def processQueryToDoImprovedLemmatization(self, posTags):
lemmas = []
wnl = WordNetLemmatizer()
ic = IndexCreation()
for word, tag in posTags:
wnTag = ic.getWordnetTag(tag)
if wnTag is None:
lemmas.append(wnl.lemmatize(word))
else:
lemmas.append(wnl.lemmatize(word, pos=wnTag))
return lemmas
def getArticleAndWordCount(self, path):
print("Number of articles:", str(len(os.listdir(path))))
indexSentenceMap = collections.OrderedDict()
wordCount = 0
ic = IndexCreation()
data = ic.readArticles(path)
data = ic.removeArticleTitle(data)
for i in range(0, len(data)):
for j in range(0, len(data[i])):
tokenizedWords = word_tokenize(data[i][j])
index = 'A' + str(i + 1) + 'S' + str(j + 1)
indexSentenceMap[index] = data[i][j]
wordCount += len(tokenizedWords)
print("Number of words in the corpus:", str(wordCount))
return indexSentenceMap
def processQueryToExtractImprovisedHolonyms(self, posTags):
holonyms = []
for word, tag in posTags:
wnTag = IndexCreation().getWordnetTag(tag)
if wnTag is not None:
synset = wn.synsets(word, pos=wnTag)
else:
synset = wn.synsets(word)
if len(synset) > 0:
if len(synset[0].part_holonyms()) > 0:
holonyms.append(
synset[0].part_holonyms()[0].name().split('.')[0])
return holonyms
def processQueryToExtractImprovisedHeadWord(self, query):
dependency_parser = CoreNLPDependencyParser('http://localhost:9000')
headWord = None
parsedSentence = list(dependency_parser.raw_parse(query))[0]
rootValue = list(
list(parsedSentence.nodes.values())[0]['deps']['ROOT'])[0]
for n in parsedSentence.nodes.values():
if n['address'] == rootValue:
headWord = n['word']
if len(headWord):
_, tag = pos_tag([headWord])[0]
wnTag = IndexCreation().getWordnetTag(tag)
if wnTag is not None:
synset = wn.synsets(headWord, pos=wnTag)
if len(synset) > 0:
headWord = synset[0].name().split('.')[0]
break
return headWord