-
Notifications
You must be signed in to change notification settings - Fork 9
/
load_data_traj.py
503 lines (388 loc) · 23.4 KB
/
load_data_traj.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
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
from pathlib import Path
import numpy as np
import pandas as pd
from random import randint
from numpy import array
from numpy import argmax
from math import sin, cos, sqrt, atan2, radians
#from geopy import distance
import matplotlib.pyplot as plt
import os
import h5py
import pickle
from PIL import Image
from datetime import date, time, datetime
from PythonCode.KF.ENUtransform import WGS84toENU, ENUtoWGS84
from itertools import groupby
from operator import itemgetter
import math
from os.path import join
interactive = True
headers=['index1', 'ship_nr','id','repeat_indicator','mmsi','nav_status','rot_over_range','rot','sog','position_accuracy','x','y','cog','true_heading',
'timestamp','special_manoeuvre','spare','raim','sync_state','slot_timeout','slot_offset', 'abs_time', 'date', 'time']
SAMPLING_TIME = 30 # actually 60 seconds, because sampling on old data is already done at 2 seconds interval
legecy_seq_len = 60 # number of data per track does not include extra 60 samples, coming from legacy
nr_of_actual_vessels = 120 #30
ROSTOCK = (12.114733, 54.145409)
def load_all_data(dim, INPUT_LEN, PRED_LEN):
############## generate data from pickle to train ANN ##
np.random.seed(10)
path = Path('/home/sing_sd/Desktop/anomaly_detection/PythonCode/Trajectory_Prediction/')
in_clm_len = dim * INPUT_LEN
out_clm_len = dim * PRED_LEN
filename = path / 'all_sampled_tracks_interpol_1minute.csv'
with open(filename, 'rb') as f:
data = pd.read_csv(f)
data = np.array(data)
filename = path / 'data_len_sampled_tracks_interpol_1minute.pkl'
with open(filename, 'rb') as f:
data_all_tracks = pickle.load(f)
overall_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels]) + (legecy_seq_len - 1) * nr_of_actual_vessels), in_clm_len), fill_value=np.nan)
# make output column same as input column with appending zeros to out_column_len
target_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels]) + (legecy_seq_len - 1) * nr_of_actual_vessels), in_clm_len), fill_value=np.nan)
# target_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels])+(timesteps-1)*nr_of_actual_vessels), out_clm_len), fill_value=np.nan)
nr_track = 0
startIndex = 0
for track_nr in range(nr_of_actual_vessels): #total number of vessels between 0 and 150, some vessels have less data so ignored
if data_all_tracks[track_nr] < 500:
continue
nr_track += 1
nr_sampled_data = data_all_tracks[track_nr]
# shift from top and put on remaining columns
# plot_trajectory(data[startIndex: startIndex + nr_sampled_data, 0:2])
overall_data[startIndex: startIndex + nr_sampled_data, 0:dim] = data[startIndex: startIndex + nr_sampled_data, 0:dim]
for clm_nr in range(1, INPUT_LEN):
overall_data[startIndex: startIndex + nr_sampled_data - clm_nr,
clm_nr * dim:(clm_nr + 1) * dim] = overall_data[
startIndex + 1: startIndex + nr_sampled_data - clm_nr + 1,
(clm_nr - 1) * dim:clm_nr * dim]
target_data[startIndex:startIndex + nr_sampled_data-INPUT_LEN-PRED_LEN+1, 0:out_clm_len] = \
overall_data[startIndex + INPUT_LEN:startIndex + nr_sampled_data - PRED_LEN+1, 0:out_clm_len]
# clear last few rows that does not have sufficient data to make ...
overall_data[startIndex + nr_sampled_data - INPUT_LEN-PRED_LEN+1 : startIndex + nr_sampled_data] = np.nan
startIndex += nr_sampled_data
overall_data = overall_data[~np.isnan(overall_data[:, 0])]
target_data = target_data[~np.isnan(target_data[:, 0])]
target_data = np.nan_to_num(target_data)
print('Overall data size= ', len(overall_data))
print("nr of tracks for training = ", nr_track)
return overall_data, target_data
# plt.rcParams.update({'font.size': 16})
# fig, ax = plt.subplots(figsize=(8, 7))
# im = Image.open('world_11_54_12-4_55.PNG') # in degrees and minutes
# ax.imshow(im, extent=(11, 12.6666, 54.0, 55), aspect='auto')
# ax.plot(overall_data[:,0], overall_data[:,1], 'w.',label='Trajectories of vessels')
# boundary_x, boundary_y = get_po_boundary()
# ax.plot(boundary_x, boundary_y, 'k-', markersize=8, label='AIS transmission reach')
# ax.plot(ROSTOCK[0], ROSTOCK[1], 'ko', markersize=12, label='Rostock location')
# plt.pause(0.01)
# ax.plot(overall_data[ind_ano[ind_po], 0], overall_data[ind_ano[ind_po], 1], 'bo', markersize=10,label='Power outage')
# ax.plot(overall_data[ind_ano[ind_ano1], 0], overall_data[ind_ano[ind_ano1], 1], 'rx', markersize=10,label='AIS anomaly')
# plt.xlabel('Longitude')
# plt.ylabel('Latitude')
# plt.title('AIS on-off switching anomaly detection')
# ax.legend()
# plt.show()
# plt.pause(0.01)
def load_all_data_ENU(dim, INPUT_LEN, PRED_LEN):
############## generate data from pickle to train ANN ##
np.random.seed(10)
path = Path('/home/sing_sd/Desktop/anomaly_detection/PythonCode/Trajectory_Prediction/')
in_clm_len = dim * INPUT_LEN
out_clm_len = dim * PRED_LEN
filename = path / 'all_sampled_tracks_interpol_1minute.csv'
with open(filename, 'rb') as f:
data = pd.read_csv(f)
data = np.array(data)
# # convert lon, lat in ENU
lon = np.transpose(data[:, 0])
lat = np.transpose(data[:, 1])
tREF = {"lon": 12.114733,
"lat": 54.145409,
"ECEF": np.array([[3660725], [785776], [514624]])
}
zENU = WGS84toENU(lon, lat, tREF, h=0.)
data[:, 0] = np.transpose(zENU[0, :])
data[:, 1] = np.transpose(zENU[1, :])
# if no z dimension then comment next line
data = np.insert(data, 2, np.transpose(zENU[2, :]), axis=1) # insert zeros on 3rd column in data for
filename = path / 'data_len_sampled_tracks_interpol_1minute.pkl'
with open(filename, 'rb') as f:
data_all_tracks = pickle.load(f)
# plt.plot(data[data_all_tracks[0]:data_all_tracks[0]+data_all_tracks[1], 0], data[data_all_tracks[0]:data_all_tracks[0]+data_all_tracks[1], 1])
# plt.show()
overall_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels]) + (legecy_seq_len - 1) * nr_of_actual_vessels), in_clm_len), fill_value=np.nan)
# make output column same as input column with appending zeros to out_column_len
target_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels]) + (legecy_seq_len - 1) * nr_of_actual_vessels), in_clm_len), fill_value=np.nan)
# target_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels])+(timesteps-1)*nr_of_actual_vessels), out_clm_len), fill_value=np.nan)
startIndex = 0
for track_nr in range(nr_of_actual_vessels): #total number of vessels between 0 and 150, some vessels have less data so ignored
nr_sampled_data = data_all_tracks[track_nr]
# shift from top and put on remaining columns
#plot_trajectory(data[startIndex: startIndex + nr_sampled_data, 0:2])
overall_data[startIndex: startIndex + nr_sampled_data, 0:dim] = data[startIndex: startIndex + nr_sampled_data, 0:dim]
for clm_nr in range(1, INPUT_LEN):
overall_data[startIndex: startIndex + nr_sampled_data - clm_nr,
clm_nr * dim:(clm_nr + 1) * dim] = overall_data[
startIndex + 1: startIndex + nr_sampled_data - clm_nr + 1,
(clm_nr - 1) * dim:clm_nr * dim]
target_data[startIndex:startIndex + nr_sampled_data-INPUT_LEN-PRED_LEN+1, 0:out_clm_len] = \
overall_data[startIndex + INPUT_LEN:startIndex + nr_sampled_data - PRED_LEN+1, 0:out_clm_len]
# clear last few rows that does not have sufficient data to make ...
overall_data[startIndex + nr_sampled_data - INPUT_LEN-PRED_LEN+1 : startIndex + nr_sampled_data] = np.nan
startIndex += nr_sampled_data
overall_data = overall_data[~np.isnan(overall_data[:, 0])]
target_data = target_data[~np.isnan(target_data[:, 0])]
target_data = np.nan_to_num(target_data)
print('Overall data size= ', len(overall_data))
return overall_data, target_data
def load_all_data_ENU_1track(dim, INPUT_LEN, PRED_LEN):
############## generate data from pickle to train ANN ##
np.random.seed(10)
path = Path('/home/sing_sd/Desktop/anomaly_detection/PythonCode/Trajectory_Prediction/')
in_clm_len = dim * INPUT_LEN
out_clm_len = dim * PRED_LEN
filename = path / 'Track167_interpolated_1min.csv'
with open(filename, 'rb') as f:
data = pd.read_csv(f)
data = np.array(data)
# # convert lon, lat in ENU
lon = np.transpose(data[:, 0])
lat = np.transpose(data[:, 1])
tREF = {"lon": 12.114733,
"lat": 54.145409,
"ECEF": np.array([[3660725], [785776], [514624]])
}
zENU = WGS84toENU(lon, lat, tREF, h=0.)
data_all_tracks = len(zENU[0,:])*np.ones(shape=nr_of_actual_vessels, dtype=int)
# plt.plot(data[data_all_tracks[0]:data_all_tracks[0]+data_all_tracks[1], 0], data[data_all_tracks[0]:data_all_tracks[0]+data_all_tracks[1], 1])
# plt.show()
overall_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels]) + (legecy_seq_len - 1) * nr_of_actual_vessels), in_clm_len), fill_value=np.nan)
# make output column same as input column with appending zeros to out_column_len
target_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels]) + (legecy_seq_len - 1) * nr_of_actual_vessels), in_clm_len), fill_value=np.nan)
# target_data = np.full(shape=(int(sum(data_all_tracks[:nr_of_actual_vessels])+(timesteps-1)*nr_of_actual_vessels), out_clm_len), fill_value=np.nan)
plt.plot(zENU[0, :], zENU[1, :], 'b.', markersize=2)
startIndex = 0
for track_nr in range(nr_of_actual_vessels): #total number of vessels between 0 and 150, some vessels have less data so ignored
nr_sampled_data = data_all_tracks[track_nr]
# shift from top and put on remaining columns
np.random.seed(track_nr)
uncertainty = np.random.uniform(-2000, 2000, nr_sampled_data)
data[:, 0] = np.transpose(zENU[0, :] + uncertainty/2)
data[:, 1] = np.transpose(zENU[1, :] + uncertainty/3)
# if no z dimension then comment next line
data = np.insert(data, 2, np.transpose(zENU[2, :]), axis=1) # insert zeros on 3rd column in data for
plt.plot(data[:, 0], data[:, 1], 'k.', markersize=2)
plt.pause(0.001)
#plot_trajectory(data[startIndex: startIndex + nr_sampled_data, 0:2])
#continue
overall_data[startIndex: startIndex + nr_sampled_data, 0:dim] = data[0: nr_sampled_data, 0:dim]
for clm_nr in range(1, INPUT_LEN):
overall_data[startIndex: startIndex + nr_sampled_data - clm_nr,
clm_nr * dim:(clm_nr + 1) * dim] = overall_data[
startIndex + 1: startIndex + nr_sampled_data - clm_nr + 1,
(clm_nr - 1) * dim:clm_nr * dim]
target_data[startIndex:startIndex + nr_sampled_data-INPUT_LEN-PRED_LEN+1, 0:out_clm_len] = \
overall_data[startIndex + INPUT_LEN:startIndex + nr_sampled_data - PRED_LEN+1, 0:out_clm_len]
# clear last few rows that does not have sufficient data to make ...
overall_data[startIndex + nr_sampled_data - INPUT_LEN-PRED_LEN+1 : startIndex + nr_sampled_data] = np.nan
startIndex += nr_sampled_data
overall_data = overall_data[~np.isnan(overall_data[:, 0])]
target_data = target_data[~np.isnan(target_data[:, 0])]
target_data = np.nan_to_num(target_data)
print('Overall data size= ', len(overall_data))
return overall_data, target_data
def plot_trajectory(data):
plt.rcParams.update({'font.size': 16})
fig, ax = plt.subplots(figsize=(8, 7))
# im = Image.open('\Sandeep\Python\AnomalyDetection\src\world_11_54_12-4_55.PNG') # in degrees and minutes
# ax.imshow(im, extent=(11, 12.6666, 54.0, 55), aspect='auto')
ax.plot(data[:,0], data[:,1], 'k.',markersize=2, label='Trajectories of vessels')
#boundary_x, boundary_y = get_po_boundary()
#ax.plot(boundary_x, boundary_y, 'k-', markersize=8, label='AIS transmission reach')
#ax.plot(ROSTOCK[0], ROSTOCK[1], 'ko', markersize=12, label='Rostock location')
plt.xlabel('Longitude')
plt.ylabel('Latitude')
# plt.title('AIS on-off switching anomaly detection')
ax.legend()
#plt.show()
plt.pause(0.01)
def load_test_data(INPUT_LEN, TARGET_LEN, features, dim, track_to_check):
in_clm_len = INPUT_LEN*dim
out_clm_len = TARGET_LEN * dim
interpolate_interval = 60 # seconds
SAMPLING_TIME = 1 # second
path = '/home/sing_sd/Desktop/anomaly_detection/PythonCode/Resources/track_pickle/'
filename = 'track{}'.format(track_to_check)
try:
data = pd.read_pickle(path + filename + '.pkl')
except IOError:
print("Error: File does not appear to exist for track ", track_to_check)
return 0, 0
# without interpolation
original_data = np.array([data.x, data.y]).transpose()
start_time = datetime.strptime(data.iloc[0]['date'] + ' ' + data.iloc[0]['time'],
'%m/%d/%Y %H:%M:%S')
end_time = datetime.strptime(data.iloc[-1]['date'] + ' ' + data.iloc[-1]['time'],
'%m/%d/%Y %H:%M:%S')
data_per_track = int((end_time - start_time).total_seconds() // interpolate_interval + 1) # interpolation interval = 2 seconds
sampling_indices = range(0, data_per_track, SAMPLING_TIME)
nr_sampled_data = len(sampling_indices)
overall_data = np.full(shape=(nr_sampled_data, in_clm_len), fill_value=np.nan)
# make output column same as input column with appending zeros to out_column_len
target_data = np.full(shape=(nr_sampled_data, in_clm_len), fill_value=np.nan)
temp_data = pd.DataFrame(index=range(data_per_track), columns=features, dtype=np.float)
for slot_index in range(0, data.shape[0]): # //
current_time = datetime.strptime(data.iloc[slot_index]['date'] + ' ' + data.iloc[slot_index]['time'],
'%m/%d/%Y %H:%M:%S')
index1 = int((current_time - start_time).total_seconds()) // interpolate_interval
temp_data.iloc[index1, 0:dim] = data.iloc[slot_index, 2:dim+2]
# interpolate
temp_data.iloc[0:sampling_indices[INPUT_LEN], :] = temp_data.iloc[0:sampling_indices[INPUT_LEN], :].interpolate(method='linear', limit_direction='forward', axis=0)
resampled_data = temp_data.iloc[sampling_indices, 0:dim]
# resampled_data.to_csv("Track167.csv", index=False)
overall_data[0:nr_sampled_data, 0:dim] = temp_data.iloc[sampling_indices, 0:dim]
temp_data_interpolated = resampled_data.interpolate(method='linear', limit_direction='forward', axis=0)
temp_data_interpolated.to_csv("Track167_interpolated_1min.csv",index=False)
temp_data_interpolated = np.array(temp_data_interpolated.iloc[:, 0:dim])
aa = temp_data.iloc[sampling_indices, 0:dim]
# aa.to_csv("Track167.csv",index=False)
# shift from top and put on remaining columns
for clm_nr in range(1, INPUT_LEN):
overall_data[0: nr_sampled_data - clm_nr, clm_nr * dim:(clm_nr + 1) * dim] = overall_data[
1: nr_sampled_data - clm_nr + 1,
(clm_nr - 1) * dim:clm_nr * dim]
return np.array(resampled_data), overall_data, temp_data_interpolated
def load_test_data_ENU(INPUT_LEN, TARGET_LEN, features, dim, track_to_check):
in_clm_len = INPUT_LEN*dim
out_clm_len = TARGET_LEN * dim
interpolate_interval = 60 # seconds
SAMPLING_TIME = 1 # second
data_features = ["x", "y", "cog", "sog"]
data_dim = len(data_features)
path = '/home/sing_sd/Desktop/anomaly_detection/PythonCode/Resources/track_pickle/'
filename = 'track{}'.format(track_to_check)
try:
data = pd.read_pickle(path + filename + '.pkl')
except IOError:
print("Error: File does not appear to exist for track ", track_to_check)
return 0, 0
# without interpolation
original_data = np.array([data.x, data.y]).transpose()
start_time = datetime.strptime(data.iloc[0]['date'] + ' ' + data.iloc[0]['time'],
'%m/%d/%Y %H:%M:%S')
end_time = datetime.strptime(data.iloc[-1]['date'] + ' ' + data.iloc[-1]['time'],
'%m/%d/%Y %H:%M:%S')
data_per_track = int((end_time - start_time).total_seconds() // interpolate_interval + 1) # interpolation interval = 2 seconds
sampling_indices = range(0, data_per_track, SAMPLING_TIME)
nr_sampled_data = len(sampling_indices)
overall_data = np.full(shape=(nr_sampled_data, in_clm_len), fill_value=np.nan)
# make output column same as input column with appending zeros to out_column_len
target_data = np.full(shape=(nr_sampled_data, in_clm_len), fill_value=np.nan)
temp_data = pd.DataFrame(index=range(data_per_track), columns=data_features, dtype=np.float)
for slot_index in range(0, data.shape[0]): # //
current_time = datetime.strptime(data.iloc[slot_index]['date'] + ' ' + data.iloc[slot_index]['time'],
'%m/%d/%Y %H:%M:%S')
index1 = int((current_time - start_time).total_seconds()) // interpolate_interval
temp_data.iloc[index1, 0:data_dim] = data.iloc[slot_index, 2:data_dim+2]
# interpolate
temp_data.iloc[0:sampling_indices[INPUT_LEN], :] = temp_data.iloc[0:sampling_indices[INPUT_LEN], :].interpolate(method='linear', limit_direction='forward', axis=0)
resampled_data = temp_data.iloc[sampling_indices, 0:data_dim]
# resampled_data.to_csv("Track167.csv", index=False)
# put lon and lat data
overall_data[0:nr_sampled_data, 0:2] = temp_data.iloc[sampling_indices, 0:2]
temp_data_interpolated = resampled_data.interpolate(method='linear', limit_direction='forward', axis=0)
# temp_data_interpolated.to_csv("Track167_interpolated_1min.csv",index=False)
temp_data_interpolated = np.array(temp_data_interpolated.iloc[:, 0:data_dim])
# aa = temp_data.iloc[sampling_indices, 0:dim]
# aa.to_csv("Track167.csv",index=False)
# convert lon, lat in ENU
lon = np.transpose(overall_data[:, 0])
lat = np.transpose(overall_data[:, 1])
tREF = {"lon": 12.114733,
"lat": 54.145409,
"ECEF": np.array([[3660725], [785776], [514624]])
}
zENU = WGS84toENU(lon, lat, tREF, h=0.)
overall_data[:, 0] = np.transpose(zENU[0, :])
overall_data[:, 1] = np.transpose(zENU[1, :])
# put z data
overall_data[:, 2] = np.transpose(zENU[2, :]) # insert zeros on 3rd column in data for
# put cog and sog
overall_data[:, 3:dim] = temp_data.iloc[sampling_indices, 2:data_dim]
# shift from top and put on remaining columns
for clm_nr in range(1, INPUT_LEN):
overall_data[0: nr_sampled_data - clm_nr, clm_nr * dim:(clm_nr + 1) * dim] = overall_data[
1: nr_sampled_data - clm_nr + 1,
(clm_nr - 1) * dim:clm_nr * dim]
return np.array(resampled_data), overall_data, temp_data_interpolated
def load_test_interpolated_data(INPUT_LEN, TARGET_LEN, features, dim, track_to_check):
in_clm_len = INPUT_LEN*dim
out_clm_len = TARGET_LEN * dim
interpolate_interval = 60 # seconds
SAMPLING_TIME = 1
path = '/home/sing_sd/Desktop/anomaly_detection/PythonCode/Resources/track_pickle/'
filename = 'track{}'.format(track_to_check)
try:
data = pd.read_pickle(path + filename + '.pkl')
except IOError:
print("Error: File does not appear to exist for track ", track_to_check)
return 0, 0
# without interpolation
original_data = np.array([data.x, data.y]).transpose()
start_time = datetime.strptime(data.iloc[0]['date'] + ' ' + data.iloc[0]['time'],
'%m/%d/%Y %H:%M:%S')
end_time = datetime.strptime(data.iloc[-1]['date'] + ' ' + data.iloc[-1]['time'],
'%m/%d/%Y %H:%M:%S')
data_per_track = int((end_time - start_time).total_seconds() // interpolate_interval + 1) # interpolation interval = 2 seconds
sampling_indices = range(0, data_per_track, SAMPLING_TIME)
nr_sampled_data = len(sampling_indices)
overall_data = np.full(shape=(nr_sampled_data, in_clm_len), fill_value=np.nan)
# make output column same as input column with appending zeros to out_column_len
target_data = np.full(shape=(nr_sampled_data, in_clm_len), fill_value=np.nan)
temp_data = pd.DataFrame(index=range(data_per_track), columns=features, dtype=np.float)
for slot_index in range(0, data.shape[0]): # //
current_time = datetime.strptime(data.iloc[slot_index]['date'] + ' ' + data.iloc[slot_index]['time'],
'%m/%d/%Y %H:%M:%S')
index1 = int((current_time - start_time).total_seconds()) // interpolate_interval
temp_data.iloc[index1, 0:dim] = data.iloc[slot_index, 2:dim+2]
# interpolate
temp_data = temp_data.interpolate(method='linear', limit_direction='forward', axis=0)
overall_data[0:nr_sampled_data, 0:dim] = temp_data.iloc[sampling_indices, 0:dim]
# shift from top and put on remaining columns
for clm_nr in range(1, INPUT_LEN):
overall_data[0: nr_sampled_data - clm_nr, clm_nr * dim:(clm_nr + 1) * dim] = overall_data[
1: nr_sampled_data - clm_nr + 1, (clm_nr - 1) * dim:clm_nr * dim]
target_data[0:nr_sampled_data - INPUT_LEN - TARGET_LEN + 1, 0:out_clm_len] = overall_data[
INPUT_LEN: nr_sampled_data - TARGET_LEN + 1, 0:out_clm_len]
# clear last few rows that does not have sufficient data to make ...
overall_data[nr_sampled_data - INPUT_LEN + 1: nr_sampled_data] = np.nan
overall_data = overall_data[~np.isnan(overall_data[:, 0])]
target_data = target_data[~np.isnan(target_data[:, 0])]
target_data = np.nan_to_num(target_data)
return original_data, overall_data, target_data
# one hot encode sequence
def one_hot_encode(sequence, n_unique):
encoding = list()
for value in sequence:
vector = [0 for _ in range(n_unique)]
vector[value] = 1
encoding.append(vector)
return array(encoding)
# decode a one hot encoded string
def one_hot_decode(encoded_seq):
return [argmax(vector) for vector in encoded_seq]
# prepare data for the LSTM
def get_input_output(dim, INPUT_LEN, PRED_LEN):
n_unique = 360
in_clm_len = dim * INPUT_LEN
out_clm_len = dim * PRED_LEN
# generate random sequence
sequence_in, sequence_out = load_all_data(dim, INPUT_LEN, PRED_LEN)
# one hot encode
X = one_hot_encode(sequence_in, n_unique)
y = one_hot_encode(sequence_out, n_unique)
# reshape as 3D
X = X.reshape((1, X.shape[0], X.shape[1]))
y = y.reshape((1, y.shape[0], y.shape[1]))
return X, y