-
Notifications
You must be signed in to change notification settings - Fork 5
/
preprocess.py
434 lines (354 loc) · 14.9 KB
/
preprocess.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
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
import os
import zipfile
import numpy as np
import torch
import pandas as pd
import datetime
import calendar
from dateutil.relativedelta import relativedelta
from math import sin, asin, cos, radians, sqrt
"""
Load NYC-Bike dataset
"""
def load_nyc_sharing_bike_data(directory="data/NYC-Sharing-Bike"):
if (not os.path.isfile(directory + "/adj_mat.npy")
or not os.path.isfile(directory + "/node_values.npy")):
if os.path.isfile(directory + "/NYC-Sharing-Bike.zip"):
with zipfile.ZipFile(directory + "/NYC-Sharing-Bike.zip", 'r') as zip_ref:
zip_ref.extractall(directory)
else:
parse_nyc_sharing_bike_data(directory, "201307-201612-citibike-tripdata.zip")
A = np.load(directory + "/adj_mat.npy")
A = A.astype(np.float32)
A = change_avg_degree(A, K=100)
# normalize adj matrix
A = A / A.sum(axis=0, keepdims=True)
# X's shape is (num_nodes, num_features, num_timesteps)
X = np.load(directory + "/node_values.npy")
X = X.astype(np.float32)
print('(num_nodes, num_features, num_timesteps) is ', X.shape)
# Normalization using Z-score method
means = np.mean(X, axis=(0, 2))
X = X - means.reshape(1, -1, 1)
stds = np.std(X, axis=(0, 2))
X = X / stds.reshape(1, -1, 1)
return A, X, means, stds
def parse_nyc_sharing_bike_data(directory, filename):
zip_path = directory + "/" + filename
if os.path.exists(zip_path):
with zipfile.ZipFile(zip_path, 'r') as zip_ref:
zip_ref.extractall(directory)
# zipfile example: 201307-201402-citibike-tripdata.zip
start_date = datetime.datetime.strptime(filename.split('-')[0], '%Y%m')
end_date = datetime.datetime.strptime(filename.split('-')[1], '%Y%m')
month_delta = (end_date.year - start_date.year) * 12 + end_date.month - start_date.month
max_timestep = (month_delta + 1) * 31 * 24
max_nodes = 1000
nodes_info = {}
# X's shape is (num_nodes, num_features, num_sequence)
X = np.zeros((max_nodes, 1, max_timestep))
# read monthly data
timestep_base = 0
for delta in range(0, month_delta + 1):
cur_date = start_date + relativedelta(months = delta)
read_monthly_tripdata(cur_date, nodes_info, X, timestep_base)
month_day = calendar.monthrange(cur_date.year, cur_date.month)[1]
timestep_base += 24 * month_day
num_nodes = len(nodes_info)
print("nodes num is %d." % num_nodes)
X = X[:num_nodes][:][:timestep_base]
A = np.zeros((num_nodes, num_nodes))
# calculate adj matrix
for (id_i, info_i) in nodes_info.items():
for (id_j, info_j) in nodes_info.items():
idx_i = info_i['index']
idx_j = info_j['index']
if idx_i != idx_j:
A[idx_i][idx_j] = 1 / calculate_distance(info_i['lon'], info_i['lat']
, info_j['lon'], info_j['lat'])
# normalize adj matrix
A = (A.T / A.sum(axis=1)).T
# save as .npy file
np.save(directory + "/nodes_info.npy", nodes_info)
np.save(directory + "/adj_mat.npy", A)
np.save(directory + "/node_values.npy", X)
return
def read_monthly_tripdata(date, nodes_info, X, timestep_base):
# path example: 2013-08 - Citi Bike trip data.csv or 201409-citibike-tripdata.csv
path = directory + "/" + date.strftime("%Y-%m") + " - Citi Bike trip data.csv"
mod = 1
if not os.path.exists(path):
path = directory + "/" + date.strftime("%Y%m") + "-citibike-tripdata.csv"
mod = 2
if not os.path.exists(path):
print("[ERROR]File %s not exists." % path)
return
data = pd.read_csv(path)
for _, row in data.iterrows():
# origin data
start_station_id = row['start station id']
end_station_id = row['end station id']
hour_index = row['stoptime'].find(":")
if mod == 1:
stoptime = datetime.datetime.strptime(row['stoptime'][:hour_index], '%Y-%m-%d %H')
elif mod == 2:
stoptime = datetime.datetime.strptime(row['stoptime'][:hour_index], '%m/%d/%Y %H')
# add station info
if not nodes_info.get(start_station_id):
index = len(nodes_info)
nodes_info[start_station_id] = {}
nodes_info[start_station_id]['index'] = index
nodes_info[start_station_id]['lon'] = row['start station longitude']
nodes_info[start_station_id]['lat'] = row['start station latitude']
if not nodes_info.get(end_station_id):
index = len(nodes_info)
nodes_info[end_station_id] = {}
nodes_info[end_station_id]['index'] = index
nodes_info[end_station_id]['lon'] = row['end station longitude']
nodes_info[end_station_id]['lat'] = row['end station latitude']
end_station_index = nodes_info[end_station_id]['index']
timestep = timestep_base + (stoptime.day - 1) * 24 + stoptime.hour
X[end_station_index][0][timestep] += 1
print("Read %s data successfully." % date.strftime("%Y-%m"))
return
"""
Load METR-LA dataset
"""
def load_metr_la_data(directory="data/METR-LA"):
if (not os.path.isfile(directory + "/adj_mat.npy")
or not os.path.isfile(directory + "/node_values.npy")):
with zipfile.ZipFile(directory + "/METR-LA.zip", 'r') as zip_ref:
zip_ref.extractall(directory)
A = np.load(directory + "/adj_mat.npy")
# X's shape is (num_nodes, num_features, num_timesteps)
X = np.load(directory + "/node_values.npy").transpose((1, 2, 0))
X = X.astype(np.float32)
print('(num_nodes, num_features, num_timesteps) is ', X.shape)
# Normalization using Z-score method
means = np.mean(X, axis=(0, 2))
X = X - means.reshape(1, -1, 1)
stds = np.std(X, axis=(0, 2))
X = X / stds.reshape(1, -1, 1)
return A, X, means, stds
"""
Load PeMS-M dataset
"""
def load_pems_m_data(directory="data/PeMS-M"):
if (not os.path.isfile(directory + "/W_228.csv")
or not os.path.isfile(directory + "/V_228.csv")):
if os.path.isfile(directory + "/PeMS-M.zip"):
with zipfile.ZipFile(directory + "/PeMS-M.zip", 'r') as zip_ref:
zip_ref.extractall(directory)
adj_path = directory + "/W_228.csv"
A = np.loadtxt(adj_path, delimiter=',')
A = A.astype(np.float32)
A = A / A.sum(axis=1)
node_path = directory + "/V_228.csv"
X = np.loadtxt(node_path, delimiter=',')
X = X.transpose(1, 0)
X = np.expand_dims(X, axis=1)
X = X.astype(np.float32)
print('(num_nodes, num_features, num_timesteps) is ', X.shape)
# Normalization using Z-score method
means = np.mean(X, axis=(0, 2))
X = X - means.reshape(1, -1, 1)
stds = np.std(X, axis=(0, 2))
X = X / stds.reshape(1, -1, 1)
return A, X, means, stds
"""
Load PeMS-D7 dataset
"""
def load_pems_d7_data(directory="data/PeMS-D7"):
if (not os.path.isfile(directory + "/adj_mat.npy")
or not os.path.isfile(directory + "/node_values.npy")):
if os.path.isfile(directory + "/PeMSD7.zip"):
with zipfile.ZipFile(directory + "/PeMSD7.zip", 'r') as zip_ref:
zip_ref.extractall(directory)
else:
parse_pems_d7_data(directory)
A = np.load(directory + "/adj_mat.npy")
A = A.astype(np.float32)
# X's shape is (num_nodes, num_features, num_timesteps)
X = np.load(directory + "/node_values.npy")
X = X.astype(np.float32)
# to avoid OOM and only load a part
percent = 0.5
X = X[:, :, :int(percent * X.shape[2])]
print('(num_nodes, num_features, num_timesteps) is ', X.shape)
# Normalization using Z-score method
means = np.mean(X, axis=(0, 2))
X = X - means.reshape(1, -1, 1)
stds = np.std(X, axis=(0, 2))
X = X / stds.reshape(1, -1, 1)
return A, X, means, stds
def parse_pems_d7_data(directory, meta_path='/d07_text_meta_2019_01_10.txt'):
if not os.path.isfile(directory + "/raw_adj_mat.npy"):
calculate_pems_adj(directory + meta_path)
node_dict = str(np.load(directory + "node_dict.npy", allow_pickle=True))
node_dict = eval(node_dict)
node_index = get_pems_node_value(directory, node_dict)
# reload and save adj for valid nodes
A = np.load(directory + "/raw_adj_mat.npy")
A = A[node_index][:, node_index]
np.save(directory + "./adj_mat.npy", A)
return
def calculate_pems_adj(meta_path):
# read metadata
data = pd.read_table(meta_path, sep='\t', usecols=['ID', 'Latitude', 'Longitude'])
data = data.dropna(axis=0, how='any').reset_index(drop=True)
num_nodes = len(data)
# calculate adj matrix
A = np.zeros((num_nodes, num_nodes))
node_dict = dict()
for i in range(num_nodes):
if i % 10 == 0:
print("percentage:", i/num_nodes)
node_i = data.loc[i]
node_dict[int(node_i['ID'])] = i
for j in range(i+1, num_nodes):
node_j = data.loc[j]
dist = calculate_distance(node_i[2], node_i[1], node_j[2], node_j[1])
A[i][j] = A[j][i] = dist
# weighted adj
A = np.exp(A**2 / -10)
A[A <= 0.05] = 0
np.save("./node_dict.npy", node_dict)
np.save("./raw_adj_mat.npy", A)
return
def get_pems_node_value(directory, node_dict, data_dir='./txt'):
start_date = datetime.datetime.strptime('2019-01-23', '%Y-%m-%d')
end_date = datetime.datetime.strptime('2019-03-22', '%Y-%m-%d')
day_delta = (end_date - start_date).days + 1
day_slots = 12 * 24
num_timesteps = day_delta * day_slots
num_nodes = len(node_dict)
num_features = 3
X = np.zeros((num_timesteps, num_nodes, num_features))
# only use weekday data
weekday_cnt = 0
for delta in range(day_delta):
cur_date = start_date + relativedelta(days = delta)
if cur_date.isoweekday() <= 5:
time_start = weekday_cnt * day_slots
time_end = time_start + day_slots
X[time_start:time_end] = read_pems_daily_data(data_dir, cur_date, node_dict)
weekday_cnt += 1
num_timesteps = weekday_cnt * day_slots
X = X[:num_timesteps]
# get valid nodes by remove nodes with label nan morn than a certain percentage
percentage = 0.2
threshold = num_timesteps * percentage
nan_cnt = np.count_nonzero(np.isnan(X[:,:,0]), axis=0)
node_index = np.arange(num_nodes)[nan_cnt<threshold]
X = X[:, node_index]
num_nodes = len(node_index)
np.save(directory + "./valid_nodes.npy", node_index)
# transpose to (time, node * feature)
X = X.reshape(num_timesteps, num_nodes * num_features)
df = pd.DataFrame(X)
# linear interpolate
df = df.interpolate(method='linear', limit_direction='both', axis=0)
X = np.array(df).reshape(num_timesteps, num_nodes, num_features)
# output shape is (num_nodes, num_features, num_timesteps)
X = X.transpose(1, 2, 0)
np.save(directory + "./node_values.npy", X)
return node_index
def read_pems_daily_data(data_dir, date, node_dict):
path = data_dir + '/d07_text_station_5min_' + date.strftime("%Y_%m_%d") + '.txt'
if not os.path.exists(path):
print("[ERROR]File %s not exists." % path)
return
# load data
cols_index = [0, 1, 9, 10, 11]
cols_name = ['Time', 'Station', 'Flow', 'Occupancy', 'Speed']
cols_order = ['Time', 'Station', 'Speed', 'Flow', 'Occupancy']
data = pd.read_table(path, header=None, sep=',', usecols=cols_index)
data.columns = cols_name
data = data[cols_order]
num_nodes = len(node_dict)
num_timesteps = 12 * 24
num_feature = 3
daily_X = np.zeros((num_nodes, num_timesteps, num_feature))
# process
for node_info in data.groupby('Station'):
if node_dict.get(node_info[0], -1) == -1:
print("[ERROR]node %d not in dict." % node_info[0])
continue
node_index = node_dict[node_info[0]]
node_feature = node_info[1].iloc[:, -3:]
daily_X[node_index][:len(node_feature)] = np.array(node_feature)
print("Read %s data successfully." % date.strftime("%Y-%m-%d"))
return daily_X.transpose(1, 0, 2)
"""
Some general function
"""
def change_avg_degree(A, K=100):
index = len(A) * K
threshold = sorted(A.flatten(), reverse=True)[index]
A[A <= threshold] = 0
return A
def hav(theta):
s = sin(theta / 2)
return s * s
def calculate_distance(lon1, lat1, lon2, lat2):
'''
Use haversine formula to calculate distance between two points by longtude and latitude.
The output is in kilometers
'''
EARTH_RADIUS=6371 # radius of the earth is 6371km
lat1 = radians(lat1)
lat2 = radians(lat2)
lon1 = radians(lon1)
lon2 = radians(lon2)
dlon = abs(lon1 - lon2)
dlat = abs(lat1 - lat2)
h = hav(dlat) + cos(lat1) * cos(lat2) * hav(dlon)
distance = 2 * EARTH_RADIUS * asin(sqrt(h))
return max(distance, 0.1)
def get_normalized_adj(A):
"""
Returns the degree normalized adjacency matrix.
"""
A = A + np.diag(np.ones(A.shape[0], dtype=np.float32))
D = np.array(np.sum(A, axis=1)).reshape((-1,))
D[D <= 10e-5] = 10e-5 # Prevent infs
diag = np.reciprocal(np.sqrt(D))
A_wave = np.multiply(np.multiply(diag.reshape((-1, 1)), A),
diag.reshape((1, -1)))
return A_wave
def generate_dataset(X, num_timesteps_input, num_timesteps_output, dataset):
"""
Takes node features for the graph and divides them into multiple samples
along the time-axis by sliding a window of size (num_timesteps_input+
num_timesteps_output) across it in steps of 1.
:param X: Node features of shape (num_vertices, num_features,
num_timesteps)
:return:
- Node features divided into multiple samples. Shape is
(num_samples, num_vertices, num_timesteps_input, num_features).
- Node targets for the samples. Shape is
(num_samples, num_vertices, num_timesteps_output).
"""
# PeMS only use weekday data and a day contains 288 slots(5min per slot)
if dataset == "pems" or dataset == "pems-m":
day_slots = 288
else:
day_slots = X.shape[2]
# Save samples
features, target = [], []
for day in range(X.shape[2] // day_slots):
day_start = day_slots * day
day_end = day_slots * (day+1)
X_day = X[:, :, day_start:day_end]
# Generate the beginning index and the ending index of a sample, which
# contains (num_points_for_training + num_points_for_predicting) points
indices = [(i, i + (num_timesteps_input + num_timesteps_output)) for i
in range(X_day.shape[2] - (
num_timesteps_input + num_timesteps_output) + 1)]
for i, j in indices:
features.append(
X_day[:, :, i: i + num_timesteps_input].transpose(
(0, 2, 1)))
target.append(X_day[:, 0, i + num_timesteps_input: j])
return torch.from_numpy(np.array(features)), \
torch.from_numpy(np.array(target))