-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy path3_candidates_elastic.py
152 lines (105 loc) · 3.38 KB
/
3_candidates_elastic.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
# coding: utf-8
# In[1]:
import json
import pandas as pd
import numpy as np
from tqdm import tqdm
import cPickle
from elasticsearch import Elasticsearch, helpers
# In[2]:
with open('tmp/df_train_folds.bin', 'rb') as f:
df_train = cPickle.load(f)
train_users = set(df_train.user_1) | set(df_train.user_2)
train_idx = sorted(train_users)
fold1 = df_train[df_train.fold == 1]
fold1_users = set(fold1.user_1) | set(fold1.user_2)
fold2 = df_train[df_train.fold == 2]
fold2_users = set(fold2.user_1) | set(fold2.user_2)
# In[11]:
components = []
uid_to_others = {}
for _, group in tqdm(df_train.groupby('component')):
users = set(group.user_1) | set(group.user_2)
components.append(users)
for uid in users:
uid_to_others[uid] = users - {uid}
# In[3]:
TRAIN_1 = 1
TRAIN_2 = 2
TEST = 3
def user_fold(uid):
if uid in fold1_users:
return TRAIN_1
if uid in fold2_users:
return TRAIN_2
return TEST
# In[8]:
es_host = '172.17.0.2'
es = Elasticsearch(host=es_host)
# In[5]:
def find_similar(uid, limit=10):
query = {
'query': {
'filtered': {
'query': {
'more_like_this': {
'like': {
'_index': 'user',
'_type': 'user_log',
'_id': int(uid),
},
'max_query_terms': 10,
'fields': ['fact.domain', 'fact.address', 'fact.param', 'fact.title^2'],
}
}
}
},
'filter': {
'bool': {
'must': [{
'term': {
'fold': user_fold(uid),
},
}],
}
},
'fields': ['_id'],
'size': limit,
}
res = es.search(index='user', doc_type='user_log', body=query)
hits = res['hits']['hits']
return [(int(d['_id']), d['_score']) for d in hits]
# In[ ]:
train_pairs = []
for uid in tqdm(train_users):
similar = find_similar(uid, limit=70)
others_truth = uid_to_others[uid]
fold = user_fold(uid)
train_pairs.extend((uid, u, score, u in others_truth, fold) for (u, score) in similar)
# In[ ]:
train_pairs = pd.DataFrame(train_pairs)
train_pairs.columns = ['user_1', 'user_2', 'es_score', 'target', 'fold']
# In[ ]:
train_pairs.user_1 = train_pairs.user_1.astype('uint32')
train_pairs.user_2 = train_pairs.user_2.astype('uint32')
train_pairs.target = train_pairs.target.astype('uint8')
train_pairs.fold = train_pairs.fold.astype('uint8')
train_pairs.es_score = train_pairs.es_score.astype('float32')
with open('tmp/es-retrieved-70.bin', 'wb') as f:
cPickle.dump(train_pairs, f)
# In[ ]:
test_users = set(range(339405)) - train_users
test_pairs = []
for uid in tqdm(test_users):
similar = find_similar(uid, limit=70)
test_pairs.extend((uid, u, score, TEST) for (u, score) in similar)
# In[ ]:
test_pairs = pd.DataFrame(test_pairs)
test_pairs.columns = ['user_1', 'user_2', 'es_score', 'fold']
test_pairs.user_1 = test_pairs.user_1.astype('uint32')
test_pairs.user_2 = test_pairs.user_2.astype('uint32')
test_pairs.fold = test_pairs.fold.astype('uint8')
test_pairs.es_score = test_pairs.es_score.astype('float32')
# In[ ]:
with open('tmp/es-retrieved-70_test.bin', 'wb') as f:
cPickle.dump(test_pairs, f)