def getMatchesOld(self, terms):
'''
returns a list of top-n matching pairs (doc-id, score)
ranked by descending scores
'''
matches = collections.defaultdict(float)
for t in terms:
if (self.index.hasTerm(t) is None):
continue
idf = self.index.getIdf(t)
for entry in self.index.getEntries(t):
docId = entry["doc-id"]
tf = entry["tf"]
matches[docId] += tf * idf
return getTopMatches([(k, v) for k, v in matches.items()], self.n)
def getMatchesNew(self, listOfTermsAndScores):
'''
returns a list of top-n matching pairs (doc-id, score)
ranked by descending scores
'''
matches = dict()
numberOfOccurences = dict()
for query in listOfTermsAndScores:
for t in query:
if self.index.hasTerm(t[0]):
idf = self.index.getIdf(t[0])
for entry in self.index.getEntries(t[0]):
docId = entry["doc-id"]
tf = entry["tf"]
d = self.index.getD(docId)
w = (idf * tf * (1 + self.k)) / (
tf + self.k *
(1 - self.b + self.b * d / self.avrgD))
matchScore = w * t[1]
if docId in matches.keys():
totalScore = matches[docId] * numberOfOccurences[
docId]
totalScore += matchScore
numberOfOccurences[docId] += 1
newScore = totalScore / numberOfOccurences[docId]
matches[docId] = newScore
'''
totalScore = matches[docId][0]*numberOfOccurences[docId]
totalScore += matchScore
numberOfOccurences[docId] += 1
newScore = totalScore/numberOfOccurences[docId]
matches[docId][0] = newScore
if matches[docId][1] < t[1]:
matches[docId][1] = t[1]
'''
else:
matches[docId] = matchScore
numberOfOccurences[docId] = 1
'''
matches[docId] = list()
matches[docId].append(matchScore)
matches[docId].append(t[1])
numberOfOccurences[docId] = 1
'''
#print(matches)
return getTopMatches([(k, v) for k, v in matches.items()], self.n)
def getMatchesNew(self, listOfTermsAndScores):
'''
returns a list of top-n matching pairs (doc-id, score)
ranked by descending scores
'''
matches = dict()
for query in listOfTermsAndScores:
for t in query:
if self.index.hasTerm(t[0]):
idf = self.index.getIdf(t[0])
for entry in self.index.getEntries(t[0]):
docId = entry["doc-id"]
tf = entry["tf"]
tf_idf = tf * idf
matchScore = tf_idf * t[1]
if docId in matches.keys():
matches[docId] += matchScore
else:
matches[docId] = matchScore
return getTopMatches([(k, v) for k, v in matches.items()], self.n)