def _translate_word(name, lang=''):
""" Helper to get word translations
Args:
name (str): Word name. Returned as the default value if not translated
lang (str): Language code, e.g. "en-us"
Returns:
str: translated version of resource name
"""
from lingua_franca.internal import resolve_resource_file
if not lang:
if lang is None:
warn(NoneLangWarning)
lang = get_default_loc()
lang_code = lang if is_supported_full_lang(lang) else \
get_full_lang_code(lang)
filename = resolve_resource_file(join("text", lang_code, name + ".word"))
if filename:
# open the file
try:
with open(filename, 'r', encoding='utf8') as f:
for line in f:
word = line.strip()
if word.startswith("#"):
continue # skip comment lines
return word
except Exception:
pass
return name # use resource name as the word
class PortugueseNormalizer(Normalizer):
with open(resolve_resource_file("text/pt-pt/normalize.json")) as f:
_default_config = json.load(f)
@staticmethod
def tokenize(utterance):
# Split things like 12%
utterance = re.sub(r"([0-9]+)([\%])", r"\1 \2", utterance)
# Split things like #1
utterance = re.sub(r"(\#)([0-9]+\b)", r"\1 \2", utterance)
# Split things like amo-te
utterance = re.sub(r"([a-zA-Z]+)(-)([a-zA-Z]+\b)", r"\1 \2 \3",
utterance)
tokens = utterance.split()
if tokens[-1] == '-':
tokens = tokens[:-1]
return tokens