o 0jW@sddlZddlZddlZddlZddlmZddlmZmZddl m Z ddl m Z m Z mZmZddlmZd gZed d Zd d ZddZddZGdddeZGdd d e ZdS)N) lru_cache)ListOptional)logging)PretrainedTokenizer _is_control_is_punctuation_is_whitespace) AddedToken CLIPTokenizercCstttdtddtttdtddtttdtdd}|dd}d }td D]}||vrI|||d ||d7}q3d d |D}tt||S) a8 Returns list of utf-8 byte and a mapping to unicode strings. We specifically avoids mapping to whitespace/control characters the bpe code barfs on. The reversible bpe codes work on unicode strings. This means you need a large # of unicode characters in your vocab if you want to avoid UNKs. When you're at something like a 10B token dataset you end up needing around 5K for decent coverage. This is a significant percentage of your normal, say, 32K bpe vocab. To avoid that, we want lookup tables between utf-8 bytes and unicode strings. !~r¡¬®ÿNrcSsg|]}t|qS)chr).0nrry/var/www/html/Deteccion_Ine/venv/lib/python3.10/site-packages/paddlex/inference/models/common/tokenizer/clip_tokenizer.py :sz$bytes_to_unicode..)listrangeordappenddictzip)bscsrbrrrbytes_to_unicode#s     r$cCs6t}|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). rrN)setadd)wordpairsZ prev_charcharrrr get_pairs>s r*cCs|dd|}|}|S)Nz\s+ )substrip)textrerrrwhitespace_cleanLsr0cCs|}|sgS|}|S)z@Runs basic whitespace cleaning and splitting on a piece of text.)r-split)r.tokensrrrwhitespace_tokenizeRs r3c@sXeZdZdZ     dddZdddZdd 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|durg}||_t||_||_||_||_dSN) do_lower_caser% never_splittokenize_chinese_chars strip_accentsdo_split_on_punc)selfr6r7r8r9r:rrr__init__ss  zBasicTokenizer.__init__cCs|r |jt|n|j}||}|jr||}td|}t|}g}|D])}||vrH|j r@| }|j dur?| |}n|j rH| |}| |||q(td|}|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. NFCFr+)r7unionr% _clean_textr8_tokenize_chinese_chars unicodedata normalizer3r6lowerr9_run_strip_accentsextend_run_split_on_puncjoin)r;r.r7Zunicode_normalized_textZ orig_tokensZ split_tokenstokenZ output_tokensrrrtokenizes,       zBasicTokenizer.tokenizecCsBtd|}g}|D]}t|}|dkrq ||q d|S)z$Strips accents from a piece of text.ZNFDZMn)rArBcategoryrrG)r;r.outputr)catrrrrDs    z!BasicTokenizer._run_strip_accentscCs|jr |dur||vr|gSt|}d}d}g}|t|krI||}t|r/||gd}n|r6|gd}|d||d7}|t|ksdd|DS) z&Splits punctuation on a piece of text.NrTFrcSsg|]}d|qS)rJ)rG)rxrrrrz5BasicTokenizer._run_split_on_punc..)r:rlenr r)r;r.r7charsiZstart_new_wordrLr)rrrrFs&     z!BasicTokenizer._run_split_on_punccCsTg}|D] }t|}||r|d|||dq||qd|S)z)Adds whitespace around any CJK character.r+rJ)r_is_chinese_charrrGr;r.rLr)cprrrr@s      z&BasicTokenizer._tokenize_chinese_charscCs|dkr|dks@|dkr|dks@|dkr|dks@|dkr |dks@|d kr(|d ks@|d kr0|d ks@|d kr8|dks@|dkrB|dkrBdSdS)z6Checks whether CP is the codepoint of a CJK character.iNii4iMiiߦii?i@ii iiiiiTFr)r;rVrrrrTs zBasicTokenizer._is_chinese_charcCsXg}|D]"}t|}|dks|dkst|rqt|r!|dq||qd|S)zBPerforms invalid character removal and whitespace cleanup on text.rir+rJ)rr r rrGrUrrrr?s   zBasicTokenizer._clean_text)TNTNTr5) __name__ __module__ __qualname____doc__r<rIrDrFr@rTr?rrrrr4\s *  r4cseZdZdZdddZiidZiZddgZ   d*d d Ze d dZ ddZ d+de e dee e de e fddZ d+ddZ d,fdd Z d+de e dee e de e fddZddZd d!Zd"d#Zd$d%Zd&d'Zd(d)ZZS)-r a Construct a CLIP tokenizer based on byte-level Byte-Pair-Encoding. This tokenizer inherits from :class:`~paddlenlp.transformers.gpt.tokenizer.GPTTokenizer`. For more information regarding those methods, please refer to this superclass. Args: vocab_file (str): Path to the vocabulary file. The vocab file contains a mapping from vocabulary strings to indices. merges_file (str): Path to the merge file. The merge file is used to split the input sentence into "subword" units. The vocab file is then used to encode those units as intices. errors (str): Paradigm to follow when decoding bytes to UTF-8. Defaults to `'replace'`. max_len (int, optional): The maximum value of the input sequence length. Defaults to `77`. bos_token (str, optional): The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token. Defaults to `"<|startoftext|>"`. eos_token (str, optional): A special token representing the end of a sequence that was used during pretraining. Defaults to `"<|endoftext|>"`. unk_token (str, optional): A special token representing the *unknown (out-of-vocabulary)* token. An unknown token is set to be `unk_token` inorder to be converted to an ID. Defaults to `"<|endoftext|>"`. pad_token (str, optional): A special token used to make arrays of tokens the same size for batching purposes. Defaults to `"<|endoftext|>"`. Examples: .. code-block:: from paddlenlp.transformers import AutoTokenizer tokenizer = AutoTokenizer.from_pretrained('openai/clip-vit-base-patch32') print(tokenizer('He was a puppeteer')) ''' {'input_ids': [49406, 797, 739, 320, 7116, 38820, 528, 49407]} ''' z vocab.jsonz merges.txt) vocab_file merges_fileZ input_idsZattention_maskreplaceM<|startoftext|> <|endoftext|>c Ksddlm} t|trt|dddn|}t|tr t|dddn|}t|tr.t|dddn|}t|tr|rPz*CLIPTokenizer.__init__..cSrkrrrlrrrrorP ricSsg|]}t|qSr)tupler1)rmergerrrrsz*CLIPTokenizer.__init__..r_r`)r_r`z[<\|startoftext\|>|<\|endoftext\|>|'s|'t|'re|'ve|'m|'ll|'d|[\p{L}]+|[\p{N}]|[^\s\p{L}\p{N}]+)(Z paddle.utilsra isinstancestrr Z"_build_special_tokens_map_extendedftfyfix_text ImportErrorrinfor4nlpr/Z _vocab_fileZ _merges_fileintmax_lenopenjsonloadencoderitemsdecodererrorsr$ byte_encoder byte_decoderreadr-r1rr rrQ bpe_rankscachecompile IGNORECASEpat)r;r[r\rr{rdrerfrgkwargsraruZ vocab_handleZ merges_handleZ bpe_mergesrrrr<>sv      zCLIPTokenizer.__init__cCs t|jS)z Returns the size of vocabulary. Returns: int: The sum of size of vocabulary and the size of special tokens. )rQrr;rrr vocab_sizes zCLIPTokenizer.vocab_sizecCst|jfi|jSr5)rrZadded_tokens_encoderrrrr get_vocabszCLIPTokenizer.get_vocabN token_ids_0 token_ids_1returncCs<|jg}|jg}|dur|||S||||||S)a Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens. A CLIP sequence has the following format: - single sequence: `<|startoftext|> X <|endoftext|>` 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 with the appropriate special tokens. N) bos_token_id eos_token_idr;rrrdrerrr build_inputs_with_special_tokenss  z.CLIPTokenizer.build_inputs_with_special_tokenscCs4|dur dg|dgSdg|ddg|dgS)aQ Build offset map from a pair of offset map by concatenating and adding offsets of special tokens. Should be overridden in a subclass if the model has a special way of building those. Args: offset_mapping_0 (List[tuple]): List of char offsets to which the special tokens will be added. offset_mapping_1 (List[tuple], optional): Optional second list of char offsets for offset mapping pairs. Returns: List[tuple]: List of char offsets with the appropriate offsets of special tokens. N)rrr)r;Zoffset_mapping_0Zoffset_mapping_1rrr(build_offset_mapping_with_special_tokenssz6CLIPTokenizer.build_offset_mapping_with_special_tokensFcsl|r tj||ddS|durdgdgt|dgSdgdgt|dgdgdgt|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)rralready_has_special_tokensNrr)superget_special_tokens_maskrQ)r;rrr __class__rrrs2z%CLIPTokenizer.get_special_tokens_maskcCsP|jg}|jg}|durt|||dgSt||||||dgS)a Create a mask from the two sequences passed. CLIP 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)rrrQrrrr$create_token_type_ids_from_sequencessz2CLIPTokenizer.create_token_type_ids_from_sequencesc sr|jvr j|St|dd|ddf}t|}|s#|dS t|fddd}|jvr4ny|\}}g}d}|t|krz|||} Wnty\|||dYn?w|||| | }|||kr|t|dkr||d|kr| |||d 7}n | |||d7}|t|ksBt|}|}t|dkrnt|}q$d |}|j|<|S) NrNTcsj|tdS)Ninf)rgetfloat)pairrrrsz#CLIPTokenizer.bpe..)keyrrr+) rrqr*minrrQindex ValueErrorrErrG) r;rHr'r(ZbigramfirstsecondZnew_wordrSjrrrbpe sJ     ,     zCLIPTokenizer.bpecsg}jdurdj|}n t|j}jj|D]!}dfdd| dD}| dd | dDq$|S)zTokenize a string.Nr+rJc3s|]}j|VqdSr5)r)rr#rrr =s  z*CLIPTokenizer._tokenize..ricss|]}|VqdSr5r)rZ bpe_tokenrrrr@s) rvrGryrIr0r/rCfindallrencoderErr1)r;r.Z bpe_tokensrHrrr _tokenize4s "zCLIPTokenizer._tokenizecCs|j||j|jS)z0Converts a token (str) in an id using the vocab.)rrrf)r;rHrrr_convert_token_to_idCsz"CLIPTokenizer._convert_token_to_idcCs |j|S)z=Converts an index (integer) in a token (str) using the vocab.)rr)r;rrrr_convert_id_to_tokenGs z"CLIPTokenizer._convert_id_to_tokencs@d|}tfdd|D}|jdjddd}|S)z:Converts a sequence of tokens (string) in a single string.rJcsg|]}j|qSr)r)rcrrrrNrPz:CLIPTokenizer.convert_tokens_to_string..ri)rrr+)rG bytearraydecoderr]r-)r;r2r.Z byte_arrayrrrconvert_tokens_to_stringKs z&CLIPTokenizer.convert_tokens_to_stringcCsX|jD]$\}}t|d|}tj||}tj|tj|kr)t||qdS)z Saves `SentencePiece `__ file (ends with '.spm') under `save_directory`. Args: save_directory (str): Directory to save files into. z_%sN) resource_files_namesrgetattrospathrGabspathshutilcopyfile)r;Zsave_directoryname file_name source_pathZ save_pathrrrsave_resourcesTs zCLIPTokenizer.save_resources)r]r^r_r`r`r`r5)NF)rWrXrYrZrZpretrained_resource_files_mapZpretrained_init_configurationZmodel_input_namesr<propertyrrrrzrrrrrrrrrrr __classcell__rrrrr sX 2   R     !  * )r}rrrA functoolsrtypingrrutilsrZtokenizer_utilsrr r r Ztokenizer_utils_baser __all__r$r*r0r3objectr4r rrrrs$     (