Merge pull request #40 from pharmaverse/dev

Dev

.github/workflows/check-standard.yaml

@@ -2,9 +2,9 @@
 # Need help debugging build failures? Start at https://github.com/r-lib/actions#where-to-find-help
 on:
   push:
-    branches: [main, master]
+    branches: [main, master, dev]
   pull_request:
-    branches: [main, master]
+    branches: [main, master, dev]
 
 name: R-CMD-check
 

.github/workflows/test-coverage.yaml

@@ -2,9 +2,9 @@
 # Need help debugging build failures? Start at https://github.com/r-lib/actions#where-to-find-help
 on:
   push:
-    branches: [main, master]
+    branches: [main, master, dev]
   pull_request:
-    branches: [main, master]
+    branches: [main, master, dev]
 
 name: test-coverage
 

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: metatools
 Type: Package
 Title: Enable the Use of 'metacore' to Help Create and Check Dataset
-Version: 0.1.2
+Version: 0.1.3
 Authors@R: c(
          person(given = "Christina",
              family = "Fillmore",
@@ -21,19 +21,19 @@ Description: Uses the metadata information stored in 'metacore' objects to check
 License: MIT + file LICENSE
 Encoding: UTF-8
 LazyData: true
-RoxygenNote: 7.2.0
+RoxygenNote: 7.2.1
 Imports: 
     dplyr,
     metacore (>= 0.0.4),
     purrr,
     rlang,
     stringr,
-    tidyr
-Suggests: 
+    tidyr,
     tibble,
+    magrittr
+Suggests:
     testthat (>= 3.0.0),
     haven,
-    magrittr,
     covr,
     safetyData,
     admiral.test

NAMESPACE

@@ -1,5 +1,6 @@
 # Generated by roxygen2: do not edit by hand
 
+export("%>%")
 export(add_labels)
 export(add_variables)
 export(build_from_derived)
@@ -27,6 +28,7 @@ importFrom(dplyr,any_of)
 importFrom(dplyr,arrange)
 importFrom(dplyr,as_tibble)
 importFrom(dplyr,bind_cols)
+importFrom(dplyr,bind_rows)
 importFrom(dplyr,case_when)
 importFrom(dplyr,distinct)
 importFrom(dplyr,everything)
@@ -41,6 +43,7 @@ importFrom(dplyr,mutate)
 importFrom(dplyr,pull)
 importFrom(dplyr,rename)
 importFrom(dplyr,select)
+importFrom(magrittr,"%>%")
 importFrom(metacore,get_control_term)
 importFrom(metacore,select_dataset)
 importFrom(purrr,discard)
@@ -69,5 +72,7 @@ importFrom(stringr,str_remove)
 importFrom(stringr,str_remove_all)
 importFrom(stringr,str_to_lower)
 importFrom(stringr,str_to_upper)
+importFrom(tibble,tibble)
 importFrom(tidyr,pivot_wider)
+importFrom(tidyr,unnest)
 importFrom(utils,capture.output)

NEWS.md

@@ -1,3 +1,10 @@
+# metatools 0.1.3
+* correct bug with `build_from_derived()` that prevented multiple from being combined 
+* removed library calls from tests 
+* remove `floating_pt_correction` from `combine_supp()` as it is best to never change the type of the main dataset 
+* Add a message to `drop_unspec_vars` to explain which variables are dropped 
+* Correct bug in `order_cols`, so it will still work when not all columns are present 
+
 # metatools 0.1.2
 * correct bug with `combine_supp()` when the data and the supp have different classes for the IDVARVAL
 * add error to `combine_supp()` to report when not all the rows in the supp have merged

R/build.R

@@ -2,27 +2,32 @@
 #'
 #' This function builds a dataset out of the columns that just need to be pulled
 #' through. So any variable that has a derivation in the format of
-#' 'dataset.variable' will be pulled through to create the new dataset. These
-#' columns are often called 'Predecessors' in ADaM, but this is not universal so
-#' that is optional to specify.
+#' 'dataset.variable' will be pulled through to create the new dataset. When
+#' there are multiple datasets present, they will be joined by the shared
+#' `key_seq` variables. These columns are often called 'Predecessors' in ADaM,
+#' but this is not universal so that is optional to specify.
 #' @param metacore metacore object that contains the specifications for the
 #'   dataset of interest.
 #' @param ds_list Named list of datasets that are needed to build the from
-#' @param dataset_name Optional string to specify the dataset. This is only
-#'   needed if the metacore object provided hasn't already been subsetted.
+#' @param dataset_name Optional string to specify the dataset that is being
+#'   built. This is only needed if the metacore object provided hasn't already
+#'   been subsetted.
 #' @param predecessor_only By default `FALSE`, but if `TRUE` will only use
 #'   derivations with the origin of 'Predecessor'
 #' @param keep Boolean to determine if the original columns should be kept. By
 #'   default `FALSE`, so only the ADaM columns are kept. If `TRUE` the resulting
 #'   dataset will have all the ADaM columns as well as any SDTM column that were
-#'   renamed in the ADaM (i.e `ARM` and `TRT01P` will be in the resulting dataset)
+#'   renamed in the ADaM (i.e `ARM` and `TRT01P` will be in the resulting
+#'   dataset)
 #'
 #' @return dataset
 #' @export
 #' @importFrom stringr str_to_lower str_detect str_extract str_to_upper
 #' @importFrom dplyr filter pull mutate group_by group_split inner_join select
-#'   full_join
+#'   full_join bind_rows
+#' @importFrom tidyr unnest
 #' @importFrom purrr map reduce
+#' @importFrom tibble tibble
 #'
 #' @examples
 #' library(metacore)
@@ -35,26 +40,29 @@
 build_from_derived <- function(metacore, ds_list, dataset_name = NULL,
                                predecessor_only = TRUE, keep = FALSE) {
    metacore <- make_lone_dataset(metacore, dataset_name)
-   derirvations <- metacore$derivations
+   derirvations <- metacore$derivations %>%
+      mutate(derivation = trimws(derivation))
+
    if (predecessor_only) {
       limited_dev_ids <- metacore$value_spec %>%
-         filter(str_detect(.data$origin, "[P|p]redecessor")) %>%
-         pull(.data$derivation_id)
+         filter(str_detect(str_to_lower(origin), "predecessor")) %>%
+         pull(derivation_id)
 
       derirvations <- derirvations %>%
-         filter(.data$derivation_id %in% limited_dev_ids)
+         filter(derivation_id %in% limited_dev_ids)
       if (nrow(derirvations) == 0) {
          stop("No presecessor variables found please check your metacore object")
       }
    }
+
    vars_to_pull_through <- derirvations %>%
-      filter(str_detect(.data$derivation, "^\\w*\\.[a-zA-Z0-9]*$"))
+      filter(str_detect(derivation, "^\\w*\\.[a-zA-Z0-9]*$"))
    # To lower so it is flexible about how people name their ds list
    vars_w_ds <- vars_to_pull_through %>%
-      mutate(ds = str_extract(.data$derivation, "^\\w*(?=\\.)") %>%
+      mutate(ds = str_extract(derivation, "^\\w*(?=\\.)") %>%
                 str_to_lower())
    ds_names <- vars_w_ds %>%
-      pull(.data$ds) %>%
+      pull(ds) %>%
       unique()
    names(ds_list) <- names(ds_list) %>%
       str_to_lower()
@@ -64,14 +72,30 @@ build_from_derived <- function(metacore, ds_list, dataset_name = NULL,
          paste0(str_to_upper(ds_names), collapse = "\n")
       ))
    }
-   join_by <- metacore$ds_vars %>%
-      filter(!is.na(.data$key_seq)) %>%
-      pull(.data$variable)
+
+   ds_keys <- metacore$ds_vars %>%
+      filter(!is.na(key_seq)) %>%
+      pull(variable)
+
+   joining_vals_to_add <- ds_list %>%
+      map(function(x){
+         names(x) %>%
+            keep(~ . %in% ds_keys)
+         })
+
+   join_by = joining_vals_to_add %>%
+      reduce(intersect)
+   additional_vals <- tibble(ds = names(ds_list),
+          variable = joining_vals_to_add) %>%
+      unnest(variable) %>%
+      mutate(col_name = variable)
+
    vars_w_ds %>%
-      mutate(col_name = str_extract(.data$derivation, "(?<=\\.).*")) %>%
+      mutate(col_name = str_extract(derivation, "(?<=\\.).*")) %>%
       inner_join(metacore$value_spec, ., by = "derivation_id") %>%
-      select(.data$variable, .data$ds, .data$col_name) %>%
-      group_by(.data$ds) %>%
+      select(variable, ds, col_name) %>%
+      bind_rows(additional_vals) %>%
+      group_by(ds) %>%
       group_split() %>%
       map(get_variables, ds_list, keep) %>%
       reduce(full_join, by = join_by)
@@ -134,13 +158,15 @@ get_variables <- function(x, ds_list, keep) {
 drop_unspec_vars <- function(dataset, metacore, dataset_name = NULL) {
    metacore <- make_lone_dataset(metacore, dataset_name)
    var_list <- metacore$ds_vars %>%
-      filter(is.na(.data$supp_flag) | !(.data$supp_flag)) %>%
-      pull(.data$variable)
+      filter(is.na(supp_flag) | !(supp_flag)) %>%
+      pull(variable)
    to_drop <- names(dataset) %>%
       discard(~ . %in% var_list)
    if (length(to_drop) > 0) {
       out <- dataset %>%
          select(-all_of(to_drop))
+      message(paste0("The following variable(s) were dropped:\n  ",
+              paste0(to_drop, collapse = "\n  ")))
    } else {
       out <- dataset
    }
@@ -182,19 +208,19 @@ drop_unspec_vars <- function(dataset, metacore, dataset_name = NULL) {
 add_variables <- function(dataset, metacore, dataset_name = NULL){
    metacore <- make_lone_dataset(metacore, dataset_name)
    var_list <- metacore$ds_vars %>%
-      filter(is.na(.data$supp_flag) | !(.data$supp_flag)) %>%
-      pull(.data$variable)
+      filter(is.na(supp_flag) | !(supp_flag)) %>%
+      pull(variable)
 
    to_add <- var_list %>%
       discard(~ . %in% names(dataset))
    if(length(to_add) > 0){
       n <- nrow(dataset)
       typing <- metacore$var_spec %>%
-         filter(.data$variable %in% to_add) %>%
-         mutate(type_fmt = str_to_lower(.data$type),
+         filter(variable %in% to_add) %>%
+         mutate(type_fmt = str_to_lower(type),
                 out_type =
                    case_when(
-                      str_detect(str_to_lower(.data$format), "date") ~ "date",
+                      str_detect(str_to_lower(format), "date") ~ "date",
                       type_fmt == "integer" ~ "integer",
                       type_fmt == "numeric" ~ "double",
                       type_fmt == "text" ~ "character",

R/checks.R

@@ -40,12 +40,12 @@ check_ct_col <- function(data, metacore, var, na_acceptable = NULL) {
   if (is.vector(ct)) {
     check <- ct
   } else if ("code" %in% names(ct)) {
-    check <- ct %>% pull(.data$code)
+    check <- ct %>% pull(code)
   } else {
     stop("We currently don't have the ability to check against external libraries")
   }
   core <- metacore$ds_vars %>%
-    filter(.data$variable == col_name_str) %>%
+    filter(variable == col_name_str) %>%
     pull(core)
   attr(core, "label") <- NULL
   test <- ifelse(is.null(na_acceptable), !identical(core, "Required"), na_acceptable)
@@ -99,18 +99,18 @@ check_ct_col <- function(data, metacore, var, na_acceptable = NULL) {
 #' check_ct_data(data, spec)
 check_ct_data <- function(data, metacore, na_acceptable = NULL) {
   codes_in_data <- metacore$value_spec %>%
-    filter(.data$variable %in% names(data), !is.na(.data$code_id)) %>%
-    pull(.data$code_id) %>%
+    filter(variable %in% names(data), !is.na(code_id)) %>%
+    pull(code_id) %>%
     unique()
   # Remove any codes that have external libraries
   codes_to_check <- metacore$codelist %>%
-    filter(.data$type != "external_library", .data$code_id %in% codes_in_data) %>%
-    select(.data$code_id)
+    filter(type != "external_library", code_id %in% codes_in_data) %>%
+    select(code_id)
   # convert list of codes to variables
   cols_to_check <- metacore$value_spec %>%
     inner_join(codes_to_check, by = "code_id") %>%
-    filter(.data$variable %in% names(data)) %>%
-    pull(.data$variable) %>%
+    filter(variable %in% names(data)) %>%
+    pull(variable) %>%
     unique()
   # send all variables through check_ct_col
   safe_chk <- safely(check_ct_col)
@@ -174,8 +174,8 @@ check_ct_data <- function(data, metacore, na_acceptable = NULL) {
 check_variables <- function(data, metacore, dataset_name = NULL) {
   metacore <- make_lone_dataset(metacore, dataset_name)
   var_list <- metacore$ds_vars %>%
-     filter(is.na(.data$supp_flag) | !(.data$supp_flag)) %>%
-    pull(.data$variable)
+     filter(is.na(supp_flag) | !(supp_flag)) %>%
+    pull(variable)
   missing <- var_list %>%
     discard(~ . %in% names(data))
   extra <- names(data) %>%

R/codelists.R

@@ -177,7 +177,7 @@ create_cat_var <- function(data, metacore, ref_var, grp_var,
     stop("Expecting 'code_decode' type of control terminology. Please check metacore object")
   }
   grp_defs <- ct %>%
-    pull(.data$code)
+    pull(code)
 
   out <- data %>%
     mutate({{ grp_var }} := create_subgrps({{ ref_var }}, grp_defs))

R/labels.R

@@ -90,7 +90,7 @@ set_variable_labels <- function(data, metacore, dataset_name = NULL) {
 
    # Grab out the var names and labels
    var_spec <- metacore$var_spec %>%
-      select(.data$variable, .data$label)
+      select(variable, label)
 
 
    ns <- var_spec$variable

R/sort.R

@@ -23,13 +23,16 @@
 #' data <- read_xpt(metatools_example("adsl.xpt"))
 #' order_cols(data, spec)
 order_cols <- function(data, metacore, dataset_name = NULL) {
-  metacore <- make_lone_dataset(metacore, dataset_name)
-  var_ord <- metacore$ds_vars %>%
-    filter(!is.na(.data$order)) %>%
-    arrange(.data$order) %>%
-    pull(.data$variable)
-  data %>%
-    select(all_of(var_ord), everything())
+   metacore <- make_lone_dataset(metacore, dataset_name)
+
+   var_ord <- metacore$ds_vars %>%
+      filter(!is.na(order)) %>%
+      arrange(order) %>%
+      pull(variable) %>%
+      keep(~. %in% names(data))
+
+   data %>%
+      select(all_of(var_ord), everything())
 }
 
 
@@ -57,12 +60,12 @@ order_cols <- function(data, metacore, dataset_name = NULL) {
 #' data <- read_xpt(metatools_example("adsl.xpt"))
 #' sort_by_key(data, spec)
 sort_by_key <- function(data, metacore, dataset_name = NULL) {
-  metacore <- make_lone_dataset(metacore, dataset_name)
-  var_ord <- metacore$ds_vars %>%
-    filter(!is.na(.data$key_seq)) %>%
-    arrange(.data$key_seq) %>%
-    pull(.data$variable)
+   metacore <- make_lone_dataset(metacore, dataset_name)
+   var_ord <- metacore$ds_vars %>%
+      filter(!is.na(key_seq)) %>%
+      arrange(key_seq) %>%
+      pull(variable)
 
-  data %>%
-    arrange(across(var_ord))
+   data %>%
+      arrange(across(var_ord))
 }