Added test for EdgeTypePool. Added some javadoc to EdgeTypePool

Author: ArnauPrat

LICENSE

@@ -168,7 +168,7 @@ of having them make modifications exclusively for you, or provide you
 with facilities for running those works, provided that you comply with
 the terms of this License in conveying all material for which you do
 not control copyright.  Those thus making or running the covered works
-for you must do so exclusively on your behalf, under your direction
+for you must do so exclusively on your behalf, under your edgeType
 and control, on terms that prohibit them from making any copies of
 your copyrighted material outside their relationship with you.
 

README.md

@@ -29,7 +29,7 @@ are correlated, and with a Zipf degree distribution.
           "name" : "country",
           "type" : "String",
           "generator" : {
-            "name" : "org.dama.datasynth.generators.CDFGenerator",
+            "name" : "org.dama.datasynth.generators.CumulativeDistributionGenerator",
             "requires" : [],
             "init" : ["/dicLocations.txt",1,5," "]
           }
@@ -38,7 +38,7 @@ are correlated, and with a Zipf degree distribution.
           "name" : "name",
           "type" : "String",
           "generator" : {
-            "name" : "org.dama.datasynth.generators.CorrellationGenerator",
+            "name" : "org.dama.datasynth.generators.UniformCorrelationGenerator",
             "requires" : ["person.country"],
             "init" : ["/namesByCountry.txt"," "]
           }
@@ -49,7 +49,7 @@ are correlated, and with a Zipf degree distribution.
   "edges" : [
   {
     "name" : "friendship",
-    "direction" : "undirected",
+    "edgeType" : "undirected",
     "source" : "person",
     "target" : "person",
     "sourceCardinality" : {
@@ -77,7 +77,7 @@ One of the most important elements in DataSynth are generators. A Generator is a
 class that representing some sort of "functor" (a callable object). This object
 is used to generate the values of a given type, for instance, for a particular attribute. For instance, the
 "country" attribute values for the persons is created using the
-"org.dama.datasynth.generators.CDFGenerator" generator, wich basically generates
+"org.dama.datasynth.generators.CumulativeDistributionGenerator" generator, wich basically generates
 values from a dictionary, containing the values and a cummulative distribution
 function. A Generator implements two methods: "initialize", and "run". The
 method "initialize", is called onece and as its name indicates is used to
@@ -177,11 +177,11 @@ So basically, given a schema specified using the json, DataSynth produces Schnap
 
 ```
 person.oid = genids(10000);
-f = init('org.dama.datasynth.generators.CDFGenerator', '/dicLocations.txt', 1, 5,' '); 
+f = init('org.dama.datasynth.generators.CumulativeDistributionGenerator', '/dicLocations.txt', 1, 5,' ');
 rparams = union(person.oid);
 person.country = map(f,rparams);
 
-f = init('org.dama.datasynth.generators.CorrellationGenerator','/namesByCountry.txt', ' '); 
+f = init('org.dama.datasynth.generators.UniformCorrelationGenerator','/namesByCountry.txt', ' ');
 rparams = union(person.oid, person.country);
 person.name = map(f,rparams);
 
@@ -240,10 +240,10 @@ The Signature specifies that this solver is used to solve vertices of type Attri
 signature : {
     @A = Edge;
     @A->source.name == @A->target.name;
-    @A.direction == 'undirected';
+    @A.edgeType == 'undirected';
 }
 ```
-This signature states that given a vertex in the dependency graph of type Edge, and this vertex source name equals vertex target name, and vertex's attribute direction equals 'undirected', then this solver can generate code to generate that vertex. The syntax "->X" specifies navigating an edge of name X (i.e. source, target, whatever) on the dependency graph, and ".Y"  specifies the value of the property with name "Y". For instance, @A->source.name means that given the Edge vertex A, we go to the source attribute (person.oid in this case) and retrieve its name. 
+This signature states that given a vertex in the dependency graph of type Edge, and this vertex source name equals vertex target name, and vertex's attribute edgeType equals 'undirected', then this solver can generate code to generate that vertex. The syntax "->X" specifies navigating an edge of name X (i.e. source, target, whatever) on the dependency graph, and ".Y"  specifies the value of the property with name "Y". For instance, @A->source.name means that given the Edge vertex A, we go to the source attribute (person.oid in this case) and retrieve its name.
 
 The same syntax is used for the place holders within the snippet. The Solver instantiator just generates code by substitution the proper values obtained from the dependency graph. 
 

src/main/java/org/dama/datasynth/test/matching/EdgeTypePool.java

@@ -5,9 +5,11 @@
 /**
  * Created by aprat on 3/03/17.
  */
-public class EdgeTypePool<XType extends Comparable<XType>, YType extends Comparable<YType>> {
+public class EdgeTypePool< XType extends Comparable<XType>,
+                           YType extends Comparable<YType>> {
 
-    public class Entry {
+    static public class Entry< XType extends Comparable<XType>,
+                               YType extends Comparable<YType>> {
 
         private XType xvalue;
         private YType yvalue;
@@ -35,9 +37,10 @@ public void setYvalue(YType yvalue) {
     }
 
     // fields
-    private LinkedList<Entry> entries                   = new LinkedList<Entry>();
+    private LinkedList<Entry> entries = new LinkedList<Entry>();
 
     public EdgeTypePool(JointDistribution<XType, YType> distribution, long numEdges, long seed) {
+
         for(JointDistribution.Entry<XType,YType> entry : distribution.getEntries()) {
             long numToInsert = (long)(entry.getProbability()*numEdges);
             for(long i = 0; i < numToInsert; i+=1) {
@@ -49,10 +52,19 @@ public EdgeTypePool(JointDistribution<XType, YType> distribution, long numEdges,
         Collections.shuffle(entries,random);
     }
 
+    /**
+     * Picks a random edge from the pool
+     * @return A randomly choosen edge. null if no remaining edges.
+     */
     public Entry pickRandomEdge() {
         return entries.removeFirst();
     };
 
+    /**
+     * Picks a random edge from the pool whose x value is the given one
+     * @param xvalue The given x value
+     * @return A random edge whose x value is the given one. null if such edge does not exist.
+     */
     public Entry pickRandomEdgeX(XType xvalue) {
         ListIterator<Entry> iterator = entries.listIterator();
         while(iterator.hasNext()) {
@@ -65,6 +77,11 @@ public Entry pickRandomEdgeX(XType xvalue) {
         return null;
     };
 
+    /**
+     * Picks a random edge from the pool whose y value is the given one
+     * @param yvalue The given y value
+     * @return A random edge whose y value is the given one. null if such edge does not exist.
+     */
     public Entry pickRandomEdgeY(YType yvalue) {
         ListIterator<Entry> iterator = entries.listIterator();
         while(iterator.hasNext()) {
@@ -77,6 +94,11 @@ public Entry pickRandomEdgeY(YType yvalue) {
         return null;
     };
 
+    /**
+     * Removes a random edge from the pool whose x and y values are the given ones
+     * @param xvalue The given x value
+     * @param yvalue The given y value
+     */
     public void removeEdge(XType xvalue, YType yvalue) {
         ListIterator<Entry> iterator = entries.listIterator();
         while(iterator.hasNext()) {

src/main/java/org/dama/datasynth/test/matching/JointDistribution.java

@@ -8,9 +8,11 @@
 /**
  * Created by aprat on 2/03/17.
  */
-public class JointDistribution<XType extends Comparable<XType>, YType extends Comparable<YType>> {
+public class JointDistribution< XType extends Comparable<XType>,
+                                YType extends Comparable<YType>> {
 
-    public static class Entry<XType extends Comparable<XType>,YType extends Comparable<YType>> {
+    public static class Entry<  XType extends Comparable<XType>,
+                                YType extends Comparable<YType>> {
         private XType    xvalue;
         private YType    yvalue;
         private double   probability;

src/main/java/org/dama/datasynth/test/matching/Matching.java

@@ -1,5 +1,6 @@
 package org.dama.datasynth.test.matching;
 
+import org.dama.datasynth.common.Types;
 import org.dama.datasynth.test.graphreader.types.Graph;
 
 import java.util.ArrayList;
@@ -9,10 +10,12 @@
  * Created by aprat on 2/03/17.
  */
 public class Matching {
-    /*public static <XType, YType> void  run(Graph graph,
+    public static <XType extends Comparable<XType>, YType extends Comparable<YType>> void  run(Graph graph,
                                            ArrayList<XType> tableX,
                                            ArrayList<YType> tableY,
                                            JointDistribution<XType,YType> distribution ) {
 
-    }*/
+        EdgeTypePool<String,String> edgeTypePool = null;
+
+    }
 }

src/test/java/org/dama/datasynth/test/matching/EdgeTypePoolTest.java

@@ -0,0 +1,59 @@
+package org.dama.datasynth.test.matching;
+
+import org.junit.Test;
+
+import java.io.ByteArrayInputStream;
+import java.io.InputStream;
+import java.util.HashMap;
+
+import static org.junit.Assert.assertTrue;
+
+/**
+ * Created by aprat on 3/03/17.
+ */
+public class EdgeTypePoolTest {
+
+    @Test
+    public void testDistribution() {
+        String str = new String( "Male Female 0.7\n" +
+                "Male Male 0.15\n" +
+                "Female Female 0.15");
+
+        InputStream stream = new ByteArrayInputStream(str.getBytes());
+        JointDistribution<String, String> distribution = new JointDistribution<String,String>();
+        distribution.load(stream," ", (String s) -> s, (String s) -> s );
+
+        EdgeTypePool<String, String> edgeTypePool = new EdgeTypePool(distribution,10000L,123456789L);
+
+        HashMap<Integer, Integer> counts = new HashMap<Integer,Integer>();
+        counts.put(0,0);
+        counts.put(1,0);
+        counts.put(2,0);
+
+        int numberObserved = 0;
+        for(int i = 0; i < 10000; i+=1, numberObserved+=1) {
+            EdgeTypePool.Entry<String,String> entry = edgeTypePool.pickRandomEdge();
+            if(entry != null) {
+                if (entry.getXvalue().compareTo("Male") == 0 && entry.getYvalue().compareTo("Female") == 0) {
+                    counts.put(0, counts.get(0) + 1);
+                    continue;
+                }
+
+                if (entry.getXvalue().compareTo("Male") == 0 && entry.getYvalue().compareTo("Male") == 0) {
+                    counts.put(1, counts.get(1) + 1);
+                    continue;
+                }
+
+                if (entry.getXvalue().compareTo("Female") == 0 && entry.getYvalue().compareTo("Female") == 0) {
+                    counts.put(2, counts.get(2) + 1);
+                    continue;
+                }
+            }
+        }
+
+        assertTrue(Math.abs(numberObserved - (double) 10000 ) <= 3);
+        assertTrue(Math.abs(counts.get(0)/(double)10000 - 0.7) < 0.001);
+        assertTrue(Math.abs(counts.get(1)/(double)10000 - 0.15) < 0.001);
+        assertTrue(Math.abs(counts.get(2)/(double)10000 - 0.15) < 0.001);
+    }
+}