refactor: update as per review

NAMESPACE

@@ -14,6 +14,7 @@ export(aggregate_assay)
 export(apply_bumpy_function)
 export(assert_choices)
 export(average_biological_replicates_dt)
+export(calc_sd)
 export(capVals)
 export(cap_xc50)
 export(convert_colData_to_json)

R/utils.R

@@ -470,19 +470,11 @@ average_biological_replicates_dt <- function(
   if (add_sd) {
     # Calculate standard deviation for both average_fields and geometric_average_fields
     sd_fields <- paste0(average_fields, "_sd")
-    geo_sd_fields <- paste0(geometric_average_fields, "_sd")
+    geom_sd_fields <- paste0(geometric_average_fields, "_sd")
 
-    data <- data[, (sd_fields) := lapply(.SD,
-                                         function(x) {
-                                           ifelse(length(x) > 1,
-                                                  stats::sd(x, na.rm = TRUE), 0)
-                                           }),
+    data <- data[, (sd_fields) := lapply(.SD, calc_sd),
                  .SDcols = average_fields, by = group_by]
-    data <- data[, (geo_sd_fields) := lapply(.SD,
-                                             function(x) {
-                                               ifelse(length(x) > 1,
-                                                      stats::sd(x, na.rm = TRUE), 0)
-                                               }),
+    data <- data[, (geom_sd_fields) := lapply(.SD, calc_sd),
                  .SDcols = geometric_average_fields, by = group_by]
 
     # Calculate count and add as a single column
@@ -751,3 +743,23 @@ get_additional_variables <- function(dt_list,
   }
 }
 
+#' Calculate Standard Deviation or Return Zero
+#'
+#' This function calculates the standard deviation of a numeric vector. If the vector has a length of 1 and it is numeric, it returns 0.
+#'
+#' @param x A numeric vector.
+#' @return The standard deviation of the vector if its length is greater than 1 or it is not numeric, otherwise 0.
+#' @examples
+#' calc_sd(c(1, 2, 3, 4, 5)) # Should return the standard deviation
+#' calc_sd(c(1)) # Should return 0
+#' calc_sd(numeric(0)) # Should return NA
+#' calc_sd(c("a", "b", "c")) # Should return NA
+#' @export
+calc_sd <- function(x) {
+  if (length(x) == 1 && is.numeric(x)) {
+    return(0)
+  } else {
+    return(stats::sd(x, na.rm = TRUE))
+  }
+}
+

tests/testthat/test-utils.R

@@ -221,6 +221,8 @@ test_that("average_biological_replicates_dt works as expected", {
                                                         add_sd = TRUE)
 
   expect_equal(dim(avg_metrics_data2), c(40, 44))
+  expect_equal(sum(grepl("_sd", names(avg_metrics_data2))), 15)
+  expect_true("count" %in% names(avg_metrics_data2))
 })
 
 test_that("get_duplicated_rows works as expected", {
@@ -564,4 +566,27 @@ test_that("capVals works as expected", {
   expect_equal(dt1c, dt2)
 
   expect_error(capVals(as.list(dt1)), "Must be a data.table")
-})
+})
+
+test_that("calc_sd returns correct standard deviation for vectors with length > 1", {
+  expect_equal(calc_sd(c(1, 2, 3, 4, 5)), sd(c(1, 2, 3, 4, 5), na.rm = TRUE))
+  expect_equal(calc_sd(c(10, 20, 30)), sd(c(10, 20, 30), na.rm = TRUE))
+})
+
+test_that("calc_sd returns 0 for vectors with length == 1 and numeric", {
+  expect_equal(calc_sd(c(1)), 0)
+})
+
+test_that("calc_sd returns NA for vectors with length == 1 and non-numeric", {
+  expect_true(is.na(calc_sd("2")))
+  expect_true(is.na(calc_sd(TRUE)))
+})
+
+test_that("calc_sd returns NA for empty vectors", {
+  expect_true(is.na(calc_sd(numeric(0))))
+})
+
+test_that("calc_sd handles NA values correctly", {
+  expect_equal(calc_sd(c(1, 2, NA, 4, 5)), sd(c(1, 2, NA, 4, 5), na.rm = TRUE))
+  expect_true(is.na(calc_sd(c(NA, NA, NA))))
+})