o 4ѹg(@sfddlZddlZddlZddlZddlmZddlmZmZm Z dZ ddZ ddZ Gd d d Z dS) N) defaultdict)ListOptionalTupleuV[’!"#$%&'()*+,-./:;<>=?@，。?★、…【】《》？“”‘’！[\]^_`{|}~\s]+cCsrg}|}t|dkr7td|}|dur|d}n|dd}||||ddd}t|dks |S)Nrz[A-Za-z!?,<>()\']+ )lowerlenrematchgroupappendreplacestrip)Z input_strtokenssr wordrZ/Users/admin/.pyenv/versions/3.10.0/lib/python3.10/site-packages/funasr/utils/kws_utils.pysplit_mixed_label s      rcCst}t}||vr||f}|||f}||fSt|}|D]`}|dks.|dks.|dkr3|d}q |dks;|dkr@|d}q |dksP|dksP|d ksP|d krU|d }q ||vr_||f}q ||vro||D]}||f}qgttd |}|D]}||f}qxq |D]c}||vr|||f}q|dkrd |vr||d f}q||df}q|dkrd|vr||df}q||df}qd|vr||df}td|dq||df}td|dq||fS)N!silz(sil)z)r)rz(noise)znoise)z(noisez)rZsilr'z)' is not in token set, replace with z+' is not in token set, replace with )tuplerr sub symbol_strlogginginfo)txtZ symbol_tableZ lexicon_tableZ tokens_strZ tokens_idxpartspartchrrrquery_token_setsP         r$c@seZdZdZdejjdedede fddZ  dd ej d ej d e de de deeee ej ff ddZddZd ej d ej fddZdej fddZdS)KwsCtcPrefixDecoderz.Decoder interface wrapper for CTCPrefixDecode.ctckeywords token_listseg_dictc s|_|_i}|D] }||||<q dh_i_|_jddd}|D].}t |||\} } ij|<| j|d<d dd| Dj|d<fd d | Dq,d S) zInitialize class. Args: ctc (torch.nn.Module): The CTC implementation. For example, :class:`espnet.nets.pytorch_backend.ctc.CTC` rrr,token_idcss|] }dt|VqdS)z%s N)str.0irrr nsz/KwsCtcPrefixDecoder.__init__..Z token_strcsg|]}j|qSr)keywords_idxsetaddr-selfrr oz0KwsCtcPrefixDecoder.__init__..N) r&r(indexr1keywords_tokenZ keywords_strrrsplitr$join) r4r&r'r(r)Z token_tabletokenZ keywords_listkeywordstrsZindexsrr3r__init__Qs  zKwsCtcPrefixDecoder.__init__Nlogitslogits_lengthskeywords_tokensetscore_beam_sizepath_beam_sizereturnc Cs|d}|}tddgffg}td|D]R} || } tdd} | |\} } g}g}t| | D])\}}|durQ|dkrP||vrP||||q6|dkr_||||q6t|dkrgq|D]}| | }|D]\}\}}}t|dkr|dnd}|dkr| |\}}}|||||}| }|||f| |<qs||krt j |dd d s| |\}}}|||}| }||dd kr||dd <| |dd <|||f| |<t j |dd d s||f}| |\}}}|||}| }|t || |d |||f| |<qs||f}| |\}}}|r6||dd kr5||dd <| |dd <n| }|t || |d |||||}|||f| |<qsqit| dddd} | d|}qdd|D}|S)a CTC prefix beam search inner implementation Args: logits (torch.Tensor): (1, max_len, vocab_size) logits_lengths (torch.Tensor): (1, ) keywords_tokenset (set): token set for filtering score score_beam_size (int): beam size for score path_beam_size (int): beam size for path Returns: List[List[int]]: nbest results r?cSs ddgfS)NrHrrrrrs z1KwsCtcPrefixDecoder.beam_search..Ng?gư>)Zabs_tolprobframe)r;rLrKcSs|dd|ddS)Nrrr)xrrrrIr6T)keyreversecSs6g|]}|d|dd|dd|ddfqS)rrr)r.yrrrr5s6z3KwsCtcPrefixDecoder.beam_search..)sizerrangerZtopkziptolistrr itemcopymathisclosedictsorteditems)r4rArBrCrDrEmaxlenZ ctc_probsZcur_hypstZprobsZ next_hypsZ top_k_probsZ top_k_indexZ filter_probsZ filter_indexrKidxrZpsprefixZpbZpnbZ cur_nodeslastZn_pbZn_pnbnodesZn_prefixhypsrrr beam_searchqs                   *zKwsCtcPrefixDecoder.beam_searchcCst|t|kr dSt|t|kr||krdSdStt|t|D]#}|||dkrGtt|D]}|||||krBq$q4|Sq$dS)NrJr)r rS)r4Z main_listZ check_listr/jrrr is_sublistszKwsCtcPrefixDecoder.is_sublistc Cs||||j}d}d}|D]P}|d}|d}t|t|ks"J|jD]+} |j| d} ||| } | dkrR| }t| | t| D] } ||| d9}qEqSq'|dur^t|}q_q|durhd||fSdS) NrGrrPr+rJrKT)FNN) rdr1r r8keysrfrSrXsqrt) r4rArBrcZ hit_keywordZ hit_scoreZone_hypZ prefix_idsZ prefix_nodesrZlaboffsetr_rrr_decode_insides.   z"KwsCtcPrefixDecoder._decode_insiderMcCs>|j|dd}t|dg}| ||S)zGet an initial state for decoding. Args: x (torch.Tensor): The encoded feature tensor Returns: decode result rrN) r&ZsoftmaxZ unsqueezedetachZsqueezecputorchZtensorrRrj)r4rMZraw_logpxlenrrrdecodes  zKwsCtcPrefixDecoder.decode)Nr?r@)__name__ __module__ __qualname____doc__rmnnModuler,listrZr>ZTensorsetintrrrdrfrjrorrrrr%NsD $ k !r%)r rrmrX collectionsrtypingrrrrrr$r%rrrrs 3