Prepare a function to compile participants data.

It combines functions from container.R

src/data/translate_usernames_into_participants_data.R

@@ -0,0 +1,52 @@
+library(RPostgres)
+library(magrittr)
+library(tidyverse)
+library(lubridate)
+
+load_container_script <- function(stream_container){
+  language <- if_else(endsWith(tolower(stream_container), "py"), "python", "r")
+  if(language == "python"){
+    library(reticulate)
+    container <- import_from_path(gsub(pattern = "\\.py$", "", basename(stream_container)), path = dirname(stream_container))
+    if(!py_has_attr(container, "pull_data"))
+      stop(paste0("The following container.py script does not have a pull_data function: ", stream_container))
+    if(!py_has_attr(container, "infer_device_os"))
+      stop(paste0("The following container.py script does not have a infer_device_os function: ", stream_container))
+    return(list("infer_device_os" = container$infer_device_os, "pull_data" = container$pull_data))
+  } else if(language == "r"){
+    source(stream_container)
+    if(!exists("pull_data"))
+      stop(paste0("The following container.R script does not have a pull_data function: ", stream_container))
+    if(!exists("infer_device_os"))
+      stop(paste0("The following container.R script does not have a infer_device_os function: ", stream_container))
+    return(list("infer_device_os" = infer_device_os, "pull_data" = pull_data))
+  }
+}
+
+prepare_participants_file <- function() {
+  #TODO Define appropriate arguments and pass them from config.yaml
+  stream_container <- snakemake@input[["stream_container"]]
+  container_functions <- load_container_script(stream_container)
+
+  pull_data_container <- container_functions$pull_data
+  # TODO Figure out how to use the functions in the container
+
+  participant_data <- pull_participants_ids("whatever", usernames, participants_container)
+  participant_data %<>%
+    rename(participant_id = id)
+
+  device_ids <- pull_participants_device_ids("whatever", participant_data$participant_id, device_id_container)
+  device_ids %<>%
+    group_by(participant_id) %>%
+    summarise(device_ids = list(unique(device_id)))
+  participant_data %<>%
+    left_join(device_ids, by = "participant_id")
+
+  start_end_datetimes <- pull_participants_start_end_dates("whatever", participant_data$participant_id, start_end_date_container)
+  participant_data %<>%
+    left_join(start_end_datetimes, by = "participant_id")
+
+  return(participant_data)
+}
+
+