nltk regex parser

>>> def ie_preprocess(document):
...    sentences = nltk.sent_tokenize(document) 
...    sentences = [nltk.word_tokenize(sent) for sent in sentences] 
...    sentences = [nltk.pos_tag(sent) for sent in sentences]