starlee
/
sldb
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

rerank is refined

This commit is contained in:
StarLee 2016-10-28 13:20:47 +08:00
parent 0afcdf9917
commit 166980260e
1 changed files with 18 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
search.py
View File

@ -227,7 +227,7 @@ import csv
if __name__ == '__main__':
########### single search
query = 'java ide'
query = 'java message queue'
# items = query.split(' ')
# items = [get_syns(term) for term in items]
combined_queries = combine_item(query)
@ -250,7 +250,8 @@ if __name__ == '__main__':
final_result = list()
for item in sorted(cans.items(),lambda x, y: cmp(x[1], y[1]),reverse = True)[:20]:
final_result.append(item[0])
print item[0]
print final_result
print_time('>>> sort done')
@ -360,16 +361,14 @@ if __name__ == '__main__':
# re rank top 10
# 把 需要重新排的 和 不需要排的分开 re_rank_prj 和 stay_rank_prj
# for 1 to to 从两个里面选，优先选 重排的，如果有冲突，优先选down的。
stay_rank_prj = [(i,final_result[i]) for i in range(0,len(final_result))]
re_rank_prj = dict() # 每一个pos 对应着 想要排到这个位置的所有prj，并且按照之前的排序顺序进行拍序
stay_rank_prj = {i:(i,final_result[i]) for i in range(0,len(final_result))} #{pos:prj}
re_rank_prj = dict() # 每一个pos 对应着 想要排到这个位置的所有prj，并且按照之前的排序顺序进行拍序 {pos: [prj] }
for rr in rerank:
if rr[1] not in re_rank_prj:
re_rank_prj[rr[1]] = list()
re_rank_prj[rr[1]].append(rr[0])
index = stay_rank_prj.index(rr[0])
if index != -1:
del stay_rank_prj[index]
del stay_rank_prj[rr[0][0]]# 从原来的位置删除
# 按照之前的排序顺序进行拍序
for key,value in re_rank_prj.items():
@ -377,21 +376,19 @@ if __name__ == '__main__':
top_10_list = list()
#merge re_rank_prj and stary_rank_prj
pos_in_stay = 0
for i in range(0,10):
#先选取要重新拍的，再排以前的
pos = 0
while True:
if (len(re_rank_prj) == 0) and (len(stay_rank_prj) == 0):
break
flag = True
for j in range(0,i+1):
if j in re_rank_prj: # 如果在re_rank_prj中比i小的还有剩余，应该先把之前的搞完吧
value = re_rank_prj[j].pop(0)
if len(re_rank_prj[j]) == 0:
del re_rank_prj[j]
flag = False
break
if flag:
value = stay_rank_prj[pos_in_stay]
pos_in_stay += 1
top_10_list.append(value[1])
if pos in re_rank_prj: # 如果在re_rank_prj中比i小的还有剩余，应该先把之前的搞完吧
top_10_list.extend(re_rank_prj[pos])
del re_rank_prj[pos]
if pos in stay_rank_prj:
top_10_list.append(stay_rank_prj[pos])
del stay_rank_prj[pos]
pos += 1
print top_10_list
# local search rejust