Cached implementation for GetLetters API

[arcturusannamalai]

We may repeatedly access letter form of words. In such cases we need to avoid recomputing and use the Cached API.

Resolution of this issue provides a CachedGetLetters mixin.

[arcturusannamalai]

All these uses can be cached within spell checker:

dom.py:26: self.letters = utf8.get_letters(word)
heuristics.py:12:def get_letters(word):
heuristics.py:16: chars = utf8.get_letters(word)
heuristics.py:36: chars = get_letters(word)
heuristics.py:87: chars = get_letters(word)
heuristics.py:109: chars = get_letters(word)
morphology.py:42: word_lett = utf8.get_letters(word)
morphology.py:50: #print(itr,utf8.get_letters(suffix))
morphology.py:54: #print('L-match-->',utf8.get_letters(longest_match))
morphology.py:59: for itr in range(len(utf8.get_letters(longest_match))):
morphology.py:84: word_lett = utf8.get_letters(word)
morphology.py:173: print(utf8.get_letters(w),'->',rval[1])
scoring.py:107: letters = utf8.get_letters(word)
suggestions.py:9:from tamil.utf8 import tamil_letters, get_letters, mei_letters, agaram_letters, sanskrit_letters,
suggestions.py:17: wordL = get_letters(word)
typographical.py:14:from tamil.utf8 import get_letters_elementary, pulli_symbols