forked from philliphartin/taut-sensoranalysis-python
-
Notifications
You must be signed in to change notification settings - Fork 0
/
outputgenerator.py
217 lines (166 loc) · 6.17 KB
/
outputgenerator.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
import csv
import time
from collections import OrderedDict
from arff import arff
out_dir = ['']
def get_timestamp():
timestr = time.strftime("%Y%m%d-%H%M")
return timestr
def create_output_folders(windowlength):
import os
newpath = 'output/' + str(windowlength)
if not os.path.exists(newpath):
os.makedirs(newpath)
out_dir[0] = newpath
return newpath
def traverse(d, sep='_', _prefix=''):
assert isinstance(d, dict)
for k, v in d.items():
if isinstance(v, dict):
yield from traverse(v, sep, "{0}{1}{2}".format(_prefix, k, sep))
else:
yield ("{0}{1}".format(_prefix, k), v)
def flatten(d):
return dict(traverse(d))
def prepare_data(master_data_set):
array = []
for user, data in master_data_set.items():
for reminder in data:
acknowledged = reminder['acknowledged']
unixtime = reminder['unixtime']
sensors = reminder['sensors']
# values to store
reminder_data_dict = {}
reminder_data_dict['acknowledged'] = acknowledged
reminder_data_dict['unixtime'] = unixtime
reminder_data_dict['userid'] = user
# if sensors contains data
if sensors:
flat_sensors = flatten(sensors) # flatten the nested dictionaries
reminder_data_dict.update(flat_sensors) # update the reminder dictionary with new data
array.append(reminder_data_dict) # append as a row in the master array
return array
def write_data_to_disk(data, windowlength):
try:
csv_raw_filepath = write_to_csv(data, windowlength)
except Exception as e:
print(e)
else:
csv_ord_filepath = convert_raw_csv_to_ordered(csv_raw_filepath)
convert_ordered_csv_to_weka(csv_ord_filepath)
def write_to_csv(data, windowlength):
filepath = create_output_folders(windowlength) + '/raw.csv'
with open(filepath, 'w') as csvfile:
ordered_fieldnames = OrderedDict(data[0])
writer = csv.DictWriter(csvfile, fieldnames=ordered_fieldnames)
writer.writeheader()
for entry in data:
writer.writerow(entry)
return filepath
def update_acknowledged_label(number):
if number == '1':
return 'True'
elif number == '0':
return 'False'
else:
return 'Off'
def convert_raw_csv_to_ordered(csv_raw_filepath):
# read in previously generated csv feature list
# get the headers, and order them
filepath = out_dir[0] + '/ordered.csv'
with open(csv_raw_filepath) as csvfile:
reader = csv.DictReader(csvfile)
header = reader.fieldnames
header_sorted = sorted(header)
with open(filepath, 'w') as ordered_file:
fieldnames = header_sorted
writer = csv.DictWriter(ordered_file, fieldnames=fieldnames)
writer.writeheader()
for row in reader:
acknowledged = row['acknowledged']
row['acknowledged'] = update_acknowledged_label(acknowledged)
if row['acknowledged'] == 'Off':
# skip the file line
continue
else:
writer.writerow(row)
return filepath
def write_weka_file_for_cohort(data, attributes):
weka_data = {
'description': '',
'relation': 'sensors',
'attributes': attributes,
'data': data,
}
f = open(out_dir[0] + '/cohort.arff', 'w')
f.write(arff.dumps(weka_data))
f.close()
def write_weka_file_for_each_user(data, attributes):
all_users = {}
for row in data:
userid = row[len(row) - 1] # User ID is the last element in list
# Get User data from row and add to user dictionary
# try to get existing key, and append to the existing list in value
try:
user_list = all_users[userid]
user_list.append(row)
all_users[userid] = user_list
except KeyError:
# If doesn't exist, create the list
user_list = [row]
all_users[userid] = user_list
for user, userdata in all_users.items():
weka_data = {
'description': 'Data for ' + user,
'relation': 'SensorRecordings',
'attributes': attributes,
'data': userdata,
}
# Write Weka formatted file for entire cohort
f = open(out_dir[0] + '/' + user + '.arff', 'w')
f.write(arff.dumps(weka_data))
f.close()
def convert_ordered_csv_to_weka(csv_ord_filepath):
headers = []
data = []
attributes = []
with open(csv_ord_filepath) as csvfile:
readcsv = csv.reader(csvfile, delimiter=',')
row_count = 0
for row in readcsv:
if row_count == 0:
# Save headers for features
headers = row
row_count += 1
else:
data.append(row)
row_count += 1
# iterate the headings to correctly format the attribute types
for attribute in headers:
if attribute == 'acknowledged':
attributes.append(('class', ['True', 'False']))
elif attribute == 'userid':
attributes.append((attribute, 'STRING'))
elif attribute == 'unixtime':
attributes.append((attribute, 'STRING'))
else:
attributes.append((attribute, 'REAL'))
# Get index of acknowledged data
count_acknowledged = 0
count_missed = 0
acknowledged_index = headers.index("acknowledged")
for row in data:
if row[acknowledged_index] == 'True':
count_acknowledged += 1
else:
count_missed += 1
print('Total Reminders Saved: ' + str(count_acknowledged + count_missed))
print('Acknowledged: ' + str(count_acknowledged) + ' / Missed: ' + str(count_missed))
write_weka_file_for_cohort(data, attributes)
# Write Weka format file for each user
write_weka_file_for_each_user(data, attributes)
# FOR DEBUGGING
# # # Use pickle to import object saved to disk
# master_data_set = pickle.load(open('pickle.p', "rb"))
# results = prepare_data(master_data_set)
# write_data_to_disk(results)