Merge pull request #533 from utkarshpawade/fix/ecdf-geom-step-259

Use geom_step() as default in ppc_ecdf_overlay() and deprecate discrete argument

Author: jgabry

NEWS.md

@@ -1,5 +1,6 @@
 # bayesplot (development version)
 
+* `ppc_ecdf_overlay()`, `ppc_ecdf_overlay_grouped()`, and `ppd_ecdf_overlay()` now always use `geom_step()`. The `discrete` argument is deprecated.
 * Fixed missing `drop = FALSE` in `nuts_params.CmdStanMCMC()`.
 * Replace `apply()` with `storage.mode()` for integer-to-numeric matrix conversion in `validate_predictions()`.
 * Fixed `is_chain_list()` to correctly reject empty lists instead of silently returning `TRUE`.

R/ppc-distributions.R

@@ -41,11 +41,10 @@
 #'   }
 #'   \item{`ppc_ecdf_overlay()`, `ppc_dens_overlay()`,
 #'          `ppc_ecdf_overlay_grouped()`, `ppc_dens_overlay_grouped()`}{
-#'    Kernel density or empirical CDF estimates of each dataset (row) in
-#'    `yrep` are overlaid, with the distribution of `y` itself on top
-#'    (and in a darker shade). When using `ppc_ecdf_overlay()` with discrete
-#'    data, set the `discrete` argument to `TRUE` for better results.
-#'    For an example of `ppc_dens_overlay()` also see Gabry et al. (2019).
+#'    Kernel density or empirical CDF estimates of each dataset (row) in `yrep`
+#'    are overlaid, with the distribution of `y` itself on top (and in a darker
+#'    shade). For an example of `ppc_dens_overlay()` also see Gabry et al.
+#'    (2019).
 #'   }
 #'   \item{`ppc_violin_grouped()`}{
 #'    The density estimate of `yrep` within each level of a grouping
@@ -85,7 +84,7 @@
 #'
 #' ppc_dens_overlay(y, yrep[1:25, ])
 #' \donttest{
-#' # ppc_ecdf_overlay with continuous data (set discrete=TRUE if discrete data)
+#' # ppc_ecdf_overlay
 #' ppc_ecdf_overlay(y, yrep[sample(nrow(yrep), 25), ])
 #'
 #' # PIT-ECDF and PIT-ECDF difference plot of the PIT values of y compared to
@@ -258,20 +257,28 @@ ppc_dens_overlay_grouped <- function(y,
 
 #' @export
 #' @rdname PPC-distributions
-#' @param discrete For `ppc_ecdf_overlay()`, should the data be treated as
-#'   discrete? The default is `FALSE`, in which case `geom="line"` is
-#'   passed to [ggplot2::stat_ecdf()]. If `discrete` is set to
-#'   `TRUE` then `geom="step"` is used.
+#' @param discrete `r lifecycle::badge("deprecated")` The `discrete` argument is
+#'   deprecated. The ECDF is a step function by definition, so `geom_step()` is
+#'   now always used.
 #' @param pad A logical scalar passed to [ggplot2::stat_ecdf()].
 #'
 ppc_ecdf_overlay <- function(y,
                              yrep,
                              ...,
-                             discrete = FALSE,
+                             discrete = deprecated(),
                              pad = TRUE,
                              size = 0.25,
                              alpha = 0.7) {
   check_ignored_arguments(...)
+
+  if (is_present(discrete)) {
+    deprecate_warn(
+      "1.16.0",
+      "ppc_ecdf_overlay(discrete)",
+      details = "The ECDF is now always plotted as a step function."
+    )
+  }
+
   data <- ppc_data(y, yrep)
 
   ggplot(data) +
@@ -291,15 +298,15 @@ ppc_ecdf_overlay <- function(y,
     stat_ecdf(
       data = function(x) dplyr::filter(x, !.data$is_y),
       mapping = aes(group = .data$rep_id, color = "yrep"),
-      geom = if (discrete) "step" else "line",
+      geom = "step",
       linewidth = size,
       alpha = alpha,
       pad = pad
     ) +
     stat_ecdf(
       data = function(x) dplyr::filter(x, .data$is_y),
       mapping = aes(color = "y"),
-      geom = if (discrete) "step" else "line",
+      geom = "step",
       linewidth = 1,
       pad = pad
     ) +
@@ -316,17 +323,24 @@ ppc_ecdf_overlay_grouped <- function(y,
                                      yrep,
                                      group,
                                      ...,
-                                     discrete = FALSE,
+                                     discrete = deprecated(),
                                      pad = TRUE,
                                      size = 0.25,
                                      alpha = 0.7) {
   check_ignored_arguments(...)
 
+  if (is_present(discrete)) {
+    deprecate_warn(
+      "1.16.0",
+      "ppc_ecdf_overlay_grouped(discrete)",
+      details = "The ECDF is now always plotted as a step function."
+    )
+  }
+
   p_overlay <- ppc_ecdf_overlay(
     y = y,
     yrep = yrep,
     ...,
-    discrete = discrete,
     pad = pad,
     size = size,
     alpha = alpha

R/ppd-distributions.R

@@ -83,12 +83,20 @@ ppd_dens_overlay <-
 ppd_ecdf_overlay <-
   function(ypred,
            ...,
-           discrete = FALSE,
+           discrete = deprecated(),
            pad = TRUE,
            size = 0.25,
            alpha = 0.7) {
     check_ignored_arguments(...)
 
+    if (is_present(discrete)) {
+      deprecate_warn(
+        "1.16.0",
+        "ppd_ecdf_overlay(discrete)",
+        details = "The ECDF is now always plotted as a step function."
+      )
+    }
+
     data <- ppd_data(ypred)
     ggplot(data, mapping = aes(x = .data$value)) +
       hline_at(
@@ -99,7 +107,7 @@ ppd_ecdf_overlay <-
       ) +
       stat_ecdf(
         mapping = aes(group = .data$rep_id, color = "ypred"),
-        geom = if (discrete) "step" else "line",
+        geom = "step",
         linewidth = size,
         alpha = alpha,
         pad = pad