def posNGram(self,
sentence,
n=2): #tokenized sentence #sentenceTokenize[1]
'''
No. POS NGrams https://cs.nyu.edu/grishman/jet/guide/PennPOS.html
'''
posDic = {}
if n % 2 == 1: #odd POS
for i in range(n - n // 2 - 1, len(sentence) - n // 2):
posTuple = []
POS = posTag(sentence[i - n // 2:i + n // 2 + 1])
for p in POS:
posTuple.append(p[1])
posTuple = tuple(posTuple)
if posTuple not in posDic:
posDic[posTuple] = 0
posDic[posTuple] = posDic[posTuple] + 1
return (posDic)
else: #even POS
for i in range(len(sentence) - n + 1):
posTuple = []
POS = posTag(sentence[i:i + n])
for p in POS:
posTuple.append(p[1])
posTuple = tuple(posTuple)
if posTuple not in posDic:
posDic[posTuple] = 0
posDic[posTuple] = posDic[posTuple] + 1
return (posDic)
def grabMySQLdocument(self, description):
db = MySQLdb.connect("localhost","dondi","","nlpText" )
cursor = db.cursor()
sql = """SELECT document FROM Documents WHERE description = '%s'""" % (description)
cursor.execute(sql)
result = (cursor.fetchall())
db.commit()
db.close()
result = result[0][0].replace('_', "'")
words = wordTokenize(result)
words = posTag(words)
return words