Replaced some use of $ with get since with $ you can get NULL if the …

…element does not exist in the named list.

R/create_data.R

@@ -15,24 +15,23 @@
 #' marginal comparisons.
 #'
 #' @export
-create_df_marginal_comparisons <- function(dat, var, percentiles, by_var) {
-  # Checks
-  if (percentiles[1] < 0 | percentiles[2] > 1) {
-    stop("The percentiles must lie between 0 and 1.",
-         call. = TRUE)
-  }
+create_df_marginal_comparisons <-
+  function(dat, var, percentiles, by_var) {
+    # Checks
+    if (percentiles[1] < 0 | percentiles[2] > 1) {
+      stop("The percentiles must lie between 0 and 1.",
+           call. = TRUE)
+    }
 
-  # Compute `low` and `high` values of the variable of interest, by group
-  ret <- dat |>
-    tidylog::group_by(.data[[by_var]]) |>
-    tidylog::mutate(
-      low = quantile(.data[[var]], percentiles[1]),
-      high = quantile(.data[[var]], percentiles[2])
-    ) |>
-    tidylog::ungroup()
+    # Compute `low` and `high` values of the variable of interest, by group
+    ret <- dat |>
+      tidylog::group_by(.data[[by_var]]) |>
+      tidylog::mutate(low = quantile(.data[[var]], percentiles[1]),
+                      high = quantile(.data[[var]], percentiles[2])) |>
+      tidylog::ungroup()
 
-  return(ret)
-}
+    return(ret)
+  }
 
 #' Create a variable's dictionary for its labels
 #'
@@ -101,11 +100,9 @@ add_metadata <- function(dat, metadat, categorical_types) {
   metadat <- metadat |>
     tidylog::mutate(
       type = as.character(type),
-      type = ifelse(
-        type %in% categorical_types,
-        "categorical",
-        type
-      )
+      type = ifelse(type %in% categorical_types,
+                    "categorical",
+                    type)
     )
 
   # Add metadata to each column of dataset
@@ -117,23 +114,23 @@ add_metadata <- function(dat, metadat, categorical_types) {
     info <- metadat |>
       tidylog::filter(variable == x) |>
       as.list()
-    info$description <- .tidy_string(info$description)
-    info$remark <- .tidy_string(info$remark)
-    info$type <- stringr::str_to_lower(info$type)
-    info$comments <- stringr::str_to_lower(info$comments)
+    info$description <- .tidy_string(get("description", info))
+    info$remark <- .tidy_string(get("remark", info))
+    info$type <- stringr::str_to_lower(get("type", info))
+    info$comments <- stringr::str_to_lower(get("comments", info))
     info <- lapply(info, as.character)
 
     # Add metadata
-    attr(dat_modified[[x]], "label") <- info$description
-    attr(dat_modified[[x]], "units") <- info$comments
-    attr(dat_modified[[x]], "remarks") <- info$remark
-    attr(dat_modified[[x]], "dag_var") <- info$dag
-    attr(dat_modified[[x]], "period") <- info$period
-    if (info$type == "categorical") {
+    attr(dat_modified[[x]], "label") <- get("description", info)
+    attr(dat_modified[[x]], "units") <- get("comments", info)
+    attr(dat_modified[[x]], "remarks") <- get("remark", info)
+    attr(dat_modified[[x]], "dag_var") <- get("dag", info)
+    attr(dat_modified[[x]], "period") <- get("period", info)
+    if (get("type", info) == "categorical") {
       dat_modified[[x]] <- labelled::labelled(
         dat_modified[[x]],
-        labels = create_mapping_labels(info$label, info$code),
-        label = info$description
+        labels = create_mapping_labels(get("label", info), get("code", info)),
+        label = get("description", info)
       ) |>
         labelled::to_factor()
     }

R/ctd.R

@@ -13,14 +13,19 @@ load_ctd <- function(path, filter_evidence) {
 
   if (filter_evidence == TRUE) {
     ctd <- ctd |>
-      tidylog::filter(DirectEvidence %in% c("marker/mechanism",
-                                            "therapeutic",
-                                            "marker/mechanism|therapeutic"))
+      tidylog::filter(
+        DirectEvidence %in% c(
+          "marker/mechanism",
+          "therapeutic",
+          "marker/mechanism|therapeutic"
+        )
+      )
   }
 
   ctd <- ctd |>
     tidyr::separate_longer_delim(DiseaseCategories,
                                  delim = "|")
+
   ctd$DiseaseCategories <- factor(ctd$DiseaseCategories,
                                   levels = sort(unique(ctd$DiseaseCategories),
                                                 decreasing = TRUE))
@@ -44,10 +49,8 @@ plot_ctd <- function(dat, group = NULL) {
   plt <- dat |>
     tidylog::filter(DiseaseCategories != "") |>
     tidylog::group_by(DiseaseCategories) |>
-    ggplot2::ggplot(mapping = ggplot2::aes(
-      x = DiseaseCategories,
-      fill = factor(ChemicalName)
-    )) +
+    ggplot2::ggplot(mapping = ggplot2::aes(x = DiseaseCategories,
+                                           fill = factor(ChemicalName))) +
     ggplot2::geom_bar() +
     ggplot2::coord_flip() +
     ggplot2::theme_minimal() +
@@ -57,7 +60,7 @@ plot_ctd <- function(dat, group = NULL) {
 
   if (!is.null(group)) {
     plt <- plt +
-      ggplot2::facet_grid(~.data[[ group ]])
+      ggplot2::facet_grid( ~ .data[[group]])
   }
 
   return(plt)

R/dags.R

@@ -46,7 +46,7 @@ minimize_missings <- function(dat, meta, adjustment_sets,
 
   # List of dataframes with covariates from adjustment sets
   dfs_covars <- lapply(adjustment_sets, function(x) {
-    mapping_covars <- meta[meta$dag %in% x, ]$variable |>
+    mapping_covars <- meta[get("dag", meta) %in% x,]$variable |>
       as.character()
     tmp <- dat |>
       tidylog::select(dplyr::all_of(c(by_var,
@@ -68,7 +68,11 @@ minimize_missings <- function(dat, meta, adjustment_sets,
     purrr::reduce(dplyr::bind_cols))
   colnames(ret_miss) <- paste0("adj_set_", 1:ncol(ret_miss))
   ret_miss <- ret_miss |>
-    tidylog::mutate({{by_var}} := levels_group_var) |>
+    tidylog::mutate({
+      {
+        by_var
+      }
+    } := levels_group_var) |>
     tidylog::relocate(.data[[by_var]])
 
   # Sum of missing values, for each adjustment set
@@ -94,9 +98,13 @@ visualize_dag <- function(dag) {
   ret <- dag |>
     ggdag::tidy_dagitty() |>
     ggdag::node_status() |>
-    ggplot2::ggplot(ggplot2::aes(x, y,
-                                 xend = xend, yend = yend,
-                                 color = status)) +
+    ggplot2::ggplot(ggplot2::aes(
+      x,
+      y,
+      xend = xend,
+      yend = yend,
+      color = status
+    )) +
     ggdag::geom_dag_edges() +
     ggdag::geom_dag_point() +
     geom_dag_text_repel() +
@@ -124,7 +132,7 @@ from_dagitty_to_ggdag <- function(dag) {
   dag <- dagitty::dagitty(dag)
 
   to_ggdag <- ggdag::tidy_dagitty(dag) |>
-    (\(x) x$data) () |>
+    (\(x) get("data", x)) () |>
     tidylog::select(name, to) |>
     tidylog::rename(from = name) |>
     tidylog::group_by(to) |>
@@ -173,13 +181,15 @@ test_npsem <- function(dag, dat, meta, params) {
           call. = TRUE)
 
   # Step 1: extract adjustment set(s)
-  dag_as <- dagitty::adjustmentSets(x = dag,
-                                    type = params$type_mas,
-                                    effect = params$type_effect)
+  dag_as <- dagitty::adjustmentSets(
+    x = dag,
+    type = get("type_mas", params),
+    effect = get("type_effect", params)
+  )
 
   res <- lapply(dag_as, function(as) {
     ret <- list()
-    ret$mapping_covars <- meta[meta$dag %in% as, ] |>
+    ret$mapping_covars <- meta[get("dag", meta) %in% as,] |>
       tidylog::distinct(dag, .keep_all = TRUE) |>
       tidylog::select(variable) |>
       c() |> unname() |> unlist() |> as.character()
@@ -189,25 +199,33 @@ test_npsem <- function(dag, dat, meta, params) {
     covariates <- dat$covariates |>
       tidylog::select(params$identifier,
                       dplyr::all_of(ret$mapping_covars))
-    colnames(covariates) <- c(params$identifier,
-                              meta[meta$variable %in% ret$mapping_covars, ]$dag |>
+    colnames(covariates) <- c(get("identifier", params),
+                              meta[get("variable", meta) %in% get("mapping_covars",
+                                                                  ret),]$dag |>
                                 as.character())
-    covariates <- covariates[, !duplicated(colnames(covariates))]
+    covariates <- covariates[,!duplicated(colnames(covariates))]
 
     # Step 3: test independencies for each exposure
-    exposure_list <- setdiff(colnames(dat$exposures), params$identifier)
+    exposure_list <- setdiff(colnames(get("exposures", dat)),
+                             get("identifier", params))
     ret$tests <- lapply(exposure_list, function(expo) {
-      dat_test <- purrr::reduce(list(covariates,
-                                     dat$exposures |>
-                                       tidylog::select(params$identifier,
-                                                       .data[[expo]]),
-                                     dat$outcome),
-                                tidylog::full_join,
-                                by = params$identifier)
-      test <- dagitty::localTests(x = dag,
-                                  data = dat_test,
-                                  type = "cis.loess",
-                                  R = 3)
+      dat_test <- purrr::reduce(
+        list(
+          covariates,
+          dat$exposures |>
+            tidylog::select(get("identifier", params),
+                            .data[[expo]]),
+          dat$outcome
+        ),
+        tidylog::full_join,
+        by = get("identifier", params)
+      )
+      test <- dagitty::localTests(
+        x = dag,
+        data = dat_test,
+        type = "cis.loess",
+        R = 3
+      )
       return(test)
     }) # End lapply over exposures