[TTS]当文本中包含全角数字的时候会报错

[arashrun]

Describe the bug
文本包含全角数字报错

期望
正常转换

错误信息

special_tokens_map.json
Traceback (most recent call last):
  File "/home/like/tts/tts.py", line 8, in <module>
    tts(text=context, output="output.wav")
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/cli/utils.py", line 328, in _warpper
    return executor_func(self, *args, **kwargs)
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/cli/tts/infer.py", line 684, in __call__
    self.infer(text=text, lang=lang, am=am, spk_id=spk_id)
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/decorator.py", line 232, in fun
    return caller(func, *(extras + args), **kw)
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddle/fluid/dygraph/base.py", line 354, in _decorate_function
    return func(*args, **kwargs)
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/cli/tts/infer.py", line 445, in infer
    frontend_dict = run_frontend(
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/exps/syn_utils.py", line 183, in run_frontend
    input_ids = frontend.get_input_ids(
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_frontend.py", line 408, in get_input_ids
    phonemes = self.get_phonemes(
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_frontend.py", line 374, in get_phonemes
    sentences = self.text_normalizer.normalize(sentence)
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_normalization/text_normlization.py", line 115, in normalize
    sentences = [self.normalize_sentence(sent) for sent in sentences]
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_normalization/text_normlization.py", line 115, in <listcomp>
    sentences = [self.normalize_sentence(sent) for sent in sentences]
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_normalization/text_normlization.py", line 86, in normalize_sentence
    sentence = RE_DATE.sub(replace_date, sentence)
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_normalization/chronology.py", line 104, in replace_date
    result += f"{verbalize_digit(year)}年"
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_normalization/num.py", line 211, in verbalize_digit
    result_symbols = [DIGITS[digit] for digit in value_string]
  File "/home/like/.virtualenvs/tts/lib/python3.9/site-packages/paddlespeech/t2s/frontend/zh_normalization/num.py", line 211, in <listcomp>
    result_symbols = [DIGITS[digit] for digit in value_string]
KeyError: '１'

[yt605155624]

感谢反馈，可以在 TN 里面找个地方加个全角转半角的逻辑，应该不难，欢迎提交修复 pr 🌹🌹🌹

[yt605155624]

@arashrun 我们有全角半角的映射表，可以尝试下

PaddleSpeech/paddlespeech/t2s/frontend/zh_normalization/constants.py

Line 19 in 8ea289a

# 全角半角转换