Make corr_hm independent from dplyr.

Author: ml-ebs-ext

R/mock_corr_hm.R

@@ -5,6 +5,14 @@
 
 mock_app_corr_hm <- function(dry_run = FALSE, update_query_string = TRUE, srv_defaults = list(), ui_defaults = list(), anlfl_flags = FALSE) {
   data <- test_data(random_bm_values = TRUE, anlfl_flags = anlfl_flags)
+  anlfl_vars <- NULL
+  if (anlfl_flags) { # drop some observations from ANLFL1
+    subjid_int <- as.integer(data$bm[["SUBJID"]])
+    drop_mask <- (16 <= subjid_int & subjid_int <= 20 & data$bm[["ANLFL1"]] == "Y")
+    data$bm <- data$bm[!drop_mask, ]
+    anlfl_vars <- c("ANLFL1", "ANLFL2")
+  }
+
   bm_dataset <- shiny::reactive({
     data[["bm"]]
   })
@@ -16,21 +24,6 @@ mock_app_corr_hm <- function(dry_run = FALSE, update_query_string = TRUE, srv_de
     ui_defaults
   )
 
-  if (anlfl_flags) {
-
-    # modifying the test data to make them asymetric so that there is visible difference in the calculated values between the two analysis flag variables
-    data$bm <- dplyr::filter(
-      data$bm,
-      !(as.numeric(as.character(.data[["SUBJID"]])) >= 16 &
-        as.numeric(as.character(.data[["SUBJID"]])) <= 20 &
-        .data[["ANLFL1"]] == "Y")
-    )
-
-    anlfl_vars <- c("ANLFL1", "ANLFL2")
-  } else {
-    anlfl_vars <- NULL
-  }
-
   srv_params <- c(
     list(
       id = "not_ebas",
@@ -74,7 +67,7 @@ mock_app_correlation_hm_mm <- function(anlfl_flags = FALSE) {
   if (anlfl_flags) { # drop some observations from ANLFL1
     subjid_int <- as.integer(bm_dataset[["SUBJID"]])
     drop_mask <- (16 <= subjid_int & subjid_int <= 20 & bm_dataset[["ANLFL1"]] == "Y")
-    bm_dataset <- bm_dataset[!drop_mask,]
+    bm_dataset <- bm_dataset[!drop_mask, ]
     anlfl_vars <- c("ANLFL1", "ANLFL2")
   }