImageOpsexif_transposeconvert)rLrb64es r+rrYsvj8*-%   I &%*:*::*FIINN78<<w+O+W+W#XYE WW^^E "IINN5)E . C(+ ((8 ws|4 E399?? + D   LL ' ' .E MM% E L  ijoipp~@~AB s2A F"" G+F<<Grcc <t|ttfrjt|rDt|dttfr+|Dcgc]}|Dcgc]}t ||c}c}}S|Dcgc]}t ||c}St ||Scc}wcc}}wcc}w)aLoads images, handling different levels of nesting. Args: images: A single image, a list of images, or a list of lists of images to load. timeout: Timeout for loading images. Returns: A single image, a list of images, a list of lists of images. rr)r7rQrRrr)rcr image_grouprLs r+ load_imagesrs&4-( v;:fQi$?eklVa[QEZw7Ql lDJK5Jug6K K&'22 RlKs B B B*BB do_rescalerescale_factor do_normalize image_mean image_stddo_padsize_divisibilitydo_center_crop crop_size do_resizesizeresamplePILImageResamplingc |r | td|r | td|r|| td|r | td| r| | tdyy)a Checks validity of typically used arguments in an `ImageProcessor` `preprocess` method. Raises `ValueError` if arguments incompatibility is caught. Many incompatibilities are model-specific. `do_pad` sometimes needs `size_divisor`, sometimes `size_divisibility`, and sometimes `size`. New models and processors added should follow existing arguments when possible. Nz=`rescale_factor` must be specified if `do_rescale` is `True`.zzDepending on the model, `size_divisibility`, `size_divisor`, `pad_size` or `size` must be specified if `do_pad` is `True`.zP`image_mean` and `image_std` must both be specified if `do_normalize` is `True`.z<`crop_size` must be specified if `do_center_crop` is `True`.zA`size` and `resample` must be specified if `do_resize` is `True`.)rI) rrrrrrrrrrrrs r+validate_preprocess_argumentsrs,n,XYY #+ I  +y/@kll)+WXXdlh&6\]]'7yr*return_tensors data_formatc t|||||| | | | dk7r td| tjk7r tdy)z Checks validity of typically used arguments in an `ImageProcessorFast` `preprocess` method. Raises `ValueError` if arguments incompatibility is caught. )rrrrrrrrptz6Only returning PyTorch tensors is currently supported.z6Only channel first data format is currently supported.N)rrIr!r')rrrrrrrrrrrrrrs r+"validate_fast_preprocess_argumentsrsZ("%!  QRR&,,,QRR-r*ceZdZdZdZddZdZdejde e e fdejfd Z dd Z d Zdd Zdd ZdZdZddZy)ImageFeatureExtractionMixinzD Mixin that contain utilities for preparing image features. ct|tjjtjfs$t |st dt|dyy)Nz Got type zS which is not supported, only `PIL.Image.Image`, `np.array` and `torch.Tensor` are.)r7r8r9rZndarrayrrIrJselfrLs r+_ensure_format_supportedz4ImageFeatureExtractionMixin._ensure_format_supportedsQ%#))//2::!>?X]H^DK=)&& I_?r*Nc|j|t|r|j}t|tj r|'t|j dtj}|jdk(r$|jddvr|jddd}|r|dz}|jtj}tjj|S|S)a" Converts `image` to a PIL Image. Optionally rescales it and puts the channel dimension back as the last axis if needed. Args: image (`PIL.Image.Image` or `numpy.ndarray` or `torch.Tensor`): The image to convert to the PIL Image format. rescale (`bool`, *optional*): Whether or not to apply the scaling factor (to make pixel values integers between 0 and 255). Will default to `True` if the image type is a floating type, `False` otherwise. rrkrjr rl)rrrAr7rZrflatfloatingrbro transposeastyper[r8r9 fromarray)rrLrescales r+ to_pil_imagez(ImageFeatureExtractionMixin.to_pil_images %%e, 5 !KKME eRZZ ($UZZ]BKK@zzQ5;;q>V#;1a0 LL*E99&&u- - r*c|j|t|tjjs|S|j dS)z Converts `PIL.Image.Image` to RGB format. Args: image (`PIL.Image.Image`): The image to convert. r)rr7r8r9rrs r+ convert_rgbz'ImageFeatureExtractionMixin.convert_rgbs8 %%e,%1L}}U##r*rLscalerWc.|j|||zS)z7 Rescale a numpy image by scale amount )r)rrLrs r+rz#ImageFeatureExtractionMixin.rescale"s %%e,u}r*c|j|t|tjjrt j |}t |r|j}|'t|jdtjn|}|r/|j|jtjd}|r"|jdk(r|jddd}|S)a Converts `image` to a numpy array. Optionally rescales it and puts the channel dimension as the first dimension. Args: image (`PIL.Image.Image` or `np.ndarray` or `torch.Tensor`): The image to convert to a NumPy array. rescale (`bool`, *optional*): Whether or not to apply the scaling factor (to make pixel values floats between 0. and 1.). Will default to `True` if the image is a PIL Image or an array/tensor of integers, `False` otherwise. channel_first (`bool`, *optional*, defaults to `True`): Whether or not to permute the dimensions of the image to put the channel dimension first. rp?rkrlr )rr7r8r9rZrfrrArintegerrrfloat32rbr)rrLr channel_firsts r+rgz*ImageFeatureExtractionMixin.to_numpy_array)s %%e, eSYY__ -HHUOE 5 !KKME;B?*UZZ]BJJ7PW LLbjj!99EE UZZ1_OOAq!,E r*c|j|t|tjjr|St |r|j d}|St j|d}|S)z Expands 2-dimensional `image` to 3 dimensions. Args: image (`PIL.Image.Image` or `np.ndarray` or `torch.Tensor`): The image to expand. r)axis)rr7r8r9r unsqueezerZ expand_dimsrs r+rz'ImageFeatureExtractionMixin.expand_dimsIs_ %%e, eSYY__ -L 5 !OOA&E NN5q1E r*c|j|t|tjjr|j |d}nw|rut|t j r0|j|jt jd}n+t|r |j|jd}t|t j rt|t j s.t j|j|j}t|t j st j|j|j}nt|rddl}t||js?t|t j r|j |}n|j"|}t||js?t|t j r|j |}n|j"|}|j$dk(r)|j&ddvr||ddddfz |ddddfz S||z |z S)a  Normalizes `image` with `mean` and `std`. Note that this will trigger a conversion of `image` to a NumPy array if it's a PIL Image. Args: image (`PIL.Image.Image` or `np.ndarray` or `torch.Tensor`): The image to normalize. mean (`List[float]` or `np.ndarray` or `torch.Tensor`): The mean (per channel) to use for normalization. std (`List[float]` or `np.ndarray` or `torch.Tensor`): The standard deviation (per channel) to use for normalization. rescale (`bool`, *optional*, defaults to `False`): Whether or not to rescale the image to be between 0 and 1. If a PIL image is provided, scaling will happen automatically. T)rrrNrkrj)rr7r8r9rgrZrrrrrfloatrfrYr@Tensor from_numpytensorrbro)rrLmeanstdrr@s r+ normalizez%ImageFeatureExtractionMixin.normalize]s %%e, eSYY__ -''t' eRZZ (dBJJ/xx~,,U[[9c2::.hhsm**5;;7 U # dELL1dBJJ/+5++D1D'5<<-Dc5<<0c2::.*%**3/C&%,,s+C ::?u{{1~7DD$//3q$}3EE EDLC' 'r*c||ntj}|j|t|tj j s|j |}t|tr t|}t|tst|dk(r|rt|tr||fn |d|df}n|j\}}||kr||fn||f\}} t|tr|n|d} || k(r|S| t| | z|z } } |.|| krtd|d|| |kDrt|| z| z |} } ||kr| | fn| | f}|j||S)a Resizes `image`. Enforces conversion of input to PIL.Image. Args: image (`PIL.Image.Image` or `np.ndarray` or `torch.Tensor`): The image to resize. size (`int` or `Tuple[int, int]`): The size to use for resizing the image. If `size` is a sequence like (h, w), output size will be matched to this. If `size` is an int and `default_to_square` is `True`, then image will be resized to (size, size). If `size` is an int and `default_to_square` is `False`, then smaller edge of the image will be matched to this number. i.e, if height > width, then image will be rescaled to (size * height / width, size). resample (`int`, *optional*, defaults to `PILImageResampling.BILINEAR`): The filter to user for resampling. default_to_square (`bool`, *optional*, defaults to `True`): How to convert `size` when it is a single int. If set to `True`, the `size` will be converted to a square (`size`,`size`). If set to `False`, will replicate [`torchvision.transforms.Resize`](https://pytorch.org/vision/stable/transforms.html#torchvision.transforms.Resize) with support for resizing only the smallest edge and providing an optional `max_size`. max_size (`int`, *optional*, defaults to `None`): The maximum allowed for the longer edge of the resized image: if the longer edge of the image is greater than `max_size` after being resized according to `size`, then the image is resized again so that the longer edge is equal to `max_size`. As a result, `size` might be overruled, i.e the smaller edge may be shorter than `size`. Only used if `default_to_square` is `False`. Returns: image: A resized `PIL.Image.Image`. r rz max_size = zN must be strictly greater than the requested size for the smaller edge size = )r)rBILINEARrr7r8r9rrQrRrnrrrIresize) rrLrrdefault_to_squaremax_sizewidthheightshortlongrequested_new_short new_shortnew_longs r+rz"ImageFeatureExtractionMixin.resizesy< (389K9T9T %%e,%1%%e,E dD !;D dC CIN '1$'<d|47DQRGBT % v16&ufovuo t.8s.Cda#// L&93?RUY?Y\a?a;b8 '#66()(4@@DvG (*.1(Y2F2Q.RT\8 05 8,hPYEZ||D8|44r*c|j|t|ts||f}t|st|tj rP|j dk(r|j|}|jddvr|jddn|jdd}n|jd|jdf}|d|dz dz}||dz}|d|dz dz}||dz}t|tjjr|j||||fS|jddvrdnd}|sKt|tj r|jddd}t|r|jddd}|dk\r!||dkr|dk\r||dkr |d||||fS|jdd t|d|dt|d|dfz} t|tj rt j || } nt|r|j#| } | d |dz dz} | |dz} | d |dz dz} | |dz}| d| | | |f<|| z }|| z }|| z }|| z }| dtd|t%| jd |td|t%| jd |f} | S) a Crops `image` to the given size using a center crop. Note that if the image is too small to be cropped to the size given, it will be padded (so the returned result has the size asked). Args: image (`PIL.Image.Image` or `np.ndarray` or `torch.Tensor` of shape (n_channels, height, width) or (height, width, n_channels)): The image to resize. size (`int` or `Tuple[int, int]`): The size to which crop the image. Returns: new_image: A center cropped `PIL.Image.Image` or `np.ndarray` or `torch.Tensor` of shape: (n_channels, height, width). rlrrjr NTF.r{)ror|)rr7rRrrZrrbrrorr8r9croprpermuter] zeros_like new_zerosr\)rrLr image_shapetopbottomleftrightr new_shape new_imagetop_pad bottom_padleft_pad right_pads r+ center_cropz'ImageFeatureExtractionMixin.center_crops %%e,$&$1a7{1~- bMKN2q8{1~- AF #wz)8I+==> w'   Qs9??2#6??QPST]TcTcdfTginPoAo o r*c|j|t|tjjr|j |}|dddddddfS)a Flips the channel order of `image` from RGB to BGR, or vice versa. Note that this will trigger a conversion of `image` to a NumPy array if it's a PIL Image. Args: image (`PIL.Image.Image` or `np.ndarray` or `torch.Tensor`): The image whose color channels to flip. If `np.ndarray` or `torch.Tensor`, the channel dimension should be first. Nr|)rr7r8r9rgrs r+flip_channel_orderz.ImageFeatureExtractionMixin.flip_channel_ordersI %%e, eSYY__ -''.ETrT1aZ  r*c||ntjj}|j|t |tjjs|j |}|j ||||||S)a Returns a rotated copy of `image`. This method returns a copy of `image`, rotated the given number of degrees counter clockwise around its centre. Args: image (`PIL.Image.Image` or `np.ndarray` or `torch.Tensor`): The image to rotate. If `np.ndarray` or `torch.Tensor`, will be converted to `PIL.Image.Image` before rotating. Returns: image: A rotated `PIL.Image.Image`. )rexpandcenter translate fillcolor)r8r9NEARESTrr7rrotate)rrLanglerrrrrs r+rz"ImageFeatureExtractionMixin.rotate0sn (389J9J %%e,%1%%e,E|| HVFicl  r*r6)NT)F)NTN)NrNNN)r$r%r&__doc__rrrrZrr rrnrrgrrrr rrr)r*r+rrsj< $RZZeSj0Abjj@(2(hA5FIV!" r*rannotation_formatsupported_annotation_formatsc||vrtdtd||tjurt |s td|tj urt |s tdyy)NzUnsupported annotation format: z must be one of zInvalid COCO detection annotations. Annotations must a dict (single image) or list of dicts (batch of images) with the following keys: `image_id` and `annotations`, with the latter being a list of annotations in the COCO format.zInvalid COCO panoptic annotations. Annotations must a dict (single image) or list of dicts (batch of images) with the following keys: `image_id`, `file_name` and `segments_info`, with the latter being a list of annotations in the COCO format.)rIformatr-r0rr1r)rrrs r+validate_annotationsrIs  <<:6(BRSoRpqrr,;;;/ <B  ,:::.{;M <;r*valid_processor_keyscaptured_kwargsct|jt|}|r+dj|}tj d|dyy)Nz, zUnused or unrecognized kwargs: ra)set differencejoinrprq)rr unused_keysunused_key_strs r+validate_kwargsr&bsJo&11#6J2KLK;/88HJKr*)rkr6) NNNNNNNNNNNN)^rriortypingrrrrrr r rArZr packagingr utilsr rrrrrrrrrrrutils.constantsrrrrrr PIL.Imager8 PIL.ImageOpsparse __version__ base_versionr9 Resamplingrtorchvision.transformsrrBOXrHAMMINGBICUBICLANCZOSpil_torch_interpolation_mappingr@ get_loggerr$rpr ImageInput VideoInputr!r-r3rrnAnnotationTyper<r>rMrOrSrVboolr^rdrgrtrxr~rrrrrrrrr'rrrr&r)r*r+r=s   NNN    w}}]W]]3??3@@A]W]]SZE[[ YY11 YY!<  & &(9(A(A  " "$5$9$9  ' '):)C)C  & &(9(A(A  & &(9(A(A  & &(9(A(A + '   H %rzz>48I3JDQSQ[Q[L\^bcq^rr   !m n    | $|$ 9\9 c5c4:!5667F  ?w  52::5$5$$D "&^+,&^T"&&*#'6:59!'+%)*. $%)/37;.>.D.D#S#SUO#S4.#Sud5k123 #S eT%[012 #S TN #S }#STN#SS#X'#S~#S 4S> "#S+,#SU3 ?34#S*+#SN\ \ ~ '"'(8#(=">d 2L$s)Ld3iLr*