def getFirstOrderKeywords(self):
keywords = []
# for each |question| in full test, get keywords
for i, question in enumerate(self.test):
print('Getting keyword for Q{}'.format(i))
if self.dataType == 'val':
questionText, answers = question
else:
questionText, answers, correctAnswer = question
keywords += util.getKeywords(questionText)
for ans in answers:
keywords += util.getKeywords(ans)
return list(set(keywords))
def __init__(self, question, N):
# print('Question:', question)
self.graph = {}
self.N = N
self.questionKeywords = util.getKeywords(question)
# print('Question keywords extracted:', self.questionKeywords)
self.importance = {kw: 1/len(self.questionKeywords) for kw in self.questionKeywords}
# self.importance = util.getImportanceDict(question)
# print('Keyword importance:', self.importance)
self.secondOrderKeywords = localKeywords[question] if question in localKeywords else self.bestWords()
initialWords = self.secondOrderKeywords + self.questionKeywords
# print('Nodes are:', initialWords)
for word in initialWords:
self.addWord(word)
def getAnswerScore(self, answer):
''' |answer| is a string '''
self.answerKeywords = util.getKeywords(answer)
if len(self.answerKeywords) == 0:
return 0
for keyword in self.answerKeywords:
self.addWord(keyword)
self.pruneGraph()
finalScore = 0
for keyword in self.answerKeywords:
finalScore += self.coherenceScore(keyword)
# print('For answer {}, final score is {}'.format(answer, finalScore))
# Get search metric
# searchScore = util.getAnswerCost(self)
# print('Search score: {}'.format(searchScore))
return finalScore #, searchScore
def takeTestW2V(self): self.reset() for num, question in enumerate(self.test): start = time.time() questionText, answers, correctAnswer = question keywords = util.getKeywords(questionText) # Get scores for each answer answerScores = [] for answer in answers: answerScores.append(util.averageSimilarity(keywords, answer)) index = answerScores.index(max(answerScores)) if self.LETTERS[index] == correctAnswer: self.correct += 1 else: self.incorrect += 1 end = time.time() self.answerReport.append((answerScores, index, correctAnswer)) self.timeReport.append(end - start)