U 9%e]@sdZddlZddlZddlZddlZddlmZmZmZddl m Z ddl m Z e eZddd Zd d d d ddddddd ZdddddZddiddiddiddidZddZddZddZddZGdd d e ZdS)!z"Tokenization classes for Flaubert.N)ListOptionalTuple)PreTrainedTokenizer)loggingz vocab.jsonz merges.txt) vocab_file merges_filezLhttps://huggingface.co/flaubert/flaubert_small_cased/resolve/main/vocab.jsonzMhttps://huggingface.co/flaubert/flaubert_base_uncased/resolve/main/vocab.jsonzKhttps://huggingface.co/flaubert/flaubert_base_cased/resolve/main/vocab.jsonzLhttps://huggingface.co/flaubert/flaubert_large_cased/resolve/main/vocab.json)zflaubert/flaubert_small_casedzflaubert/flaubert_base_uncasedzflaubert/flaubert_base_casedzflaubert/flaubert_large_casedzLhttps://huggingface.co/flaubert/flaubert_small_cased/resolve/main/merges.txtzMhttps://huggingface.co/flaubert/flaubert_base_uncased/resolve/main/merges.txtzKhttps://huggingface.co/flaubert/flaubert_base_cased/resolve/main/merges.txtzLhttps://huggingface.co/flaubert/flaubert_large_cased/resolve/main/merges.txti do_lowercaseFTcCsddd}||dddS)zQ Converts `text` to Unicode (if it's not already), assuming UTF-8 input. utf-8strictcSs<t|tr|||St|tr$|Stdt|ddS)Nznot expecting type '') isinstancebytesdecodestr TypeErrortype)sencodingerrorsrq/var/www/html/Darija-Ai-API/env/lib/python3.8/site-packages/transformers/models/flaubert/tokenization_flaubert.py ensure_textRs    z'convert_to_unicode..ensure_textignore)rr)r r r)textrrrrconvert_to_unicodeMs rcCs6t}|d}|ddD]}|||f|}q|S)z Return set of symbol pairs in a word. word is represented as tuple of symbols (symbols being variable-length strings) rN)setadd)wordpairsZ prev_charcharrrr get_pairs^s r#cCs|dd}tdd|}|dd}|dd}|dd}|d d }|d d }|d d }|dd}|dd}|dd}|dd}|dd}|dd}|dd}|dd}|dd}|dd}|dd}|d d!}|d"d#}|d$d%}|d&d'}|d(d)}|d*d+}|d,d-}td.d|}|d/d0}|d1d2}|d3d4}|d5d6}|d7d8}|d9d:}|d;d<}|d=d>}|d?d@}|S)Azz Port of https://github.com/moses-smt/mosesdecoder/blob/master/scripts/tokenizer/replace-unicode-punctuation.perl u，,u。\s*z. u、u”"u“u∶:u：u？?u《u》u）)u！!u（(u；;u１1u」u「u０0u３3u２2u５5u６6u９9u７7u８8u４4u．\s*u～~u’r u…z...u━-u〈u【[u】]u％%)replaceresub)rrrrreplace_unicode_punctlsJ                                  r@cCs8g}|D]$}t|}|dr"q||qd|S)zw Port of https://github.com/moses-smt/mosesdecoder/blob/master/scripts/tokenizer/remove-non-printing-char.perl C) unicodedatacategory startswithappendjoin)routputr"catrrrremove_non_printing_chars   rJcsjeZdZdZeZeZeZ e Z dddddddddd d d d d dddg ddf fdd Z e ddZddZddZddZddZe ddZd d!Zd"d#Zd$d%Zd=d&d'Zd(d)Zd*d+Zd,d-Zd>eeeeeeed.d/d0Zd?eeeeeeeed1fd2d3 Z d@eeeeeeed.d4d5Z!dAe"ee"e#e"d6d7d8Z$d9d:Z%d;d<Z&Z'S)BFlaubertTokenizera Construct a Flaubert tokenizer. Based on Byte-Pair Encoding. The tokenization process is the following: - Moses preprocessing and tokenization. - Normalizing all inputs text. - The arguments `special_tokens` and the function `set_special_tokens`, can be used to add additional symbols (like "__classify__") to a vocabulary. - The argument `do_lowercase` controls lower casing (automatically set for pretrained vocabularies). 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`): Vocabulary file. merges_file (`str`): Merges file. do_lowercase (`bool`, *optional*, defaults to `False`): Controls lower casing. 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. bos_token (`str`, *optional*, defaults to `""`): The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token. 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 `cls_token`. sep_token (`str`, *optional*, defaults to `""`): 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 (`str`, *optional*, defaults to `""`): The token used for padding, for example when batching sequences of different lengths. cls_token (`str`, *optional*, defaults to `""`): 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 (`str`, *optional*, defaults to `""`): 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. additional_special_tokens (`List[str]`, *optional*, defaults to `["","","","","","","","","",""]`): List of additional special tokens. lang2id (`Dict[str, int]`, *optional*): Dictionary mapping languages string identifiers to their IDs. id2lang (`Dict[int, str]`, *optional*): Dictionary mapping language IDs to their string identifiers. Fzzzzz z z z z z z z z z Nc  sl| dd}|dk rtdd|_||_z ddl}Wntk rRtdYnX||_i|_i|_ dddh|_ | |_ | |_ | dk r| dk rt | t | kstd|_d|_t|d d }t||_W5QRXd d |jD|_t|d d }|d dd}W5QRXdd|D}tt|tt ||_i|_tjf|||||| | | | d | dS)Ndo_lowercase_and_remove_accentz`do_lowercase_and_remove_accent` is passed as a keyword argument, but this won't do anything. `FlaubertTokenizer` will always set it to `False`.FrzsYou need to install sacremoses to use FlaubertTokenizer. See https://pypi.org/project/sacremoses/ for installation.zhthjar rcSsi|]\}}||qSrr).0kvrrr sz.FlaubertTokenizer.__init__.. cSs g|]}t|ddqS)N)tuplesplit)rQmergerrr !sz.FlaubertTokenizer.__init__..) unk_token bos_token sep_token pad_token cls_token mask_tokenadditional_special_tokenslang2idid2lang) poploggerwarningrLr sacremoses ImportErrorsmcache_moses_punct_normalizercache_moses_tokenizerZlang_with_custom_tokenizerrcrdlenAssertionErrorja_word_tokenizerZzh_word_tokenizeropenjsonloadencoderitemsdecoderreadrYdictziprange bpe_rankscachesuper__init__)selfrr r r\r]r^r_r`rarbrcrdkwargsrLrhZ vocab_handleZ merges_handleZmerges __class__rrr}sX      zFlaubertTokenizer.__init__cCs|jSN)rLr~rrr do_lower_case2szFlaubertTokenizer.do_lower_casecCs8||jkr$|jj|d}||j|<n |j|}||S)Nlang)rkrjZMosesPunctNormalizer normalize)r~rrZpunct_normalizerrrrmoses_punct_norm8s    z"FlaubertTokenizer.moses_punct_normcCs>||jkr$|jj|d}||j|<n |j|}|j|dddS)NrF)Z return_strescape)rlrjZMosesTokenizertokenize)r~rrZmoses_tokenizerrrrmoses_tokenizeAs    z FlaubertTokenizer.moses_tokenizecCs t|}|||}t|}|Sr)r@rrJ)r~rrrrrmoses_pipelineJs z FlaubertTokenizer.moses_pipelinec Cs|jdkrz(ddl}|dtjdd|_WnVttfk rtdtdtdtdtd td YnXt |j |S) Nrz-model r6z/local/share/kytea/model.binzMake sure you install KyTea (https://github.com/neubig/kytea) and it's python wrapper (https://github.com/chezou/Mykytea-python) with the following stepsz81. git clone git@github.com:neubig/kytea.git && cd kyteaz2. autoreconf -iz#3. ./configure --prefix=$HOME/localz4. make && make installz5. pip install kytea) roMykyteaospath expanduserAttributeErrorrirferrorlistZgetWS)r~rrrrr ja_tokenizeQs"       zFlaubertTokenizer.ja_tokenizecCs t|jSr)rmrsrrrr vocab_sizefszFlaubertTokenizer.vocab_sizecCst|jf|jSr)rwrsZadded_tokens_encoderrrrr get_vocablszFlaubertTokenizer.get_vocabc st|dd|ddf}|jkr2j|St|}|sF|dSt|fddd}|jkrhqf|\}}g}d}|t|krcsj|tdS)Ninf)rzgetfloat)pairrrrzz'FlaubertTokenizer.bpe..keyrrrW z z ) rXr{r#minrzrmindex ValueErrorextendrFrG) r~tokenr r!ZbigramfirstsecondZnew_wordijrrrbpepsF    2      zFlaubertTokenizer.bpecCs:|dddd}t|}td|}|jr6|}|S)Nz``r%z''NFC)r=rrCrr lower)r~rrrrpreprocess_texts  z!FlaubertTokenizer.preprocess_textcCsd}|r"|jr"||jkr"td|r0|}n&||}|j||d}|j||d}g}|D]"}|r^|t| |dq^|S)a Tokenize a string given language code using Moses. Details of tokenization: - [sacremoses](https://github.com/alvations/sacremoses): port of Moses - Install with `pip install sacremoses` Args: - bypass_tokenizer: Allow users to preprocess and tokenize the sentences externally (default = False) (bool). If True, we only apply BPE. Returns: List of tokens. frzSupplied language code not found in lang2id mapping. Please check that your language is supported by the loaded pretrained model.rr) rcrfrrYrrrrrr)r~rZbypass_tokenizerrZ split_tokensrrrr _tokenizes  zFlaubertTokenizer._tokenizecCs|j||j|jS)z0Converts a token (str) in an id using the vocab.)rsrr\)r~rrrr_convert_token_to_idsz&FlaubertTokenizer._convert_token_to_idcCs|j||jS)z=Converts an index (integer) in a token (str) using the vocab.)rurr\)r~rrrr_convert_id_to_tokensz&FlaubertTokenizer._convert_id_to_tokencCsd|dd}|S)z:Converts a sequence of tokens (string) in a single string.rBrr)rGr=strip)r~tokensZ out_stringrrrconvert_tokens_to_stringsz*FlaubertTokenizer.convert_tokens_to_string) 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 XLM 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. N)Z bos_token_id sep_token_id)r~rrZbosseprrr build_inputs_with_special_tokenss  z2FlaubertTokenizer.build_inputs_with_special_tokens)rralready_has_special_tokensrcsf|rtj||ddS|dk rLdgdgt|dgdgt|dgSdgdgt|dgS)a Retrieve 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` method. Args: token_ids_0 (`List[int]`): List of IDs. token_ids_1 (`List[int]`, *optional*): Optional second list of IDs for sequence pairs. already_has_special_tokens (`bool`, *optional*, defaults to `False`): Whether or not the token list is already formatted with special tokens for the model. Returns: `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token. T)rrrNrr)r|get_special_tokens_maskrm)r~rrrrrrrs.z)FlaubertTokenizer.get_special_tokens_maskcCsV|jg}|jg}|dkr.t|||dgSt|||dgt||dgS)a Create a mask from the two sequences passed to be used in a sequence-pair classification task. An XLM 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`, this method only returns the first portion of the mask (0s). 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 [token type IDs](../glossary#token-type-ids) according to the given sequence(s). Nrr)rZ cls_token_idrm)r~rrrclsrrr$create_token_type_ids_from_sequencess z6FlaubertTokenizer.create_token_type_ids_from_sequences)save_directoryfilename_prefixrc Cstj|s"td|ddStj||r6|dndtd}tj||rX|dndtd}t|ddd $}|t j |j d d d d dW5QRXd}t|ddd `}t |j dddD]B\}} || krtd|d| }|d|d|d7}qW5QRX||fS)NzVocabulary path (z) should be a directoryr7rBrr wr rPrWTF)indent sort_keys ensure_asciirUrcSs|dS)Nrr)kvrrrrCrz3FlaubertTokenizer.save_vocabulary..rzSaving vocabulary to zZ: BPE merge indices are not consecutive. Please check that the tokenizer is not corrupted!rr)rrisdirrfrrGVOCAB_FILES_NAMESrpwriterqdumpsrssortedrzrtrg) r~rrrZ merge_filefrwriterZ bpe_tokensZ token_indexrrrsave_vocabulary3s0 ( z!FlaubertTokenizer.save_vocabularycCs|j}d|d<|S)Nrj)__dict__copy)r~staterrr __getstate__Ps zFlaubertTokenizer.__getstate__cCs:||_z ddl}Wntk r.tdYnX||_dS)NrznYou need to install sacremoses to use XLMTokenizer. See https://pypi.org/project/sacremoses/ for installation.)rrhrirj)r~drhrrr __setstate__Vs  zFlaubertTokenizer.__setstate__)F)N)NF)N)N)(__name__ __module__ __qualname____doc__rZvocab_files_namesPRETRAINED_VOCAB_FILES_MAPZpretrained_vocab_files_mapPRETRAINED_INIT_CONFIGURATIONZpretrained_init_configuration&PRETRAINED_POSITIONAL_EMBEDDINGS_SIZESZmax_model_input_sizesr}propertyrrrrrrrrrrrrrrintrrboolrrrrrrr __classcell__rrrrrKs4 S    , &     rK)rrqrr>rCtypingrrrZtokenization_utilsrutilsrZ get_loggerrrfrrrrrr#r@rJrKrrrrsJ   ,