import re, json from collections import defaultdict from nltk.tokenize import word_tokenize, sent_tokenize
index = defaultdict(list)
.