From 74cf4ada1c20c093648fd9cf5061640c456e5a18 Mon Sep 17 00:00:00 2001 From: = <=> Date: Tue, 12 Apr 2022 14:00:44 +0000 Subject: [PATCH] Cr-feat window length for all empaticas sensors. --- config.yaml | 36 ++++++++++++------- .../empatica_accelerometer/{cf => cr}/main.py | 35 ++++++++++++------ .../{cf => cr}/main.py | 28 ++++++++++----- .../{cf => cr}/main.py | 27 +++++++++----- .../empatica_temperature/{cf => cr}/main.py | 29 +++++++++------ 5 files changed, 105 insertions(+), 50 deletions(-) rename src/features/empatica_accelerometer/{cf => cr}/main.py (55%) rename src/features/empatica_blood_volume_pulse/{cf => cr}/main.py (60%) rename src/features/empatica_electrodermal_activity/{cf => cr}/main.py (61%) rename src/features/empatica_temperature/{cf => cr}/main.py (59%) diff --git a/config.yaml b/config.yaml index bc43d0d7..77d36d68 100644 --- a/config.yaml +++ b/config.yaml @@ -477,10 +477,10 @@ EMPATICA_ACCELEROMETER: CONTAINER: ACC PROVIDERS: DBDP: - COMPUTE: True + COMPUTE: False FEATURES: ["maxmagnitude", "minmagnitude", "avgmagnitude", "medianmagnitude", "stdmagnitude"] SRC_SCRIPT: src/features/empatica_accelerometer/dbdp/main.py - CF: + CR: COMPUTE: True FEATURES: ["fqHighestPeakFreqs", "fqHighestPeaks", "fqEnergyFeat", "fqEntropyFeat", "fqHistogramBins","fqAbsMean", "fqSkewness", "fqKurtosis", "fqInterquart", # Freq features "meanLow", "areaLow", "totalAbsoluteAreaBand", "totalMagnitudeBand", "entropyBand", "skewnessBand", "kurtosisBand", @@ -490,7 +490,10 @@ EMPATICA_ACCELEROMETER: "peaksDataLow", "sumPerComponentBand", "velocityBand", "meanKineticEnergyBand", "totalKineticEnergyBand", "squareSumOfComponent", "sumOfSquareComponents", "averageVectorLength", "averageVectorLengthPower", "rollAvgLow", "pitchAvgLow", "rollStdDevLow", "pitchStdDevLow", "rollMotionAmountLow", "rollMotionRegularityLow", "manipulationLow", "rollPeaks", "pitchPeaks", "rollPitchCorrelation"] # Acc features - SRC_SCRIPT: src/features/empatica_accelerometer/cf/main.py + WINDOWS: + COMPUTE: False + WINDOW_LENGTH: 10 # specify window length in seconds + SRC_SCRIPT: src/features/empatica_accelerometer/cr/main.py # See https://www.rapids.science/latest/features/empatica-heartrate/ @@ -507,48 +510,57 @@ EMPATICA_TEMPERATURE: CONTAINER: TEMP PROVIDERS: DBDP: - COMPUTE: True + COMPUTE: False FEATURES: ["maxtemp", "mintemp", "avgtemp", "mediantemp", "modetemp", "stdtemp", "diffmaxmodetemp", "diffminmodetemp", "entropytemp"] SRC_SCRIPT: src/features/empatica_temperature/dbdp/main.py - CF: + CR: COMPUTE: True FEATURES: ["autocorrelations", "countAboveMean", "countBelowMean", "maximum", "minimum", "meanAbsChange", "longestStrikeAboveMean", "longestStrikeBelowMean", "stdDev", "median", "meanChange", "numberOfZeroCrossings", "absEnergy", "linearTrendSlope", "ratioBeyondRSigma", "binnedEntropy", "numOfPeaksAutocorr", "numberOfZeroCrossingsAutocorr", "areaAutocorr", "calcMeanCrossingRateAutocorr", "countAboveMeanAutocorr", "sumPer", "sumSquared", "squareSumOfComponent", "sumOfSquareComponents"] - SRC_SCRIPT: src/features/empatica_temperature/cf/main.py + WINDOWS: + COMPUTE: False + WINDOW_LENGTH: 90 # specify window length in seconds + SRC_SCRIPT: src/features/empatica_temperature/cr/main.py # See https://www.rapids.science/latest/features/empatica-electrodermal-activity/ EMPATICA_ELECTRODERMAL_ACTIVITY: CONTAINER: EDA PROVIDERS: DBDP: - COMPUTE: True + COMPUTE: False FEATURES: ["maxeda", "mineda", "avgeda", "medianeda", "modeeda", "stdeda", "diffmaxmodeeda", "diffminmodeeda", "entropyeda"] SRC_SCRIPT: src/features/empatica_electrodermal_activity/dbdp/main.py - CF: + CR: COMPUTE: True FEATURES: ['mean', 'std', 'q25', 'q75', 'qd', 'deriv', 'power', 'numPeaks', 'ratePeaks', 'powerPeaks', 'sumPosDeriv', 'propPosDeriv', 'derivTonic', 'sigTonicDifference', 'freqFeats','maxPeakAmplitudeChangeBefore', 'maxPeakAmplitudeChangeAfter', 'avgPeakAmplitudeChangeBefore', 'avgPeakAmplitudeChangeAfter', 'avgPeakChangeRatio', 'maxPeakIncreaseTime', 'maxPeakDecreaseTime', 'maxPeakDuration', 'maxPeakChangeRatio', 'avgPeakIncreaseTime', 'avgPeakDecreaseTime', 'avgPeakDuration', 'maxPeakResponseSlopeBefore', 'maxPeakResponseSlopeAfter', 'signalOverallChange', 'changeDuration', 'changeRate', 'significantIncrease', 'significantDecrease'] - SRC_SCRIPT: src/features/empatica_electrodermal_activity/cf/main.py + WINDOWS: + COMPUTE: False + WINDOW_LENGTH: 80 # specify window length in seconds + SRC_SCRIPT: src/features/empatica_electrodermal_activity/cr/main.py # See https://www.rapids.science/latest/features/empatica-blood-volume-pulse/ EMPATICA_BLOOD_VOLUME_PULSE: CONTAINER: BVP PROVIDERS: DBDP: - COMPUTE: True + COMPUTE: False FEATURES: ["fqHighestPeakFreqs", "fqHighestPeaks", "fqEnergyFeat", "fqEntropyFeat", "fqHistogramBins","fqAbsMean", "fqSkewness", "fqKurtosis", "fqInterquart", # Freq features "maxbvp", "minbvp", "avgbvp", "medianbvp", "modebvp", "stdbvp", "diffmaxmodebvp", "diffminmodebvp", "entropybvp"] # HRV features SRC_SCRIPT: src/features/empatica_blood_volume_pulse/dbdp/main.py - CF: + CR: COMPUTE: True FEATURES: ['meanHr', 'ibi', 'sdnn', 'sdsd', 'rmssd', 'pnn20', 'pnn50', 'sd', 'sd2', 'sd1/sd2', 'numRR'] - SRC_SCRIPT: src/features/empatica_blood_volume_pulse/cf/main.py + WINDOWS: + COMPUTE: False + WINDOW_LENGTH: 4 # specify window length in seconds + SRC_SCRIPT: src/features/empatica_blood_volume_pulse/cr/main.py # See https://www.rapids.science/latest/features/empatica-inter-beat-interval/ EMPATICA_INTER_BEAT_INTERVAL: diff --git a/src/features/empatica_accelerometer/cf/main.py b/src/features/empatica_accelerometer/cr/main.py similarity index 55% rename from src/features/empatica_accelerometer/cf/main.py rename to src/features/empatica_accelerometer/cr/main.py index 6ebf2068..43812eb8 100644 --- a/src/features/empatica_accelerometer/cf/main.py +++ b/src/features/empatica_accelerometer/cr/main.py @@ -14,7 +14,7 @@ def getSampleRate(data): return 1000/timestamps_diff -def extractAccFeaturesFromIntradayData(acc_intraday_data, features, time_segment, filter_data_by_segment): +def extractAccFeaturesFromIntradayData(acc_intraday_data, features, window_length, time_segment, filter_data_by_segment): acc_intraday_features = pd.DataFrame(columns=["local_segment"] + features) if not acc_intraday_data.empty: @@ -27,12 +27,20 @@ def extractAccFeaturesFromIntradayData(acc_intraday_data, features, time_segment acc_intraday_features = pd.DataFrame() # apply methods from calculate features module - acc_intraday_features = \ - acc_intraday_data.groupby('local_segment').apply(lambda x: calculateFeatures( \ - convertInputInto2d(x['double_values_0'], x.shape[0]), \ - convertInputInto2d(x['double_values_1'], x.shape[0]), \ - convertInputInto2d(x['double_values_2'], x.shape[0]), \ - fs=int(sample_rate), featureNames=features)) + if window_length is None: + acc_intraday_features = \ + acc_intraday_data.groupby('local_segment').apply(lambda x: calculateFeatures( \ + convertInputInto2d(x['double_values_0'], x.shape[0]), \ + convertInputInto2d(x['double_values_1'], x.shape[0]), \ + convertInputInto2d(x['double_values_2'], x.shape[0]), \ + fs=int(sample_rate), featureNames=features)) + else: + acc_intraday_features = \ + acc_intraday_data.groupby('local_segment').apply(lambda x: calculateFeatures( \ + convertInputInto2d(x['double_values_0'], window_length*int(sample_rate)), \ + convertInputInto2d(x['double_values_1'], window_length*int(sample_rate)), \ + convertInputInto2d(x['double_values_2'], window_length*int(sample_rate)), \ + fs=int(sample_rate), featureNames=features)) acc_intraday_features.reset_index(inplace=True) @@ -40,18 +48,23 @@ def extractAccFeaturesFromIntradayData(acc_intraday_data, features, time_segment -def cf_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): +def cr_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): acc_intraday_data = pd.read_csv(sensor_data_files["sensor_data"]) requested_intraday_features = provider["FEATURES"] + + if provider["WINDOWS"]["COMPUTE"]: + requested_window_length = provider["WINDOWS"]["WINDOW_LENGTH"] + else: + requested_window_length = None + # name of the features this function can compute base_intraday_features_names = accelerometerFeatureNames + frequencyFeatureNames # 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 - acc_intraday_features = extractAccFeaturesFromIntradayData(acc_intraday_data, - intraday_features_to_compute, time_segment, - filter_data_by_segment) + acc_intraday_features = extractAccFeaturesFromIntradayData(acc_intraday_data, intraday_features_to_compute, + requested_window_length, time_segment, filter_data_by_segment) return acc_intraday_features \ No newline at end of file diff --git a/src/features/empatica_blood_volume_pulse/cf/main.py b/src/features/empatica_blood_volume_pulse/cr/main.py similarity index 60% rename from src/features/empatica_blood_volume_pulse/cf/main.py rename to src/features/empatica_blood_volume_pulse/cr/main.py index 3423a770..8b7f0c81 100644 --- a/src/features/empatica_blood_volume_pulse/cf/main.py +++ b/src/features/empatica_blood_volume_pulse/cr/main.py @@ -15,7 +15,7 @@ def getSampleRate(data): return 1000/timestamps_diff -def extractBVPFeaturesFromIntradayData(bvp_intraday_data, features, time_segment, filter_data_by_segment): +def extractBVPFeaturesFromIntradayData(bvp_intraday_data, features, window_length, time_segment, filter_data_by_segment): bvp_intraday_features = pd.DataFrame(columns=["local_segment"] + features) if not bvp_intraday_data.empty: @@ -27,28 +27,38 @@ def extractBVPFeaturesFromIntradayData(bvp_intraday_data, features, time_segment bvp_intraday_features = pd.DataFrame() - # apply methods from calculate features module - bvp_intraday_features = \ - bvp_intraday_data.groupby('local_segment').apply(\ - lambda x: calculateFeatures(convertInputInto2d(x['blood_volume_pulse'], x.shape[0]), fs=int(sample_rate), featureNames=features)) + # apply methods from calculate features module + if window_length is None: + bvp_intraday_features = \ + bvp_intraday_data.groupby('local_segment').apply(\ + lambda x: calculateFeatures(convertInputInto2d(x['blood_volume_pulse'], x.shape[0]), fs=int(sample_rate), featureNames=features)) + else: + bvp_intraday_features = \ + bvp_intraday_data.groupby('local_segment').apply(\ + lambda x: calculateFeatures(convertInputInto2d(x['blood_volume_pulse'], window_length*int(sample_rate)), fs=int(sample_rate), featureNames=features)) bvp_intraday_features.reset_index(inplace=True) return bvp_intraday_features -def cf_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): +def cr_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): bvp_intraday_data = pd.read_csv(sensor_data_files["sensor_data"]) requested_intraday_features = provider["FEATURES"] + + if provider["WINDOWS"]["COMPUTE"]: + requested_window_length = provider["WINDOWS"]["WINDOW_LENGTH"] + else: + requested_window_length = None + # name of the features this function can compute base_intraday_features_names = hrvFeatureNames + frequencyFeatureNames # 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 - bvp_intraday_features = extractBVPFeaturesFromIntradayData(bvp_intraday_data, - intraday_features_to_compute, time_segment, - filter_data_by_segment) + bvp_intraday_features = extractBVPFeaturesFromIntradayData(bvp_intraday_data, intraday_features_to_compute, + requested_window_length, time_segment, filter_data_by_segment) return bvp_intraday_features \ No newline at end of file diff --git a/src/features/empatica_electrodermal_activity/cf/main.py b/src/features/empatica_electrodermal_activity/cr/main.py similarity index 61% rename from src/features/empatica_electrodermal_activity/cf/main.py rename to src/features/empatica_electrodermal_activity/cr/main.py index 1c7d1735..c4f11349 100644 --- a/src/features/empatica_electrodermal_activity/cf/main.py +++ b/src/features/empatica_electrodermal_activity/cr/main.py @@ -13,7 +13,7 @@ def getSampleRate(data): return 1000/timestamps_diff -def extractEDAFeaturesFromIntradayData(eda_intraday_data, features, time_segment, filter_data_by_segment): +def extractEDAFeaturesFromIntradayData(eda_intraday_data, features, window_length, time_segment, filter_data_by_segment): eda_intraday_features = pd.DataFrame(columns=["local_segment"] + features) if not eda_intraday_data.empty: @@ -26,27 +26,38 @@ def extractEDAFeaturesFromIntradayData(eda_intraday_data, features, time_segment eda_intraday_features = pd.DataFrame() # apply methods from calculate features module - eda_intraday_features = \ - eda_intraday_data.groupby('local_segment').apply(\ - lambda x: calculateFeatures(convertInputInto2d(x['electrodermal_activity'], x.shape[0]), fs=int(sample_rate), featureNames=features)) + if window_length is None: + eda_intraday_features = \ + eda_intraday_data.groupby('local_segment').apply(\ + lambda x: calculateFeatures(convertInputInto2d(x['electrodermal_activity'], x.shape[0]), fs=int(sample_rate), featureNames=features)) + else: + eda_intraday_features = \ + eda_intraday_data.groupby('local_segment').apply(\ + lambda x: calculateFeatures(convertInputInto2d(x['electrodermal_activity'], window_length*int(sample_rate)), fs=int(sample_rate), featureNames=features)) + eda_intraday_features.reset_index(inplace=True) return eda_intraday_features -def cf_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): +def cr_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): eda_intraday_data = pd.read_csv(sensor_data_files["sensor_data"]) requested_intraday_features = provider["FEATURES"] + + if provider["WINDOWS"]["COMPUTE"]: + requested_window_length = provider["WINDOWS"]["WINDOW_LENGTH"] + else: + requested_window_length = None + # name of the features this function can compute base_intraday_features_names = gsrFeatureNames # 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 - eda_intraday_features = extractEDAFeaturesFromIntradayData(eda_intraday_data, - intraday_features_to_compute, time_segment, - filter_data_by_segment) + eda_intraday_features = extractEDAFeaturesFromIntradayData(eda_intraday_data, intraday_features_to_compute, + requested_window_length, time_segment, filter_data_by_segment) return eda_intraday_features \ No newline at end of file diff --git a/src/features/empatica_temperature/cf/main.py b/src/features/empatica_temperature/cr/main.py similarity index 59% rename from src/features/empatica_temperature/cf/main.py rename to src/features/empatica_temperature/cr/main.py index 04cbeb0e..b804298c 100644 --- a/src/features/empatica_temperature/cf/main.py +++ b/src/features/empatica_temperature/cr/main.py @@ -4,6 +4,7 @@ from scipy.stats import entropy from CalculatingFeatures.helper_functions import convert1DEmpaticaToArray, convertInputInto2d, genericFeatureNames from CalculatingFeatures.calculate_features import calculateFeatures +import sys def getSampleRate(data): try: @@ -13,7 +14,7 @@ def getSampleRate(data): return 1000/timestamps_diff -def extractTempFeaturesFromIntradayData(temperature_intraday_data, features, time_segment, filter_data_by_segment): +def extractTempFeaturesFromIntradayData(temperature_intraday_data, features, window_length, time_segment, filter_data_by_segment): temperature_intraday_features = pd.DataFrame(columns=["local_segment"] + features) if not temperature_intraday_data.empty: @@ -25,28 +26,36 @@ def extractTempFeaturesFromIntradayData(temperature_intraday_data, features, tim temperature_intraday_features = pd.DataFrame() - # apply methods from calculate features module - temperature_intraday_features = \ - temperature_intraday_data.groupby('local_segment').apply(\ - lambda x: calculateFeatures(convertInputInto2d(x['temperature'], x.shape[0]), fs=int(sample_rate), featureNames=features)) + # apply methods from calculate features module + if window_length is None: + temperature_intraday_features = \ + temperature_intraday_data.groupby('local_segment').apply(\ + lambda x: calculateFeatures(convertInputInto2d(x['temperature'], x.shape[0]), fs=int(sample_rate), featureNames=features)) + else: + temperature_intraday_features = \ + temperature_intraday_data.groupby('local_segment').apply(\ + lambda x: calculateFeatures(convertInputInto2d(x['temperature'], window_length*int(sample_rate)), fs=int(sample_rate), featureNames=features)) temperature_intraday_features.reset_index(inplace=True) return temperature_intraday_features -def cf_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): +def cr_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs): temperature_intraday_data = pd.read_csv(sensor_data_files["sensor_data"]) requested_intraday_features = provider["FEATURES"] + if provider["WINDOWS"]["COMPUTE"]: + requested_window_length = provider["WINDOWS"]["WINDOW_LENGTH"] + else: + requested_window_length = None + # name of the features this function can compute base_intraday_features_names = genericFeatureNames # 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 - temperature_intraday_features = extractTempFeaturesFromIntradayData(temperature_intraday_data, - intraday_features_to_compute, time_segment, - filter_data_by_segment) - + temperature_intraday_features = extractTempFeaturesFromIntradayData(temperature_intraday_data, intraday_features_to_compute, + requested_window_length, time_segment, filter_data_by_segment) return temperature_intraday_features \ No newline at end of file