Merge pull request #1735 from danielinteractive/1734_fix_step_mismatch

Align step() with Stan and add int_step()

NEWS.md

@@ -5,6 +5,8 @@
 * Fit extended-support Beta models via family `xbeta` 
 thanks to Ioannis Kosmidis. (#1698)
 * Add a `seed` argument to `loo_R2` thanks to Marco Colombo. (#1713)
+* Add the `int_step` R function to match the corresponding Stan 
+function. (#1734)
 
 ### Bug Fixes
 
@@ -15,6 +17,8 @@ in parallel even on Unix systems. To avoid potential speed loss for small
 models, `log_lik` will not use `option(mc.cores)` anymore.
 These changes may be reverted once the underlying causes of this
 issue have been fixed. (#1658)
+* Align the definition of the R function `step()` with the definition in Stan,
+such that `step(0) == 1` (#1734)
 
 ### Other Changes
 

R/numeric-helpers.R

@@ -65,6 +65,16 @@ log1m <- function(x) {
 }
 
 step <- function(x) {
+  if (is.logical(x)) {
+    warning2(
+      "Consider using int_step() instead of step(), ",
+      "because step(FALSE) = step(TRUE) = 1 in Stan."
+    )
+  }
+  ifelse(x >= 0, 1, 0)
+}
+
+int_step <- function(x) {
   ifelse(x > 0, 1, 0)
 }