a h7;@sdZddlZddlZddlZddlZddlmZddlmZm Z m Z m Z ddl m Z e eZddd Zd d ZGd d d ZddZddZGdddeZdgZdS)z$Tokenization classes for OpenAI GPT.N)Optional)PreTrainedTokenizer _is_control_is_punctuation_is_whitespace)loggingz vocab.jsonz merges.txt) vocab_file merges_filecCs|}|sgS|}|S)z@Runs basic whitespace cleaning and splitting on a piece of text.)stripsplit)texttokensrj/var/www/html/assistant/venv/lib/python3.9/site-packages/transformers/models/openai/tokenization_openai.pywhitespace_tokenize$s rc@sNeZdZdZdddZdddZdd Zdd d Zd d ZddZ ddZ dS)BasicTokenizera Constructs a BasicTokenizer that will run basic tokenization (punctuation splitting, lower casing, etc.). Args: do_lower_case (`bool`, *optional*, defaults to `True`): Whether or not to lowercase the input when tokenizing. never_split (`Iterable`, *optional*): Collection of tokens which will never be split during tokenization. Only has an effect when `do_basic_tokenize=True` tokenize_chinese_chars (`bool`, *optional*, defaults to `True`): Whether or not to tokenize Chinese characters. This should likely be deactivated for Japanese (see this [issue](https://github.com/huggingface/transformers/issues/328)). strip_accents (`bool`, *optional*): Whether or not to strip all accents. If this option is not specified, then it will be determined by the value for `lowercase` (as in the original BERT). do_split_on_punc (`bool`, *optional*, defaults to `True`): In some instances we want to skip the basic punctuation splitting so that later tokenization can capture the full context of the words, such as contractions. TNcCs2|dur g}||_t||_||_||_||_dSN) do_lower_caseset never_splittokenize_chinese_chars strip_accentsdo_split_on_punc)selfrrrrrrrr__init__Es zBasicTokenizer.__init__cCs|r|jt|n|j}||}|jr4||}td|}t|}g}|D]R}||vr|j r| }|j dur| |}n|j r| |}| |||qPtd|}|S)aj Basic Tokenization of a piece of text. For sub-word tokenization, see WordPieceTokenizer. Args: never_split (`List[str]`, *optional*) Kept for backward compatibility purposes. Now implemented directly at the base class level (see [`PreTrainedTokenizer.tokenize`]) List of token not to split. NFCF )runionr _clean_textr_tokenize_chinese_chars unicodedata normalizerrlowerr_run_strip_accentsextend_run_split_on_puncjoin)rr rZunicode_normalized_textZ orig_tokens split_tokenstokenZ output_tokensrrrtokenizeUs$       zBasicTokenizer.tokenizecCsBtd|}g}|D]"}t|}|dkr,q||qd|S)z$Strips accents from a piece of text.ZNFDZMn)r!r"categoryappendr')rr outputcharcatrrrr${s   z!BasicTokenizer._run_strip_accentscCs|jr|dur||vr|gSt|}d}d}g}|t|kr||}t|r^||gd}n |rl|gd}|d||d7}q0dd|DS) z&Splits punctuation on a piece of text.NrTFcSsg|]}d|qS)r+)r').0xrrr z5BasicTokenizer._run_split_on_punc..)rlistlenrr-)rr rcharsiZstart_new_wordr.r/rrrr&s"    z!BasicTokenizer._run_split_on_punccCsTg}|D]@}t|}||r>|d|||dq||qd|S)z)Adds whitespace around any CJK character.rr+)ord_is_chinese_charr-r'rr r.r/cprrrr s     z&BasicTokenizer._tokenize_chinese_charscCs|dkr|dks|dkr |dks|dkr0|dks|dkr@|dks|d krP|d ks|d kr`|d ks|d krp|dks|dkr|dkrdSdS)z6Checks whether CP is the codepoint of a CJK character.iNii4iMiiߦii?i@ii iiiiiTFr)rr>rrrr<sD  zBasicTokenizer._is_chinese_charcCsXg}|D]D}t|}|dks|dkst|r.qt|rB|dq||qd|S)zBPerforms invalid character removal and whitespace cleanup on text.rirr+)r;rrr-r'r=rrrrs  zBasicTokenizer._clean_text)TNTNT)N)N) __name__ __module__ __qualname____doc__rr*r$r&r r<rrrrrr.s  &  rcCs6t}|d}|ddD]}|||f|}q|S)z Return set of symbol pairs in a word. word is represented as tuple of symbols (symbols being variable-length strings) rr2N)radd)wordpairsZ prev_charr/rrr get_pairss rFcCsn|dd}|dd}|dd}|dd}|dd}td d |}td d |}td d|}|S)zm fixes some issues the spacy tokenizer had on books corpus also does some whitespace standardization u—-u–u―u…z...´'zD(-+|~+|!+|"+|;+|\?+|\++|,+|\)+|\(+|\\+|\/+|\*+|\[+|\]+|}+|{+|\|+|_+)z \1 z\s*\n\s*z z[^\S\n]+r)replaceresubr )r rrrtext_standardizes     rMcseZdZdZeZddgZdfdd ZeddZ ed d Z d d Z d dZ ddZ ddZddZddZdeeeeedddZZS)OpenAIGPTTokenizera( Construct a GPT Tokenizer. Based on Byte-Pair-Encoding with the following peculiarities: - lowercases all inputs, - uses `SpaCy` tokenizer and `ftfy` for pre-BPE tokenization if they are installed, fallback to BERT's `BasicTokenizer` if not. 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`): Path to the vocabulary file. merges_file (`str`): Path to the merges file. 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. Z input_idsZattention_maskc  s4z.ddl}ddlm}|}|j|_|j|_Wn.ty\tdt dd|_d|_Yn0t |dd}t ||_ Wdn1s0Ydd |j D|_t |dd&} | d d d } Wdn1s0Yd d| D} tt| tt| |_i|_tjfd|i|dS)Nr)EnglishzQftfy or spacy is not installed using BERT BasicTokenizer instead of SpaCy & ftfy.T)rutf-8encodingcSsi|]\}}||qSrr)r3kvrrr r6z/OpenAIGPTTokenizer.__init__.. r2r1cSsg|]}t|qSr)tupler )r3mergerrrr5r6z/OpenAIGPTTokenizer.__init__.. unk_token)ftfyZ spacy.lang.enrP tokenizernlpfix_text ImportErrorloggerwarningropenjsonloadencoderitemsdecoderreadr dictzipranger8 bpe_rankscachesuperr) rr r rZkwargsr[rPZ_nlpZ vocab_handleZ merges_handleZmerges __class__rrrs&      *4zOpenAIGPTTokenizer.__init__cCsdS)NTrrrrrrsz OpenAIGPTTokenizer.do_lower_casecCs t|jSr)r8rerrrrr vocab_sizeszOpenAIGPTTokenizer.vocab_sizecCst|jfi|jSr)rireZadded_tokens_encoderrrrrr get_vocab#szOpenAIGPTTokenizer.get_vocabc st|dd|ddf}|jvr2j|St|}|sF|dSt|fddd}|jvrhqd|\}}g}d}|t|kr:z|||} Wn*ty|||dYq:Yn0|||| | }|||kr"|t|dkr"||d|kr"| |||d7}qx| |||d7}qxt|}|}t|dkrZqdqFt|}qFd |}|d kr|d }|j|<|S) Nr1csj|tdS)Ninf)rlgetfloat)pairrrrr0r6z(OpenAIGPTTokenizer.bpe..keyrr2rz z ) rXrmrFminrlr8index ValueErrorr%r-r') rr)rDrEZbigramfirstsecondZnew_wordr:jrrrrbpe&sF     2      zOpenAIGPTTokenizer.bpecCsg}|jdur@|j|}|D]}|t||dqn>|t||}|D]$}|t||j dqX|S)zTokenize a string.Nr) r^r]r*r%r7rr rMr r#)rr r(r)rrr _tokenizeRs  "zOpenAIGPTTokenizer._tokenizecCs|j||j|jS)z0Converts a token (str) in an id using the vocab.)rerwrZ)rr)rrr_convert_token_to_idasz'OpenAIGPTTokenizer._convert_token_to_idcCs|j||jS)z0Converts an id in a token (BPE) using the vocab.)rgrwrZ)rrrrr_convert_id_to_tokenesz'OpenAIGPTTokenizer._convert_id_to_tokencCsd|dd}|S)z:Converts a sequence of tokens (string) in a single string.r+rur)r'rJr )rrZ out_stringrrrconvert_tokens_to_stringisz+OpenAIGPTTokenizer.convert_tokens_to_stringN)save_directoryfilename_prefixreturnc CsTtj|s"td|ddStj||r6|dndtd}tj||rX|dndtd}t|ddd .}|t j |j d d d d dWdn1s0Yd}t|ddd v}|dt |j dddD]D\}} || krtd|d| }|d|d|d7}qWdn1sB0Y||fS)NzVocabulary path (z) should be a directoryrGr+r r wrQrRr}TF)indent sort_keys ensure_asciirWrz#version: 0.2 cSs|dS)Nr2r)kvrrrrzr6z4OpenAIGPTTokenizer.save_vocabulary..r{zSaving vocabulary to zZ: BPE merge indices are not consecutive. Please check that the tokenizer is not corrupted!rr2)ospathisdirr`errorr'VOCAB_FILES_NAMESrbwritercdumpsresortedrlrfra) rrrr Z merge_filefrwriterZ bpe_tokensZ token_indexrrrsave_vocabularyns. <   *z"OpenAIGPTTokenizer.save_vocabulary)rO)N)r?r@rArBrZvocab_files_namesZmodel_input_namesrpropertyrrsrtrrrrrstrrrXr __classcell__rrrprrNs  ,rN)rBrcrrKr!typingrZtokenization_utilsrrrrutilsrZ get_loggerr?r`rrrrFrMrN__all__rrrrs&    " "