Merge branch 'master' into skip_tests

.github/workflows/pkgdown.yaml

@@ -41,7 +41,7 @@ jobs:
 
       - name: Deploy to GitHub pages 🚀
         if: github.event_name != 'pull_request'
-        uses: JamesIves/[email protected].1
+        uses: JamesIves/[email protected].3
         with:
           clean: false
           branch: gh-pages

DESCRIPTION

@@ -1,8 +1,8 @@
 Package: ranger
 Type: Package
 Title: A Fast Implementation of Random Forests
-Version: 0.16.2
-Date: 2024-05-16
+Version: 0.16.3
+Date: 2024-08-20
 Author: Marvin N. Wright [aut, cre], Stefan Wager [ctb], Philipp Probst [ctb]
 Maintainer: Marvin N. Wright <[email protected]>
 Description: A fast implementation of Random Forests, particularly suited for high
@@ -19,7 +19,7 @@ Suggests:
     survival,
     testthat
 Encoding: UTF-8
-RoxygenNote: 7.3.1
+RoxygenNote: 7.3.2
 URL: https://imbs-hl.github.io/ranger/,
     https://github.com/imbs-hl/ranger
 BugReports: https://github.com/imbs-hl/ranger/issues

NEWS.md

@@ -1,4 +1,7 @@
 
+# ranger 0.16.3
+* Fix a bug for always.split.variables (for some settings)
+
 # ranger 0.16.2
 * Add Poisson splitting rule for regression trees
 

cpp_version/src/version.h

@@ -1,3 +1,3 @@
 #ifndef RANGER_VERSION
-#define RANGER_VERSION "0.16.2"
+#define RANGER_VERSION "0.16.3"
 #endif

src/Forest.cpp

@@ -953,8 +953,8 @@ void Forest::setAlwaysSplitVariables(const std::vector<std::string>& always_spli
     }
   }
 
-  // Sort in reverse order for removing with erase later
-  std::sort(deterministic_varIDs.rbegin(), deterministic_varIDs.rend());
+  // Sort for removing later
+  std::sort(deterministic_varIDs.begin(), deterministic_varIDs.end());
 }
 
 void Forest::showProgress(std::string operation, size_t max_progress) {

src/utility.cpp

@@ -171,8 +171,8 @@ void drawWithoutReplacementFisherYates(std::vector<size_t>& result, std::mt19937
   std::iota(result.begin(), result.end(), 0);
 
   // Skip indices
-  for (size_t i = 0; i < skip.size(); ++i) {
-    result.erase(result.begin() + skip[i]);
+  for (auto it = skip.rbegin(); it != skip.rend(); it++) {
+    result.erase(result.begin() + *it);
   }
 
   // Draw without replacement using Fisher Yates algorithm

src/utility.h

@@ -181,7 +181,8 @@ void drawWithoutReplacementSimple(std::vector<size_t>& result, std::mt19937_64&
     size_t num_samples);
 
 /**
- * Simple algorithm for sampling without replacement (skip values), faster for smaller num_samples
+ * Simple algorithm for sampling without replacement (skip values), faster for smaller num_samples. 
+ * skip values are expected to be sorted in ascending order.
  * @param result Vector to add results to. Will not be cleaned before filling.
  * @param random_number_generator Random number generator
  * @param range_length Length of range. Interval to draw from: 0..max-1
@@ -203,6 +204,7 @@ void drawWithoutReplacementFisherYates(std::vector<size_t>& result, std::mt19937
 
 /**
  * Fisher Yates algorithm for sampling without replacement (skip values).
+ * skip values are expected to be sorted in ascending order.
  * @param result Vector to add results to. Will not be cleaned before filling.
  * @param random_number_generator Random number generator
  * @param max Length of range. Interval to draw from: 0..max-1

tests/testthat/test_poissonsplit.R

@@ -34,8 +34,8 @@ test_that("poisson splitting works on poisson distributed data", {
   n_test = (max(n_train)+1):n
   df_train = df[n_train, ]
   df_test = df[n_test, ]
-  rf_poi <- ranger(y ~ ., df_train, splitrule = "poisson", num.trees = 50, min.node.size = 50, poisson.tau = 1, seed = 12)
-  rf_mse <- ranger(y ~ ., df_train, splitrule = "variance", num.trees = 50, min.node.size = 50, seed = 13)
+  rf_poi <- ranger(y ~ ., df_train, splitrule = "poisson", num.trees = 50, min.node.size = 50, poisson.tau = 1, seed = 123)
+  rf_mse <- ranger(y ~ ., df_train, splitrule = "variance", num.trees = 50, min.node.size = 50, seed = 123)
 
   expect_is(rf_poi, "ranger")
   # deviance on test set

tests/testthat/test_regularization.R

@@ -132,7 +132,7 @@ test_that("Fewer variables used with regularization, survival extratrees", {
   rf_noreg <- ranger(Surv(time, status) ~ ., dat_surv, num.trees = 5, min.node.size = 10, mtry = 4, splitrule = "extratrees")
   rf_reg <- ranger(Surv(time, status) ~ ., dat_surv, num.trees = 5, min.node.size = 10, mtry = 4, splitrule = "extratrees", num.threads = 1, 
                    regularization.factor = .0001, regularization.usedepth = TRUE)
-  expect_lt(get_num_splitvars(rf_reg), 
+  expect_lte(get_num_splitvars(rf_reg), 
             get_num_splitvars(rf_noreg))
 })