qiK dZddlmZddlZddlmZmZmZddl m Z m Z m Z ddl mZmZmZmZmZmZmZmZmZmZmZddlmZdd lmZmZmZmZm Z m!Z!dd l"m#Z#erddl$Z$erddl%Z%e jLe'Z(Gd d ed Z)dZ* ddejVde,dzde,dzdee,zdzfdZ-e#dGddeZ.dgZ/y)z%Image processor class for LayoutLMv3.)IterableN)BaseImageProcessor BatchFeature get_size_dict)resizeto_channel_dimension_format to_pil_image) IMAGENET_STANDARD_MEANIMAGENET_STANDARD_STDChannelDimension ImageInputPILImageResamplinginfer_channel_dimension_formatis_scaled_imagemake_flat_list_of_imagesto_numpy_array valid_imagesvalidate_preprocess_arguments) ImagesKwargs) TensorTypefilter_out_non_signature_kwargsis_pytesseract_availableis_vision_availableloggingrequires_backends)requiresc<eZdZUdZeed<edzed<edzed<y)LayoutLMv3ImageProcessorKwargsa apply_ocr (`bool`, *optional*, defaults to `True`): Whether to apply the Tesseract OCR engine to get words + normalized bounding boxes. Can be overridden by the `apply_ocr` parameter in the `preprocess` method. ocr_lang (`str`, *optional*): The language, specified by its ISO code, to be used by the Tesseract OCR engine. By default, English is used. Can be overridden by the `ocr_lang` parameter in the `preprocess` method. tesseract_config (`str`, *optional*): Any additional custom configuration flags that are forwarded to the `config` parameter when calling Tesseract. For example: '--psm 6'. Can be overridden by the `tesseract_config` parameter in the `preprocess` method. apply_ocrNocr_langtesseract_config)__name__ __module__ __qualname____doc__bool__annotations__strl/opt/pipecat/venv/lib/python3.12/site-packages/transformers/models/layoutlmv3/image_processing_layoutlmv3.pyrr9s" ODjDj r+rF)totalctd|d|z ztd|d|z ztd|d|z ztd|d|z zgS)Nirr)int)boxwidthheights r, normalize_boxr5Ls` DCFUN #$ DCFVO $% DCFUN #$ DCFVO $%  r+imagelangr"input_data_formatct||}|j\}}tj||d|}|d|d|d|d|df\}} } } } t |D cgc]\} }|j r| }} }t |D cgc] \} }| |vs |}} }t | D cgc] \} }| |vs |} } }t | D cgc] \} }| |vs |} } }t | D cgc] \} }| |vs |} } }t | D cgc] \} }| |vs |} } }g}t | | | | D]$\}}}}||||z||zg}|j|&g}|D]}|jt||| t|t|k(sJd ||fScc}} wcc}} wcc}} wcc}} wcc}} wcc}} w) zdApplies Tesseract OCR on a document image, and returns recognized words + normalized bounding boxes.r8dict)r7 output_typeconfigtextlefttopr3r4z-Not as many words as there are bounding boxes) r size pytesseract image_to_data enumeratestripzipappendr5len)r6r7r"r8 pil_image image_width image_heightdatawordsr?r@r3r4idxwordirrelevant_indicescoord actual_boxesxywh actual_boxnormalized_boxesr2s r,apply_tesseractrYUsU6GHI )K  $ $YTvVf gD&*6lDL$u+tT[}^bck^l&l#E4eV09/?T)#ttzz|#TT#,U#3 Uic4sBT7TT UE U$-dO Ujc5sBT7TE UD U#,S> SZS%S@R5R5 SC S%.u%5 WzsEDV9VU WE W&/&7 Y U3FX;Xe YF YL$UF3( 1aAE1q5) J'( O c; MNO u:-. ._0__ . " "")U U U S W YsH F(9F( F.F.0 F4=F4 F:F:4 GG G#G)vision)backendsc xeZdZdZdgZeZddejddddddddf de de e e fdzd ed e d e d e d e ee zdzde ee zdzde de dzde dzddffd Zejddfdej"de e e fd ede ezdzde ezdzdej"f dZeddddddddddddej*dfdede dzde e e fdzd e dzd e dzd e dzd e ee zdzde ee zdzde dzde dzde dzde ezdzdede ezdzdej2j2fdZxZS)LayoutLMv3ImageProcessora Constructs a LayoutLMv3 image processor. Args: do_resize (`bool`, *optional*, defaults to `True`): Whether to resize the image's (height, width) dimensions to `(size["height"], size["width"])`. Can be overridden by `do_resize` in `preprocess`. size (`dict[str, int]` *optional*, defaults to `{"height": 224, "width": 224}`): Size of the image after resizing. Can be overridden by `size` in `preprocess`. resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BILINEAR`): Resampling filter to use if resizing the image. Can be overridden by `resample` in `preprocess`. do_rescale (`bool`, *optional*, defaults to `True`): Whether to rescale the image's pixel values by the specified `rescale_value`. Can be overridden by `do_rescale` in `preprocess`. rescale_factor (`float`, *optional*, defaults to 1 / 255): Value by which the image's pixel values are rescaled. Can be overridden by `rescale_factor` in `preprocess`. do_normalize (`bool`, *optional*, defaults to `True`): Whether to normalize the image. Can be overridden by the `do_normalize` parameter in the `preprocess` method. image_mean (`Iterable[float]` or `float`, *optional*, defaults to `IMAGENET_STANDARD_MEAN`): Mean to use if normalizing the image. This is a float or list of floats the length of the number of channels in the image. Can be overridden by the `image_mean` parameter in the `preprocess` method. image_std (`Iterable[float]` or `float`, *optional*, defaults to `IMAGENET_STANDARD_STD`): Standard deviation to use if normalizing the image. This is a float or list of floats the length of the number of channels in the image. Can be overridden by the `image_std` parameter in the `preprocess` method. apply_ocr (`bool`, *optional*, defaults to `True`): Whether to apply the Tesseract OCR engine to get words + normalized bounding boxes. Can be overridden by the `apply_ocr` parameter in the `preprocess` method. ocr_lang (`str`, *optional*): The language, specified by its ISO code, to be used by the Tesseract OCR engine. By default, English is used. Can be overridden by the `ocr_lang` parameter in the `preprocess` method. tesseract_config (`str`, *optional*): Any additional custom configuration flags that are forwarded to the `config` parameter when calling Tesseract. For example: '--psm 6'. Can be overridden by the `tesseract_config` parameter in the `preprocess` method. pixel_valuesTNgp? do_resizerAresample do_rescalerescale_factor do_normalize image_mean image_stdr r!r"returnc  &t |di| ||nddd}t|}||_||_||_||_| jd||_||_ ||nt|_ ||nt|_ | |_| |_| |_y)N)r4r3 rescale_valuer*)super__init__rr`rArarbgetrcrdr rer rfr r!r")selfr`rArarbrcrdrerfr r!r"kwargs __class__s r,rlz!LayoutLMv3ImageProcessor.__init__s "6"'tc-JT""   $%jj.I((2(>*DZ&/&;AV"  0r+r6 data_formatr8c t|}d|vsd|vrtd|j|d|df}t|f||||d|S)a Resize an image to `(size["height"], size["width"])`. Args: image (`np.ndarray`): Image to resize. size (`dict[str, int]`): Dictionary in the format `{"height": int, "width": int}` specifying the size of the output image. resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BILINEAR`): `PILImageResampling` filter to use when resizing the image e.g. `PILImageResampling.BILINEAR`. data_format (`ChannelDimension` or `str`, *optional*): The channel dimension format for the output image. If unset, the channel dimension format of the input image is used. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - `"none"` or `ChannelDimension.NONE`: image in (height, width) format. input_data_format (`ChannelDimension` or `str`, *optional*): The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - `"none"` or `ChannelDimension.NONE`: image in (height, width) format. Returns: `np.ndarray`: The resized image. r4r3zFThe `size` dictionary must contain the keys `height` and `width`. Got )rArarqr8)r ValueErrorkeysr)rnr6rArarqr8ro output_sizes r,rzLayoutLMv3ImageProcessor.resizesyFT" 4 7$#6efjfofofqerst tH~tG}5   #/     r+imagesreturn_tensorsc ||n |j}||n |j}t|}||n |j}||n |j}||n |j }||n |j }||n |j}| | n |j} | | n |j} | | n |j} | | n |j} t|}t|s tdt||||| ||||Dcgc] }t!|}}|r#t#|drt$j'd|t)|d}| rKt+|dg}g}|D]6}t-|| | |\}}|j/||j/|8|r"|Dcgc]}|j1||||}}|r!|Dcgc]}|j3|||}}|r"|Dcgc]}|j5||| | }}|Dcgc]}t7||| }}t9d |i| }| r |d <|d<|Scc}wcc}wcc}wcc}wcc}w)ai 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`. do_resize (`bool`, *optional*, defaults to `self.do_resize`): Whether to resize the image. size (`dict[str, int]`, *optional*, defaults to `self.size`): Desired size of the output image after applying `resize`. resample (`int`, *optional*, defaults to `self.resample`): Resampling filter to use if resizing the image. This can be one of the `PILImageResampling` filters. 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 pixel values between [0, 1]. rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`): Rescale factor to apply to the image pixel values. Only has an effect 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 `Iterable[float]`, *optional*, defaults to `self.image_mean`): Mean values to be used for normalization. Only has an effect if `do_normalize` is set to `True`. image_std (`float` or `Iterable[float]`, *optional*, defaults to `self.image_std`): Standard deviation values to be used for normalization. Only has an effect if `do_normalize` is set to `True`. apply_ocr (`bool`, *optional*, defaults to `self.apply_ocr`): Whether to apply the Tesseract OCR engine to get words + normalized bounding boxes. ocr_lang (`str`, *optional*, defaults to `self.ocr_lang`): The language, specified by its ISO code, to be used by the Tesseract OCR engine. By default, English is used. tesseract_config (`str`, *optional*, defaults to `self.tesseract_config`): Any additional custom configuration flags that are forwarded to the `config` parameter when calling Tesseract. return_tensors (`str` or `TensorType`, *optional*): The type of tensors to return. Can be one of: - Unset: Return a list of `np.ndarray`. - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`. - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.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. zSInvalid image type. Must be of type PIL.Image.Image, numpy.ndarray, or torch.Tensor)rbrcrdrerfr`rArarzIt 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.rBr:)r6rArar8)r6scaler8)r6meanstdr8)input_channel_dimr^)rL tensor_typerMboxes)r`rArrarbrcrdrerfr r!r"rrrsrrrlogger warning_oncerrrYrGrrescale normalizer r)rnrvr`rArarbrcrdrerfr r!r"rwrqr8r6 words_batch boxes_batchrMr~rLs r, preprocessz#LayoutLMv3ImageProcessor.preprocesssH"+!6IDNN 'tTYYT"'38#-#9Zt +9+E4K^K^'3'?|TEVEV #-#9Zt !*!6IDNN !*!6IDNN '38/?/K+QUQfQf)&1F#rs s%!)%! 6<vay I   dM 2KK *.uh@Pduv u""5)""5) * $ %dXYj kF $ 5Rc dF $U^opF ou ej '{N_ `  .&!9~V 'DM'DM c=.    s4H-H24H7H<9I)r#r$r%r&model_input_namesr valid_kwargsrBILINEARr'r;r)r1floatrrlnpndarrayr rrFIRSTrrPILImager __classcell__)rps@r,r]r]{s$L((1L&*'9'B'B '!5948#')1138nt#1% 1  1  11HUO+d218E?*T111*1*1 1L(:'B'B59;? . zz. 38n. % . ++d2 . !11D8 .  . `%&"&&*"&'+$(5948!%#'+26(8(>(>;?!PP$;P38nt# P 4K P PTkPHUO+d2P8E?*T1P$;P*P*Pj(4/P&P !11D8!P" #P'Pr+r])N)0r&collections.abcrnumpyrimage_processing_utilsrrrimage_transformsrr r image_utilsr r r rrrrrrrrprocessing_utilsrutilsrrrrrrutils.import_utilsrrrB get_loggerr#rrr5rr)rYr]__all__r*r+r,rs,$UUQQ    -+   H %!\!&8< ## ::## *##Dj##(#-4 ##L ;M1M M` & &r+