Cleanup (#2)

* final args and logger * rm comments * add author javadoc * rename variables * rm redundant constructor * add logger * bump version
kkoutsilis · Nov 22, 2022 · 8639302 · 8639302
1 parent 02348db
commit 8639302
Show file tree

Hide file tree

Showing 19 changed files with 151 additions and 99 deletions.
diff --git a/README.md b/README.md
@@ -20,7 +20,7 @@ It requires 6 arguments in this order.
 Example:
 
 ```bash
-$ java −jar SACT−1.0−SNAPSHOT.jar ./exp1.csv ./res.csv 3 1 1 2
+$ java −jar SACT−1.1−SNAPSHOT.jar ./exp1.csv ./res.csv 3 1 1 2
 ```
 
 It will print the created clusters in the console as well as create a csv file with the clusters.

diff --git a/pom.xml b/pom.xml
@@ -5,7 +5,7 @@
 
   <groupId>com.kkoutsilis</groupId>
   <artifactId>SACT</artifactId>
-  <version>1.0-SNAPSHOT</version>
+  <version>1.1-SNAPSHOT</version>
 
   <name>SACT</name>
 

diff --git a/src/main/java/com/kkoutsilis/App.java b/src/main/java/com/kkoutsilis/App.java
@@ -1,5 +1,11 @@
 package com.kkoutsilis;
 
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+import java.util.logging.Level;
+import java.util.logging.Logger;
+
 import com.kkoutsilis.algorithms.ClusteringAlgorithm;
 import com.kkoutsilis.algorithms.HCNN;
 import com.kkoutsilis.graphs.Graph;
@@ -8,18 +14,18 @@
 import com.kkoutsilis.utilities.CsvHandler;
 import com.kkoutsilis.utilities.NearestNeighbour;
 
-import java.util.List;
-import java.util.Map;
-import java.util.Set;
-
+/**
+ * @author kkoutsilis
+ */
 public class App {
     public static void main(String[] args) throws Exception {
-        String inputFilePath = args[0];
-        String outputFilePath = args[1];
-        int k = Integer.parseInt(args[2]);
-        int fistNearestNeighbourAlgorithm = Integer.parseInt(args[3]);
-        int secondNearestNeighbourAlgorithm = Integer.parseInt(args[4]);
-        int n = Integer.parseInt(args[5]);
+        final Logger logger = Logger.getLogger(App.class.getName());
+        final String inputFilePath = args[0];
+        final String outputFilePath = args[1];
+        final int k = Integer.parseInt(args[2]);
+        final int fistNearestNeighbourAlgorithm = Integer.parseInt(args[3]);
+        final int secondNearestNeighbourAlgorithm = Integer.parseInt(args[4]);
+        final int n = Integer.parseInt(args[5]);
 
         long startTime = System.nanoTime();
 
@@ -70,9 +76,10 @@ public static void main(String[] args) throws Exception {
             s.forEach(v -> System.out.print(v.getLabel() + " "));
             System.out.println();
         }
-        System.out.println("Clustering execution time in milliseconds: " + elapsedTime / 1000000);
+        logger.log(Level.INFO,
+                () -> "Clustering execution time in milliseconds: " + elapsedTime / 1000000);
 
         SilhouetteCoefficient quality = new SilhouetteCoefficient(result, graph);
-        System.out.printf("Silhouette Coefficient: %.03f", quality.calculate());
+        logger.log(Level.INFO, () -> String.format("Silhouette Coefficient: %.03f", quality.calculate()));
     }
 }
diff --git a/src/main/java/com/kkoutsilis/algorithms/ClusteringAlgorithm.java b/src/main/java/com/kkoutsilis/algorithms/ClusteringAlgorithm.java
@@ -5,6 +5,9 @@
 import java.util.List;
 import java.util.Set;
 
+/**
+ * @author kkoutsilis
+ */
 public interface ClusteringAlgorithm {
     List<Set<Vertex>> fit() throws Exception;
 }
diff --git a/src/main/java/com/kkoutsilis/algorithms/HCNN.java b/src/main/java/com/kkoutsilis/algorithms/HCNN.java
@@ -10,20 +10,23 @@
 import com.kkoutsilis.utilities.Distance;
 
 import java.util.*;
+import java.util.logging.Level;
+import java.util.logging.Logger;
 
-
+/**
+ * @author kkoutsilis
+ */
 public class HCNN implements ClusteringAlgorithm {
     private Graph graph;
     private List<Integer> indexes;
-    private List<Vertex> G;
+    private List<Vertex> data;
     private int n;
     private Map<Vertex, Set<Vertex>> fistNnAlgorithm;
     private Map<Vertex, Set<Vertex>> secondNnAlgorithm;
+    private final Logger logger = Logger.getLogger(HCNN.class.getName());
 
-    public HCNN() {
-    }
-
-    public HCNN(Graph graph, int n, Map<Vertex, Set<Vertex>> fistNnAlgorithm, Map<Vertex, Set<Vertex>> secondNnAlgorithm) {
+    public HCNN(Graph graph, int n, Map<Vertex, Set<Vertex>> fistNnAlgorithm,
+            Map<Vertex, Set<Vertex>> secondNnAlgorithm) {
         this.graph = graph;
         this.n = n;
         this.fistNnAlgorithm = fistNnAlgorithm;
@@ -32,76 +35,71 @@ public HCNN(Graph graph, int n, Map<Vertex, Set<Vertex>> fistNnAlgorithm, Map<Ve
         for (int i = 0; i < this.graph.getVertices().size(); i++) {
             this.indexes.add(i);
         }
-        this.G = new ArrayList<>(this.graph.getVertices().keySet());
+        this.data = new ArrayList<>(this.graph.getVertices().keySet());
     }
 
     @Override
     public List<Set<Vertex>> fit() throws Exception {
+        this.logger.log(Level.INFO, "Fitting...");
         ClustersAndOutliers clustersAndOutliers = this.initializeClustering();
-        List<Set<Vertex>> C = clustersAndOutliers.getClusters();
-        Set<Vertex> D = clustersAndOutliers.getOutliers();
+        List<Set<Vertex>> clusters = clustersAndOutliers.getClusters();
+        Set<Vertex> outliers = clustersAndOutliers.getOutliers();
 
-        while (C.size() > this.n) {
-            int L = C.size();
-            // line 4
-            float tmax = 0;
-            for (int i = 0; i < L; i++) {
-                for (int j = i + 1; j < L; j++) {
-                    float similarity = sim(C.get(i), C.get(j));
-                    if (tmax < similarity) {
-                        tmax = similarity;
+        while (clusters.size() > this.n) {
+            int l = clusters.size();
+            float tMax = 0;
+            for (int i = 0; i < l; i++) {
+                for (int j = i + 1; j < l; j++) {
+                    float similarity = sim(clusters.get(i), clusters.get(j));
+                    if (tMax < similarity) {
+                        tMax = similarity;
                     }
                 }
             }
-            // line 5
-            if (tmax == 0) {
+            if (tMax == 0) {
                 break;
             }
-            // line 7
             Set<MSP> msp = new LinkedHashSet<>();
-            for (int i = 0; i < L; i++) {
-                for (int j = i + 1; j < L; j++) {
-                    float similarity = sim(C.get(i), C.get(j));
-                    if (similarity == tmax) {
+            for (int i = 0; i < l; i++) {
+                for (int j = i + 1; j < l; j++) {
+                    float similarity = sim(clusters.get(i), clusters.get(j));
+                    if (similarity == tMax) {
                         msp.add(new MSP(i, j));
                     }
                 }
             }
-            // line 8-10
-            DisjointSets DS = new DisjointSets();
-            for (Set<Vertex> vertices : C) {
-                DS.makeSet(vertices);
+            DisjointSets disjointSets = new DisjointSets();
+            for (Set<Vertex> vertices : clusters) {
+                disjointSets.makeSet(vertices);
             }
-            // line 11-12
             for (MSP msp1 : msp) {
                 int i = msp1.getI();
                 int j = msp1.getJ();
-                DS.union(C.get(i), C.get(j));
+                disjointSets.union(clusters.get(i), clusters.get(j));
             }
-            // line 13-16
-            for (int i = 0; i < L; i++) {
-                int index = DS.findSet(C.get(i));
-                Set<Vertex> repr = DS.getRepresentative(index).getHead().getData();
-                if (repr != C.get(i)) {
-                    repr.addAll(C.get(i));
-                    C.set(i, Collections.emptySet());
+            for (int i = 0; i < l; i++) {
+                int index = disjointSets.findSet(clusters.get(i));
+                Set<Vertex> repr = disjointSets.getRepresentative(index).getHead().getData();
+                if (repr != clusters.get(i)) {
+                    repr.addAll(clusters.get(i));
+                    clusters.set(i, Collections.emptySet());
                 }
             }
-            // line 17-20
-            List<Set<Vertex>> Ctemp = new ArrayList<>(C);
-            C = new ArrayList<>();
-            for (Set<Vertex> c : Ctemp) {
+            List<Set<Vertex>> tempClusters = new ArrayList<>(clusters);
+            clusters = new ArrayList<>();
+            for (Set<Vertex> c : tempClusters) {
                 if (!c.isEmpty()) {
-                    C.add(c);
+                    clusters.add(c);
                 }
             }
         }
-        C = assignOutliers(C, D);
+        clusters = assignOutliers(clusters, outliers);
 
-        return C;
+        return clusters;
     }
 
     private List<Set<Vertex>> assignOutliers(List<Set<Vertex>> clusters, Set<Vertex> outliers) {
+        this.logger.log(Level.INFO, "Assigning outliers...");
         float[] label = new float[this.indexes.size() + 1];
         Set<Vertex> loners = new HashSet<>();
         for (int i = 0; i <= this.indexes.size(); i++) {
@@ -150,7 +148,7 @@ private List<Set<Vertex>> assignOutliers(List<Set<Vertex>> clusters, Set<Vertex>
     }
 
     private ClustersAndOutliers initializeClustering() {
-        // initializing
+        this.logger.log(Level.INFO, "Initializing clusters...");
         int nOfIndexes = this.indexes.size();
         CorePair[] corePairs = new CorePair[nOfIndexes];
         int[] label = new int[nOfIndexes];
@@ -162,8 +160,7 @@ private ClustersAndOutliers initializeClustering() {
         Set<Vertex> outliers = new LinkedHashSet<>();
         float[][] structSimZ = new float[nOfIndexes][nOfIndexes];
 
-        // lines 2-7
-        for (Vertex i : this.G) {
+        for (Vertex i : this.data) {
             Set<Vertex> Q = this.fistNnAlgorithm.get(i);
             Set<Vertex> tmp = new HashSet<>(this.fistNnAlgorithm.get(i));
             for (Vertex j : tmp) {
@@ -174,12 +171,11 @@ private ClustersAndOutliers initializeClustering() {
             }
         }
 
-        // line 8
         List<PairSim> pairSims = new ArrayList<>();
         for (int i = 0; i < nOfIndexes; i++) {
             for (int j = i + 1; j < nOfIndexes; j++) {
-                Vertex vI = this.G.get(i);
-                Vertex vJ = this.G.get(j);
+                Vertex vI = this.data.get(i);
+                Vertex vJ = this.data.get(j);
                 float structSim = z(vI, vJ);
                 if (structSim > 0) {
                     pairSims.add(new PairSim(vI.getLabel(), vJ.getLabel(), structSim));
@@ -195,14 +191,14 @@ private ClustersAndOutliers initializeClustering() {
             int j = pairSim.getJ() - 1;
             if (label[i] == 0 && label[j] == 0) {
                 last += 1;
-                label[i] = label[j] = last;// Not sure about that, line 14
+                label[i] = label[j] = last;
                 corePairs[last] = new CorePair(i, j);
-                Vertex vI = this.G.get(i);
-                Vertex vJ = this.G.get(j);
+                Vertex vI = this.data.get(i);
+                Vertex vJ = this.data.get(j);
                 Set<Vertex> union = new LinkedHashSet<>(fistNnAlgorithm.get(vI));
                 union.addAll(this.fistNnAlgorithm.get(vJ));
                 for (Vertex u : union) {
-                    int ui = this.G.indexOf(u); // wrong ??
+                    int ui = this.data.indexOf(u);
                     if (label[ui] == 0) {
                         label[ui] = last;
                     }
@@ -211,14 +207,14 @@ private ClustersAndOutliers initializeClustering() {
                         int h = corePair.getI();
                         int k = corePair.getJ();
 
-                        Vertex vuI = this.G.get(ui);
+                        Vertex vuI = this.data.get(ui);
                         float max = z(vuI, vI);
                         float temp = z(vuI, vJ);
                         if (max < temp) {
                             max = temp;
                         }
-                        Vertex vH = this.G.get(h);
-                        Vertex vK = this.G.get(k);
+                        Vertex vH = this.data.get(h);
+                        Vertex vK = this.data.get(k);
 
                         float min = z(vuI, vH);
                         temp = z(vuI, vK);
@@ -234,21 +230,20 @@ private ClustersAndOutliers initializeClustering() {
             }
         }
 
-        // line 23-28
         List<Set<Vertex>> clusters = new ArrayList<>();
-        for (int i : this.indexes) {
+        for (int i = 0; i <= this.indexes.size(); i++) {
             clusters.add(new HashSet<>());
         }
         for (int i : this.indexes) {
-            Vertex vI = this.G.get(i);
+            Vertex vI = this.data.get(i);
             if (label[i] > 0) {
                 clusters.get(label[i]).add(vI);
             }
             if (label[i] == 0) {
                 outliers.add(vI);
             }
         }
-        clusters.removeIf(Set::isEmpty); // Not the best but it works for now.
+        clusters.removeIf(Set::isEmpty); // Not the best but works for now.
         return new ClustersAndOutliers(clusters, outliers);
 
     }
@@ -292,6 +287,8 @@ private int link(Set<Vertex> clusterA, Set<Vertex> clusterB) {
     }
 
     private float sim(Set<Vertex> clusterA, Set<Vertex> clusterB) {
-        return ((float) conn(clusterA, clusterB) / (float) (clusterA.size() * clusterB.size())) * ((float) link(clusterA, clusterB) / (float) clusterA.size()) * ((float) link(clusterB, clusterA) / (float) clusterB.size());
+        return ((float) conn(clusterA, clusterB) / (float) (clusterA.size() * clusterB.size()))
+                * ((float) link(clusterA, clusterB) / (float) clusterA.size())
+                * ((float) link(clusterB, clusterA) / (float) clusterB.size());
     }
 }
diff --git a/src/main/java/com/kkoutsilis/algorithms/helpers/ClustersAndOutliers.java b/src/main/java/com/kkoutsilis/algorithms/helpers/ClustersAndOutliers.java
@@ -1,11 +1,13 @@
 package com.kkoutsilis.algorithms.helpers;
 
-
 import com.kkoutsilis.graphs.Vertex;
 
 import java.util.List;
 import java.util.Set;
 
+/**
+ * @author kkoutsilis
+ */
 public class ClustersAndOutliers {
     private List<Set<Vertex>> clusters;
     private Set<Vertex> outliers;

diff --git a/src/main/java/com/kkoutsilis/algorithms/helpers/CorePair.java b/src/main/java/com/kkoutsilis/algorithms/helpers/CorePair.java
@@ -2,6 +2,9 @@
 
 import java.util.Objects;
 
+/**
+ * @author kkoutsilis
+ */
 public class CorePair {
     private int i;
     private int j;
@@ -38,8 +41,10 @@ public void setPair(int i, int j) {
 
     @Override
     public boolean equals(Object o) {
-        if (this == o) return true;
-        if (!(o instanceof CorePair)) return false;
+        if (this == o)
+            return true;
+        if (!(o instanceof CorePair))
+            return false;
         CorePair corePair = (CorePair) o;
         return getI() == corePair.getI() && getJ() == corePair.getJ();
     }

diff --git a/src/main/java/com/kkoutsilis/algorithms/helpers/MSP.java b/src/main/java/com/kkoutsilis/algorithms/helpers/MSP.java
@@ -1,5 +1,8 @@
 package com.kkoutsilis.algorithms.helpers;
 
+/**
+ * @author kkoutsilis
+ */
 public class MSP {
     private int i;
     private int j;

diff --git a/src/main/java/com/kkoutsilis/algorithms/helpers/PairSim.java b/src/main/java/com/kkoutsilis/algorithms/helpers/PairSim.java
@@ -2,6 +2,9 @@
 
 import java.util.Objects;
 
+/**
+ * @author kkoutsilis
+ */
 public class PairSim {
     private int i;
     private int j;
@@ -48,8 +51,10 @@ public void setPairSim(int i, int j, float similarity) {
 
     @Override
     public boolean equals(Object o) {
-        if (this == o) return true;
-        if (!(o instanceof PairSim)) return false;
+        if (this == o)
+            return true;
+        if (!(o instanceof PairSim))
+            return false;
         PairSim pairSim = (PairSim) o;
         return getI() == pairSim.getI() && getJ() == pairSim.getJ() && getSimilarity() == pairSim.getSimilarity();
     }