rhash

exp2/README.md

@@ -4,6 +4,8 @@
 exp2/
 |----data/
  ...
+|----rhash/
+ ...
 |----openke_transE/
  |----base/
  |----config/

exp2/rhash/ent_map.py

@@ -0,0 +1,120 @@
+ent_map = {}
+
+freq_origin = {}
+
+f_train_origin = open('FB15k-237/train.tsv' ,'r')
+lines = f_train_origin.readlines()
+for line in lines:
+ temp = line.split()
+ word = temp[0]
+ if word not in freq_origin:
+ freq_origin[word] = [1,0,0,0,0]
+ else:
+ freq_origin[word][0] += 1
+ word = temp[2]
+ if word not in freq_origin:
+ freq_origin[word] = [0,1,0,0,0]
+ else:
+ freq_origin[word][1] += 1
+
+
+f_dev_origin = open('FB15k-237/dev.tsv' ,'r')
+lines = f_dev_origin.readlines()
+for line in lines:
+ temp = line.split()
+ word = temp[0]
+ if word not in freq_origin:
+ freq_origin[word] = [0,0,1,0,0]
+ else:
+ freq_origin[word][2] += 1
+ word = temp[2]
+ if word not in freq_origin:
+ freq_origin[word] = [0,0,0,1,0]
+ else:
+ freq_origin[word][3] += 1
+
+
+f_test_origin = open('FB15k-237/test.tsv' ,'r')
+lines = f_test_origin.readlines()
+for line in lines:
+ temp = line.split()
+ word = temp[0]
+ if word not in freq_origin:
+ freq_origin[word] = [0,0,0,0,1]
+ else:
+ freq_origin[word][4] += 1
+ word = temp[2]
+
+
+freq_ta = {}
+
+f_train_ta = open('data/train.txt' ,'r')
+lines = f_train_ta.readlines()
+for line in lines:
+ temp = line.split()
+ word = temp[0]
+ if word not in freq_ta:
+ freq_ta[word] = [1,0,0,0,0]
+ else:
+ freq_ta[word][0] += 1
+ word = temp[2]
+ if word not in freq_ta:
+ freq_ta[word] = [0,1,0,0,0]
+ else:
+ freq_ta[word][1] += 1
+
+
+f_dev_ta = open('data/dev.txt' ,'r')
+lines = f_dev_ta.readlines()
+for line in lines:
+ temp = line.split()
+ word = temp[0]
+ if word not in freq_ta:
+ freq_ta[word] = [0,0,1,0,0]
+ else:
+ freq_ta[word][2] += 1
+ word = temp[2]
+ if word not in freq_ta:
+ freq_ta[word] = [0,0,0,1,0]
+ else:
+ freq_ta[word][3] += 1
+
+
+f_test_ta = open('data/test.txt' ,'r')
+lines = f_test_ta.readlines()
+for line in lines:
+ temp = line.split()
+ word = temp[0]
+ if word not in freq_ta:
+ freq_ta[word] = [0,0,0,0,1]
+ else:
+ freq_ta[word][4] += 1
+
+count = 0
+
+for word in freq_origin:
+ freq_word = freq_origin[word]
+ first = True
+ for id in freq_ta:
+ freq_id = freq_ta[id]
+ if (freq_word[0] == freq_id[0] and freq_word[1] == freq_id[1] and freq_word[2] == freq_id[2] \
+ and freq_word[3] == freq_id[3] and freq_word[4] == freq_id[4]):
+ if(first):
+ ent_map[word] = id
+ first = False
+ count += 1
+ else:
+ # ent_map[word].append(id)
+ ent_map.pop(word)
+ count -= 1
+ break
+
+
+f = open('./refresh/ent_map.txt', 'w')
+for key in ent_map:
+ f.write(str(key))
+ f.write('\t')
+ f.write(str(ent_map[key]))
+ f.write('\n')
+
+print(count)

exp2/rhash/ent_map_3.py

@@ -0,0 +1,81 @@
+rel_map = {}
+f = open('refresh/rel_map.txt', 'r')
+lines = f.readlines()
+for line in lines:
+ temp = line.split()
+ rel_map[temp[0]] = temp[1]
+f.close()
+
+ent_map = {}
+f = open('refresh/ent_certain_map_3.txt', 'r')
+lines = f.readlines()
+for line in lines:
+ temp = line.split()
+ ent_map[temp[0]] = temp[1]
+f.close()
+
+##########################################
+# num = 20466
+# f1 = open('data/test.txt' ,'r')
+# f2 = open('FB15k-237/test.tsv','r')
+# for i in range(20466):
+# line1 = f1.readline()
+# line2 = f2.readline()
+# temp1 = line1.split()
+# temp2 = line2.split()
+# if temp2[0] not in ent_map:
+# ent_map[temp2[0]] = temp1[0]
+
+###################################\
+
+f = open('FB15k-237/train.tsv' ,'r')
+f1 = open('data/train.txt' ,'r')
+lines = f.readlines()
+lines_ta = f1.readlines()
+
+count = 0
+for line in lines:
+ count += 1
+ if(count % 100 == 0):
+ print(count)
+ temp = line.split()
+ if(temp[1] not in rel_map):
+ print('error')
+ exit(-1)
+ if(temp[0] in ent_map and temp[2] in ent_map):
+ continue
+ if(temp[0] in ent_map):
+ ent1 = ent_map[temp[0]]
+ rel = rel_map[temp[1]]
+ first = True
+ for line_ta in lines_ta:
+ temp_ta = line_ta.split()
+ if(ent1 == temp_ta[0] and rel == temp_ta[1]):
+ if(first):
+ ent_map[temp[2]] = temp_ta[2]
+ first = False
+ else:
+ ent_map.pop(temp[2])
+ break
+
+ elif(temp[2] in ent_map):
+ ent2 = ent_map[temp[2]]
+ rel = rel_map[temp[1]]
+ first = True
+ for line_ta in lines_ta:
+ temp_ta = line_ta.split()
+ if(ent2 == temp_ta[2] and rel == temp_ta[1]):
+ if(first):
+ ent_map[temp[0]] = temp_ta[0]
+ first = False
+ else:
+ ent_map.pop(temp[0])
+ break
+
+
+f = open('./refresh/ent_certain_map_3.txt', 'w')
+for key in ent_map:
+ f.write(str(key))
+ f.write('\t')
+ f.write(str(ent_map[key]))
+ f.write('\n')