Merge pull request #7 from certara/data_validation

Data validation function

R/cov_search.R

@@ -13,6 +13,12 @@
 #'
 MLCovSearch <- function(tab, list_pop_param, cov_continuous, cov_factors, seed = 123) {
 
+  # Check that covariates supplied by user exist in the data
+  errors <- data_validation(tab, list_pop_param, cov_continuous, cov_factors)
+  if (length(errors) > 0) {
+    stop(paste0(errors, sep = "\n"), call. = FALSE)
+  }
+
   stopifnot(is.numeric(seed))
   set.seed(seed)
 
@@ -247,7 +253,25 @@ MLCovSearch <- function(tab, list_pop_param, cov_continuous, cov_factors, seed =
   )
 } 
 
-
+data_validation <- function(tab, list_pop_param, cov_continuous, cov_factors) {
+  errors <- c()
+
+  for (i in 1:3) {
+    vectors <- list(list_pop_param, cov_continuous, cov_factors)
+    vector_names <- c("list_pop_param", "cov_continuous", "cov_factors")
+
+    missing_values <- setdiff(vectors[[i]], colnames(tab))
+
+    if (length(missing_values) > 0) {
+      error_message <-
+        paste( "The following values from", vector_names[[i]], "are missing in the dataset:", toString(missing_values))
+
+      errors <- c(errors, error_message)
+    }
+  }
+  return(errors)
+
+}
 
 
 
@@ -282,6 +306,13 @@ MLCovSearch <- function(tab, list_pop_param, cov_continuous, cov_factors, seed =
 #' }
 #' @export
 generate_residualsplots <- function(tab, list_pop_param, cov_continuous, cov_factors, result_ML, result_5folds, i, seed = 123 ) {
+
+  # Check that covariates supplied by user exist in the data
+  errors <- data_validation(tab, list_pop_param, cov_continuous, cov_factors)
+  if (length(errors) > 0) {
+    stop(paste0(errors, sep = "\n"), call. = FALSE)
+  }
+
   stopifnot(is.numeric(seed))
   set.seed(seed)
 

R/generate_residualsplots2.R

@@ -36,6 +36,12 @@ generate_residualsplots2 <- function(data, result, i, seed = NULL) {
   result_ML <- result$result_ML
   result_5folds <- result$result_5folds
 
+  # Check that covariates supplied by user exist in the data
+  errors <- data_validation(data, list_pop_param, cov_continuous, cov_factors)
+  if (length(errors) > 0) {
+    stop(paste0(errors, sep = "\n"), call. = FALSE)
+  }
+
   # Selection of columns required
   data <- data %>%
     dplyr::select(ID, dplyr::all_of(list_pop_param), dplyr::all_of(cov_continuous), dplyr::all_of(cov_factors))

tests/testthat/test-generate_residualsplots.R

@@ -64,3 +64,14 @@ testthat::test_that("test-generate_residualsplots returns the same plots each ti
   vdiffr::expect_doppelganger("SEX Residuals", plot_V1$SEX)
 })
 
+testthat::test_that("Error messages will be generated when values supplied to the function are absent in the data.frame", {
+
+  testthat::expect_error(generate_residualsplots(tab = data,
+                                                 list_pop_param = c("clearance"),
+                                                 cov_continuous = c("weight"),
+                                                 cov_factors = c("dIaB"),
+                                                 result_ML = result$result_ML, #selected covariates for parameter of interest after the voting
+                                                 result_5folds = result$result_5folds, #selected covariates for parameter of interest for each folds
+                                                 i=c('CL')))
+})
+

tests/testthat/test-generateresidualsplots2.R

@@ -42,3 +42,9 @@ testthat::test_that("generate_residualsplots does not return plots for Cl", {
 testthat::test_that("generate_residualsplots does not return plots for V1", {
   testthat::expect_null(plot_V1)
 })
+
+testthat::test_that("Error messages will be generated when values supplied to the function are absent in the data.frame", {
+
+  #Since residualsplots2 function pulls these values from the results object, it should always run correctly. 
+  testthat::expect_no_error(generate_residualsplots2(data, result, i = c('V1')))
+})

tests/testthat/test-mlcovsearch.R

@@ -25,6 +25,14 @@ result <- suppressWarnings(MLCovSearch(data, #NONMEM output
                                          "HGB","HCT","PLT"),
                       cov_factors = c("SEX","RACE","DIAB","ALQ","WACT","SMQ")))
 
+testthat::test_that("Error messages will be generated when values supplied to the MLCovSearch are absent in the data.frame", {
+
+  testthat::expect_error(MLCovSearch(data, #NONMEM output
+    list_pop_param = c("clearance"),
+    cov_continuous = c("weight"),
+    cov_factors = c("dIaB")))
+})
+
 testthat::test_that("MLCovSearch result_ML is a data.frame with only one covariate selection", {
 
   testthat::expect_true(is.data.frame(result$result_ML))