def get_k(queryAddress,documentAddress,e):
document1, length_by_docid1 = make_invidx(documentAddress, e)
query1=calculate_query(get_query(queryAddress))
simi1= cosine_distance(document1,query1,100)
result=readFile(resultAddress)
for num in simi1:
for weight1,docid in simi1[num]:
if docid in result:
print(docid)
def userLogExpan(extend,documentAddress,queryAddress):
ses=userFeedback("C://TweetStudy//12//12//userFeedback3.txt")
# print(ses)
ret={}
for e in extend:
document1, length_by_docid1 = make_invidx(documentAddress, e)
tf,atid=read_file(documentAddress,e)
# query1=calculate_query(get_query(queryAddress))
dic=get_query(queryAddress)
for num in dic:
re=coWeight(ses,nltk.word_tokenize(dic[num]),document1,tf,atid)
ret=selectTopTerm(re,10)
return ret
def get_topK(queryAddress,documentAddress,extend):
for e in extend:
document1, length_by_docid1 = make_invidx(documentAddress, e)
document2 = make_invidx_by_OkapiBM25(documentAddress, e)
#print(document)
#print(length_by_docid)
#saveIndex(index, (document, length_by_docid))
#document, length_by_docid = loadIndex(index)
#get query
query1=calculate_query(get_query(queryAddress))
for num in query1:
document3= make_index_AIG(documentAddress,e,query1[num])
simi3=sum_score(document3,10)
simi2=sum_keyword(document2,query1[num],10)
#print(simi)
#document= make_index_AIG(documentAddress,e,query1[num])
simi1= cosine_distance(document1,query1,10)
# simi=sum_keyword(document, query1[num],10)
#print(query1)
#cosine similarity
#simi=score(document,query1,10)
#simi= cosine_distance(document,query1,10)
#print(simi)
for num in simi1:
for weight1,docid in simi1[num]:
#print(docid)
for weight2,docid2 in simi2:
if docid==docid2:
print(docid2," : ",weight2)
for weight3,docid3 in simi3:
if docid==docid3:
print(docid3,":",weight3)
for weight2,docid2 in simi2:
for weight3,docid3 in simi3:
if docid2==docid3:
print(docid3,":",weight3)
