X-Git-Url: https://git.njae.me.uk/?a=blobdiff_plain;f=language_models.py;fp=language_models.py;h=19f886fcefcb4384184e0bbad108e6925f029bbf;hb=8c4e8509ebd603f878a844b7cec8d0e2375ec8f9;hp=929746888d036fb54de3f1fbf228e296e0bcd027;hpb=4015126014c197871c657e7f746234f87c5f8e59;p=cipher-tools.git

diff --git a/language_models.py b/language_models.py
index 9297468..19f886f 100644
--- a/language_models.py
+++ b/language_models.py
@@ -6,6 +6,8 @@ import unicodedata
 import itertools
 from math import log10
 
+unaccent_specials = ''.maketrans({"’": "'"})
+
 def letters(text):
     """Remove all non-alphabetic characters from a text
     >>> letters('The Quick')
@@ -31,7 +33,8 @@ def unaccent(text):
     >>> unaccent('HÉLLÖ')
     'HELLO'
     """
-    return unicodedata.normalize('NFKD', text).\
+    translated_text = text.translate(unaccent_specials)
+    return unicodedata.normalize('NFKD', translated_text).\
         encode('ascii', 'ignore').\
         decode('utf-8')