lenML/ChatTTS-Forge

1# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.2#3# Licensed under the Apache License, Version 2.0 (the "License");4# you may not use this file except in compliance with the License.5# You may obtain a copy of the License at6#7#     http://www.apache.org/licenses/LICENSE-2.08#9# Unless required by applicable law or agreed to in writing, software10# distributed under the License is distributed on an "AS IS" BASIS,11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.12# See the License for the specific language governing permissions and13# limitations under the License.14import re15from typing import List16 17from .char_convert import tranditional_to_simplified18from .chronology import (19    RE_DATE,20    RE_DATE2,21    RE_TIME,22    RE_TIME_RANGE,23    replace_date,24    replace_date2,25    replace_time,26)27from .constants import F2H_ASCII_LETTERS, F2H_DIGITS, F2H_SPACE28from .num import (29    RE_DECIMAL_NUM,30    RE_DEFAULT_NUM,31    RE_FRAC,32    RE_INTEGER,33    RE_NUMBER,34    RE_PERCENTAGE,35    RE_POSITIVE_QUANTIFIERS,36    RE_RANGE,37    replace_default_num,38    replace_frac,39    replace_negative_num,40    replace_number,41    replace_percentage,42    replace_positive_quantifier,43    replace_range,44)45from .phonecode import (46    RE_MOBILE_PHONE,47    RE_NATIONAL_UNIFORM_NUMBER,48    RE_TELEPHONE,49    replace_mobile,50    replace_phone,51)52from .quantifier import RE_TEMPERATURE, replace_measure, replace_temperature53 54 55class TextNormalizer:56    def __init__(self):57        self.SENTENCE_SPLITOR = re.compile(r"([：、，；。？！,;?!][”’]?)")58 59    def _split(self, text: str, lang="zh") -> List[str]:60        """Split long text into sentences with sentence-splitting punctuations.61        Args:62            text (str): The input text.63        Returns:64            List[str]: Sentences.65        """66        # Only for pure Chinese here67        if lang == "zh":68            text = text.replace(" ", "")69            # 过滤掉特殊字符70            text = re.sub(r"[——《》【】<=>{}()（）#&@“”^_|…\\]", "", text)71        text = self.SENTENCE_SPLITOR.sub(r"\1\n", text)72        text = text.strip()73        sentences = [sentence.strip() for sentence in re.split(r"\n+", text)]74        return sentences75 76    def _post_replace(self, sentence: str) -> str:77        # sentence = sentence.replace('/', '每')78        # sentence = sentence.replace('~', '至')79        # sentence = sentence.replace('～', '至')80        sentence = sentence.replace("①", "一")81        sentence = sentence.replace("②", "二")82        sentence = sentence.replace("③", "三")83        sentence = sentence.replace("④", "四")84        sentence = sentence.replace("⑤", "五")85        sentence = sentence.replace("⑥", "六")86        sentence = sentence.replace("⑦", "七")87        sentence = sentence.replace("⑧", "八")88        sentence = sentence.replace("⑨", "九")89        sentence = sentence.replace("⑩", "十")90        sentence = sentence.replace("α", "阿尔法")91        sentence = sentence.replace("β", "贝塔")92        sentence = sentence.replace("γ", "伽玛").replace("Γ", "伽玛")93        sentence = sentence.replace("δ", "德尔塔").replace("Δ", "德尔塔")94        sentence = sentence.replace("ε", "艾普西龙")95        sentence = sentence.replace("ζ", "捷塔")96        sentence = sentence.replace("η", "依塔")97        sentence = sentence.replace("θ", "西塔").replace("Θ", "西塔")98        sentence = sentence.replace("ι", "艾欧塔")99        sentence = sentence.replace("κ", "喀帕")100        sentence = sentence.replace("λ", "拉姆达").replace("Λ", "拉姆达")101        sentence = sentence.replace("μ", "缪")102        sentence = sentence.replace("ν", "拗")103        sentence = sentence.replace("ξ", "克西").replace("Ξ", "克西")104        sentence = sentence.replace("ο", "欧米克伦")105        sentence = sentence.replace("π", "派").replace("Π", "派")106        sentence = sentence.replace("ρ", "肉")107        sentence = (108            sentence.replace("ς", "西格玛")109            .replace("Σ", "西格玛")110            .replace("σ", "西格玛")111        )112        sentence = sentence.replace("τ", "套")113        sentence = sentence.replace("υ", "宇普西龙")114        sentence = sentence.replace("φ", "服艾").replace("Φ", "服艾")115        sentence = sentence.replace("χ", "器")116        sentence = sentence.replace("ψ", "普赛").replace("Ψ", "普赛")117        sentence = sentence.replace("ω", "欧米伽").replace("Ω", "欧米伽")118        # re filter special characters, have one more character "-" than line 68119        # sentence = re.sub(r'[-——《》【】<=>{}()（）#&@“”^_|…\\]', '', sentence)120        return sentence121 122    def normalize_sentence(self, sentence: str) -> str:123        # basic character conversions124        sentence = tranditional_to_simplified(sentence)125        sentence = (126            sentence.translate(F2H_ASCII_LETTERS)127            .translate(F2H_DIGITS)128            .translate(F2H_SPACE)129        )130 131        # number related NSW verbalization132        sentence = RE_DATE.sub(replace_date, sentence)133        sentence = RE_DATE2.sub(replace_date2, sentence)134 135        # range first136        sentence = RE_TIME_RANGE.sub(replace_time, sentence)137        sentence = RE_TIME.sub(replace_time, sentence)138 139        sentence = RE_TEMPERATURE.sub(replace_temperature, sentence)140        sentence = replace_measure(sentence)141        sentence = RE_FRAC.sub(replace_frac, sentence)142        sentence = RE_PERCENTAGE.sub(replace_percentage, sentence)143        sentence = RE_MOBILE_PHONE.sub(replace_mobile, sentence)144 145        sentence = RE_TELEPHONE.sub(replace_phone, sentence)146        sentence = RE_NATIONAL_UNIFORM_NUMBER.sub(replace_phone, sentence)147 148        sentence = RE_RANGE.sub(replace_range, sentence)149        sentence = RE_INTEGER.sub(replace_negative_num, sentence)150        sentence = RE_DECIMAL_NUM.sub(replace_number, sentence)151        sentence = RE_POSITIVE_QUANTIFIERS.sub(replace_positive_quantifier, sentence)152        sentence = RE_DEFAULT_NUM.sub(replace_default_num, sentence)153        sentence = RE_NUMBER.sub(replace_number, sentence)154        sentence = self._post_replace(sentence)155 156        return sentence157 158    def normalize(self, text: str, lang="") -> List[str]:159        sentences = self._split(text, lang)160        sentences = [self.normalize_sentence(sent) for sent in sentences]161        return sentences162