o ei @sdZddlZddlmZddlmZmZddlZddlm Z ddl m Z e e Z dd e jd ed eeeee jffd d Zd e jd eefddZddeeefded e jfddZde jd e jfddZdS)zUtilities for k2 integration with SpeechBrain. This code was adjusted from icefall (https://github.com/k2-fsa/icefall). Authors: * Pierre Champion 2023 * Zeyu Zhao 2023 * Georgios Karakasidis 2023 N)Path)ListUnion) get_logger)k2F best_paths return_raggedreturncCst|jtjr(|jd}|j|j}|d}|d}t||j }n|jd}t||j}|d}|j dksCJ|rG|S| S)aJ Extract the texts (as word IDs) from the best-path FSAs. Arguments --------- best_paths: k2.Fsa A k2.Fsa with best_paths.arcs.num_axes() == 3, i.e. containing multiple FSAs, which is expected to be the result of k2.shortest_path (otherwise the returned values won't be meaningful). return_ragged: bool True to return a ragged tensor with two axes [utt][word_id]. False to return a list-of-list word IDs. Returns ------- Returns a list of lists of int, containing the label sequences we decoded. rr) isinstance aux_labelsr RaggedTensorremove_values_leqarcsshapecompose remove_axisvaluesnum_axestolist)rr r aux_shaper^/home/ubuntu/transcripts/venv/lib/python3.10/site-packages/speechbrain/k2_integration/utils.pylattice_path_to_textids    rcs<t|dd}g}|D]}|dfdd|Dq |S)a Convert the best path to a list of strings. Arguments --------- best_paths: k2.Fsa It is the path in the lattice with the highest score for a given utterance. word_table: List[str] or Dict[int,str] It is a list or dict that maps word IDs to words. Returns ------- texts: List[str] A list of strings, each of which is the decoding result of the corresponding utterance. F)r  csg|]}|qSrr).0wid word_tablerr ^sz)lattice_paths_to_text..)rappendjoin)rrhypstextswidsrrrlattice_paths_to_textGs r&TpathcachecCst|}tj|ddr*|r*td|dtj t j |dddd}|St d|tj |s@td|d t|d d !}tjj|d d }t ||dddWd|S1sjwY|S)a load a lm to be used in the decoding graph creation (or lm rescoring). Arguments --------- path: str The path to an FST LM (ending with .fst.txt) or a k2-converted LM (in pytorch .pt format). cache: bool Whether or not to load/cache the LM from/to the .pt format (in the same dir). Returns ------- G: k2.Fsa An FSA representing the LM. z.fst.txtz.ptz Loading 'z\' from its cached .pt format. Set 'caching: False' in the yaml if this is not what you want.cpu) map_locationzLoading G LM: zFile z2 not found. You need to run arpa_to_fst to get it.zutf-8)encodingF)acceptorNi)strosr'existsreplaceloggerwarningrFsa from_dicttorchloadinfoisfileFileNotFoundErroropen from_openfstreadsaveas_dict)r'r(Gfrrrload_Gbs*    rAr?cCsdd|jvr d|jd<|`tj|gd}t|}t|}t|}t|ds0|j |_ |S)a Prepare a LM with the purpose of using it for LM rescoring. For instance, in the librispeech recipe this is a 4-gram LM (while a 3gram LM is used for HLG construction). Arguments --------- G: k2.Fsa An FSA representing the LM. Returns ------- G: k2.Fsa An FSA representing the LM, with the following modifications: - G.aux_labels is removed - G.lm_scores is set to G.scores - G is arc-sorted _propertiesNr) lm_scores) __dict__r rr3 from_fsastoarc_sortadd_epsilon_self_loopshasattrscoresclonerC)r?rrrprepare_rescoring_Gs       rL)F)T)__doc__r.pathlibrtypingrrr5speechbrain.utils.loggerrr__name__r1r3boolintrrr-r&rArLrrrrs&    .")