-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy path0_download_dataset.py
231 lines (216 loc) · 14.1 KB
/
0_download_dataset.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
import requests
import os
from pathlib import Path
import pickle
from shutil import unpack_archive
urls = dict()
# urls['ecg']=['http://www.cs.ucr.edu/~eamonn/discords/ECG_data.zip',
# 'http://www.cs.ucr.edu/~eamonn/discords/mitdbx_mitdbx_108.txt',
# 'http://www.cs.ucr.edu/~eamonn/discords/qtdbsele0606.txt',
# 'http://www.cs.ucr.edu/~eamonn/discords/chfdbchf15.txt',
# 'http://www.cs.ucr.edu/~eamonn/discords/qtdbsel102.txt']
# urls['gesture']=['http://www.cs.ucr.edu/~eamonn/discords/ann_gun_CentroidA']
# urls['space_shuttle']=['http://www.cs.ucr.edu/~eamonn/discords/TEK16.txt',
# 'http://www.cs.ucr.edu/~eamonn/discords/TEK17.txt',
# 'http://www.cs.ucr.edu/~eamonn/discords/TEK14.txt']
# urls['respiration']=['http://www.cs.ucr.edu/~eamonn/discords/nprs44.txt',
# 'http://www.cs.ucr.edu/~eamonn/discords/nprs43.txt']
# urls['power_demand']=['http://www.cs.ucr.edu/~eamonn/discords/power_data.txt']
# for dataname in urls:
# raw_dir = Path('dataset', dataname, 'raw')
# raw_dir.mkdir(parents=True, exist_ok=True)
# for url in urls[dataname]:
# filename = raw_dir.joinpath(Path(url).name)
# print('Downloading', url)
# resp =requests.get(url)
# filename.write_bytes(resp.content)
# if filename.suffix=='':
# filename.rename(filename.with_suffix('.txt'))
# print('Saving to', filename.with_suffix('.txt'))
# if filename.suffix=='.zip':
# print('Extracting to', filename)
# unpack_archive(str(filename), extract_dir=str(raw_dir))
# for filepath in raw_dir.glob('*.txt'):
# with open(str(filepath)) as f:
# # Label anomaly points as 1 in the dataset
# labeled_data=[]
# for i, line in enumerate(f):
# tokens = [float(token) for token in line.split()]
# if raw_dir.parent.name== 'ecg':
# # Remove time-step channel
# tokens.pop(0)
# if filepath.name == 'chfdbchf15.txt':
# tokens.append(1.0) if 2250 < i < 2400 else tokens.append(0.0)
# elif filepath.name == 'xmitdb_x108_0.txt':
# tokens.append(1.0) if 4020 < i < 4400 else tokens.append(0.0)
# elif filepath.name == 'mitdb__100_180.txt':
# tokens.append(1.0) if 1800 < i < 1990 else tokens.append(0.0)
# elif filepath.name == 'chfdb_chf01_275.txt':
# tokens.append(1.0) if 2330 < i < 2600 else tokens.append(0.0)
# elif filepath.name == 'ltstdb_20221_43.txt':
# tokens.append(1.0) if 650 < i < 780 else tokens.append(0.0)
# elif filepath.name == 'ltstdb_20321_240.txt':
# tokens.append(1.0) if 710 < i < 850 else tokens.append(0.0)
# elif filepath.name == 'chfdb_chf13_45590.txt':
# tokens.append(1.0) if 2800 < i < 2960 else tokens.append(0.0)
# elif filepath.name == 'stdb_308_0.txt':
# tokens.append(1.0) if 2290 < i < 2550 else tokens.append(0.0)
# elif filepath.name == 'qtdbsel102.txt':
# tokens.append(1.0) if 4230 < i < 4430 else tokens.append(0.0)
# elif filepath.name == 'ann_gun_CentroidA.txt':
# tokens.append(1.0) if 2070 < i < 2810 else tokens.append(0.0)
# elif filepath.name == 'TEK16.txt':
# tokens.append(1.0) if 4270 < i < 4370 else tokens.append(0.0)
# elif filepath.name == 'TEK17.txt':
# tokens.append(1.0) if 2100 < i < 2145 else tokens.append(0.0)
# elif filepath.name == 'TEK14.txt':
# tokens.append(1.0) if 1100 < i < 1200 or 1455 < i < 1955 else tokens.append(0.0)
# elif filepath.name == 'nprs44.txt':
# tokens.append(1.0) if 16192 < i < 16638 or 20457 < i < 20911 else tokens.append(0.0)
# elif filepath.name == 'nprs43.txt':
# tokens.append(1.0) if 12929 < i < 13432 or 14877 < i < 15086 or 15729 < i < 15924 else tokens.append(0.0)
# elif filepath.name == 'power_data.txt':
# tokens.append(1.0) if 8254 < i < 8998 or 11348 < i < 12143 or 33883 < i < 34601 else tokens.append(0.0)
# labeled_data.append(tokens)
# # Fill in the point where there is no signal value.
# if filepath.name == 'ann_gun_CentroidA.txt':
# for i, datapoint in enumerate(labeled_data):
# for j,channel in enumerate(datapoint[:-1]):
# if channel == 0.0:
# labeled_data[i][j] = 0.5 * labeled_data[i - 1][j] + 0.5 * labeled_data[i + 1][j]
# # Save the labeled dataset as .pkl extension
# labeled_whole_dir = raw_dir.parent.joinpath('labeled', 'whole')
# labeled_whole_dir.mkdir(parents=True, exist_ok=True)
# with open(str(labeled_whole_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data, pkl)
# # Divide the labeled dataset into trainset and testset, then save them
# labeled_train_dir = raw_dir.parent.joinpath('labeled','train')
# labeled_train_dir.mkdir(parents=True,exist_ok=True)
# labeled_test_dir = raw_dir.parent.joinpath('labeled','test')
# labeled_test_dir.mkdir(parents=True,exist_ok=True)
# if filepath.name == 'chfdb_chf13_45590.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[:2439], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[2439:3726], pkl)
# elif filepath.name == 'chfdb_chf01_275.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[:1833], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[1833:3674], pkl)
# elif filepath.name == 'chfdbchf15.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[3381:14244], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[33:3381], pkl)
# elif filepath.name == 'qtdbsel102.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[10093:44828], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[211:10093], pkl)
# elif filepath.name == 'mitdb__100_180.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[2328:5271], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[73:2328], pkl)
# elif filepath.name == 'stdb_308_0.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[2986:5359], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[265:2986], pkl)
# elif filepath.name == 'ltstdb_20321_240.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[1520:3531], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[73:1520], pkl)
# elif filepath.name == 'xmitdb_x108_0.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[424:3576], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[3576:5332], pkl)
# elif filepath.name == 'ltstdb_20221_43.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[1121:3731], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[0:1121], pkl)
# elif filepath.name == 'ann_gun_CentroidA.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[3000:], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[:3000], pkl)
# elif filepath.name == 'nprs44.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[363:12955], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[12955:24082], pkl)
# elif filepath.name == 'nprs43.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[4285:10498], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[10498:17909], pkl)
# elif filepath.name == 'power_data.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[15287:33432], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[501:15287], pkl)
# elif filepath.name == 'TEK17.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[2469:4588], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[1543:2469], pkl)
# elif filepath.name == 'TEK16.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[521:3588], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[3588:4539], pkl)
# elif filepath.name == 'TEK14.txt':
# with open(str(labeled_train_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[2089:4098], pkl)
# with open(str(labeled_test_dir.joinpath(filepath.name).with_suffix('.pkl')), 'wb') as pkl:
# pickle.dump(labeled_data[97:2089], pkl)
# nyc_taxi_raw_path = Path('dataset/nyc_taxi/raw/nyc_taxi.csv')
# labeled_data = []
# with open(str(nyc_taxi_raw_path),'r') as f:
# for i, line in enumerate(f):
# tokens = [float(token) for token in line.strip().split(',')[1:]]
# tokens.append(1) if 150 < i < 250 or \
# 5970 < i < 6050 or \
# 8500 < i < 8650 or \
# 8750 < i < 8890 or \
# 10000 < i < 10200 or \
# 14700 < i < 14800 \
# else tokens.append(0)
# labeled_data.append(tokens)
# nyc_taxi_train_path = nyc_taxi_raw_path.parent.parent.joinpath('labeled','train',nyc_taxi_raw_path.name).with_suffix('.pkl')
# nyc_taxi_train_path.parent.mkdir(parents=True, exist_ok=True)
# with open(str(nyc_taxi_train_path),'wb') as pkl:
# pickle.dump(labeled_data[:13104], pkl)
# nyc_taxi_test_path = nyc_taxi_raw_path.parent.parent.joinpath('labeled','test',nyc_taxi_raw_path.name).with_suffix('.pkl')
# nyc_taxi_test_path.parent.mkdir(parents=True, exist_ok=True)
# with open(str(nyc_taxi_test_path),'wb') as pkl:
# pickle.dump(labeled_data[13104:], pkl)
bearing_sets = []
import pdb
for root, dir, files in os.walk('dataset'):
for file in files:
if("combinedfiles" in file and ".csv" in file):
bearing_sets.append(root + '/' + file)
for set in bearing_sets:
labeled_data = []
raw_path = Path(set)
with open(str(raw_path),'r') as f:
for i, line in enumerate(f):
if(i==0): #skip header
continue
tokens = [float(token) for token in line.strip().split(',')[1:]]
tokens.append(0)
labeled_data.append(tokens)
train_len = int(0.15 * len(labeled_data))
train_path = raw_path.parent.joinpath('labeled','train',raw_path.name).with_suffix('.pkl')
train_path.parent.mkdir(parents=True, exist_ok=True)
with open(str(train_path),'wb') as pkl:
pickle.dump(labeled_data[:train_len], pkl)
test_path = raw_path.parent.joinpath('labeled','test',raw_path.name).with_suffix('.pkl')
test_path.parent.mkdir(parents=True, exist_ok=True)
with open(str(test_path),'wb') as pkl:
pickle.dump(labeled_data, pkl) #Test data is just the full length file. Test overlaps train.