matching.py

#! /usr/bin/env python

import nltk
import string
import question_analysis
import preprocessing
import MicrosoftNgram
import itertools

from nltk.corpus import wordnet as wn


list_tuples = preprocessing.actual_tuples

list_tuples.sort()

keywords = question_analysis.keywords_analysis 

tuples_needed = []

concept = []
for i in range(len(keywords)):
	for j in range(len(list_tuples)):
		if( list_tuples[j].startswith(keywords[i]+",")):
			tuples_needed.append(list_tuples[j])

#print tuples_needed

concept2 = []

for i in range(len(tuples_needed)):
	a = tuples_needed[i].split(",")
	b = nltk.word_tokenize(a[1])
	
	for j in range(len(b)):
		concept2.append(b[j])



tuples_normalized = []
	
s = MicrosoftNgram.LookupService()

string_perm = itertools.permutations(keywords)

string = []

for i in string_perm:
	string.append(" ".join(i))


results = []

for i in range(len(string)):
	for t in s.Generate(string[i], maxgen=5): results.append(t)

answers = []

for i in range(len(results)):
	 answers.append(results[i][0])

concept2 = list(set(concept2))

list_answers =  list( set(concept2) & set(answers) )

print "the most probable answers for your question"
print list_answers 

"""


# Needs some editing to filter some answers from the list_answers to get the exact answer(s)

synsets = []

for i in keywords:
	 synsets.append(wn.synsets(i))


wordnet_words = []


for i in range(len(synsets)):
	myString = str(synsets[i])
	wordnet_words.append(myString[myString.find("'")+1:myString.find("')")])

lemma_words = []

for i in wordnet_words:
	w = wn.synset(i).hyponyms()
	for j in w:
		lemma_words.append(j.lemma_names)

print lemma_words 

"""