U &cU7 @sdZddlZddlZddlZddlmZddlmZmZddl m Z e e Z ddiZdd d d d d ddddiZdddddddddZdZGddde ZdS)z' Tokenization classes for ALBERT model.N)copyfile)ListOptional)PreTrainedTokenizer vocab_filez spiece.modelzOhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-base-v1-spiece.modelzPhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-large-v1-spiece.modelzQhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-xlarge-v1-spiece.modelzRhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-xxlarge-v1-spiece.modelzOhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-base-v2-spiece.modelzPhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-large-v2-spiece.modelzQhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-xlarge-v2-spiece.modelzRhttps://s3.amazonaws.com/models.huggingface.co/bert/albert-xxlarge-v2-spiece.model)zalbert-base-v1zalbert-large-v1zalbert-xlarge-v1zalbert-xxlarge-v1zalbert-base-v2zalbert-large-v2zalbert-xlarge-v2zalbert-xxlarge-v2iu▁c seZdZdZeZeZeZ d(fd d Z e d d Z d dZ ddZddZddZd)ddZddZddZddZd*eeeeeeeddd Zd+eeeeeeeed!d"d#Zd,eeeeeeedd$d%Zd&d'ZZS)-AlbertTokenizerar Constructs an ALBERT tokenizer. Based on `SentencePiece `__ This tokenizer inherits from :class:`~transformers.PreTrainedTokenizer` which contains most of the methods. Users should refer to the superclass for more information regarding methods. Args: vocab_file (:obj:`string`): `SentencePiece `__ file (generally has a .spm extension) that contains the vocabulary necessary to instantiate a tokenizer. do_lower_case (:obj:`bool`, `optional`, defaults to :obj:`True`): Whether to lowercase the input when tokenizing. remove_space (:obj:`bool`, `optional`, defaults to :obj:`True`): Whether to strip the text when tokenizing (removing excess spaces before and after the string). keep_accents (:obj:`bool`, `optional`, defaults to :obj:`False`): Whether to keep accents when tokenizing. bos_token (:obj:`string`, `optional`, defaults to "[CLS]"): The beginning of sequence token that was used during pre-training. Can be used a sequence classifier token. .. note:: When building a sequence using special tokens, this is not the token that is used for the beginning of sequence. The token used is the :obj:`cls_token`. eos_token (:obj:`string`, `optional`, defaults to "[SEP]"): The end of sequence token. .. note:: When building a sequence using special tokens, this is not the token that is used for the end of sequence. The token used is the :obj:`sep_token`. unk_token (:obj:`string`, `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. sep_token (:obj:`string`, `optional`, defaults to "[SEP]"): The separator token, which is used when building a sequence from multiple sequences, e.g. two sequences for sequence classification or for a text and a question for question answering. It is also used as the last token of a sequence built with special tokens. pad_token (:obj:`string`, `optional`, defaults to ""): The token used for padding, for example when batching sequences of different lengths. cls_token (:obj:`string`, `optional`, defaults to "[CLS]"): The classifier token which is used when doing sequence classification (classification of the whole sequence instead of per-token classification). It is the first token of the sequence when built with special tokens. mask_token (:obj:`string`, `optional`, defaults to "[MASK]"): The token used for masking values. This is the token used when training this model with masked language modeling. This is the token which the model will try to predict. Attributes: sp_model (:obj:`SentencePieceProcessor`): The `SentencePiece` processor that is used for every conversion (string, tokens and IDs). TF[CLS][SEP][MASK]c  stjf||||| | | d| z ddl} Wn tk rNtdYnX||_||_||_||_ | |_ |j |dS)N) bos_token eos_token unk_token sep_token pad_token cls_token mask_tokenrzYou need to install SentencePiece to use AlbertTokenizer: https://github.com/google/sentencepiecepip install sentencepiece) super__init__ sentencepiece ImportErrorloggerwarning do_lower_case remove_space keep_accentsrSentencePieceProcessorsp_modelLoad)selfrrrrrrrrrrrkwargsspm __class__D/tmp/pip-unpacked-wheel-ymerj3tt/transformers/tokenization_albert.pyrrs0   zAlbertTokenizer.__init__cCs t|jS)N)lenr r"r'r'r( vocab_sizeszAlbertTokenizer.vocab_sizecs(fddtjD}|j|S)Ncsi|]}||qSr')Zconvert_ids_to_tokens).0ir*r'r( sz-AlbertTokenizer.get_vocab..)ranger+updateZadded_tokens_encoder)r"Zvocabr'r*r( get_vocabs zAlbertTokenizer.get_vocabcCs|j}d|d<|S)Nr )__dict__copy)r"stater'r'r( __getstate__s zAlbertTokenizer.__getstate__cCsP||_z ddl}Wn tk r2tdYnX||_|j|jdS)Nrr) r2rrrrrr r!r)r"dr$r'r'r( __setstate__s  zAlbertTokenizer.__setstate__cCsj|jrd|}n|}|dddd}|jsXtd|}ddd|D}|jrf| }|S) N z``"z''ZNFKDcSsg|]}t|s|qSr') unicodedata combining)r,cr'r'r( s z3AlbertTokenizer.preprocess_text..) rjoinstripsplitreplacerr; normalizerlower)r"inputsoutputsr'r'r(preprocess_texts zAlbertTokenizer.preprocess_textcCs||}|s|j|}n|j|dd}g}|D]}t|dkr|dtdkr|dr|j|ddtd}|d tkr|d d tkrt|d dkr|dd}n|d dd|d <| |d| |q4| |q4|S) z Tokenize a string. @g?r,Nr:r) rGr ZEncodeAsPiecesZSampleEncodeAsPiecesr)strisdigitrBSPIECE_UNDERLINEappendextend)r"textsamplepiecesZ new_piecesZpieceZ cur_piecesr'r'r( _tokenizes  (  zAlbertTokenizer._tokenizecCs |j|S)z2 Converts a token (str) in an id using the vocab. )r Z PieceToId)r"tokenr'r'r(_convert_token_to_idsz$AlbertTokenizer._convert_token_to_idcCs |j|S)z=Converts an index (integer) in a token (str) using the vocab.)r Z IdToPiece)r"indexr'r'r(_convert_id_to_tokensz$AlbertTokenizer._convert_id_to_tokencCsd|td}|S)Nr:r8)r?rBrNr@)r"tokensZ out_stringr'r'r(convert_tokens_to_stringsz(AlbertTokenizer.convert_tokens_to_stringN) token_ids_0 token_ids_1returncCs8|jg}|jg}|dkr$|||S|||||S)a Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens. An ALBERT sequence has the following format: - single sequence: ``[CLS] X [SEP]`` - pair of sequences: ``[CLS] A [SEP] B [SEP]`` Args: token_ids_0 (:obj:`List[int]`): List of IDs to which the special tokens will be added token_ids_1 (:obj:`List[int]`, `optional`, defaults to :obj:`None`): Optional second list of IDs for sequence pairs. Returns: :obj:`List[int]`: list of `input IDs <../glossary.html#input-ids>`__ with the appropriate special tokens. N sep_token_id cls_token_idr"r[r\sepclsr'r'r( build_inputs_with_special_tokenss  z0AlbertTokenizer.build_inputs_with_special_tokens)r[r\already_has_special_tokensr]csz|r*|dk rtdttfdd|S|dk r`dgdgt|dgdgt|dgSdgdgt|dgS)a  Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding special tokens using the tokenizer ``prepare_for_model`` or ``encode_plus`` methods. Args: token_ids_0 (:obj:`List[int]`): List of ids. token_ids_1 (:obj:`List[int]`, `optional`, defaults to :obj:`None`): Optional second list of IDs for sequence pairs. already_has_special_tokens (:obj:`bool`, `optional`, defaults to :obj:`False`): Set to True if the token list is already formatted with special tokens for the model Returns: :obj:`List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token. NzYou should not supply a second sequence if the provided sequence of ids is already formatted with special tokens for the model.cs|jjfkrdSdS)Nrrr^)xr*r'r(z9AlbertTokenizer.get_special_tokens_mask..rr) ValueErrorlistmapr))r"r[r\rer'r*r(get_special_tokens_masks.z'AlbertTokenizer.get_special_tokens_maskcCsV|jg}|jg}|dkr.t|||dgSt|||dgt||dgS)a Creates a mask from the two sequences passed to be used in a sequence-pair classification task. An ALBERT sequence pair mask has the following format: :: 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 | first sequence | second sequence | if token_ids_1 is None, only returns the first portion of the mask (0s). Args: token_ids_0 (:obj:`List[int]`): List of ids. token_ids_1 (:obj:`List[int]`, `optional`, defaults to :obj:`None`): Optional second list of IDs for sequence pairs. Returns: :obj:`List[int]`: List of `token type IDs <../glossary.html#token-type-ids>`_ according to the given sequence(s). Nrr)r_r`r)rar'r'r($create_token_type_ids_from_sequences$s z4AlbertTokenizer.create_token_type_ids_from_sequencescCs^tj|s td|dStj|td}tj|j tj|krXt |j ||fS)a2 Save the sentencepiece vocabulary (copy original file) and special tokens file to a directory. Args: save_directory (:obj:`str`): The directory in which to save the vocabulary. Returns: :obj:`Tuple(str)`: Paths to the files saved. z*Vocabulary path ({}) should be a directoryNr) ospathisdirrerrorformatr?VOCAB_FILES_NAMESabspathrr)r"Zsave_directoryZout_vocab_filer'r'r(save_vocabularyCs  zAlbertTokenizer.save_vocabulary) TTFr r r r r r r )F)N)NF)N)__name__ __module__ __qualname____doc__rsZvocab_files_namesPRETRAINED_VOCAB_FILES_MAPZpretrained_vocab_files_map&PRETRAINED_POSITIONAL_EMBEDDINGS_SIZESZmax_model_input_sizesrpropertyr+r1r5r7rGrTrVrXrZrintrrdboolrlrmru __classcell__r'r'r%r(r9s\4+         r)ryloggingrnr;shutilrtypingrrZtokenization_utilsr getLoggerrvrrsrzr{rNrr'r'r'r(s<