mirror of
https://github.com/storytold/storyteller-ml.git
synced 2026-10-18 07:41:03 +00:00
Brandon Thomas
16 lines
587 B
Python
16 lines
587 B
Python
|
|
import re
|
|
from typing import List
|
|
from text.pipeline_components.phonetization.cmudict import CMUDict
|
|
|
|
# Regex taken from open source tacotron2 extensions
|
|
_arpabet_split_re = re.compile(r'[^\s]*{.*?}[^\s]*|[\w]+|[^\w\s]+')
|
|
|
|
class BasePhonetization:
|
|
def __init__(self, phoneme_dictionary: CMUDict):
|
|
self.phoneme_dictionary = phoneme_dictionary
|
|
|
|
def _split_grapheme_sentence_to_words(self, text_sequence: str) -> List[str]:
|
|
"""Split a grapheme word sequence to individual grapheme words"""
|
|
return [word for word in _arpabet_split_re.findall(text_sequence)]
|