add support for parallelism with future (#208)

* add support for parallelism with future
* use future's RNG when using doFuture

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: stacks
 Title: Tidy Model Stacking
-Version: 1.0.3.9000
+Version: 1.0.3.9001
 Authors@R: c(
     person("Simon", "Couch", , "[email protected]", role = c("aut", "cre")),
     person("Max", "Kuhn", , "[email protected]", role = "aut"),
@@ -19,8 +19,10 @@ Depends:
 Imports:
     butcher (>= 0.1.3),
     cli,
+    doFuture,
     dplyr (>= 1.1.0),
     foreach,
+    future,
     generics,
     ggplot2,
     glmnet,
@@ -57,4 +59,4 @@ Config/testthat/edition: 3
 Encoding: UTF-8
 LazyData: true
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.2.3.9000
+RoxygenNote: 7.3.1

NEWS.md

@@ -1,5 +1,7 @@
 # stacks (development version)
 
+* Introduced support for parallel processing using the [future](https://www.futureverse.org/) framework. The stacks package previously supported parallelism with foreach, and users can use either framework for now. In a future release, stacks will begin the deprecation cycle for parallelism with foreach, so we encourage users to begin migrating their code now. See [the _Parallel Processing_ section in the tune package's "Optimizations" article](https://tune.tidymodels.org/articles/extras/optimizations.html#parallel-processing) to learn more (#866).
+
 * Improved error message for unsupported model modes (#152).
 
 # stacks 1.0.3

R/fit_members.R

@@ -15,8 +15,8 @@
 #' model contains the necessary components to predict on new data.
 #' 
 #' @details 
-#' To fit members in parallel, please register a parallel backend function. 
-#' See the documentation of [foreach::foreach()] for examples.
+#' To fit members in parallel, please create a plan with the future package. 
+#' See the documentation of [future::plan()] for examples.
 #' 
 #' @template note_example_data
 #' 
@@ -117,15 +117,24 @@ fit_members <- function(model_stack, ...) {
       dplyr::full_join(metrics_dict, by = c("value" = ".config"), multiple = "all")
   }
 
-  if (foreach::getDoParWorkers() > 1) {
-    `%do_op%` <- foreach::`%dopar%`
+  if (foreach::getDoParWorkers() > 1 || future::nbrOfWorkers() > 1) {
+    `%do_op%` <- switch(
+      # note some backends can return +Inf
+      min(future::nbrOfWorkers(), 2),
+      foreach::`%dopar%`,
+      doFuture::`%dofuture%`
+    )
   } else {
     `%do_op%` <- foreach::`%do%`
   }
 
   # fit each of them
   member_fits <- 
-    foreach::foreach(mem = member_names, .inorder = FALSE) %do_op% {
+    foreach::foreach(
+      mem = member_names,
+      .inorder = FALSE,
+      .options.future = list(seed = TRUE)
+    ) %do_op% {
       asNamespace("stacks")$fit_member(
         name = mem,
         wflows = model_stack[["model_defs"]],