o i(@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) input_strtokenssr wordrJ/home/ubuntu/.local/lib/python3.10/site-packages/funasr/utils/kws_utils.pysplit_mixed_label s      rcCst}t}||vr||f}|||f}||fSt|}|D]a}|dks.|dks.|dkr3|d}q |dks;|dkr@|d}q |dksP|dksP|d ksP|d krU|d }q ||vr_||f}q ||vrp||D]}||f}qgq ttd |}|D]}||f}qyq |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)rsilr'z)' is not in token set, replace with z+' is not in token set, replace with )tuplerr sub symbol_strlogginginfo)txt symbol_table lexicon_table tokens_str tokens_idxpartspartchrrrquery_token_setsR         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__.. token_strcsg|]}j|qSr)keywords_idxsetaddr3selfrr oz0KwsCtcPrefixDecoder.__init__..N) r,r.indexr8keywords_token keywords_strrrsplitr*join) r;r,r-r.r/ token_tabletoken keywords_listkeywordstrsindexsrr:r__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)NrSrrrrrs z1KwsCtcPrefixDecoder.beam_search..Ng?gư>)abs_tolprobframe)rDrXrWcSs|dd|ddS)Nrrr)xrrrrTr=T)keyreversecSs6g|]}|d|dd|dd|ddfqS)rrr)r4yrrrr<s6z3KwsCtcPrefixDecoder.beam_search..)sizerrangertopkziptolistrr itemcopymathisclosedictsorteditems)r;rLrMrNrOrPmaxlen ctc_probscur_hypstprobs next_hyps top_k_probs top_k_index filter_probs filter_indexrWidxrpsprefixpbpnb cur_nodeslastn_pbn_pnbnodesn_prefixhypsrrr beam_searchqs                   *zKwsCtcPrefixDecoder.beam_searchcCst|t|kr dSt|t|kr||krdSdStt|t|D]#}|||dkrGtt|D]}|||||krBnq4|Sq$dS)NrUr)r r_)r; main_list check_listr5jrrr is_sublistszKwsCtcPrefixDecoder.is_sublistc Cs||||j}d}d}|D]P}|d}|d}t|t|ks"J|jD]+} |j| d} ||| } | dkrR| }t| | t| D] } ||| d9}qEnq'|dur^t|}nq|durhd||fSdS) NrRrr\r1rUrWT)FNN) rr8r r?keysrr_resqrt) r;rLrMr hit_keyword hit_scoreone_hyp prefix_ids prefix_nodesrlaboffsetrtrrr_decode_insides.   z"KwsCtcPrefixDecoder._decode_insiderYcCs>|j|dd}t|dg}| ||S)zGet an initial state for decoding. Args: x (torch.Tensor): The encoded feature tensor Returns: decode result rr) r,softmax unsqueezedetachsqueezecputorchtensorr^r)r;rYraw_logpxlenrrrdecodes  zKwsCtcPrefixDecoder.decode)NrJrK)__name__ __module__ __qualname____doc__rnnModuler2listrgrITensorsetintrrrrrrrrrrr+NsD $ k !r+)r r rre collectionsrtypingrrrrrr*r+rrrrs 3