Remove the French OCR cleanup backend for now, a patchset seems to be missing.

mayan/apps/ocr/lang/fra.py

@@ -1,41 +0,0 @@
-# -*- coding: utf-8 -*-
-import re
-
-
-def check_word(word):
-    return word
-
-    ALL_ALPHANUM = re.compile('([0-9a-záéíóúüñ])', re.I)
-    NON_ALPHANUM = re.compile('([^0-9a-záéíóúüñ])', re.I)
-
-    TOO_MANY_VOWELS = re.compile('[aáeéiíoóuúü]{3}', re.I)
-    TOO_MANY_CONSONANTS = re.compile('[bcdfghjklmnñpqrstvwxyz]{5}', re.I)
-    ALL_ALPHA = re.compile('^[a-z]+$', re.I)
-    SINGLE_LETTER_WORDS = re.compile('^[aeoóuy]$', re.I)
-
-    #(L) If a string is longer than 20 characters, it is garbage
-    if len(word) > 20:
-        return None
-
-    #(A) If a string’s ratio of alphanumeric characters to total
-    #characters is less than 50%, the string is garbage
-    if len(ALL_ALPHANUM.findall(word)) < len(word) / 2:
-        return None
-
-    #Remove word if all the letters in the word are non alphanumeric
-    if len(NON_ALPHANUM.findall(word)) == len(word):
-        return None
-
-    #Removed words with too many consecutie vowels
-    if TOO_MANY_VOWELS.findall(word):
-        return None
-
-    #Removed words with too many consecutie consonants
-    if TOO_MANY_CONSONANTS.findall(word):
-        return None
-
-    #Only allow specific single letter words
-    if len(word) == 1 and not SINGLE_LETTER_WORDS.findall(word):
-        return None
-
-    return word