qi7dZddlZddlZddlZddlZddlmZddlmZm Z ddl Z ddl m Z ddlmZerddl mZdd lmZmZdd lmZej,eZd d iZd ZedGddeZdgZy)z$Tokenization class for SigLIP model.N)copyfile) TYPE_CHECKINGAny) AddedToken)SentencePieceBackend) TextInput)loggingrequires_backends)requires vocab_filez spiece.modelu▁) sentencepiece)backendsc eZdZdZeZddgZ d"deee fdzddffd Z e dZ d Z d#d eed eedzd edeeffd ZdeedeefdZ d$d eed eedzdeefdZ d$d eed eedzdeefdZdZdZdedefdZdddZd%dddeeffd Ze dZdZdZdZdZd$ded edzdeefd!Z xZ!S)&SiglipTokenizera Construct a Siglip tokenizer. Based on [SentencePiece](https://github.com/google/sentencepiece). This tokenizer inherits from [`PreTrainedTokenizer`] which contains most of the main methods. Users should refer to this superclass for more information regarding those methods. Args: vocab_file (`str`): [SentencePiece](https://github.com/google/sentencepiece) file (generally has a *.spm* extension) that contains the vocabulary necessary to instantiate a tokenizer. eos_token (`str`, *optional*, defaults to `""`): The end of sequence token. unk_token (`str`, *optional*, defaults to `""`): The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this token instead. pad_token (`str`, *optional*, defaults to `""`): The token used for padding, for example when batching sequences of different lengths. additional_special_tokens (`list[str]`, *optional*): Additional special tokens used by the tokenizer. sp_model_kwargs (`dict`, *optional*): Will be passed to the `SentencePieceProcessor.__init__()` method. The [Python wrapper for SentencePiece](https://github.com/google/sentencepiece/tree/master/python) can be used, among other things, to set: - `enable_sampling`: Enable subword regularization. - `nbest_size`: Sampling parameters for unigram. Invalid for BPE-Dropout. - `nbest_size = {0,1}`: No sampling is performed. - `nbest_size > 1`: samples from the nbest_size results. - `nbest_size < 0`: assuming that nbest_size is infinite and samples from the all hypothesis (lattice) using forward-filtering-and-backward-sampling algorithm. - `alpha`: Smoothing parameter for unigram sampling, and dropout probability of merge operations for BPE-dropout. model_max_length (`int`, *optional*, defaults to 64): The maximum length (in number of tokens) for model inputs. do_lower_case (`bool`, *optional*, defaults to `True`): Whether or not to lowercase the input when tokenizing. input_idsattention_maskNsp_model_kwargsreturnc Rt|dt|trt|ddddn|}t|trt|ddddn|}t|trt|ddddn|}|in||_||_t |d||||||j||d| y)NprotobufTF)rstriplstrip normalizedspecial)r eos_token unk_token pad_tokenadditional_special_tokensrmodel_max_length do_lower_case)r isinstancestrrrr!super__init__) selfr rrrrrr r!kwargs __class__s `/opt/pipecat/venv/lib/python3.12/site-packages/transformers/models/siglip/tokenization_siglip.pyr&zSiglipTokenizer.__init__Xs $ +)S) yduVZ [ )S) yduVZ [ )S) yduVZ [  &5%A )B-43D3D"D MM,T^^,<=++   1 122 2r+ct|jg}|t||zdgzSt||z|z|zdgzS)a Create a mask from the two sequences passed to be used in a sequence-pair classification task. T5 does not make use of token type ids, therefore a list of zeros is returned. Args: token_ids_0 (`list[int]`): List of IDs. token_ids_1 (`list[int]`, *optional*): Optional second list of IDs for sequence pairs. Returns: `list[int]`: List of zeros. r)rDr@)r'r:r;eoss r*$create_token_type_ids_from_sequencesz4SiglipTokenizer.create_token_type_ids_from_sequencessP   !  {S()QC/ /;${2S89QC??r+cX|j|}||S|j|}||zS)a Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens. A sequence has the following format: - single sequence: `X ` - pair of sequences: `A B ` Args: token_ids_0 (`list[int]`): List of IDs to which the special tokens will be added. token_ids_1 (`list[int]`, *optional*): Optional second list of IDs for sequence pairs. Returns: `list[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens. )rG)r'r:r;s r* build_inputs_with_special_tokensz0SiglipTokenizer.build_inputs_with_special_tokenss;&22;?   66{CK, ,r+cD|jj}d|d<|S)Nr.)__dict__copy)r'states r* __getstate__zSiglipTokenizer.__getstate__s# ""$ j r+c||_t|dsi|_tjdi|j|_|j j |jy)Nrr")rNhasattrrspmSentencePieceProcessorr.Loadr )r'ds r* __setstate__zSiglipTokenizer.__setstate__sO t./#%D 22JT5I5IJ  4??+r+textcj|jtjddtjS)N) translater$ maketransstring punctuation)r'rYs r*remove_punctuationz"SiglipTokenizer.remove_punctuations$~~cmmBF4F4FGHHr+keep_punctuation_exact_stringcjr|j}|r*|jfd|j|D}nj |}t j dd|}|j}|S)aReturns canonicalized `text` (puncuation removed). Args: text (`str`): String to be canonicalized. keep_punctuation_exact_string (`str`, *optional*): If provided, then this exact string is kept. For example providing '{}' will keep any occurrences of '{}' (but will still remove '{' and '}' that appear separately). c3@K|]}j|ywr-)r`).0partr's r* z4SiglipTokenizer.canonicalize_text..s!626''-6sz\s+ )r!lowerjoinsplitr`resubstrip)r'rYrbs` r*canonicalize_textz!SiglipTokenizer.canonicalize_textsw   ::**Eb:c6D**40Dvvfc4(zz| r+r c t|t|jtdzfi|}t |dkDr"|dtk(r|d|j vr|dd}|S)z8 Converts a string to a list of tokens. rhr>rN)r%tokenizeSPIECE_UNDERLINEreplacer@all_special_tokens)r'rYadd_special_tokensr(tokensr)s r*rqzSiglipTokenizer.tokenizese!"2T\\BRTW5X"Xc\bc v;?vay,<<dNeNeAeABZF r+cpt|jjt|jSr-)r@r.encoder$rr0s r*unk_token_lengthz SiglipTokenizer.unk_token_lengths%4==''DNN(;<==r+c |j|d}|jj|t}|jj|j|zt}t ||j k\r||j dS|S)u* Returns a tokenized string. We de-activated the `add_dummy_prefix` option, thus the sentencepiece internals will always strip any SPIECE_UNDERLINE. For example: `self.sp_model.encode(f"{SPIECE_UNDERLINE}Hey", out_type = str)` will give `['H', 'e', 'y']` instead of `['▁He', 'y']`. Thus we always encode `f"{unk_token}text"` and strip the `unk_token`. Here is an example with `unk_token = ""` and `unk_token_length = 4`. `self.tokenizer.sp_model.encode(" Hey", out_type = str)[4:]`. Nra)out_type)ror.rxr$rr@ry)r'rYr(rvs r* _tokenizezSiglipTokenizer._tokenizes%%d$%O%%dS%9%%dnnt&;c%J25f+AVAV2Vvd++-.b\bbr+c8|jj|S)z0Converts a token (str) in an id using the vocab.)r. piece_to_id)r'tokens r*_convert_token_to_idz$SiglipTokenizer._convert_token_to_id2s}}((//r+c<|jj|}|S)z=Converts an index (integer) in a token (str) using the vocab.)r. IdToPiece)r'indexrs r*_convert_id_to_tokenz$SiglipTokenizer._convert_id_to_token6s ''. r+cg}d}d}|D]P}||jvr-|s|dz }||jj||zz }d}g}>|j|d}R||jj|z }|j S)z:Converts a sequence of tokens (string) in a single string.r[FrhT)rtr.decodeappendrn)r'rvcurrent_sub_tokens out_stringprev_is_specialrs r*convert_tokens_to_stringz(SiglipTokenizer.convert_tokens_to_string;s  (E///&#%Jdmm223EFNN "&%'""))%0"' ( dmm**+=>> !!r+save_directoryfilename_prefixctjj|stj d|dytjj ||r|dzndt dz}tjj|jtjj|k7rBtjj|jrt|j||fStjj|jsCt|d5}|jj}|j|ddd|fS|fS#1swY|fSxYw)NzVocabulary path (z) should be a directory-r[r wb)ospathisdirloggererrorrjVOCAB_FILES_NAMESabspathr isfileropenr.serialized_model_protowrite)r'rrout_vocab_fileficontent_spiece_models r*save_vocabularyzSiglipTokenizer.save_vocabularyNs%ww}}^, LL,^,<zrNN@T)NFr-)F)"__name__ __module__ __qualname____doc__rvocab_files_namesmodel_input_namesdictr$rr&propertyr1r9listintboolr?rGrJrLrQrXr`rorqryr|rrrtupler __classcell__)r)s@r*rr+s&P*$&67 "&15+ c3h$.+  + Z.. puO9O379t3COhlO cO8 3c 3tCy 3GK@9@379t3C@ c@.GK-9-379t3C- c-4 ,IsIsIHL0[QUVYQZ>>c(0 "&!c!C$J!Z_`cZd!r+r)rrrlr^rEshutilrtypingrrrrTtokenization_utils_baser tokenization_utils_sentencepiecerr utilsr r utils.import_utilsr get_loggerrrrrrr__all__r"r+r*rs+ %1D4/*   H %!>2 %&q!*q!'q!h  r+