2020-04-21 00:05:54 +02:00
|
|
|
def optional_ar_input(wildcards):
|
|
|
|
with open("data/external/"+wildcards.pid, encoding="ISO-8859-1") as external_file:
|
|
|
|
external_file_content = external_file.readlines()
|
|
|
|
platform = external_file_content[1].strip()
|
|
|
|
if platform == "android":
|
|
|
|
return ["data/raw/{pid}/plugin_google_activity_recognition_with_datetime_unified.csv",
|
|
|
|
"data/processed/{pid}/plugin_google_activity_recognition_deltas.csv"]
|
|
|
|
else:
|
|
|
|
return ["data/raw/{pid}/plugin_ios_activity_recognition_with_datetime_unified.csv",
|
|
|
|
"data/processed/{pid}/plugin_ios_activity_recognition_deltas.csv"]
|
|
|
|
|
2020-04-08 22:02:58 +02:00
|
|
|
rule sms_features:
|
2019-10-24 22:27:43 +02:00
|
|
|
input:
|
|
|
|
"data/raw/{pid}/messages_with_datetime.csv"
|
|
|
|
params:
|
|
|
|
sms_type = "{sms_type}",
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-08 22:02:58 +02:00
|
|
|
features = lambda wildcards: config["SMS"]["FEATURES"][wildcards.sms_type]
|
2019-10-24 22:27:43 +02:00
|
|
|
output:
|
2019-11-06 21:38:08 +01:00
|
|
|
"data/processed/{pid}/sms_{sms_type}_{day_segment}.csv"
|
2019-10-24 22:27:43 +02:00
|
|
|
script:
|
2020-04-08 22:02:58 +02:00
|
|
|
"../src/features/sms_features.R"
|
2019-10-25 16:21:09 +02:00
|
|
|
|
2020-04-08 17:51:18 +02:00
|
|
|
rule call_features:
|
2019-10-25 16:21:09 +02:00
|
|
|
input:
|
2019-11-12 21:12:43 +01:00
|
|
|
"data/raw/{pid}/calls_with_datetime_unified.csv"
|
2019-10-25 16:21:09 +02:00
|
|
|
params:
|
|
|
|
call_type = "{call_type}",
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-08 17:51:18 +02:00
|
|
|
features = lambda wildcards: config["CALLS"]["FEATURES"][wildcards.call_type]
|
2019-10-25 16:21:09 +02:00
|
|
|
output:
|
2019-11-06 20:47:33 +01:00
|
|
|
"data/processed/{pid}/call_{call_type}_{day_segment}.csv"
|
2019-10-25 16:21:09 +02:00
|
|
|
script:
|
2020-04-08 17:51:18 +02:00
|
|
|
"../src/features/call_features.R"
|
2019-11-05 16:47:55 +01:00
|
|
|
|
|
|
|
rule battery_deltas:
|
|
|
|
input:
|
2019-12-05 22:24:05 +01:00
|
|
|
"data/raw/{pid}/battery_with_datetime_unified.csv"
|
2019-11-05 16:47:55 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/battery_deltas.csv"
|
|
|
|
script:
|
2019-11-05 21:17:20 +01:00
|
|
|
"../src/features/battery_deltas.R"
|
|
|
|
|
2019-11-27 21:27:48 +01:00
|
|
|
rule screen_deltas:
|
|
|
|
input:
|
2020-03-03 01:26:20 +01:00
|
|
|
screen = "data/raw/{pid}/screen_with_datetime.csv",
|
|
|
|
participant_info = "data/external/{pid}"
|
2019-11-27 21:27:48 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/screen_deltas.csv"
|
|
|
|
script:
|
|
|
|
"../src/features/screen_deltas.R"
|
|
|
|
|
2019-11-27 21:53:37 +01:00
|
|
|
rule google_activity_recognition_deltas:
|
|
|
|
input:
|
2020-04-21 00:05:54 +02:00
|
|
|
"data/raw/{pid}/plugin_google_activity_recognition_with_datetime_unified.csv"
|
2019-11-27 21:53:37 +01:00
|
|
|
output:
|
2019-11-28 00:06:31 +01:00
|
|
|
"data/processed/{pid}/plugin_google_activity_recognition_deltas.csv"
|
2019-11-27 21:53:37 +01:00
|
|
|
script:
|
2020-04-21 00:05:54 +02:00
|
|
|
"../src/features/activity_recognition_deltas.R"
|
|
|
|
|
|
|
|
rule ios_activity_recognition_deltas:
|
|
|
|
input:
|
|
|
|
"data/raw/{pid}/plugin_ios_activity_recognition_with_datetime_unified.csv"
|
|
|
|
output:
|
|
|
|
"data/processed/{pid}/plugin_ios_activity_recognition_deltas.csv"
|
|
|
|
script:
|
|
|
|
"../src/features/activity_recognition_deltas.R"
|
2019-11-27 21:53:37 +01:00
|
|
|
|
2020-04-09 19:20:39 +02:00
|
|
|
rule location_barnett_features:
|
2019-11-05 21:17:20 +01:00
|
|
|
input:
|
2019-12-10 01:15:10 +01:00
|
|
|
raw = "data/raw/{pid}/locations_raw.csv",
|
|
|
|
fused = rules.resample_fused_location.output
|
2019-11-05 21:17:20 +01:00
|
|
|
params:
|
2020-04-09 19:20:39 +02:00
|
|
|
features = config["BARNETT_LOCATION"]["FEATURES"],
|
2019-12-10 01:15:10 +01:00
|
|
|
locations_to_use = config["BARNETT_LOCATION"]["LOCATIONS_TO_USE"],
|
2019-11-05 21:17:20 +01:00
|
|
|
accuracy_limit = config["BARNETT_LOCATION"]["ACCURACY_LIMIT"],
|
2020-03-05 16:32:38 +01:00
|
|
|
timezone = config["BARNETT_LOCATION"]["TIMEZONE"],
|
|
|
|
day_segment = "{day_segment}"
|
2019-11-05 21:17:20 +01:00
|
|
|
output:
|
2020-03-05 16:32:38 +01:00
|
|
|
"data/processed/{pid}/location_barnett_{day_segment}.csv"
|
2019-11-05 21:17:20 +01:00
|
|
|
script:
|
2020-04-09 19:20:39 +02:00
|
|
|
"../src/features/location_barnett_features.R"
|
2019-11-06 18:19:30 +01:00
|
|
|
|
2020-04-03 19:07:09 +02:00
|
|
|
rule bluetooth_features:
|
2019-11-06 18:19:30 +01:00
|
|
|
input:
|
|
|
|
"data/raw/{pid}/bluetooth_with_datetime.csv"
|
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-03 19:07:09 +02:00
|
|
|
features = config["BLUETOOTH"]["FEATURES"]
|
2019-11-06 18:19:30 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/bluetooth_{day_segment}.csv"
|
|
|
|
script:
|
2020-04-03 19:07:09 +02:00
|
|
|
"../src/features/bluetooth_features.R"
|
2020-04-21 00:05:54 +02:00
|
|
|
|
2020-04-08 20:36:36 +02:00
|
|
|
rule activity_features:
|
2019-11-06 19:34:47 +01:00
|
|
|
input:
|
2020-04-21 00:05:54 +02:00
|
|
|
optional_ar_input
|
2019-11-18 20:22:08 +01:00
|
|
|
params:
|
|
|
|
segment = "{day_segment}",
|
2020-04-21 00:05:54 +02:00
|
|
|
features = config["ACTIVITY_RECOGNITION"]["FEATURES"]
|
2019-11-06 19:34:47 +01:00
|
|
|
output:
|
2020-04-21 00:05:54 +02:00
|
|
|
"data/processed/{pid}/activity_recognition_{day_segment}.csv"
|
2019-11-06 19:34:47 +01:00
|
|
|
script:
|
2020-04-21 00:05:54 +02:00
|
|
|
"../src/features/activity_recognition.py"
|
2019-11-08 18:18:21 +01:00
|
|
|
|
2020-04-09 22:06:25 +02:00
|
|
|
rule battery_features:
|
2019-11-08 18:18:21 +01:00
|
|
|
input:
|
|
|
|
"data/processed/{pid}/battery_deltas.csv"
|
2019-11-25 18:53:32 +01:00
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-09 22:06:25 +02:00
|
|
|
features = config["BATTERY"]["FEATURES"]
|
2019-11-08 18:18:21 +01:00
|
|
|
output:
|
2019-11-25 18:53:32 +01:00
|
|
|
"data/processed/{pid}/battery_{day_segment}.csv"
|
2019-11-08 18:18:21 +01:00
|
|
|
script:
|
2020-04-09 22:06:25 +02:00
|
|
|
"../src/features/battery_features.py"
|
2019-11-27 20:25:17 +01:00
|
|
|
|
2020-04-08 17:05:16 +02:00
|
|
|
rule screen_features:
|
2019-11-27 20:25:17 +01:00
|
|
|
input:
|
2019-12-04 22:04:37 +01:00
|
|
|
screen_deltas = "data/processed/{pid}/screen_deltas.csv",
|
|
|
|
phone_sensed_bins = "data/interim/{pid}/phone_sensed_bins.csv"
|
2019-11-27 20:25:17 +01:00
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-03-04 18:21:36 +01:00
|
|
|
reference_hour_first_use = config["SCREEN"]["REFERENCE_HOUR_FIRST_USE"],
|
2020-04-08 17:05:16 +02:00
|
|
|
features_deltas = config["SCREEN"]["FEATURES_DELTAS"],
|
2020-03-03 23:31:15 +01:00
|
|
|
episode_types = config["SCREEN"]["EPISODE_TYPES"],
|
2019-12-05 17:07:40 +01:00
|
|
|
bin_size = config["PHONE_VALID_SENSED_DAYS"]["BIN_SIZE"]
|
2019-11-27 20:25:17 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/screen_{day_segment}.csv"
|
|
|
|
script:
|
2020-04-08 17:05:16 +02:00
|
|
|
"../src/features/screen_features.py"
|
2020-01-14 15:51:39 +01:00
|
|
|
|
2020-04-08 20:02:53 +02:00
|
|
|
rule light_features:
|
2020-01-14 15:51:39 +01:00
|
|
|
input:
|
|
|
|
"data/raw/{pid}/light_with_datetime.csv",
|
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-08 20:02:53 +02:00
|
|
|
features = config["LIGHT"]["FEATURES"],
|
2020-01-14 15:51:39 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/light_{day_segment}.csv"
|
|
|
|
script:
|
2020-04-08 20:02:53 +02:00
|
|
|
"../src/features/light_features.py"
|
2020-01-15 20:15:24 +01:00
|
|
|
|
2020-04-02 23:36:28 +02:00
|
|
|
rule accelerometer_features:
|
2020-01-15 20:15:24 +01:00
|
|
|
input:
|
|
|
|
"data/raw/{pid}/accelerometer_with_datetime.csv",
|
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-02 23:36:28 +02:00
|
|
|
features = config["ACCELEROMETER"]["FEATURES"],
|
2020-01-15 20:15:24 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/accelerometer_{day_segment}.csv"
|
|
|
|
script:
|
2020-04-02 23:36:28 +02:00
|
|
|
"../src/features/accelerometer_features.py"
|
2020-01-29 22:22:53 +01:00
|
|
|
|
2020-04-08 19:31:43 +02:00
|
|
|
rule applications_foreground_features:
|
2020-02-07 17:52:55 +01:00
|
|
|
input:
|
|
|
|
"data/interim/{pid}/applications_foreground_with_datetime_with_genre.csv",
|
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
|
|
|
single_categories = config["APPLICATIONS_FOREGROUND"]["SINGLE_CATEGORIES"],
|
|
|
|
multiple_categories = config["APPLICATIONS_FOREGROUND"]["MULTIPLE_CATEGORIES"],
|
|
|
|
single_apps = config["APPLICATIONS_FOREGROUND"]["SINGLE_APPS"],
|
|
|
|
excluded_categories = config["APPLICATIONS_FOREGROUND"]["EXCLUDED_CATEGORIES"],
|
|
|
|
excluded_apps = config["APPLICATIONS_FOREGROUND"]["EXCLUDED_APPS"],
|
2020-04-08 19:31:43 +02:00
|
|
|
features = config["APPLICATIONS_FOREGROUND"]["FEATURES"],
|
2020-02-07 17:52:55 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/applications_foreground_{day_segment}.csv"
|
|
|
|
script:
|
2020-04-08 19:31:43 +02:00
|
|
|
"../src/features/applications_foreground_features.py"
|
2020-02-07 17:52:55 +01:00
|
|
|
|
2020-05-15 23:51:00 +02:00
|
|
|
rule wifi_features:
|
|
|
|
input:
|
|
|
|
"data/raw/{pid}/wifi_with_datetime.csv"
|
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
|
|
|
features = config["WIFI"]["FEATURES"]
|
|
|
|
output:
|
|
|
|
"data/processed/{pid}/wifi_{day_segment}.csv"
|
|
|
|
script:
|
|
|
|
"../src/features/wifi_features.R"
|
|
|
|
|
2020-04-08 21:31:43 +02:00
|
|
|
rule fitbit_heartrate_features:
|
2020-02-07 17:35:15 +01:00
|
|
|
input:
|
2020-05-15 23:51:00 +02:00
|
|
|
heartrate_summary_data = "data/raw/{pid}/fitbit_heartrate_summary_with_datetime.csv",
|
|
|
|
heartrate_intraday_data = "data/raw/{pid}/fitbit_heartrate_intraday_with_datetime.csv"
|
2020-02-07 17:35:15 +01:00
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-08 21:31:43 +02:00
|
|
|
features = config["HEARTRATE"]["FEATURES"],
|
2020-05-15 23:51:00 +02:00
|
|
|
daily_features_from_summary_data = config["HEARTRATE"]["DAILY_FEATURES_FROM_SUMMARY_DATA"]
|
2020-02-07 17:35:15 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/fitbit_heartrate_{day_segment}.csv"
|
|
|
|
script:
|
2020-04-08 21:31:43 +02:00
|
|
|
"../src/features/fitbit_heartrate_features.py"
|
2020-02-07 17:35:15 +01:00
|
|
|
|
2020-04-03 23:03:45 +02:00
|
|
|
rule fitbit_step_features:
|
2020-01-29 22:22:53 +01:00
|
|
|
input:
|
|
|
|
steps_data = "data/raw/{pid}/fitbit_steps_with_datetime.csv",
|
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-04-03 23:03:45 +02:00
|
|
|
features_all_steps = config["STEP"]["FEATURES"]["ALL_STEPS"],
|
|
|
|
features_sedentary_bout = config["STEP"]["FEATURES"]["SEDENTARY_BOUT"],
|
|
|
|
features_active_bout = config["STEP"]["FEATURES"]["ACTIVE_BOUT"],
|
2020-03-09 20:59:51 +01:00
|
|
|
threshold_active_bout = config["STEP"]["THRESHOLD_ACTIVE_BOUT"],
|
|
|
|
include_zero_step_rows = config["STEP"]["INCLUDE_ZERO_STEP_ROWS"]
|
2020-01-29 22:22:53 +01:00
|
|
|
output:
|
|
|
|
"data/processed/{pid}/fitbit_step_{day_segment}.csv"
|
|
|
|
script:
|
2020-04-03 23:03:45 +02:00
|
|
|
"../src/features/fitbit_step_features.py"
|
2020-04-13 19:24:52 +02:00
|
|
|
|
2020-05-15 23:51:00 +02:00
|
|
|
rule fitbit_sleep_features:
|
|
|
|
input:
|
|
|
|
sleep_summary_data = "data/raw/{pid}/fitbit_sleep_summary_with_datetime.csv",
|
|
|
|
sleep_intraday_data = "data/raw/{pid}/fitbit_sleep_intraday_with_datetime.csv"
|
2020-04-13 19:24:52 +02:00
|
|
|
params:
|
|
|
|
day_segment = "{day_segment}",
|
2020-05-15 23:51:00 +02:00
|
|
|
sleep_types = config["SLEEP"]["SLEEP_TYPES"],
|
|
|
|
daily_features_from_summary_data = config["SLEEP"]["DAILY_FEATURES_FROM_SUMMARY_DATA"]
|
2020-04-13 19:24:52 +02:00
|
|
|
output:
|
2020-05-15 23:51:00 +02:00
|
|
|
"data/processed/{pid}/fitbit_sleep_{day_segment}.csv"
|
2020-04-13 19:24:52 +02:00
|
|
|
script:
|
2020-05-15 23:51:00 +02:00
|
|
|
"../src/features/fitbit_sleep_features.py"
|