NATURESPEAK-ML-UTTER/utterance/utils.py

import string, regex

def clean(text: str) -> str:

	s = text.split('\n')

	if(len(s) > 0):
		tok_1 = s[0].split(' ')
		if len(tok_1) > 0 and tok_1[0].strip() in string.punctuation:
				s_1 = ' '.join(tok_1[1:])
				s[0] = s_1.capitalize()	
		else:
			s[0] = s[0].capitalize()

	return '\n'.join(s)

def format(text: str) -> str:

	return text.replace('\r\n', '\n').replace('\n\n', '\n').replace('‘', "'").replace("’", "'")

def fragments(utterance: str):
	frags = []
	sentences = utterance.splitlines()

	PUNCT_RE = regex.compile(r'(\p{Punctuation})') 

	skip_punctuation = ["'"] 

	for s in sentences:
		sf = PUNCT_RE.split(s)
		cum = ""
		for k in sf:
			if len(k) < 1:
				continue
			elif len(k) > 1:
				cum += k
			elif k not in string.punctuation:
				cum += k
			else:
				cum += k
				if k in skip_punctuation:
					continue
				else:
					frags.append(cum)
					cum = ""
		cum += '\n'
		frags.append(cum)

	return frags