fix: invalid chars in french.txt

2026-06-16 11:31:50 +02:00
parent 61086e0b9b
commit d866b8fbb2
1 changed files with 2 additions and 4 deletions
@@ -1,5 +1,6 @@
 import logging
 import pathlib
+import re
 import typing

 import unidecode
@@ -22,10 +23,7 @@ class WordList:
        self.logger.info("[%s] Loaded %d words", self.__name, len(self.words))

    def __escape_word(self, word: str) -> str:
-        word = unidecode.unidecode(word.strip()).lower()
-        if " " in word:
-            return ""
-        return word
+        return re.sub(r"[^a-z0-9-]", "", unidecode.unidecode(word.strip()).lower())

    def close(self) -> None:
        pass