cosine similarity returns 0 if one of the strings is shorter than k

Author: tdebatty

pom.xml

@@ -10,7 +10,7 @@
 
     <name>${project.artifactId}</name>
     <url>https://github.com/tdebatty/java-string-similarity</url>
-    <description>Implementation of various string similarity and distance algorithms: Levenshtein, Jaro, Jaro-winkler, n-Gram, Q-Gram (Jaccard index), Longest Common Subsequence edit distance,...</description>
+    <description>Implementation of various string similarity and distance algorithms: Levenshtein, Jaro-winkler, n-Gram, Q-Gram, Jaccard index, Longest Common Subsequence edit distance, cosine similarity...</description>
 
     <properties>
         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
@@ -34,8 +34,8 @@
         <connection>scm:git:[email protected]:tdebatty/java-string-similarity.git</connection>
         <developerConnection>scm:git:[email protected]:tdebatty/java-string-similarity.git</developerConnection>
         <url>[email protected]:tdebatty/java-string-similarity.git</url>
-      <tag>v0.7</tag>
-  </scm>
+        <tag>v0.7</tag>
+    </scm>
 
     <distributionManagement>
         <snapshotRepository>
@@ -120,7 +120,7 @@
                 <groupId>org.apache.maven.plugins</groupId>
                 <artifactId>maven-release-plugin</artifactId>
                 <version>2.5.1</version>
-		        <configuration>
+                <configuration>
                     <tagNameFormat>v@{project.version}</tagNameFormat>
                 </configuration>
             </plugin>
@@ -154,9 +154,9 @@
                     </instrumentation>
                 </configuration>
             </plugin>
-
         </plugins>
     </build>
+    
     <dependencies>
         <dependency>
             <groupId>junit</groupId>

src/main/java/info/debatty/java/stringsimilarity/Cosine.java

@@ -44,15 +44,22 @@ public class Cosine extends ShingleBased implements
      *
      * @param k
      */
-    public Cosine(int k) {
+    public Cosine(final int k) {
         super(k);
     }
 
+    /**
+     * 
+     */
     public Cosine() {
         super();
     }
 
     public double similarity(String s1, String s2) {
+
+        if (s1.length() < k || s2.length() < k) {
+            return 0;
+        }
         KShingling ks = new KShingling(k);
         int[] profile1 = ks.getArrayProfile(s1);
         int[] profile2 = ks.getArrayProfile(s2);
@@ -61,7 +68,7 @@ public double similarity(String s1, String s2) {
     }
 
     /**
-     * Compute the norm L2 : sqrt(Sum_i( v_i²))
+     * Compute the norm L2 : sqrt(Sum_i( v_i²)).
      *
      * @param profile
      * @return L2 norm

src/test/java/info/debatty/java/stringsimilarity/CosineTest.java

@@ -24,10 +24,6 @@
 
 package info.debatty.java.stringsimilarity;
 
-import org.junit.After;
-import org.junit.AfterClass;
-import org.junit.Before;
-import org.junit.BeforeClass;
 import org.junit.Test;
 import static org.junit.Assert.*;
 
@@ -36,34 +32,26 @@
  * @author Thibault Debatty
  */
 public class CosineTest {
-    
-    public CosineTest() {
-    }
-    
-    @BeforeClass
-    public static void setUpClass() {
-    }
-    
-    @AfterClass
-    public static void tearDownClass() {
-    }
-    
-    @Before
-    public void setUp() {
-    }
-    
-    @After
-    public void tearDown() {
-    }
 
     /**
      * Test of similarity method, of class Cosine.
      */
     @Test
-    public void testSimilarity() {
+    public final void testSimilarity() {
         System.out.println("similarity");
         Cosine instance = new Cosine();
         double result = instance.similarity("ABC", "ABCE");
         assertEquals(0.71, result, 0.01);
     }
+
+    /**
+     * If one of the strings is smaller than k, the similarity should be 0.
+     */
+    @Test
+    public final void testSmallString() {
+        System.out.println("test small string");
+        Cosine instance = new Cosine(3);
+        double result = instance.similarity("AB", "ABCE");
+        assertEquals(0.0, result, 0.00001);
+    }
 }