From e0681763ef61a974b82a52d38ebdcbe758092215 Mon Sep 17 00:00:00 2001
From: Joel Therrien <joel@joeltherrien.ca>
Date: Mon, 10 Sep 2018 12:31:35 -0700
Subject: [PATCH] Add convenience methods to improve R interface performance

---
 .../randomforest/tree/Forest.java             | 17 ++++++++--
 .../randomforest/utils/RUtils.java            | 34 +++++++++++++++++++
 .../randomforest/utils/Utils.java             |  2 +-
 .../competingrisk/TestCompetingRisk.java      |  4 +--
 4 files changed, 52 insertions(+), 5 deletions(-)
 create mode 100644 src/main/java/ca/joeltherrien/randomforest/utils/RUtils.java

diff --git a/src/main/java/ca/joeltherrien/randomforest/tree/Forest.java b/src/main/java/ca/joeltherrien/randomforest/tree/Forest.java
index 5835e89..9d6913f 100644
--- a/src/main/java/ca/joeltherrien/randomforest/tree/Forest.java
+++ b/src/main/java/ca/joeltherrien/randomforest/tree/Forest.java
@@ -5,6 +5,7 @@ import lombok.Builder;
 
 import java.util.Collection;
 import java.util.Collections;
+import java.util.List;
 import java.util.stream.Collectors;
 
 @Builder
@@ -16,17 +17,29 @@ public class Forest<O, FO> { // O = output of trees, FO = forest output. In prac
     public FO evaluate(CovariateRow row){
 
         return treeResponseCombiner.combine(
-                trees.parallelStream()
+                trees.stream()
                 .map(node -> node.evaluate(row))
                 .collect(Collectors.toList())
         );
 
     }
 
+    /**
+     * Used primarily in the R package interface to avoid R loops; and for easier parallelization.
+     *
+     * @param rowList List of CovariateRows to evaluate
+     * @return A List of predictions.
+     */
+    public List<FO> evaluate(List<CovariateRow> rowList){
+        return rowList.parallelStream()
+                .map(this::evaluate)
+                .collect(Collectors.toList());
+    }
+
     public FO evaluateOOB(CovariateRow row){
 
         return treeResponseCombiner.combine(
-          trees.parallelStream()
+          trees.stream()
           .filter(tree -> !tree.idInBootstrapSample(row.getId()))
           .map(node -> node.evaluate(row))
           .collect(Collectors.toList())
diff --git a/src/main/java/ca/joeltherrien/randomforest/utils/RUtils.java b/src/main/java/ca/joeltherrien/randomforest/utils/RUtils.java
new file mode 100644
index 0000000..419f20f
--- /dev/null
+++ b/src/main/java/ca/joeltherrien/randomforest/utils/RUtils.java
@@ -0,0 +1,34 @@
+package ca.joeltherrien.randomforest.utils;
+
+import java.util.List;
+
+/**
+ * These static methods are designed to make the R interface more performant; and to avoid using R loops.
+ *
+ */
+public final class RUtils {
+
+    public static double[] extractTimes(final MathFunction function){
+        final List<Point> pointList = function.getPoints();
+        final double[] times = new double[pointList.size()];
+
+        for(int i=0; i<pointList.size(); i++){
+            times[i] = pointList.get(i).getTime();
+        }
+
+        return times;
+    }
+
+    public static double[] extractY(final MathFunction function){
+        final List<Point> pointList = function.getPoints();
+        final double[] times = new double[pointList.size()];
+
+        for(int i=0; i<pointList.size(); i++){
+            times[i] = pointList.get(i).getY();
+        }
+
+        return times;
+    }
+
+
+}
diff --git a/src/main/java/ca/joeltherrien/randomforest/utils/Utils.java b/src/main/java/ca/joeltherrien/randomforest/utils/Utils.java
index 2342851..247ec3f 100644
--- a/src/main/java/ca/joeltherrien/randomforest/utils/Utils.java
+++ b/src/main/java/ca/joeltherrien/randomforest/utils/Utils.java
@@ -3,7 +3,7 @@ package ca.joeltherrien.randomforest.utils;
 import java.util.*;
 import java.util.concurrent.ThreadLocalRandom;
 
-public class Utils {
+public final class Utils {
 
     public static MathFunction estimateOneMinusECDF(final double[] times){
         final Point defaultPoint = new Point(0.0, 1.0);
diff --git a/src/test/java/ca/joeltherrien/randomforest/competingrisk/TestCompetingRisk.java b/src/test/java/ca/joeltherrien/randomforest/competingrisk/TestCompetingRisk.java
index 8bbd4ce..db6d0c4 100644
--- a/src/test/java/ca/joeltherrien/randomforest/competingrisk/TestCompetingRisk.java
+++ b/src/test/java/ca/joeltherrien/randomforest/competingrisk/TestCompetingRisk.java
@@ -325,8 +325,8 @@ public class TestCompetingRisk {
         */
 
         // Consistency results
-        closeEnough(0.395, errorRates[0], 0.01);
-        closeEnough(0.345, errorRates[1], 0.01);
+        closeEnough(0.395, errorRates[0], 0.02);
+        closeEnough(0.345, errorRates[1], 0.02);
 
         System.out.println(errorRateCalculator.calculateNaiveMortalityError(new int[]{1,2}));