def drawing(tree):
if not (tree is None):
tree.draw()
def drawing(tree):
tree.draw()
print("Example sentences")
print(str_sentence1)
print(str_sentence2)
tokens1 = nltk.word_tokenize(str_sentence1)
tokens2 = nltk.word_tokenize(str_sentence2)
# Create the Chart and Viterbi parsers, with the input grammar
chart_parser = nltk.ChartParser(grammar)
viterbi_parser = nltk.ViterbiParser(grammar)
# Results for the Chart Parser
print("Parse trees obtained with the Chart parser")
print("Sentence 1")
for tree in chart_parser.parse(tokens1):
print(tree)
tree.draw()
print("Sentence 2")
for tree in chart_parser.parse(tokens2):
print(tree)
tree.draw()
# Results for the Viterbi Parser
print("Parse trees obtained with the Viterbi parser")
print("Sentence 1")
for tree in viterbi_parser.parse(tokens1):
print(tree)
tree.draw()
print("Sentence 2")
for tree in viterbi_parser.parse(tokens2):
def namedEntityRecognition(self,sentence):
tokens = nltk.word_tokenize(sentence)
pos_tags = nltk.pos_tag(tokens)
#print nltk.ne_chunk(pos_tags, binary=True)
grammar = "NP: {<DT|PP\$>?<JJ>*<NNP>+}"
"""
NP: {<DT|PP\$>?<JJ>*<NN>} # chunk determiner/possessive, adjectives and proper nouns
{<NNP>+} # chunk sequences of proper nouns
"""
cp = nltk.RegexpParser(grammar)
tree = cp.parse(pos_tags)
possibleTitles = []
possibleTitle = ""
possibleDelimTitle = ""
delimCounter = 0
delimBool = False
tree.draw()
for subtree in tree:
if str(type (subtree)) == "<type 'tuple'>":
if delimBool ==True:
delimCounter+=1
possibleDelimTitle +=subtree[0]+" "
if delimCounter > numWrdDelimiter:
delimBool = False
delimCounter = 0
possibleDelimTitle = ""
possibleTitle = ""
else:
if subtree.node == "NP":
firstLeaf = subtree.leaves()[0]
firstString = firstLeaf[0]
#print subtree
if firstString.endswith("."):
#check first leaf to see if a leaf of a
#subtree begins with a string with a period
#and the end. This happens sometimes when a
#PN begins a sentence, the previouse tuple gets
#included as a PN
subtree.pop(0)
if len(subtree)>1:
for leaf in subtree.leaves():
possibleTitle +=leaf[0]+" "
possibleTitles.append(possibleTitle.strip())
if len(subtree)>=1:
#try to concat a movie title if its seperated by non nouns
for leaf in subtree.leaves():
possibleDelimTitle +=leaf[0]+" "
if delimCounter > 0:
#the tail end of the split proper noun
possibleTitles.append(possibleDelimTitle.strip())
#reset the counter
delimBool = True
delimCounter = 0
for title in possibleTitles:
#if we have a title with all caps we can discard this with high confidence
titleClean = re.sub(r'\W+', ' ', title)
if titleClean.isupper():
possibleTitles = []
if "." in title:
#sometimes the NLTK will say "Jump Street. I"
#instead of "Jump Street", this alters this case
possibleTitles[possibleTitles.index(title)]= title.split(".")[0]
if " 's" in title:
possibleTitles[possibleTitles.index(title)]= title.replace(" 's","'s")
return possibleTitles
def drawTree():
tree = nltk.Tree.parse('(Tweet (Adj old) (NP (N men) (Conj and) (N women)))')
cmutree = nltk.Tree.parse('(Tweet (Adj old) (NP (N men) (Conj and) (N women)))')
tree.draw()
def main(tigerFile):
tree = ptbToTree(tigerFile)
tree.draw()