2022-03-22 13:48:43 +01:00
|
|
|
import pandas as pd
|
2022-05-20 15:18:45 +02:00
|
|
|
import numpy as np
|
2022-03-22 13:48:43 +01:00
|
|
|
from scipy.stats import entropy
|
|
|
|
|
2022-04-19 15:24:46 +02:00
|
|
|
from cr_features.helper_functions import convert_to2d, gsr_features
|
|
|
|
from cr_features.calculate_features import calculate_features
|
2022-06-03 14:34:36 +02:00
|
|
|
from cr_features.gsr import extractGsrFeatures2D
|
2022-05-10 13:36:49 +02:00
|
|
|
from cr_features_helper_methods import extract_second_order_features
|
2022-03-23 16:13:53 +01:00
|
|
|
|
2022-05-20 15:18:45 +02:00
|
|
|
import sys
|
|
|
|
|
|
|
|
#pd.set_option('display.max_columns', None)
|
|
|
|
#pd.set_option('display.max_rows', None)
|
|
|
|
#np.seterr(invalid='ignore')
|
|
|
|
|
2022-03-22 13:48:43 +01:00
|
|
|
|
2022-04-19 15:24:46 +02:00
|
|
|
def extract_eda_features_from_intraday_data(eda_intraday_data, features, window_length, time_segment, filter_data_by_segment):
|
2022-03-22 13:48:43 +01:00
|
|
|
eda_intraday_features = pd.DataFrame(columns=["local_segment"] + features)
|
2022-03-23 16:13:53 +01:00
|
|
|
|
2022-03-28 15:50:08 +02:00
|
|
|
if not eda_intraday_data.empty:
|
2022-05-10 13:36:49 +02:00
|
|
|
sample_rate = 4
|
2022-03-28 15:50:08 +02:00
|
|
|
|
2022-03-22 13:48:43 +01:00
|
|
|
eda_intraday_data = filter_data_by_segment(eda_intraday_data, time_segment)
|
|
|
|
|
2022-05-20 15:18:45 +02:00
|
|
|
if not eda_intraday_data.empty:
|
2022-03-23 16:13:53 +01:00
|
|
|
|
2022-03-22 13:48:43 +01:00
|
|
|
eda_intraday_features = pd.DataFrame()
|
|
|
|
|
2022-03-28 14:37:02 +02:00
|
|
|
# apply methods from calculate features module
|
2022-04-12 16:00:44 +02:00
|
|
|
if window_length is None:
|
|
|
|
eda_intraday_features = \
|
|
|
|
eda_intraday_data.groupby('local_segment').apply(\
|
2022-06-30 17:15:37 +02:00
|
|
|
lambda x: extractGsrFeatures2D(convert_to2d(x['electrodermal_activity'], x.shape[0]), sampleRate=sample_rate, featureNames=features,
|
|
|
|
threshold=.01, offset=1, riseTime=5, decayTime=15))
|
2022-04-12 16:00:44 +02:00
|
|
|
else:
|
|
|
|
eda_intraday_features = \
|
|
|
|
eda_intraday_data.groupby('local_segment').apply(\
|
2022-06-30 17:15:37 +02:00
|
|
|
lambda x: extractGsrFeatures2D(convert_to2d(x['electrodermal_activity'], window_length*sample_rate), sampleRate=sample_rate, featureNames=features,
|
|
|
|
threshold=.01, offset=1, riseTime=5, decayTime=15))
|
2022-04-12 16:00:44 +02:00
|
|
|
|
2022-03-22 13:48:43 +01:00
|
|
|
eda_intraday_features.reset_index(inplace=True)
|
|
|
|
|
|
|
|
return eda_intraday_features
|
|
|
|
|
|
|
|
|
2022-04-12 16:00:44 +02:00
|
|
|
def cr_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs):
|
2022-03-22 13:48:43 +01:00
|
|
|
eda_intraday_data = pd.read_csv(sensor_data_files["sensor_data"])
|
|
|
|
|
|
|
|
requested_intraday_features = provider["FEATURES"]
|
2022-04-12 16:00:44 +02:00
|
|
|
|
2022-04-13 15:18:23 +02:00
|
|
|
calc_windows = kwargs.get('calc_windows', False)
|
|
|
|
|
|
|
|
if provider["WINDOWS"]["COMPUTE"] and calc_windows:
|
2022-04-12 16:00:44 +02:00
|
|
|
requested_window_length = provider["WINDOWS"]["WINDOW_LENGTH"]
|
|
|
|
else:
|
|
|
|
requested_window_length = None
|
|
|
|
|
2022-03-22 13:48:43 +01:00
|
|
|
# name of the features this function can compute
|
2022-04-19 15:24:46 +02:00
|
|
|
base_intraday_features_names = gsr_features
|
2022-03-22 13:48:43 +01:00
|
|
|
# the subset of requested features this function can compute
|
|
|
|
intraday_features_to_compute = list(set(requested_intraday_features) & set(base_intraday_features_names))
|
|
|
|
|
|
|
|
# extract features from intraday data
|
2022-04-19 15:24:46 +02:00
|
|
|
eda_intraday_features = extract_eda_features_from_intraday_data(eda_intraday_data, intraday_features_to_compute,
|
2022-04-12 16:00:44 +02:00
|
|
|
requested_window_length, time_segment, filter_data_by_segment)
|
2022-03-22 13:48:43 +01:00
|
|
|
|
2022-04-25 15:07:03 +02:00
|
|
|
if calc_windows:
|
2022-06-21 16:09:49 +02:00
|
|
|
if provider["WINDOWS"]["IMPUTE_NANS"]:
|
|
|
|
eda_intraday_features[eda_intraday_features["numPeaks"] == 0] = \
|
|
|
|
eda_intraday_features[eda_intraday_features["numPeaks"] == 0].fillna(0)
|
|
|
|
pd.set_option('display.max_columns', None)
|
|
|
|
|
2022-04-25 15:07:03 +02:00
|
|
|
so_features_names = provider["WINDOWS"]["SECOND_ORDER_FEATURES"]
|
|
|
|
eda_second_order_features = extract_second_order_features(eda_intraday_features, so_features_names)
|
2022-05-20 15:18:45 +02:00
|
|
|
|
2022-04-25 15:07:03 +02:00
|
|
|
return eda_intraday_features, eda_second_order_features
|
|
|
|
|
2022-03-22 13:48:43 +01:00
|
|
|
return eda_intraday_features
|