U 9%eT@sddlZddlmZddlmZmZmZmZmZddl Z ddl m Z m Z mZddlmZeeZdZdd d Zd d d dddddddddddddiZdddddddddddddddZdddgddddd d!d"gd#Zddddd d!d"d$ZGd%d&d&eZdS)'N)copyfile)AnyDictListOptionalTuple) AddedToken BatchEncodingPreTrainedTokenizer)loggingu▁zsentencepiece.bpe.modelztokenizer.json) vocab_filetokenizer_filer zOhttps://huggingface.co/uclanlp/plbart-base/resolve/main/sentencepiece.bpe.modelzahttps://huggingface.co/uclanlp/plbart-c-cpp-defect-detection/resolve/main/sentencepiece.bpe.modelzRhttps://huggingface.co/uclanlp/plbart-cs-java/resolve/main/sentencepiece.bpe.modelzUhttps://huggingface.co/uclanlp/plbart-en_XX-java/resolve/main/sentencepiece.bpe.modelzShttps://huggingface.co/uclanlp/plbart-go-en_XX/resolve/main/sentencepiece.bpe.modelz_https://huggingface.co/uclanlp/plbart-java-clone-detection/resolve/main/sentencepiece.bpe.modelzRhttps://huggingface.co/uclanlp/plbart-java-cs/resolve/main/sentencepiece.bpe.modelzUhttps://huggingface.co/uclanlp/plbart-java-en_XX/resolve/main/sentencepiece.bpe.modelz[https://huggingface.co/uclanlp/plbart-javascript-en_XX/resolve/main/sentencepiece.bpe.modelzThttps://huggingface.co/uclanlp/plbart-php-en_XX/resolve/main/sentencepiece.bpe.modelzWhttps://huggingface.co/uclanlp/plbart-python-en_XX/resolve/main/sentencepiece.bpe.modelz]https://huggingface.co/uclanlp/plbart-refine-java-medium/resolve/main/sentencepiece.bpe.modelz\https://huggingface.co/uclanlp/plbart-refine-java-small/resolve/main/sentencepiece.bpe.modelzUhttps://huggingface.co/uclanlp/plbart-ruby-en_XX/resolve/main/sentencepiece.bpe.model)zuclanlp/plbart-basez%uclanlp/plbart-c-cpp-defect-detectionzuclanlp/plbart-cs-javazuclanlp/plbart-en_XX-javazuclanlp/plbart-go-en_XXz#uclanlp/plbart-java-clone-detectionzuclanlp/plbart-java-cszuclanlp/plbart-java-en_XXzuclanlp/plbart-javascript-en_XXzuclanlp/plbart-php-en_XXzuclanlp/plbart-python-en_XXz!uclanlp/plbart-refine-java-mediumz uclanlp/plbart-refine-java-smallzuclanlp/plbart-ruby-en_XXiZ__java__Z __python__ __en_XX__Z__javascript__Z__php__Z__ruby__Z__go__)basemulti)javapythonen_XX javascriptphprubygoc seZdZUdZeZeZeZ ddgZ gZ e e ed<gZe e ed<dEeeeefd fdd ZddZddZeddZeedddZejed dddZdFe e ee e ee e dfdd ZdGe e ee e e e dd d!ZdHe e ee e e e dd"d#Zeeeeed$d%d&Zd'd(Z ee ed)d*d+Z!d,d-Z"d.d/Z#d0d1Z$dIeeee%ed2d3d4Z&dJe eeee eee'd7fd8d9 Z(d:d;Z)dd?Z+ed d@dAdBZ,eed@dCdDZ-Z.S)KPLBartTokenizera Construct an PLBART tokenizer. Adapted from [`RobertaTokenizer`] and [`XLNetTokenizer`]. Based on [SentencePiece](https://github.com/google/sentencepiece). The tokenization method is ` ` for source language documents, and ` ` for target language documents. Args: vocab_file (`str`): Path to the vocabulary file. src_lang (`str`, *optional*): A string representing the source language. tgt_lang (`str`, *optional*): A string representing the target language. bos_token (`str`, *optional*, defaults to `""`): The start of sequence token. eos_token (`str`, *optional*, defaults to `""`): The end of sequence 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. cls_token (`str`, *optional*, defaults to `""`): The cls token, which is a special token used as the first token for all tasks. 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. mask_token(`str`, *optional*, defaults to `""`): The token used for masking values. This is the token used when training this model with masking tasks. This is only used in the `"base"` tokenizer type. For `"multi"` tokenizer, masking is never done for the downstream tasks. language_codes (`str`, *optional*, defaults to `"base"`): What language codes to use. Should be one of `"base"` or `"multi"`. 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. Examples: ```python >>> from transformers import PLBartTokenizer >>> tokenizer = PLBartTokenizer.from_pretrained("uclanlp/plbart-python-en_XX", src_lang="python", tgt_lang="en_XX") >>> example_python_phrase = "def maximum(a,b,c):NEW_LINE_INDENTreturn max([a,b,c])" >>> expected_translation_english = "Returns the maximum value of a b c." >>> inputs = tokenizer(example_python_phrase, text_target=expected_translation_english, return_tensors="pt") ```Z input_idsZattention_mask prefix_tokens suffix_tokensrN)sp_model_kwargsc st|trt|dddn|}| dkr(in| _| } | } tjfj_jt||_ | _ t j }ddddd_ d_ tj_fd d t|D_d d jD_j d krtjtjj j d <j jdd j D_tj|dk rFfdd|Dj d krz| _jdk rpjjnj_n"| dk r| nd_jj_tjf|||||||| | | | jd || _jdS)NTF)lstriprstriprr)rrrrcs"i|]\}}|j|jqS) sp_model_sizefairseq_offset).0icodeselfr&m/var/www/html/Darija-Ai-API/env/lib/python3.8/site-packages/transformers/models/plbart/tokenization_plbart.py sz,PLBartTokenizer.__init__..cSsi|]\}}||qSr&r&r)kvr&r&r.r/srr cSsi|]\}}||qSr&r&r0r&r&r.r/scsg|]}|kr|qSr&r&)r)t)_additional_special_tokensr&r. sz,PLBartTokenizer.__init__..r) bos_token eos_token unk_token sep_token cls_token pad_token mask_tokenlanguage_codesrsrc_langtgt_langadditional_special_tokensr!) isinstancestrr r!!_convert_lang_code_special_formatspmSentencePieceProcessorsp_modelLoadr r=FAIRSEQ_LANGUAGE_CODESfairseq_tokens_to_idsr(lenr' enumeratelang_code_to_iditemsZid_to_lang_codeupdatefairseq_ids_to_tokenslistkeysextend _src_langZcur_lang_code_idsuper__init__r?set_src_lang_special_tokens)r-r r6r7r9r:r8r;r<r=rr>r?r!r@kwargsZfairseq_language_codes __class__)r4r-r.rUsd         zPLBartTokenizer.__init__cCs$|j}d|d<|j|d<|S)NrFsp_model_proto)__dict__copyrFserialized_model_proto)r-stater&r&r. __getstate__ s zPLBartTokenizer.__getstate__cCs8||_t|dsi|_tjf|j|_|j|jdS)Nr!)r[hasattrr!rDrErFZLoadFromSerializedProtorZ)r-dr&r&r. __setstate__s  zPLBartTokenizer.__setstate__cCsF|jdkr(t|jt|j|jdSt|jt|j|jSdS)Nrr$)r=rJrFrLr(r,r&r&r. vocab_sizes zPLBartTokenizer.vocab_size)returncCs|jSN)rSr,r&r&r.r>%szPLBartTokenizer.src_lang) new_src_langrdcCs ||}||_||jdSre)rCrSrV)r-rfr&r&r.r>)s F) token_ids_0 token_ids_1already_has_special_tokensrdcsx|rtj||ddSdgt|j}dgt|j}|dkrT|dgt||S|dgt|dgt||S)a Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding special tokens using the tokenizer `prepare_for_model` method. Args: token_ids_0 (`List[int]`): List of IDs. token_ids_1 (`List[int]`, *optional*): Optional second list of IDs for sequence pairs. already_has_special_tokens (`bool`, *optional*, defaults to `False`): Whether or not the token list is already formatted with special tokens for the model. Returns: `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token. T)rgrhrir$Nr)rTget_special_tokens_maskrJrr)r-rgrhriZ prefix_onesZ suffix_onesrXr&r.rj/sz'PLBartTokenizer.get_special_tokens_mask)rgrhrdcCs,|dkr|j||jS|j|||jS)ac Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens. An PLBART sequence has the following format, where `X` represents the sequence: - `input_ids` (for encoder) `X [eos, src_lang_code]` - `decoder_input_ids`: (for decoder) `X [eos, tgt_lang_code]` BOS is never used. Pairs of sequences are not the expected use case, but they will be handled without a separator. 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)rr)r-rgrhr&r&r. build_inputs_with_special_tokensMsz0PLBartTokenizer.build_inputs_with_special_tokenscCsP|jg}|jg}|dkr.t|||dgSt||||||dgS)a Create a mask from the two sequences passed to be used in a sequence-pair classification task. PLBart 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. Nr)Z sep_token_idZ cls_token_idrJ)r-rgrhsepclsr&r&r.$create_token_type_ids_from_sequenceshs z4PLBartTokenizer.create_token_type_ids_from_sequences)return_tensorsr>r?cKs^|dks|dkrtd|||_|||_||fd|d|}||j}||d<|S)zIUsed by translation pipeline, to prepare inputs for the generate functionNzATranslation requires a `src_lang` and a `tgt_lang` for this modelT)Zadd_special_tokensroZforced_bos_token_id) ValueErrorrCr>r?Zconvert_tokens_to_ids)r-Z raw_inputsror>r? extra_kwargsinputsZ tgt_lang_idr&r&r._build_translation_inputss   z)PLBartTokenizer._build_translation_inputscs(fddtjD}|j|S)Ncsi|]}||qSr&)Zconvert_ids_to_tokens)r)r*r,r&r.r/sz-PLBartTokenizer.get_vocab..)rangercrNZadded_tokens_encoder)r-Zvocabr&r,r. get_vocabs zPLBartTokenizer.get_vocab)textrdcCs|jj|tdS)N)Zout_type)rFencoderB)r-rvr&r&r. _tokenizeszPLBartTokenizer._tokenizecCs4||jkr|j|S|j|}|r.||jS|jS)z0Converts a token (str) in an id using the vocab.)rIrFZ PieceToIdr(Z unk_token_id)r-tokenZspm_idr&r&r._convert_token_to_ids   z$PLBartTokenizer._convert_token_to_idcCs&||jkr|j|S|j||jS)z=Converts an index (integer) in a token (str) using the vocab.)rOrFZ IdToPiecer()r-indexr&r&r._convert_id_to_tokens  z$PLBartTokenizer._convert_id_to_tokencCsd|td}|S)zIConverts a sequence of tokens (strings for sub-words) in a single string. )joinreplaceSPIECE_UNDERLINEstrip)r-tokensZ out_stringr&r&r.convert_tokens_to_stringsz(PLBartTokenizer.convert_tokens_to_string)save_directoryfilename_prefixrdc Cstj|s"td|ddStj||r6|dndtd}tj|jtj|krztj |jrzt |j|n8tj |jst |d}|j }||W5QRX|fS)NzVocabulary path (z) should be a directory-r}r wb)ospathisdirloggererrorrVOCAB_FILES_NAMESabspathr isfileropenrFr]write)r-rrZout_vocab_filefiZcontent_spiece_modelr&r&r.save_vocabularys (  zPLBartTokenizer.save_vocabularyrr) src_textsr> tgt_textsr?rdc s*|||_|||_tj||f|Sre)rCr>r?rTprepare_seq2seq_batch)r-rr>rr?rWrXr&r.rs  z%PLBartTokenizer.prepare_seq2seq_batchcCs ||jSre)rVr>r,r&r&r._switch_to_input_modesz%PLBartTokenizer._switch_to_input_modecCs ||jSre)set_tgt_lang_special_tokensr?r,r&r&r._switch_to_target_modesz&PLBartTokenizer._switch_to_target_modecCsP||}|dk r|j|nd|_g|_|jdk rB|j|jg|_n |jg|_dS)z_Reset the special tokens to the source lang setting. No prefix and suffix=[eos, src_lang_code].NrCrLZ cur_lang_coderZ eos_token_idr)r-r>r&r&r.rVs   z+PLBartTokenizer.set_src_lang_special_tokens)langrdcCsP||}|dk r|j|nd|_g|_|jdk rB|j|jg|_n |jg|_dS)zcReset the special tokens to the target language setting. No prefix and suffix=[eos, tgt_lang_code].Nrr-rr&r&r.rs   z+PLBartTokenizer.set_tgt_lang_special_tokenscCs|tkrt|n|}|S)z;Convert Language Codes to format tokenizer uses if required)FAIRSEQ_LANGUAGE_CODES_MAPrQrr&r&r.rCsz1PLBartTokenizer._convert_lang_code_special_format) rrrrrrr rNNNNN)NF)N)N)N)rNr)/__name__ __module__ __qualname____doc__rZvocab_files_names&PRETRAINED_POSITIONAL_EMBEDDINGS_SIZESZmax_model_input_sizesPRETRAINED_VOCAB_FILES_MAPZpretrained_vocab_files_mapZmodel_input_namesrrint__annotations__rrrrBrrUr_rbpropertyrcr>setterboolrjrkrnrsrurxrzr|rrrr rrrrVrrC __classcell__r&r&rXr.rjs =\             r)rshutilrtypingrrrrrZ sentencepiecerDZtokenization_utilsr r r utilsr Z get_loggerrrrrrrrHrrr&r&r&r.sh    *