BAEL-980 lsh code (eugenp#2047)

* BAEL-980 lsh code * BAEL-980 rename test
saihong · Jun 18, 2017 · 0144471 · 0144471
1 parent 4ca15c9
commit 0144471
Show file tree

Hide file tree

Showing 2 changed files with 55 additions and 0 deletions.
diff --git a/libraries/pom.xml b/libraries/pom.xml
@@ -345,6 +345,12 @@
             <artifactId>opennlp-tools</artifactId>
             <version>1.8.0</version>
         </dependency>
+        <dependency>
+            <groupId>info.debatty</groupId>
+            <artifactId>java-lsh</artifactId>
+            <version>${java-lsh.version}</version>
+        </dependency>
+
 
     </dependencies>
     <properties>
@@ -371,6 +377,7 @@
         <serenity.plugin.version>1.4.0</serenity.plugin.version>
         <jUnitParams.version>1.1.0</jUnitParams.version>
         <netty.version>4.1.10.Final</netty.version>
+        <java-lsh.version>0.10</java-lsh.version>
     </properties>
 
 </project>
diff --git a/libraries/src/test/java/com/baeldung/lsh/LocalSensitiveHashingUnitTest.java b/libraries/src/test/java/com/baeldung/lsh/LocalSensitiveHashingUnitTest.java
@@ -0,0 +1,48 @@
+package com.baeldung.lsh;
+
+import info.debatty.java.lsh.LSHMinHash;
+import org.junit.Ignore;
+import org.junit.Test;
+
+import java.util.Arrays;
+
+import static org.assertj.core.api.Assertions.assertThat;
+
+
+public class LocalSensitiveHashingUnitTest {
+
+    @Ignore("for simplicity of the example number of input vectors is very low, that's why LSH may yield non deterministic results")
+    @Test()
+    public void givenNVectors_whenPerformLSH_thenShouldCalculateSameHashForSimilarVectors() {
+        //given
+        boolean[] vector1 = new boolean[]{true, true, true, true, true};
+        boolean[] vector2 = new boolean[]{false, false, false, true, false};
+        boolean[] vector3 = new boolean[]{false, false, true, true, false};
+
+        int sizeOfVectors = 5;
+        int numberOfBuckets = 10;
+        int stages = 4;
+
+        LSHMinHash lsh = new LSHMinHash(stages, numberOfBuckets, sizeOfVectors);
+
+        //when
+        int[] firstHash = lsh.hash(vector1);
+        int[] secondHash = lsh.hash(vector2);
+        int[] thirdHash = lsh.hash(vector3);
+
+        System.out.println(Arrays.toString(firstHash));
+        System.out.println(Arrays.toString(secondHash));
+        System.out.println(Arrays.toString(thirdHash));
+
+        //then
+        int lastIndexOfResult = stages - 1;
+        assertThat(firstHash[lastIndexOfResult]).isNotEqualTo(secondHash[lastIndexOfResult]);
+        assertThat(firstHash[lastIndexOfResult]).isNotEqualTo(thirdHash[lastIndexOfResult]);
+        assertThat(isCloseOrEqual(secondHash[lastIndexOfResult], thirdHash[lastIndexOfResult], numberOfBuckets)).isTrue();
+    }
+
+    private boolean isCloseOrEqual(int secondHash, int thirdHash, int numberOfBuckets) {
+        return Math.abs(secondHash - thirdHash) < numberOfBuckets / 2;
+    }
+}
+