def processLine(line, tlds):
rawFields = line.split("|")
url = rawFields[0]
urlFields = parseUrl(url, tlds)
timestamp = rawFields[5]
tld=urlFields[0]
hostname=urlFields[1]
port=urlFields[2]
subdomain=urlFields[3]
path1=urlFields[4]
path2=urlFields[5]
path3=urlFields[6]
path4=urlFields[7]
path5=urlFields[8]
path6=urlFields[9]
ret = [timestamp, tld, hostname, port, subdomain, path1, path2, path3, path4, path5, path6]
# ret.append(url)
return ret
def processOneUrl(url):
timestamp = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%s")
urlFields = parseUrl(url, tlds)
# Create dict: timestamp,tld,hostname,port,subdomain,path1,path2,path3,path4,path5,path6
modelInput = dict(timestamp=timestamp, url=url,
tld=urlFields[0],
hostname=urlFields[1],
port=urlFields[2],
subdomain=urlFields[3],
path1=urlFields[4],
path2=urlFields[5],
path3=urlFields[6],
path4=urlFields[7],
path5=urlFields[8],
path6=urlFields[9])
result = model.run(modelInput)
table = texttable.Texttable()
oneStepPredictions = result.inferences['multiStepPredictions'][1]
threeStepPredictions = result.inferences['multiStepPredictions'][3]
printTopNPredictions(oneStepPredictions, "next", table, 3)
printTopNPredictions(threeStepPredictions, "third", table, 3)
print table.draw()