Small enhancements to benchmarks and EmojiUtils

felldo · Nov 27, 2023 · aa343f4 · aa343f4
1 parent b77573d
commit aa343f4
Show file tree

Hide file tree

Showing 3 changed files with 68 additions and 3 deletions.
diff --git a/lib/src/jmh/java/benchmark/EmojiManagerBenchmark.java b/lib/src/jmh/java/benchmark/EmojiManagerBenchmark.java
@@ -3,6 +3,7 @@
 import net.fellbaum.jemoji.Emoji;
 import net.fellbaum.jemoji.EmojiManager;
 import net.fellbaum.jemoji.EmojiManagerTest;
+import net.fellbaum.jemoji.IndexedEmoji;
 import org.openjdk.jmh.annotations.Benchmark;
 import org.openjdk.jmh.annotations.OutputTimeUnit;
 
@@ -66,6 +67,11 @@ public List<Emoji> extractEmojisInOrder() {
         return EmojiManager.extractEmojisInOrder(TEXT);
     }
 
+    @Benchmark
+    public List<IndexedEmoji> extractEmojisInOrderWithIndex() {
+        return EmojiManager.extractEmojisInOrderWithIndex(TEXT);
+    }
+
     @Benchmark
     public List<Emoji> extractEmojisInOrderOnlyEmojisLengthDescending() {
         return EmojiManager.extractEmojisInOrder(EmojiManagerTest.ALL_EMOJIS_STRING);

diff --git a/lib/src/jmh/java/benchmark/excluded/CodePointBenchmark.java b/lib/src/jmh/java/benchmark/excluded/CodePointBenchmark.java
@@ -1,15 +1,38 @@
 package benchmark.excluded;
 
+import benchmark.EmojiManagerBenchmark;
 import org.openjdk.jmh.annotations.Benchmark;
+import org.openjdk.jmh.annotations.OutputTimeUnit;
 import org.openjdk.jmh.annotations.Param;
 import org.openjdk.jmh.annotations.Scope;
 import org.openjdk.jmh.annotations.State;
 
+import java.io.BufferedReader;
+import java.io.InputStreamReader;
+import java.util.Objects;
+import java.util.concurrent.TimeUnit;
+import java.util.stream.Collectors;
+
 @State(Scope.Benchmark)
+@OutputTimeUnit(TimeUnit.NANOSECONDS)
 public class CodePointBenchmark {
+
+    private static final String TEXT = new BufferedReader(new InputStreamReader(Objects.requireNonNull(EmojiManagerBenchmark.class.getClassLoader().getResourceAsStream("ExampleTextFileWithEmojis.txt"))))
+            .lines().collect(Collectors.joining("\n"));
+
     @Param("\uD83D\uDC4D\uD83C\uDFFC")
     private String alias;
 
+    /*
+    Benchmark                                            (alias)  Mode  Cnt   Score   Error  Units
+    CodePointBenchmark.codePointArrayLength                👍🏼    avgt   10  20,336 ± 0,062  ns/op
+    CodePointBenchmark.codePointCount                      👍🏼    avgt   10   2,838 ± 0,011  ns/op
+    CodePointBenchmark.codePointStreamCount                👍🏼    avgt   10  22,974 ± 0,236  ns/op
+    CodePointBenchmark.codepointsStreamToArray             👍🏼    avgt   10  1707,719 ±  11,200  us/op
+    CodePointBenchmark.codepointsArrayForIndex             👍🏼    avgt   10   610,829 ± 112,007  us/op
+    CodePointBenchmark.codepointsStreamToArrayIterator     👍🏼    avgt   10  1604,416 ±   7,238  us/op
+     */
+
     @Benchmark
     public int codePointArrayLength() {
         return alias.codePoints().toArray().length;
@@ -24,4 +47,38 @@ public long codePointStreamCount() {
     public int codePointCount() {
         return alias.codePointCount(0, alias.length());
     }
+
+    @Benchmark
+    @OutputTimeUnit(TimeUnit.MICROSECONDS)
+    public int[] codepointsStreamToArray() {
+        return TEXT.codePoints().toArray();
+    }
+
+    @Benchmark
+    @OutputTimeUnit(TimeUnit.MICROSECONDS)
+    public int[] codepointsStreamToArrayIterator() {
+        int[] codePoints = new int[getCodePointCount(TEXT)];
+        int j = 0;
+        for (final int codepoint : (Iterable<Integer>) TEXT.codePoints()::iterator) {
+            codePoints[j++] = codepoint;
+        }
+        return codePoints;
+    }
+
+    @Benchmark
+    @OutputTimeUnit(TimeUnit.MICROSECONDS)
+    public int[] codepointsArrayForIndex() {
+        int[] codePoints = new int[getCodePointCount(TEXT)];
+        int j = 0;
+        for (int i = 0; i < TEXT.length(); ) {
+            final int codePoint = TEXT.codePointAt(i);
+            codePoints[j++] = codePoint;
+            i += Character.charCount(codePoint);
+        }
+        return codePoints;
+    }
+
+    private static int getCodePointCount(final String text) {
+        return text.codePointCount(0, text.length());
+    }
 }
diff --git a/lib/src/main/java/net/fellbaum/jemoji/EmojiUtils.java b/lib/src/main/java/net/fellbaum/jemoji/EmojiUtils.java
@@ -5,7 +5,9 @@
 
 class EmojiUtils {
 
-    public static int getCodePointCount(String string) {
+    private EmojiUtils() {}
+
+    public static int getCodePointCount(final String string) {
         return string.codePointCount(0, string.length());
     }
 
@@ -29,8 +31,8 @@ public static <K, V> Optional<V> findEmojiByEitherAlias(final Map<K, V> map, fin
         return Optional.empty();
     }
 
-    public static int[] stringToCodePoints(String text) {
-        int[] codePoints = new int[getCodePointCount(text)];
+    public static int[] stringToCodePoints(final String text) {
+        final int[] codePoints = new int[getCodePointCount(text)];
         int j = 0;
         for (int i = 0; i < text.length();) {
             final int codePoint = text.codePointAt(i);