Merge branch 'master' of github.com:X-XG/WebInfo

exp3/src/MusicBase/MatSim.py

@@ -1,3 +1,4 @@
+from math import sqrt
 import numpy as np
 
 MusicNum = 21602
@@ -6,26 +7,19 @@
 output_path = '../../output/'
 
 
-def JaccobiSim(list1:list, list2:list):
+def JaccobiSim(list1:list, list2:list, mode = 1):
  intersection = len(set(list1).intersection(set(list2)))
  union = len(list1) + len(list2) - intersection
- return intersection/union
- # intersection = 0
- # i = 0
- # j = 0
- # while i < len(list1) and j < len(list2):
- # if list1[i] == list2[j]:
- # intersection += 1
- # i += 1
- # j += 1
- # elif list1[i] < list2[j]:
- # i += 1
- # else:
- # j += 1
- # union = len(list1) + len(list2) - intersection
- # return intersection/union
-
-def MatSimGen():
+ if mode == 0:
+ return intersection
+ elif mode == 0.5:
+ return intersection/sqrt(union)
+ elif mode == 1:
+ return intersection/union
+ else:
+ exit(-1)
+
+def MatSimGen(mode = 1):
  MapMusicID = {}
  MatSim = np.zeros((MusicNum, MusicNum))
 
@@ -49,8 +43,8 @@ def MatSimGen():
  if num %10 == 0:
  print(num)
  for Music2 in MapMusicID:
- MatSim[Music1][Music2] = JaccobiSim(MapMusicID[Music1], MapMusicID[Music2])
- np.save('MatJaccobiSim.npy', MatSim)
+ MatSim[Music1][Music2] = JaccobiSim(MapMusicID[Music1], MapMusicID[Music2], mode)
+ np.save('MatJaccobiSim'+str(mode) +'.npy', MatSim)
 
 def MatDiagMinimize():
  MatSim = np.load('MatJaccobiSim.npy')
@@ -59,7 +53,5 @@ def MatDiagMinimize():
  np.save('MatJaccobiSim_DiagMinimized.npy', MatSim)
 
 if __name__ == '__main__':
- MatDiagMinimize()
- # MatSimGen()
-
-
+ # MatDiagMinimize()
+ MatSimGen(mode = 1)

exp3/src/MusicBase/main.py

@@ -1,8 +1,8 @@
 import numpy as np
 
-MatSimPath = 'MatJaccobiSim_DiagMinimized.npy'
+MatSimPath = 'MatJaccobiSim1.npy'
 data_path = '../../data/DoubanMusic.txt'
-result_path = 'MusicBase.txt'
+result_path = 'MusicBaseMode1.txt'
 MusicNum = 21602
 
 def main(CarryOn = False):
@@ -37,6 +37,8 @@ def main(CarryOn = False):
  continue
  predict_list = []
  for MusicID in range(MusicNum):
+ if MusicID in UserMap[UserID]:
+ continue
  predict = 0
  for UserMusic in UserMap[UserID]:
  predict += MatSim[UserMusic][MusicID]
@@ -54,4 +56,4 @@ def main(CarryOn = False):
  print(count)
 
 if __name__ == '__main__':
- main(CarryOn=True)
+ main(CarryOn=False)

exp3/src/UserBase/MatSim.py

@@ -0,0 +1,27 @@
+import sys
+sys.path.append('..')
+from utils.getMaps import getUserMap
+from utils.JacoobiSim import JaccobiSim
+sys.path.append('UserBase')
+import numpy as np
+
+UserNum = 23599
+data_path = '../../data/DoubanMusic.txt'
+output_path = '../../output/'
+
+def MatSimGen(mode = 1):
+ UserMap = getUserMap()
+
+ MatSim = np.zeros((UserNum, UserNum))
+
+ num = 0
+ for User1 in UserMap:
+ num += 1
+ if num %10 == 0:
+ print(num)
+ for User2 in UserMap:
+ MatSim[User1][User2] = JaccobiSim(UserMap[User1], UserMap[User2], mode)
+ np.save('MatJaccobiSim'+str(mode) +'.npy', MatSim)
+
+if __name__ == '__main__':
+ MatSimGen(mode = 1)

exp3/src/UserBase/main.py

@@ -0,0 +1,53 @@
+import sys
+sys.path.append('..')
+from utils.getMaps import getUserMap, getMusicMap
+sys.path.append('UserBase')
+import numpy as np
+
+MatSimPath = 'MatJaccobiSim1.npy'
+data_path = '../../data/DoubanMusic.txt'
+result_path = 'UserBaseMode1.txt'
+MusicNum = 21602
+UserNum = 23599
+
+def main(CarryOn = False):
+ MatSim = np.load(MatSimPath)
+
+ UserMap = getUserMap()
+
+ MusicMap = getMusicMap()
+
+ if CarryOn:
+ f = open(result_path, 'r')
+ count = len(f.readlines())
+ f.close()
+ f = open(result_path, 'a')
+ else:
+ f = open(result_path, 'w')
+ count = 0
+
+ for UserID in UserMap:
+ if UserID < count:
+ continue
+ predict_list = []
+ for MusicID in range(MusicNum):
+ if MusicID in UserMap[UserID]:
+ continue
+ predict = 0
+ for OtherUser in MusicMap[MusicID]:
+ predict += MatSim[UserID][OtherUser]
+ predict_list.append((predict,MusicID))
+ predict_list.sort(reverse=True)
+ f.write(str(UserID))
+ f.write('\t')
+ f.write(str(predict_list[0][1]))
+ for i in range(1,100):
+ f.write(',')
+ f.write(str(predict_list[i][1]))
+ f.write('\n')
+ count += 1
+ if count % 1 == 0:
+ print(count)
+
+if __name__ == '__main__':
+ main(CarryOn=True)

exp3/src/utils/JacoobiSim.py

@@ -0,0 +1,13 @@
+from math import sqrt
+
+def JaccobiSim(list1:list, list2:list, mode = 1):
+ intersection = len(set(list1).intersection(set(list2)))
+ union = len(list1) + len(list2) - intersection
+ if mode == 0:
+ return intersection
+ elif mode == 0.5:
+ return intersection/sqrt(union)
+ elif mode == 1:
+ return intersection/union
+ else:
+ exit(-1)

exp3/src/utils/getMaps.py

@@ -0,0 +1,35 @@
+def getUserMap(data_path = '../../data/DoubanMusic.txt'):
+ UserMap = {}
+
+ f = open(data_path, 'r')
+ lines = f.readlines()
+ f.close()
+
+ for line in lines:
+ temp = line.split()
+ UserID = int(temp[0])
+ for pair in temp[1:]:
+ MusicID = int(pair.split(',')[0])
+ if UserID not in UserMap:
+ UserMap[UserID] = [MusicID]
+ else:
+ UserMap[UserID].append(MusicID)
+ return UserMap
+
+def getMusicMap(data_path = '../../data/DoubanMusic.txt'):
+ MapMusicID = {}
+
+ f = open(data_path, 'r')
+ lines = f.readlines()
+ f.close()
+
+ for line in lines:
+ temp = line.split()
+ UserID = int(temp[0])
+ for pair in temp[1:]:
+ MusicID = int(pair.split(',')[0])
+ if MusicID not in MapMusicID:
+ MapMusicID[MusicID] = [UserID]
+ else:
+ MapMusicID[MusicID].append(UserID)
+ return MapMusicID