fix(normalization): fix for short strings

OpenPecha · Feb 21, 2024 · 85a3155 · 85a3155
1 parent 84e7d3e
commit 85a3155
Showing 1 changed file with 2 additions and 0 deletions.
diff --git a/botok/utils/unicode_normalization.py b/botok/utils/unicode_normalization.py
@@ -182,6 +182,8 @@ def is_suffix(char):
 
 
 def normalize_invalid_start_string(s):
+    if len(s) < 2:
+        return s
     # we put the vowel in second place if the string starts with a vowel
     if is_vowel(s[0]) and not is_vowel(s[1]) and not is_suffix(s[1]):
         return s[1] + s[0] + (s[2:] if len(s) > 2 else "")