summerofdata/preproc.py at main · juojuol/summerofdata · GitHub

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
import os
import sys
from datetime import date
import pandas as pd
import datetime as dt
import logging
import glob
from wearables import watchoff
import matplotlib.pyplot as plt

def preproc(in_file, device, sr='1T', truncate=True, write=True, plot=True, recording_period_min=7, interpolate_limit=10, interpolate_method='linear'):


    data = []

    try:
        today = date.today()

        out_dir = os.path.dirname(in_file) + '/preproc/'

        if not os.path.isdir(out_dir):
            os.mkdir(out_dir)
            print("created output directory %s" % (out_dir))

        devices = ['fitbit', 'actiwatch']

        logger = logging.getLogger(__name__)
        f_handler = logging.FileHandler(out_dir + str(recording_period_min) + '_days.log')
        logger.addHandler(f_handler)
        #logging.basicConfig(filename=log_file, filemode='x', format='%(asctime)s - %(message)s', level=logging.INFO)


        if device == 'actiwatch':

            record_id = os.path.basename(in_file).str.split('_')[0] # check this

            with open(in_file) as f:
                for i, l in enumerate(f):
                    if ' Epoch-by-Epoch Data ' in l:
                        try:
                            data = pd.read_csv(in_file, skiprows = i+11, usecols = [1,2,3])
                            print('successfully read Actiware data file')
                        except:
                            try:
                                data = pd.read_csv(in_file, skiprows = i+12, usecols = [1,2,3])
                                print('successfully read Actiware data file')
                            except:
                                print('unable to read Actiware data file')

                        break

            data['Time'] = data['Date'] + ' ' + data['Time']
            data['Time'] = pd.to_datetime(data['Time'])

        elif device == 'fitbit':

            record_id = os.path.basename(in_file).split("WA_")[1][0:5]

            data = pd.read_csv(in_file)
            data.columns = ['Time', 'Activity']
            data['Time'] = pd.to_datetime(data['Time'])

        else:
            raise ValueError("Invalid device type. Expected one of: %s" % devices)

        print('record %s' % (record_id))

        data.index = data['Time']
        data = data.resample(sr).sum()
        data = data['Activity']

        if device == 'fitbit':
            data = watchoff.watchoff(record_id, data, in_file, out_dir)

        start_time = data.first_valid_index() # TO DO: find first non-zero activity value
        end_time = data.last_valid_index()
        period = end_time - start_time

        raw = data
        raw.to_csv(out_dir + '/' + record_id + '.csv', index=True, index_label=None, header=None, na_rep='NaN')

        missingNum = data.isnull().sum()
        error = 0
        logging.info('%s processing' % record_id)

        if missingNum > 0:
            # remove trailing and leading activity values
            length_init = len(data)
            data = data.loc[start_time:end_time]

            logging.info('----- removed leading and trailing NaN activity values')
            missingNum = data.isnull().sum()

        if missingNum > 0:
            # interpolate
            data.interpolate(method=interpolate_method, limit=interpolate_limit, inplace=True, limit_area='inside')
            logging.info('----- interpolated with %s, limit = %s' % (interpolate_method, interpolate_limit))
            if not os.path.isdir(out_dir + '/interpolated/'):
                os.makedirs(out_dir + '/interpolated/')
            data.to_csv(out_dir + '/interpolated/%s_interpolated-method-%s_lim-%s-epoch.csv' % (record_id, interpolate_method, interpolate_limit), index=True, index_label=None, header=None, na_rep='NaN')
            missingNum = data.isnull().sum()

        # truncating to first ndays of data
        if truncate == True:
            data = data[data.index <= (start_time + dt.timedelta(seconds=30) +
                        dt.timedelta(days=recording_period_min))]
            end_time = data.last_valid_index()
            period = end_time - start_time
            logging.info('----- truncated recording period to %s days' % recording_period_min)
            missingNum = data.isnull().sum()
            if not os.path.isdir(out_dir + '/truncated/'):
                os.makedirs(out_dir + '/truncated/')
            data.to_csv(out_dir + '/truncated/%s_truncated-%s_d.csv' % (record_id, recording_period_min), index=True, index_label=None, header=None, na_rep='NaN')

        if plot == True:
            f, axs = plt.subplots(2, 1, sharex=True)
            axs[0].plot(raw.index, raw, color = 'blue')
            axs[0].set_title(record_id + ', ' + str(recording_period_min) + ' days')
            axs[0].xaxis.set_visible(False)

            axs[1].plot(data.index, data, color = 'red')
            plt.xticks(rotation=45)
            plt.tight_layout()

            if not os.path.isdir(out_dir + '/figures/'):
                os.makedirs(out_dir + '/figures/')
            plt.savefig(out_dir + '/figures/' + record_id + '_' + str(recording_period_min) + '_d_interpolate-' + interpolate_method + '.png', dpi = 300)

        if missingNum > 0.10 * len(data):
            print('----- error: missing values = %.2f percent' %
                  (100*(missingNum / len(data))))
            logging.warning(
                '----- discard: missing more than 10 percent of data, %.2f percent missing' % (missingNum / len(data)))
            error = error + 1

        if period < dt.timedelta(days=recording_period_min):
            print('----- error: less than %s days actigraphy data - recording period is %s ' %
                  (str(recording_period_min), str(period)))
            logging.warning('----- discard: insufficient recording period %s' %
                            (str(period)))
            error = error + 1

        if missingNum > 0:
            print('... error: after processing, still missing %.2f percent data' %
                  (100*(missingNum/len(data))))
            logging.warning(
                '----- error: missing %.2f percent after processing' % (100*(missingNum/len(data))))
            error = error + 1

        if error == 0:
            logging.info('----- success: %.2f percent NaN, %s recording period' %
                         (100*(missingNum / len(data)), str(period)))

            print('----- success: %.2f percent NaN, %s recording period' %
                  (100*(missingNum / len(data)), str(period)))
        else:
            print('----- exclude from analysis')

    except Exception as e:
        exc_type, exc_obj, exc_tb = sys.exc_info()
        fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1]
        print('unable to preprocess subject %s' % record_id)
        print(e)
        print(exc_type, fname, exc_tb.tb_lineno)

    return data