much faster (c++) implementation of wilkinson binning

DESCRIPTION

@@ -34,7 +34,8 @@ Imports:
     distributional (>= 0.3.2),
     numDeriv,
     glue,
-    quadprog
+    quadprog,
+    Rcpp
 Suggests:
     knitr,
     testthat (>= 3.0.0),
@@ -70,6 +71,7 @@ Collate:
     "ggdist-package.R"
     "util.R"
     "rd.R"
+    "RcppExports.R"
     "abstract_geom.R"
     "abstract_stat.R"
     "abstract_stat_slabinterval.R"
@@ -123,3 +125,5 @@ Collate:
     "deprecated.R"
 Roxygen: list(markdown = TRUE)
 Config/testthat/edition: 3
+LinkingTo: 
+    Rcpp

NAMESPACE

@@ -246,6 +246,7 @@ export(weighted_quantile_fun)
 import(ggplot2)
 import(grid)
 import(vctrs)
+importFrom(Rcpp,sourceCpp)
 importFrom(cli,cli_abort)
 importFrom(cli,cli_warn)
 importFrom(distributional,cdf)
@@ -338,3 +339,4 @@ importFrom(tibble,as_tibble)
 importFrom(tibble,tibble)
 importFrom(tidyselect,eval_select)
 importFrom(utils,packageVersion)
+useDynLib(ggdist, .registration = TRUE)

NEWS.md

@@ -10,6 +10,8 @@ New features and enhancements:
   to `bandwidth_nrd0()` when they fail, with a warning that suggests trying 
   a dotplot or histogram (as these failures tend to happen on data that is not
   a good candidate for a density plot in the first place) (#196).
+* Much faster (C++) implementation of Wilkinson dotplot binning, especially
+  for large dotplots.
 
 Bug fixes:
 

R/RcppExports.R

@@ -0,0 +1,7 @@
+# Generated by using Rcpp::compileAttributes() -> do not edit by hand
+# Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
+
+wilkinson_bin_to_right_ <- function(x, width) {
+    .Call(`_ggdist_wilkinson_bin_to_right_`, x, width)
+}
+

R/binning_methods.R

@@ -410,21 +410,8 @@ wilkinson_bin_to_right = function(x, width) {
     ))
   }
 
-  # determine bins and midpoints of bins
-  bins = rep(NA_integer_, length(x))
-  bins[[1]] = 1L
-  current_bin = 1L
-  first_x = x[[1]]
-  n = 1
-  for (i in seq_along(x)[-1]) {
-    # This is equivalent to x[[i]] - first_x >= width but it accounts for machine precision.
-    # If we instead used `>=` directly some things that should be symmetric will not be
-    if (x[[i]] - first_x - width >= -.Machine$double.eps) {
-      current_bin = current_bin + 1L
-      first_x = x[[i]]
-    }
-    bins[[i]] = current_bin
-  }
+  # determine bins
+  bins = wilkinson_bin_to_right_(x, width)
 
   # determine bin positions
   # can take advantage of the fact that bins is sorted runs of numbers to

R/ggdist-package.R

@@ -1,3 +1,9 @@
+## usethis namespace: start
+#' @importFrom Rcpp sourceCpp
+#' @useDynLib ggdist, .registration = TRUE
+## usethis namespace: end
+NULL
+
 #' Visualizations of Distributions and Uncertainty
 #'
 #' @docType package

src/.gitignore

@@ -0,0 +1,3 @@
+*.o
+*.so
+*.dll

src/RcppExports.cpp

@@ -0,0 +1,34 @@
+// Generated by using Rcpp::compileAttributes() -> do not edit by hand
+// Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
+
+#include <Rcpp.h>
+
+using namespace Rcpp;
+
+#ifdef RCPP_USE_GLOBAL_ROSTREAM
+Rcpp::Rostream<true>&  Rcpp::Rcout = Rcpp::Rcpp_cout_get();
+Rcpp::Rostream<false>& Rcpp::Rcerr = Rcpp::Rcpp_cerr_get();
+#endif
+
+// wilkinson_bin_to_right_
+IntegerVector wilkinson_bin_to_right_(const NumericVector& x, double width);
+RcppExport SEXP _ggdist_wilkinson_bin_to_right_(SEXP xSEXP, SEXP widthSEXP) {
+BEGIN_RCPP
+    Rcpp::RObject rcpp_result_gen;
+    Rcpp::RNGScope rcpp_rngScope_gen;
+    Rcpp::traits::input_parameter< const NumericVector& >::type x(xSEXP);
+    Rcpp::traits::input_parameter< double >::type width(widthSEXP);
+    rcpp_result_gen = Rcpp::wrap(wilkinson_bin_to_right_(x, width));
+    return rcpp_result_gen;
+END_RCPP
+}
+
+static const R_CallMethodDef CallEntries[] = {
+    {"_ggdist_wilkinson_bin_to_right_", (DL_FUNC) &_ggdist_wilkinson_bin_to_right_, 2},
+    {NULL, NULL, 0}
+};
+
+RcppExport void R_init_ggdist(DllInfo *dll) {
+    R_registerRoutines(dll, NULL, CallEntries, NULL, NULL);
+    R_useDynamicSymbols(dll, FALSE);
+}

src/binning_methods.cpp

@@ -0,0 +1,24 @@
+#include <Rcpp.h>
+#include <float.h> // for DBL_EPSILON
+using namespace Rcpp;
+
+// [[Rcpp::export]]
+IntegerVector wilkinson_bin_to_right_(const NumericVector& x, double width) {
+  int n = x.size();
+  IntegerVector bins(n);
+  bins[0] = 1;
+  int current_bin = 1;
+  double first_x = x[0];
+
+  for(int i = 1; i < n; i++) {
+    // This is equivalent to x[i] - first_x >= width but it accounts for machine precision.
+    // If we instead used `>=` directly some things that should be symmetric will not be
+    if (x[i] - first_x - width >= -DBL_EPSILON) {
+      current_bin = current_bin + 1;
+      first_x = x[i];
+    }
+    bins[i] = current_bin;
+  }
+
+  return bins;
+}