Terminology 'measure' used in place of 'metric' to reduce confusion w…

…ith mathematical metric concept.

src/main/java/uk/ac/standrews/cs/utilities/JSONReader.java

@@ -558,14 +558,14 @@ private void stringToValue(final String s) {
             }
             try {
                 if (s.indexOf('.') > -1 || s.indexOf('e') > -1 || s.indexOf('E') > -1) {
-                    theDouble = Double.valueOf(s);
+                    theDouble = Double.parseDouble(s);
                     setNextSymbol(DOUBLE);
                     return;
                 }
 
-                final Long myLong = new Long(s);
-                if (myLong == myLong.intValue()) {
-                    theInteger = myLong.intValue();
+                final long myLong = Long.parseLong(s);
+                if (myLong == (int) myLong) {
+                    theInteger = (int) myLong;
                     setNextSymbol(INTEGER);
                     return;
                 }

src/main/java/uk/ac/standrews/cs/utilities/all_pairs/AllPairs.java

@@ -16,9 +16,9 @@
  */
 package uk.ac.standrews.cs.utilities.all_pairs;
 
-import uk.ac.standrews.cs.utilities.metrics.Cosine;
-import uk.ac.standrews.cs.utilities.metrics.implementation.FeatureVector;
-import uk.ac.standrews.cs.utilities.metrics.implementation.KeyFreqPair;
+import uk.ac.standrews.cs.utilities.measures.Cosine;
+import uk.ac.standrews.cs.utilities.measures.implementation.FeatureVector;
+import uk.ac.standrews.cs.utilities.measures.implementation.KeyFreqPair;
 
 import java.util.ArrayList;
 import java.util.HashMap;

src/main/java/uk/ac/standrews/cs/utilities/all_pairs/AllPairs1.java

@@ -16,8 +16,8 @@
  */
 package uk.ac.standrews.cs.utilities.all_pairs;
 
-import uk.ac.standrews.cs.utilities.metrics.implementation.FeatureVector;
-import uk.ac.standrews.cs.utilities.metrics.implementation.KeyFreqPair;
+import uk.ac.standrews.cs.utilities.measures.implementation.FeatureVector;
+import uk.ac.standrews.cs.utilities.measures.implementation.KeyFreqPair;
 
 import java.util.ArrayList;
 import java.util.HashMap;

src/main/java/uk/ac/standrews/cs/utilities/all_pairs/DocumentFeaturePair.java

@@ -16,7 +16,7 @@
  */
 package uk.ac.standrews.cs.utilities.all_pairs;
 
-import uk.ac.standrews.cs.utilities.metrics.implementation.FeatureVector;
+import uk.ac.standrews.cs.utilities.measures.implementation.FeatureVector;
 
 /**
  * Created by al on 27/09/2017.

src/main/java/uk/ac/standrews/cs/utilities/lsh/MinHash.java

@@ -16,7 +16,7 @@
  */
 package uk.ac.standrews.cs.utilities.lsh;
 
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.StringMetric;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.StringMeasure;
 
 import java.util.*;
 
@@ -77,7 +77,7 @@ private static int hashFunction(Object inputData, int seedOne, int seedTwo) {
      */
     public static Integer[] createMinHashSignature(String src, int sig_size, int shingle_size) {
 
-        Set<String> set_ngrams = StringMetric.extractNGrams(src, shingle_size);
+        Set<String> set_ngrams = StringMeasure.extractNGrams(src, shingle_size);
 
         // Create a min hash array initialized to all int max values
         Integer[] signature = new Integer[sig_size];

src/main/java/uk/ac/standrews/cs/utilities/lsh/MinHashExample1.java

@@ -16,8 +16,8 @@
  */
 package uk.ac.standrews.cs.utilities.lsh;
 
-import uk.ac.standrews.cs.utilities.metrics.Jaccard;
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.StringMetric;
+import uk.ac.standrews.cs.utilities.measures.Jaccard;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.StringMeasure;
 
 import java.util.Arrays;
 import java.util.Set;
@@ -32,8 +32,8 @@ public static void main(String[] args) {
         String input1 = "The attribute to awe and majesty.";
         String input2 = "But there's but one in all doth hold his place.";
 
-        Set<String> input1_2grams = StringMetric.extractNGrams(input1, 2);
-        Set<String> input2_2grams = StringMetric.extractNGrams(input2, 2);
+        Set<String> input1_2grams = StringMeasure.extractNGrams(input1, 2);
+        Set<String> input2_2grams = StringMeasure.extractNGrams(input2, 2);
 
         Integer[] input1_minHashSignature = MinHash.createMinHashSignature(input1, 50, 2);
         Integer[] input2_minHashSignature = MinHash.createMinHashSignature(input2, 50, 2);

src/main/java/uk/ac/standrews/cs/utilities/lsh/MinHashExample2.java

@@ -16,8 +16,8 @@
  */
 package uk.ac.standrews.cs.utilities.lsh;
 
-import uk.ac.standrews.cs.utilities.metrics.Jaccard;
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.StringMetric;
+import uk.ac.standrews.cs.utilities.measures.Jaccard;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.StringMeasure;
 
 import java.util.Arrays;
 import java.util.HashMap;
@@ -89,8 +89,8 @@ public static void showSimilarities(MinHash mh) {
         for (String sentence : ozs_words) {
             for (String key : map.keySet()) {
 
-                Set<String> sentence_2grams = StringMetric.extractNGrams(sentence, 2);
-                Set<String> key_2grams = StringMetric.extractNGrams(key, 2);
+                Set<String> sentence_2grams = StringMeasure.extractNGrams(sentence, 2);
+                Set<String> key_2grams = StringMeasure.extractNGrams(key, 2);
 
                 Integer[] sentence_minHashSignature = MinHash.createMinHashSignature(sentence, 50, 2);
                 Integer[] key_minHashSignature = MinHash.createMinHashSignature(key, 50, 2);

src/main/java/uk/ac/standrews/cs/utilities/m_tree/MTree.java

@@ -17,8 +17,8 @@
 package uk.ac.standrews.cs.utilities.m_tree;
 
 import uk.ac.standrews.cs.utilities.archive.Diagnostic;
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.DataDistance;
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.Metric;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.DataDistance;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.Measure;
 
 import java.util.ArrayList;
 import java.util.List;
@@ -47,15 +47,15 @@ public class MTree<T> {
     public Node root = null;
     int number_of_entries = 0;
     private final int max_level_size; // size of a level
-    final Metric<T> distance_wrapper;
+    final Measure<T> distance_wrapper;
 
-    public MTree(final Metric<T> d, final int max_level_size) {
+    public MTree(final Measure<T> d, final int max_level_size) {
 
         distance_wrapper = d;
         this.max_level_size = max_level_size;
     }
 
-    public MTree(final Metric<T> d) {
+    public MTree(final Measure<T> d) {
 
         this(d, DEFAULT_MAX_LEVEL_SIZE);
     }

src/main/java/uk/ac/standrews/cs/utilities/m_tree/experiments/MTreeSizeCheck.java

@@ -36,7 +36,7 @@ public class MTreeSizeCheck {
     private int increment = 5000000; // 5 million
     private int max = 30000000; // 30 million.
 
-    private EuclideanDistance distance_metric = new EuclideanDistance();
+    private EuclideanDistance distance_measure = new EuclideanDistance();
 
     private void loadTest() {
 
@@ -50,7 +50,7 @@ private void loadTest() {
      */
     private void createTree(int size) {
 
-        MTree<Point> tree = new MTree<>(distance_metric);
+        MTree<Point> tree = new MTree<>(distance_measure);
 
         long time = System.currentTimeMillis();
         System.out.println("Creating tree of size " + size);

src/main/java/uk/ac/standrews/cs/utilities/m_tree/experiments/MTreeStringDictionaryCheck.java

@@ -18,9 +18,8 @@
 
 import uk.ac.standrews.cs.utilities.FileManipulation;
 import uk.ac.standrews.cs.utilities.m_tree.MTree;
-import uk.ac.standrews.cs.utilities.metrics.Levenshtein;
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.DataDistance;
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.Metric;
+import uk.ac.standrews.cs.utilities.measures.Levenshtein;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.DataDistance;
 
 import java.io.IOException;
 import java.nio.file.Paths;

src/main/java/uk/ac/standrews/cs/utilities/m_tree/experiments/euclidean/EuclideanDistance.java

@@ -16,20 +16,25 @@
  */
 package uk.ac.standrews.cs.utilities.m_tree.experiments.euclidean;
 
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.Metric;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.Measure;
 
-public class EuclideanDistance extends Metric<Point> {
+public class EuclideanDistance extends Measure<Point> {
 
     public double calculateDistance(Point p1, Point p2) {
 
         double x_distance = p1.x - p2.x;
         double y_distance = p1.y - p2.y;
 
-        return normaliseArbitraryPositiveDistance(Math.sqrt((x_distance * x_distance) + (y_distance * y_distance)));
+        return Math.sqrt((x_distance * x_distance) + (y_distance * y_distance));
     }
 
     @Override
-    public String getMetricName() {
+    public String getMeasureName() {
         return "EuclideanDistance (2D)";
     }
+
+    @Override
+    public boolean maxDistanceIsOne() {
+        return false;
+    }
 }

src/main/java/uk/ac/standrews/cs/utilities/metrics/BagDistance.java → src/main/java/uk/ac/standrews/cs/utilities/measures/BagDistance.java

@@ -14,9 +14,9 @@
  * You should have received a copy of the GNU General Public License along with utilities. If not, see
  * <http://www.gnu.org/licenses/>.
  */
-package uk.ac.standrews.cs.utilities.metrics;
+package uk.ac.standrews.cs.utilities.measures;
 
-import uk.ac.standrews.cs.utilities.metrics.coreConcepts.StringMetric;
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.StringMeasure;
 
 import java.util.ArrayList;
 import java.util.List;
@@ -30,43 +30,40 @@
  *         "String Matching with Metric Trees Using an Approximate Distance"
  *         Ilaria Bartolini, Paolo Ciaccia and Marco Patella,
  *         in Proceedings of the 9th International Symposium on String Processing
- *         and Information Retrieval, Lisbone, Purtugal, September 2002.
+ *         and Information Retrieval, Lisbone, Portugal, September 2002.
  */
-public class BagDistance extends StringMetric {
+public class BagDistance extends StringMeasure {
 
     @Override
-    public String getMetricName() {
+    public String getMeasureName() {
         return "BagDistance";
     }
 
-    public double calculateStringDistance(String str1, String str2) {
-        return 1 - similarity(str1, str2);
-    }
-
-    public double similarity(String str1, String str2) {
+    @Override
+    public boolean maxDistanceIsOne() { return true; }
 
-        int n = str1.length();
-        int m = str2.length();
+    @Override
+    public double calculateDistance(final String x, final String y) {
 
-        List<Character> list1 = toList(str1);
-        List<Character> list2 = toList(str2);
+        final List<Character> list1 = toList(x);
+        final List<Character> list2 = toList(y);
 
         for (Character ch : list1) {
             list2.remove(ch);                 // only removes if in the list
         }
 
         // ch must be typed as Character not char, since otherwise we call List.remove(index)...
 
-        for (Character ch : toList(str2)) {   // note make a copy of list2 because we have removed characters from original above.
+        for (Character ch : toList(y)) {      // note make a copy of list2 because we have removed characters from original above.
             list1.remove(ch);                 // only removes if in the list
         }
 
-        return 1.0 - (((double) Math.max(list1.size(), list2.size())) / Math.max(n, m));
+        return (((double) Math.max(list1.size(), list2.size())) / Math.max(x.length(), y.length()));
     }
 
-    private List<Character> toList(String input) {
+    private static List<Character> toList(String input) {
 
-        List<Character> result = new ArrayList<>();
+        final List<Character> result = new ArrayList<>();
         for (char ch : input.toCharArray()) {
             result.add(ch);
         }

src/main/java/uk/ac/standrews/cs/utilities/measures/Cosine.java

@@ -0,0 +1,75 @@
+/*
+ * Copyright 2021 Systems Research Group, University of St Andrews:
+ * <https://github.com/stacs-srg>
+ *
+ * This file is part of the module utilities.
+ *
+ * utilities is free software: you can redistribute it and/or modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later
+ * version.
+ *
+ * utilities is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied
+ * warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along with utilities. If not, see
+ * <http://www.gnu.org/licenses/>.
+ */
+package uk.ac.standrews.cs.utilities.measures;
+
+import uk.ac.standrews.cs.utilities.measures.coreConcepts.StringMeasure;
+import uk.ac.standrews.cs.utilities.measures.implementation.FeatureVector;
+import uk.ac.standrews.cs.utilities.measures.implementation.QgramDistribution;
+import uk.ac.standrews.cs.utilities.measures.implementation.SparseDistribution;
+
+public class Cosine extends StringMeasure {
+
+    @Override
+    public String getMeasureName() {
+        return "Cosine";
+    }
+
+    @Override
+    public boolean maxDistanceIsOne() { return true; }
+
+    @Override
+    protected double calculateDistance(final String x, final String y) {
+
+        return distance(new SparseDistribution(topAndTail(x)), new SparseDistribution(topAndTail(y)));
+    }
+
+    public double distance(final FeatureVector x, final FeatureVector y) {
+
+        return distance(new SparseDistribution(x), new SparseDistribution(y));
+    }
+
+    private double distance(final SparseDistribution x, final SparseDistribution y) {
+
+        x.convertToProbabilityBased();
+        y.convertToProbabilityBased();
+
+        double dot_product = 0.0d;
+
+        for (QgramDistribution qgram : x) {
+
+            QgramDistribution qi = y.getEntry(qgram.key);
+
+            if (qi != null) {
+                dot_product += qgram.count * qi.count;
+            }
+        }
+
+        final double cosine_similarity = dot_product / (x.magnitude() * y.magnitude());
+        final double angular_distance = 2.0 * Math.acos(Math.min(cosine_similarity, 1d)) / Math.PI; // Truncate at 1.0 in case of rounding error.
+
+        if (Double.isNaN(angular_distance)) {
+            throw new RuntimeException("Cosine.distance returned Nan");
+        }
+
+        return angular_distance;
+    }
+
+    public static void main(String[] args) {
+
+        new Cosine().printExamples();
+    }
+}