def _get_all_speakers(l: PreDataList) -> Tuple[SpeakersDict, SpeakersLogDict]:
all_speakers: List[str] = [x.speaker_name for x in l.items()]
all_speakers_count = Counter(all_speakers)
speakers_log = SpeakersLogDict.fromcounter(all_speakers_count)
all_speakers = remove_duplicates_list_orderpreserving(all_speakers)
speakers_dict = SpeakersDict.fromlist(all_speakers)
return speakers_dict, speakers_log
def _get_ds_data(l: PreDataList, speakers_dict: SpeakersDict,
accents: AccentsDict, symbols: SymbolIdDict) -> DsDataList:
result = [
DsData(entry_id=i,
basename=values.name,
speaker_name=values.speaker_name,
speaker_id=speakers_dict[values.speaker_name],
text=values.text,
serialized_symbols=symbols.get_serialized_ids(values.symbols),
serialized_accents=accents.get_serialized_ids(values.accents),
wav_path=values.wav_path,
lang=values.lang,
gender=values.gender) for i, values in enumerate(l.items())
]
return DsDataList(result)
def _get_symbols_id_dict(l: PreDataList) -> SymbolIdDict:
symbols = set()
for x in l.items():
symbols = symbols.union(set(x.symbols))
return SymbolIdDict.init_from_symbols(symbols)
def _get_all_accents(l: PreDataList) -> AccentsDict:
accents = set()
for x in l.items():
accents = accents.union(set(x.accents))
return AccentsDict.init_from_accents(accents)