Change the way we compute battery episodes

rules/features.smk

@@ -54,13 +54,13 @@ rule calls_r_features:
     script:
         "../src/features/calls/calls_entry.R"
 
-rule battery_deltas:
+rule battery_episodes:
     input:
-        expand("data/raw/{{pid}}/{sensor}_with_datetime_unified.csv", sensor=config["BATTERY"]["DB_TABLE"])
+        expand("data/raw/{{pid}}/{sensor}_raw.csv", sensor=config["BATTERY"]["DB_TABLE"])
     output:
-        "data/processed/{pid}/battery_deltas.csv"
+        "data/interim/{pid}/battery_episodes.csv"
     script:
-        "../src/features/battery_deltas.R"
+        "../src/features/battery/episodes/battery_episodes.R"
 
 rule screen_episodes:
     input:
@@ -172,7 +172,7 @@ rule activity_features:
 
 rule battery_features:
     input:
-        "data/processed/{pid}/battery_deltas.csv"
+        "data/interim/{pid}/battery_episodes.csv"
     params:
         day_segment = "{day_segment}",
         features = config["BATTERY"]["FEATURES"]

src/features/battery/episodes/battery_episodes.R

@@ -0,0 +1,24 @@
+source("renv/activate.R")
+library("dplyr")
+
+battery <- read.csv(snakemake@input[[1]])
+
+if(nrow(battery) > 0){
+  # TODO expose this in the config file
+  threshold_between_rows = 30
+  battery_episodes <- battery %>% 
+    filter(battery_status >= 2 ) %>% # discard unknown states
+    mutate(start_timestamp = timestamp,
+          end_timestamp = lead(start_timestamp) - 1,
+          time_diff = (end_timestamp - start_timestamp) / 1000 / 60,
+          time_diff = if_else(time_diff > threshold_between_rows, threshold_between_rows, time_diff),
+          episode_id = 1:n()) %>%
+    select(episode_id, start_timestamp, end_timestamp, battery_level)
+} else {
+  battery_episodes <- data.frame(episode_id = numeric(), 
+                            start_timestamp = numeric(),
+                            end_timestamp = character(),
+                            battery_level = character())
+}
+
+write.csv(battery_episodes, snakemake@output[[1]], row.names = FALSE)

src/features/battery_deltas.R

@@ -1,34 +0,0 @@
-source("renv/activate.R")
-
-library("tidyverse")
-
-battery <- read.csv(snakemake@input[[1]])
-
-if(nrow(battery) > 0){
-  consumption <- battery %>%
-    mutate(group = ifelse(lag(battery_status) != battery_status, 1, 0) %>% coalesce(0),
-           group_id = cumsum(group) + 1) %>%
-    filter(battery_status == 2 | battery_status == 3) %>%
-    group_by(group_id) %>%
-    summarize(battery_diff = first(battery_level) - last(battery_level),
-              time_diff = (last(timestamp) - first(timestamp)) / (1000 * 60),
-              local_start_date_time = first(local_date_time),
-              local_end_date_time = last(local_date_time),
-              local_start_date = first(local_date),
-              local_end_date = last(local_date),
-              local_start_day_segment = first(local_day_segment),
-              local_end_day_segment = last(local_day_segment)) %>%
-    select(-group_id) %>%
-    filter(time_diff > 6) # Avoids including quick cycles
-} else {
-  consumption <- data.frame(battery_diff = numeric(), 
-                            time_diff = numeric(),
-                            local_start_date_time = character(),
-                            local_end_date_time = character(),
-                            local_start_date = character(),
-                            local_end_date = character(),
-                            local_start_day_segment = character(),
-                            local_end_day_segment = character())
-}
-
-write.csv(consumption, snakemake@output[[1]], row.names = FALSE)

src/features/screen_features.py

@@ -1,18 +0,0 @@
-import pandas as pd
-import itertools
-from screen.screen_base import base_screen_features
-
-screen_data = pd.read_csv(snakemake.input["screen_deltas"], parse_dates=["local_start_date_time", "local_end_date_time", "local_start_date", "local_end_date"])
-phone_sensed_bins = pd.read_csv(snakemake.input["phone_sensed_bins"], parse_dates=["local_date"], index_col="local_date")
-phone_sensed_bins[phone_sensed_bins > 0] = 1
-day_segment = snakemake.params["day_segment"]
-screen_features = pd.DataFrame(columns=["local_date"])
-
-requested_features_deltas = ["firstuseafter" + "{0:0=2d}".format(snakemake.params["reference_hour_first_use"]) if feature_name == "firstuseafter" else feature_name for feature_name in snakemake.params["features_deltas"]]
-requested_features = ["".join(feature) for feature in itertools.product(requested_features_deltas, snakemake.params["episode_types"])]
-
-screen_features = screen_features.merge(base_screen_features(screen_data, phone_sensed_bins, day_segment, snakemake.params), on="local_date", how="outer")
-
-assert len(requested_features) + 1 == screen_features.shape[1], "The number of features in the output dataframe (=" + str(screen_features.shape[1]) + ") does not match the expected value (=" + str(len(requested_features)) + " + 1). Verify your screen feature extraction functions"
-
-screen_features.to_csv(snakemake.output[0], index=False)