2019-10-24 18:11:24 +02:00
|
|
|
rule download_dataset:
|
|
|
|
input:
|
|
|
|
"data/external/{pid}"
|
|
|
|
params:
|
2019-10-24 23:27:00 +02:00
|
|
|
group = config["DOWNLOAD_DATASET"]["GROUP"],
|
2019-10-24 22:08:05 +02:00
|
|
|
table = "{sensor}"
|
2019-10-24 18:11:24 +02:00
|
|
|
output:
|
2019-10-24 22:08:05 +02:00
|
|
|
"data/raw/{pid}/{sensor}_raw.csv"
|
2019-10-24 18:11:24 +02:00
|
|
|
script:
|
2019-10-24 22:08:05 +02:00
|
|
|
"../src/data/download_dataset.R"
|
|
|
|
|
|
|
|
rule readable_datetime:
|
|
|
|
input:
|
|
|
|
sensor_input = rules.download_dataset.output
|
|
|
|
params:
|
|
|
|
timezones = None,
|
2019-10-24 23:27:00 +02:00
|
|
|
fixed_timezone = config["READABLE_DATETIME"]["FIXED_TIMEZONE"]
|
2020-01-15 23:18:10 +01:00
|
|
|
wildcard_constraints:
|
2020-01-16 23:20:04 +01:00
|
|
|
sensor = '(' + '|'.join([re.escape(x) for x in config["SENSORS"]]) + ')' # only process smartphone sensors, not fitbit
|
2019-10-24 22:08:05 +02:00
|
|
|
output:
|
|
|
|
"data/raw/{pid}/{sensor}_with_datetime.csv"
|
|
|
|
script:
|
2019-11-05 18:34:22 +01:00
|
|
|
"../src/data/readable_datetime.R"
|
|
|
|
|
|
|
|
rule phone_valid_sensed_days:
|
|
|
|
input:
|
|
|
|
all_sensors = expand("data/raw/{{pid}}/{sensor}_with_datetime.csv", sensor=config["SENSORS"])
|
|
|
|
params:
|
|
|
|
bin_size = config["PHONE_VALID_SENSED_DAYS"]["BIN_SIZE"],
|
|
|
|
min_valid_hours = config["PHONE_VALID_SENSED_DAYS"]["MIN_VALID_HOURS"],
|
|
|
|
min_bins_per_hour = config["PHONE_VALID_SENSED_DAYS"]["MIN_BINS_PER_HOUR"]
|
|
|
|
output:
|
|
|
|
"data/interim/{pid}/phone_valid_sensed_days.csv"
|
|
|
|
script:
|
2019-11-12 20:57:27 +01:00
|
|
|
"../src/data/phone_valid_sensed_days.R"
|
|
|
|
|
2019-12-04 17:33:25 +01:00
|
|
|
rule phone_sensed_bins:
|
|
|
|
input:
|
|
|
|
all_sensors = expand("data/raw/{{pid}}/{sensor}_with_datetime.csv", sensor=config["SENSORS"])
|
|
|
|
params:
|
|
|
|
bin_size = config["PHONE_VALID_SENSED_DAYS"]["BIN_SIZE"]
|
|
|
|
output:
|
|
|
|
"data/interim/{pid}/phone_sensed_bins.csv"
|
|
|
|
script:
|
|
|
|
"../src/data/phone_sensed_bins.R"
|
|
|
|
|
2019-11-12 20:57:27 +01:00
|
|
|
rule unify_ios_android:
|
|
|
|
input:
|
|
|
|
sensor_data = "data/raw/{pid}/{sensor}_with_datetime.csv",
|
|
|
|
participant_info = "data/external/{pid}"
|
|
|
|
params:
|
|
|
|
sensor = "{sensor}"
|
|
|
|
output:
|
|
|
|
"data/raw/{pid}/{sensor}_with_datetime_unified.csv"
|
|
|
|
script:
|
2019-12-10 00:23:00 +01:00
|
|
|
"../src/data/unify_ios_android.R"
|
|
|
|
|
|
|
|
rule resample_fused_location:
|
|
|
|
input:
|
|
|
|
locations = "data/raw/{pid}/locations_raw.csv",
|
|
|
|
phone_sensed_bins = rules.phone_sensed_bins.output
|
|
|
|
params:
|
|
|
|
bin_size = config["PHONE_VALID_SENSED_DAYS"]["BIN_SIZE"],
|
|
|
|
timezone = config["RESAMPLE_FUSED_LOCATION"]["TIMEZONE"],
|
|
|
|
consecutive_threshold = config["RESAMPLE_FUSED_LOCATION"]["CONSECUTIVE_THRESHOLD"],
|
|
|
|
time_since_valid_location = config["RESAMPLE_FUSED_LOCATION"]["TIME_SINCE_VALID_LOCATION"]
|
|
|
|
output:
|
|
|
|
"data/raw/{pid}/locations_resampled.csv"
|
|
|
|
script:
|
2020-01-15 23:18:10 +01:00
|
|
|
"../src/data/resample_fused_location.R"
|
|
|
|
|
2020-01-16 00:28:56 +01:00
|
|
|
rule application_genres:
|
|
|
|
input:
|
|
|
|
"data/raw/{pid}/applications_foreground_with_datetime.csv"
|
|
|
|
params:
|
|
|
|
catalogue_source = config["APPLICATION_GENRES"]["CATALOGUE_SOURCE"],
|
|
|
|
catalogue_file = config["APPLICATION_GENRES"]["CATALOGUE_FILE"],
|
|
|
|
update_catalogue_file = config["APPLICATION_GENRES"]["UPDATE_CATALOGUE_FILE"],
|
|
|
|
scrape_missing_genres = config["APPLICATION_GENRES"]["SCRAPE_MISSING_GENRES"]
|
|
|
|
output:
|
|
|
|
"data/interim/{pid}/applications_foreground_with_datetime_with_genre.csv"
|
|
|
|
script:
|
|
|
|
"../src/data/application_genres.R"
|
|
|
|
|
2020-01-16 23:20:04 +01:00
|
|
|
rule fitbit_with_datetime:
|
2020-01-15 23:18:10 +01:00
|
|
|
input:
|
|
|
|
"data/raw/{pid}/fitbit_data_raw.csv"
|
|
|
|
params:
|
|
|
|
local_timezone = config["READABLE_DATETIME"]["FIXED_TIMEZONE"],
|
2020-01-16 23:20:04 +01:00
|
|
|
fitbit_sensor = "{fitbit_sensor}"
|
2020-01-15 23:18:10 +01:00
|
|
|
output:
|
2020-01-16 23:20:04 +01:00
|
|
|
"data/raw/{pid}/fitbit_{fitbit_sensor}_with_datetime.csv"
|
2020-01-15 23:18:10 +01:00
|
|
|
script:
|
2020-01-16 23:20:04 +01:00
|
|
|
"../src/data/fitbit_readable_datetime.py"
|
2020-01-16 00:28:56 +01:00
|
|
|
|