rapids/rules/preprocessing.smk

184 lines
8.0 KiB
Plaintext
Raw Normal View History

rule restore_sql_file:
input:
sql_file = "data/external/rapids_example.sql",
db_credentials = ".env"
params:
group = config["DOWNLOAD_PARTICIPANTS"]["GROUP"]
output:
touch("data/interim/restore_sql_file.done")
script:
"../src/data/restore_sql_file.py"
2020-08-03 23:30:15 +02:00
rule create_example_participant_files:
output:
expand("data/external/{pid}", pid = ["example01", "example02"])
shell:
"echo 'a748ee1a-1d0b-4ae9-9074-279a2b6ba524\nandroid\ntest01\n2020/04/23,2020/05/04\n' >> ./data/external/example01 && echo '13dbc8a3-dae3-4834-823a-4bc96a7d459d\nios\ntest02\n2020/04/23,2020/05/04\n' >> ./data/external/example02"
rule download_participants:
params:
group = config["DOWNLOAD_PARTICIPANTS"]["GROUP"],
2020-02-18 22:41:10 +01:00
ignored_device_ids = config["DOWNLOAD_PARTICIPANTS"]["IGNORED_DEVICE_IDS"],
timezone = config["TIMEZONE"]
priority: 1
script:
"../src/data/download_participants.R"
2019-10-24 18:11:24 +02:00
rule download_dataset:
input:
"data/external/{pid}"
params:
group = config["DOWNLOAD_DATASET"]["GROUP"],
2020-02-20 21:51:22 +01:00
table = "{sensor}",
timezone = config["TIMEZONE"],
aware_multiplatform_tables = config["ACTIVITY_RECOGNITION"]["DB_TABLE"]["ANDROID"] + "," + config["ACTIVITY_RECOGNITION"]["DB_TABLE"]["IOS"] + "," + config["CONVERSATION"]["DB_TABLE"]["ANDROID"] + "," + config["CONVERSATION"]["DB_TABLE"]["IOS"],
2020-08-11 22:18:06 +02:00
unifiable_sensors = {"calls": config["CALLS"]["DB_TABLE"], "battery": config["BATTERY"]["DB_TABLE"], "screen": config["SCREEN"]["DB_TABLE"], "ios_activity_recognition": config["ACTIVITY_RECOGNITION"]["DB_TABLE"]["IOS"], "ios_conversation": config["CONVERSATION"]["DB_TABLE"]["IOS"]}
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 compute_day_segments:
input:
config["DAY_SEGMENTS"]["FILE"]
params:
2020-09-14 20:21:36 +02:00
day_segments_type = config["DAY_SEGMENTS"]["TYPE"],
pid = "{pid}"
output:
2020-09-14 20:21:36 +02:00
segments_file = "data/interim/day_segments/{pid}_day_segments.csv",
segments_labels_file = "data/interim/day_segments/{pid}_day_segments_labels.csv",
script:
"../src/data/compute_day_segments.py"
PHONE_SENSORS = []
PHONE_SENSORS.extend([config["MESSAGES"]["DB_TABLE"], config["CALLS"]["DB_TABLE"], config["LOCATIONS"]["DB_TABLE"], config["BLUETOOTH"]["DB_TABLE"], config["BATTERY"]["DB_TABLE"], config["SCREEN"]["DB_TABLE"], config["LIGHT"]["DB_TABLE"], config["ACCELEROMETER"]["DB_TABLE"], config["APPLICATIONS_FOREGROUND"]["DB_TABLE"], config["CONVERSATION"]["DB_TABLE"]["ANDROID"], config["CONVERSATION"]["DB_TABLE"]["IOS"], config["ACTIVITY_RECOGNITION"]["DB_TABLE"]["ANDROID"], config["ACTIVITY_RECOGNITION"]["DB_TABLE"]["IOS"]])
PHONE_SENSORS.extend(config["PHONE_VALID_SENSED_BINS"]["DB_TABLES"])
2020-07-24 22:35:41 +02:00
if len(config["WIFI"]["DB_TABLE"]["VISIBLE_ACCESS_POINTS"]) > 0:
PHONE_SENSORS.append(config["WIFI"]["DB_TABLE"]["VISIBLE_ACCESS_POINTS"])
2020-07-24 22:35:41 +02:00
if len(config["WIFI"]["DB_TABLE"]["CONNECTED_ACCESS_POINTS"]) > 0:
PHONE_SENSORS.append(config["WIFI"]["DB_TABLE"]["CONNECTED_ACCESS_POINTS"])
2020-07-24 22:35:41 +02:00
2019-10-24 22:08:05 +02:00
rule readable_datetime:
input:
sensor_input = "data/raw/{pid}/{sensor}_raw.csv",
2020-09-14 20:21:36 +02:00
day_segments = "data/interim/day_segments/{pid}_day_segments.csv"
2019-10-24 22:08:05 +02:00
params:
timezones = None,
fixed_timezone = config["READABLE_DATETIME"]["FIXED_TIMEZONE"],
2020-09-16 20:54:20 +02:00
day_segments_type = config["DAY_SEGMENTS"]["TYPE"],
include_past_periodic_segments = config["DAY_SEGMENTS"]["INCLUDE_PAST_PERIODIC_SEGMENTS"]
2020-01-15 23:18:10 +01:00
wildcard_constraints:
sensor = '.*(' + '|'.join([re.escape(x) for x in PHONE_SENSORS]) + ').*' # only process smartphone sensors, not fitbit
2019-10-24 22:08:05 +02:00
output:
2020-07-23 18:00:51 +02:00
"data/raw/{pid}/{sensor}_with_datetime.csv"
2019-10-24 22:08:05 +02:00
script:
2019-11-05 18:34:22 +01:00
"../src/data/readable_datetime.R"
2020-07-09 19:01:50 +02:00
rule phone_sensed_bins:
2019-11-05 18:34:22 +01:00
input:
all_sensors = optional_phone_sensed_bins_input
2019-11-05 18:34:22 +01:00
params:
2020-07-09 19:01:50 +02:00
bin_size = config["PHONE_VALID_SENSED_BINS"]["BIN_SIZE"]
2019-11-05 18:34:22 +01:00
output:
2020-07-09 19:01:50 +02:00
"data/interim/{pid}/phone_sensed_bins.csv"
2019-11-05 18:34:22 +01:00
script:
2020-07-09 19:01:50 +02:00
"../src/data/phone_sensed_bins.R"
2020-07-09 19:01:50 +02:00
rule phone_valid_sensed_days:
input:
2020-07-09 19:01:50 +02:00
phone_sensed_bins = "data/interim/{pid}/phone_sensed_bins.csv"
params:
min_valid_hours_per_day = "{min_valid_hours_per_day}",
min_valid_bins_per_hour = "{min_valid_bins_per_hour}"
output:
"data/interim/{pid}/phone_valid_sensed_days_{min_valid_hours_per_day}hours_{min_valid_bins_per_hour}bins.csv"
script:
2020-07-09 19:01:50 +02:00
"../src/data/phone_valid_sensed_days.R"
rule unify_ios_android:
input:
sensor_data = "data/raw/{pid}/{sensor}_with_datetime.csv",
participant_info = "data/external/{pid}"
params:
sensor = "{sensor}",
2020-08-11 22:18:06 +02:00
unifiable_sensors = {"calls": config["CALLS"]["DB_TABLE"], "battery": config["BATTERY"]["DB_TABLE"], "screen": config["SCREEN"]["DB_TABLE"], "ios_activity_recognition": config["ACTIVITY_RECOGNITION"]["DB_TABLE"]["IOS"], "ios_conversation": config["CONVERSATION"]["DB_TABLE"]["IOS"]}
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 process_location_types:
2019-12-10 00:23:00 +01:00
input:
locations = "data/raw/{pid}/{sensor}_with_datetime.csv",
phone_sensed_bins = rules.phone_sensed_bins.output,
2020-09-14 20:21:36 +02:00
day_segments = "data/interim/day_segments/{pid}_day_segments.csv"
2019-12-10 00:23:00 +01:00
params:
2020-07-09 19:01:50 +02:00
bin_size = config["PHONE_VALID_SENSED_BINS"]["BIN_SIZE"],
timezone = config["LOCATIONS"]["TIMEZONE"],
consecutive_threshold = config["LOCATIONS"]["FUSED_RESAMPLED_CONSECUTIVE_THRESHOLD"],
time_since_valid_location = config["LOCATIONS"]["FUSED_RESAMPLED_TIME_SINCE_VALID_LOCATION"],
day_segments_type = config["DAY_SEGMENTS"]["TYPE"],
locations_to_use = "{locations_to_used}"
2019-12-10 00:23:00 +01:00
output:
"data/raw/{pid}/{sensor}_processed_{locations_to_used}.csv"
2019-12-10 00:23:00 +01:00
script:
"../src/data/process_location_types.R"
2020-01-15 23:18:10 +01:00
rule application_genres:
input:
"data/raw/{pid}/{sensor}_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/raw/{pid}/{sensor}_with_datetime_with_genre.csv"
script:
"../src/data/application_genres.R"
rule fitbit_heartrate_with_datetime:
2020-01-15 23:18:10 +01:00
input:
expand("data/raw/{{pid}}/{fitbit_table}_raw.csv", fitbit_table=config["HEARTRATE"]["DB_TABLE"])
2020-01-15 23:18:10 +01:00
params:
local_timezone = config["READABLE_DATETIME"]["FIXED_TIMEZONE"],
fitbit_sensor = "heartrate"
2020-01-15 23:18:10 +01:00
output:
summary_data = "data/raw/{pid}/fitbit_heartrate_summary_with_datetime.csv",
intraday_data = "data/raw/{pid}/fitbit_heartrate_intraday_with_datetime.csv"
2020-01-15 23:18:10 +01:00
script:
"../src/data/fitbit_readable_datetime.py"
rule fitbit_step_with_datetime:
input:
expand("data/raw/{{pid}}/{fitbit_table}_raw.csv", fitbit_table=config["STEP"]["DB_TABLE"])
params:
local_timezone = config["READABLE_DATETIME"]["FIXED_TIMEZONE"],
fitbit_sensor = "steps"
output:
intraday_data = "data/raw/{pid}/fitbit_step_intraday_with_datetime.csv"
script:
"../src/data/fitbit_readable_datetime.py"
rule fitbit_sleep_with_datetime:
input:
expand("data/raw/{{pid}}/{fitbit_table}_raw.csv", fitbit_table=config["SLEEP"]["DB_TABLE"])
params:
local_timezone = config["READABLE_DATETIME"]["FIXED_TIMEZONE"],
fitbit_sensor = "sleep"
output:
summary_data = "data/raw/{pid}/fitbit_sleep_summary_with_datetime.csv",
intraday_data = "data/raw/{pid}/fitbit_sleep_intraday_with_datetime.csv"
script:
"../src/data/fitbit_readable_datetime.py"
2020-09-01 00:51:06 +02:00
rule join_wifi_tables:
input:
unpack(optional_wifi_input)
output:
"data/raw/{pid}/wifi_with_datetime_visibleandconnected.csv"
script:
"../src/data/join_visible_and_connected_wifi.R"