-
Notifications
You must be signed in to change notification settings - Fork 3
/
input_data.py
174 lines (135 loc) · 6.89 KB
/
input_data.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
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
#!/usr/bin/python3
# coding=utf-8
import tensorflow as tf
import pandas as pd
import numpy as np
import random
from sklearn.model_selection import train_test_split
from data_augmentation import DataAugmentation
class Load_data():
def __init__(self,city):
self.city = city
self.data_aug = DataAugmentation(city)
def data_size(self):
query_list = []
query_data = pd.read_csv('./train_data/'+self.city+'-query.csv')
for i in query_data.index:
query_list.append(query_data.loc[i].tolist())
return len(query_list)
def random_embedding(self):
poi_list = []
poi_data = pd.read_csv('./self-embedding/'+self.city+'_poi_weight.csv')
em_size = poi_data.shape[1]
for i in poi_data.index:
poi_list.append(np.random.randn(em_size).tolist())
poi_size = len(poi_list)
embedding = tf.constant(poi_list,dtype='float64')
return embedding, poi_size
def self_embedding(self):
poi_list = []
poi_data = pd.read_csv('./self-embedding/'+self.city+'_poi_weight.csv')
for i in poi_data.index:
poi_list.append(poi_data.loc[i].tolist())
poi_size = len(poi_list)
embedding = tf.constant(poi_list,dtype='float64')
return embedding, poi_size
def load_dataset(self,BATCH_SIZE):
query_list = []
trajs_list = []
query_data = pd.read_csv('./train_data/'+self.city+'-query.csv')
trajs_data = open('./train_data/'+self.city+'-trajs.dat','r')
for i in query_data.index:
query_list.append(query_data.loc[i].tolist())
for line in trajs_data.readlines():
tlist = [eval(i) for i in line.split()]
trajs_list.append(tlist)
print('total number:',len(query_list),len(trajs_list))
trajs_list = tf.keras.preprocessing.sequence.pad_sequences(trajs_list, padding='post')
query_train, query_val, trajs_train, trajs_val = train_test_split(query_list,trajs_list,test_size=0.2)
print('train_set:',len(query_train),len(trajs_train))
print('test_set:', len(query_val), len(trajs_val))
dt_train = tf.data.Dataset.from_tensor_slices((query_train, trajs_train)).shuffle(len(query_train))
dt_train = dt_train.batch(BATCH_SIZE, drop_remainder=True)
dt_val = tf.data.Dataset.from_tensor_slices((query_val, trajs_val)).shuffle(len(query_val))
dt_val = dt_val.batch(BATCH_SIZE, drop_remainder=True)
return dt_train, dt_val, int(len(query_train)/BATCH_SIZE), int(len(query_val)/BATCH_SIZE)
def load_pretrain_dataset(self, que, traj):
pre_que = que
r1 = random.randint(1, 4)
r2 = random.randint(1, 4)
sample1 = self.gen_random_sample(r1, traj)
sample2 = self.gen_random_sample(r2, traj)
return pre_que, sample1, sample2
def gen_random_sample(self, rand_num, traj):
if (rand_num == 0):
return tf.nn.embedding_lookup(self.data_aug.original(), traj)
if (rand_num == 1):
return tf.nn.embedding_lookup(self.data_aug.token_cutoff(), traj)
if (rand_num == 2):
trajs = traj.numpy()
sample = []
for traj in trajs:
traj_aug = tf.nn.embedding_lookup(self.data_aug.token_shuffing(traj), traj)
sample.append(traj_aug.numpy())
return tf.constant(sample, dtype='double')
if (rand_num == 3):
return tf.nn.embedding_lookup(self.data_aug.feature_cutoff(), traj)
if(rand_num == 4):
return tf.nn.embedding_lookup(self.data_aug.dropout(), traj)
def load_dataset_one(self, index, BATCH_SIZE):
query_list = []
trajs_list = []
query_data = pd.read_csv('./train_data/'+self.city+'-query.csv')
trajs_data = open('./train_data/'+self.city+'-trajs.dat','r')
for i in query_data.index:
query_list.append(query_data.loc[i].tolist())
for line in trajs_data.readlines():
tlist = [eval(i) for i in line.split()]
trajs_list.append(tlist)
print('total number：',len(query_list),len(trajs_list))
trajs_list = tf.keras.preprocessing.sequence.pad_sequences(trajs_list, padding='post')
trajs_list = trajs_list.tolist()
query_val = []
trajs_val = []
query_val.append(query_list.pop(index))
trajs_val.append(trajs_list.pop(index))
query_train = query_list
trajs_train = trajs_list
print('train set:',len(query_train),len(trajs_train))
print('test set:', len(query_val), len(trajs_val))
dt_train = tf.data.Dataset.from_tensor_slices((query_train, trajs_train)).shuffle(len(query_train))
dt_train = dt_train.batch(BATCH_SIZE, drop_remainder=True)
# print(dt_train)
dt_val = tf.data.Dataset.from_tensor_slices((query_val, trajs_val)).shuffle(len(query_val))
dt_val = dt_val.batch(1, drop_remainder=True)
return dt_train, dt_val, int(len(query_train)/BATCH_SIZE), 1
def load_dataset_train(self, BATCH_SIZE):
query_train = []
trajs_train = []
query_data = pd.read_csv('./train_data/' + self.city + '-query-train.csv')
trajs_data = open('./train_data/' + self.city + '-trajs-train.dat', 'r')
for i in query_data.index:
query_train.append(query_data.loc[i].tolist())
for line in trajs_data.readlines():
tlist = [eval(i) for i in line.split()]
trajs_train.append(tlist)
print('训练集总量：', len(query_train), len(trajs_train))
trajs_train = tf.keras.preprocessing.sequence.pad_sequences(trajs_train, padding='post')
dt_train = tf.data.Dataset.from_tensor_slices((query_train, trajs_train)).shuffle(len(query_train))
dt_train = dt_train.batch(BATCH_SIZE, drop_remainder=True)
return dt_train, int(len(query_train) / BATCH_SIZE)
def load_dataset_test(self, BATCH_SIZE):
query_test = []
trajs_test = []
query_data = pd.read_csv('./train_data/' + self.city + '-query-test.csv')
trajs_data = open('./train_data/' + self.city + '-trajs-test.dat', 'r')
for i in query_data.index:
query_test.append(query_data.loc[i].tolist())
for line in trajs_data.readlines():
tlist = [eval(i) for i in line.split()]
trajs_test.append(tlist)
print('train set number:', len(query_test), len(trajs_test))
trajs_test = tf.keras.preprocessing.sequence.pad_sequences(trajs_test, padding='post')
dt_test = tf.data.Dataset.from_tensor_slices((query_test, trajs_test)).shuffle(len(query_test))
dt_test = dt_test.batch(BATCH_SIZE, drop_remainder=True)
return dt_test, int(len(query_test) / BATCH_SIZE)