-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathdataprocess.py
146 lines (124 loc) · 5.01 KB
/
dataprocess.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
import numpy as np
import os
import pandas as pd
import scipy.sparse as sp
from sklearn.preprocessing import OneHotEncoder
def normalize_adj(adj):
"""Symmetrically normalize adjacency matrix."""
# a_hat = D^(-0.5)*a^*D^(-0.5)
# adj = sp.coo_matrix(adj)
rowsum = np.array(adj.sum(1)) # D
# D = diag(row_sum(a^))
d_inv_sqrt = np.power(rowsum, -0.5).flatten() # D^-0.5
d_inv_sqrt[np.isinf(d_inv_sqrt)] = 0.
d_mat_inv_sqrt = sp.diags(d_inv_sqrt) # D^-0.5
return adj.dot(d_mat_inv_sqrt).transpose().dot(d_mat_inv_sqrt).tocoo() # D^-0.5AD^0.5
def transforTosparsematrix(edge, hasvalue, shape):
# transfer sparse matrix to normal matrix
# input: source, target nodes index, edge_value(=1 if hasvalue=0)
# output: normal matrix
# e.g.
# # source index
# i = [0, 0, 0, 1, 1, 1]
# # target index
# j = [0, 1, 2, 0, 3, 4]
# value on the edge
edge = edge.astype(int)
if hasvalue:
value = edge[:, 2]
else:
value = np.ones_like(edge[:, 0], dtype=np.float)
# print(value.shape)
# print(edge.shape)
A = sp.coo_matrix((value, (edge[:, 0], edge[:, 1])), shape)
def loadinput(city, seq_len, start_NORM, end_NORM,loop_num):
# city: capital/nyc
graph_path = ''
if city == 'nyc':
graph_path = 'D:\my\dataset\citibike\data\day\graph/new_station/'
start_demand = pd.read_csv('D:\my\dataset\citibike\data\day\demand\demand_eaStart.csv')
end_demand = pd.read_csv('D:\my\dataset\citibike\data\day\demand\demand_eaEnd.csv')
# weather_type temp wind_speed
wea_df = pd.read_csv('D:\my\dataset\citibike\data\day/weather_enc.csv')
else:
graph_path = 'D:\my\dataset\capitalbike\data\day\graph/new_station'
start_demand = pd.read_csv('D:\my\dataset\capitalbike\data\day\demand\demand_eaStart.csv')
end_demand = pd.read_csv('D:\my\dataset\capitalbike\data\day\demand\demand_eaEnd.csv')
wea_df = pd.read_csv('D:\my\dataset\weather\washington/Washington-day.csv')
calendar = pd.read_csv('D:\my\dataset\calendar/calendar_enc.csv') # weekday[1-6] workday[0/1]
start_demand = start_NORM.fit_transform(start_demand)
end_demand = end_NORM.fit_transform(end_demand)
os.chdir(graph_path)
file_list = os.listdir(graph_path)
# print(start_demand.columns)
adj_list = []
feat_list = []
# end_list=[]
label_list = []
timefeat_list = []
label_date = ''
station_num = start_demand['Station_id'].values.shape[0]
# batch
loop=0
i=0
while i < (len(file_list) - seq_len):
# print(i)
if i + seq_len == len(file_list): break
label_date = file_list[i + seq_len].split('.')[0]
# label
label_value = np.reshape(end_demand[label_date].values, [-1, 1])
label_list.append(label_value)
for j in range(seq_len):
filename = file_list[i + j]
date = filename.split('.')[0]
start_value = np.reshape(start_demand[date].values, [-1, 1])
end_value = np.reshape(end_demand[date].values, [-1, 1])
feat_value = np.concatenate((start_value, end_value), axis=1)
weather_v = wea_df[wea_df['date'] == date].values[:, 2:]
calendar_v = calendar[calendar['Date'] == date].values[:, 2:]
other_feat = np.concatenate([weather_v, calendar_v], axis=1).astype(float)
# end_value = end_demand[date]
graph_df = pd.read_csv(filename)[['source', 'target']]
edge = graph_df[(graph_df['source'] > 0) & (graph_df['target'] > 0)].values
adj = transforTosparsematrix(edge, hasvalue=False, shape=[station_num, station_num])
# adj normalization
# adj_list.append(normalize_adj(adj))
adj_list.append(adj.toarray())
feat_list.append(feat_value)
timefeat_list.append(other_feat[0])
# end_list.append((end_value))
yield adj_list, feat_list, timefeat_list, label_date, label_list
if loop<loop_num-1 and i+1>0.95*(len(file_list) - seq_len):
loop+=1
i=0
adj_list = []
feat_list = []
label_list = []
timefeat_list = []
i+=1
class MinMaxNormalization(object):
'''MinMax Normalization --> [0, 1]
x = (x - min) / (max - min).
'''
def __init__(self):
pass
def fit(self, X):
x_v = X.values[:, 1:]
self._min = x_v.min()
self._max = x_v.max()
# print("min:", self._min, "max:", self._max)
def transform(self, X):
X = 1. * (X - self._min) / (self._max - self._min)
# X = X * 2. - 1.
return X
def fit_transform(self, X):
self.fit(X)
return self.transform(X)
def inverse_transform(self, X):
# X = (X + 1.) / 2.
X = 1. * X * (self._max - self._min) + self._min
return X
def station_file_num(path):
df = pd.read_csv(path + '/demand/demand_eaEnd.csv')
file_n = path + 'graph/new_station/'
return len(df['Station_id'].unique()), len(os.listdir(file_n))