def genConferenceIdFeature(instances, paperList, maxConferenceId):
sys.stderr.write("genConferenceIdFeature\n")
d = {}
for line in paperList:
paperId = int(line[0])
conferenceId = int(line[3])
d[paperId] = conferenceId
feature = Feature(maxConferenceId)
for instance in instances:
authorId, paperId = instance[0], instance[1]
conferenceId = d[paperId] + 1 # -1
feature.addLine([[conferenceId, 1.0]])
feature.fix()
return feature
def genJournalIdFeature(instances, paperList, maxJournalId):
sys.stderr.write("genJournalIdFeature\n")
d = {}
for line in paperList:
paperId = int(line[0])
journalId = int(line[4])
d[paperId] = journalId
feature = Feature(maxJournalId)
for instance in instances:
authorId, paperId = instance[0], instance[1]
journalId = d[paperId] + 1 # -1
feature.addLine([[journalId, 1.0]])
feature.fix()
return feature
def genCoauthorFeature(instances, pathFname, maxAuthorId):
'''
return Feature(sparse, [features])
'''
sys.stderr.write("genCoauthorFeature\n")
paperAuthorDict = {}
csvReader = csv.reader(file(pathFname))
csvReader.next()
counter = Counter("paperAuthorDict")
for line in csvReader:
counter.inc()
authorId, paperId = int(line[0]), int(line[1])
paperAuthorDict.setdefault(authorId, set())
paperAuthorDict[authorId].add(paperId)
feature = Feature(maxAuthorId)
counter = Counter("instance", 1000)
for line in instances:
counter.inc()
authorId, paperId = line[0], line[1]
feature.addLine(map(lambda x: [int(x), 1.0], paperAuthorDict[paperId]))
feature.fix()
return feature