Add uk-ua symbols configuration to include "`" in normalized utterances

OpenVoiceOS · Jun 18, 2024 · d76e6a6 · d76e6a6
1 parent 96b0203
commit d76e6a6
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 1 deletion.
diff --git a/lingua_franca/lang/parse_uk.py b/lingua_franca/lang/parse_uk.py
@@ -1642,7 +1642,8 @@ def extract_numbers_uk(text, short_scale=True, ordinals=False):
 
 
 class UkrainianNormalizer(Normalizer):
- with open(resolve_resource_file("text/uk-ua/normalize.json"), encoding='utf8') as f:
+ with open(resolve_resource_file("text/uk-ua/normalize.json"),
+ encoding='utf8') as f:
  _default_config = json.load(f)
 
 

diff --git a/lingua_franca/res/text/uk-ua/normalize.json b/lingua_franca/res/text/uk-ua/normalize.json
@@ -69,6 +69,8 @@
  "тисяча": "1000",
  "тисяч": "1000"
  },
+ "symbols": [".", ",", ";", "_", "!", "?", "<", ">", "|", "(", ")", "=", "[",
+ "]", "{", "}", "»", "«", "*", "~", "^", "\""],
  "stopwords": [],
  "articles": []
 }