Merge pull request #17 from DataSketches/ExtendedUnion

Extended union

Author: leerho

src/main/java/com/yahoo/sketches/theta/DirectUnion.java

@@ -147,6 +147,37 @@ else if (serVer == 3) {
     }
     else throw new IllegalArgumentException("SerVer is unknown: "+serVer);
   }
+  
+  @Override
+  public void update(long datum) {
+    gadget_.update(datum);
+  }
+  
+  @Override
+  public void update(double datum) {
+    gadget_.update(datum);
+  }
+  
+  @Override
+  public void update(String datum) {
+    gadget_.update(datum);
+  }
+  
+  @Override
+  public void update(byte[] data) {
+    gadget_.update(data);
+  }
+  
+  @Override
+  public void update(int[] data) {
+    gadget_.update(data);
+  }
+  
+  @Override
+  public void update(long[] data) {
+    gadget_.update(data);
+  }
+  
   //must trust seed, no seedhash. No p, can't be empty, can only be compact, ordered, size > 24
   private void processVer1(Memory skMem) {
     //unionEmpty_ flag is merged with the gadget

src/main/java/com/yahoo/sketches/theta/HeapUnion.java

@@ -25,7 +25,6 @@ class HeapUnion extends SetOperation implements Union {
   private long unionThetaLong_;
   private boolean unionEmpty_;
 
-  
   /**
    * Construct a new Union SetOperation on the java heap.  Called by SetOperation.Builder.
    * 
@@ -129,6 +128,36 @@ else if (serVer == 3) {
     else throw new IllegalArgumentException("SerVer is unknown: "+serVer);
   }
 
+  @Override
+  public void update(long datum) {
+    gadget_.update(datum);
+  }
+  
+  @Override
+  public void update(double datum) {
+    gadget_.update(datum);
+  }
+  
+  @Override
+  public void update(String datum) {
+    gadget_.update(datum);
+  }
+  
+  @Override
+  public void update(byte[] data) {
+    gadget_.update(data);
+  }
+  
+  @Override
+  public void update(int[] data) {
+    gadget_.update(data);
+  }
+  
+  @Override
+  public void update(long[] data) {
+    gadget_.update(data);
+  }
+  
   //must trust seed, no seedhash. No p, can't be empty, can only be compact, ordered, size > 24
   private void processVer1(Memory skMem) {
     unionEmpty_ = false; //Empty rule: AND the empty states

src/main/java/com/yahoo/sketches/theta/Union.java

@@ -31,6 +31,57 @@ public interface Union {
    */
   void update(Memory mem);
 
+  /**
+   * Present this union with a long.
+   * 
+   * @param datum The given long datum.
+   */
+  public void update(long datum);
+  
+  /**
+   * Present this union with the given double (or float) datum. 
+   * The double will be converted to a long using Double.doubleToLongBits(datum), 
+   * which normalizes all NaN values to a single NaN representation. 
+   * Plus and minus zero will be normalized to plus zero. 
+   * The special floating-point values NaN and +/- Infinity are treated as distinct.
+   * 
+   * @param datum The given double datum.
+   */
+  public void update(double datum);
+  
+  /**
+   * Present this union with the given String. 
+   * The string is converted to a byte array using UTF8 encoding. 
+   * If the string is null or empty no update attempt is made and the method returns.
+   * 
+   * @param datum The given String.
+   */
+  public void update(String datum);
+  
+  /**
+   * Present this union with the given byte array. 
+   * If the byte array is null or empty no update attempt is made and the method returns.
+   * 
+   * @param data The given byte array.
+   */
+  public void update(byte[] data);
+  
+  /**
+   * Present this union with the given integer array. 
+   * If the integer array is null or empty no update attempt is made and the method returns.
+   * 
+   * @param data The given int array.
+   */
+  public void update(int[] data);
+  
+  /**
+   * Present this union with the given long array. 
+   * If the long array is null or empty no update attempt is made and the method returns.
+   * 
+   * @param data The given long array.
+   */
+  public void update(long[] data);
+  
   /**
    * Gets the result of this operation as a CompactSketch of the chosen form
    * @param dstOrdered 

src/main/javadoc/resources/dictionary.html

@@ -10,8 +10,8 @@ <h3><a name="accuracy">Sketch Accuracy</a></h3>
 About sketch accuracy...
 
 <h3><a name="alphaTCF">Alpha TCF</a></h3>
-The Alpha Theta Choosing Function (TCF) and the theory behind it is fully described in this 
-<a href="ThetaSketchFramework.pdf">paper</a>. 
+The Alpha Theta Choosing Function (TCF) and the theory behind it is fully described in the 
+<a href="https://github.com/DataSketches/DataSketches.github.io/blob/master/docs/ThetaSketchFramework.pdf">Theta Sketch Framework</a> paper. 
 The alpha algorithm is optimized for speed and accuracy in a real-time sketch
 building / estimating environment.
 <p>One of the properties of the Alpha Algorithm used for cache management within a sketch is that
@@ -189,8 +189,8 @@ <h3><a name="seedHash">Seed Hash</a></h3>
 or (2) that when deserializing or wrapping a sketch image that the caller has the correct seed.
 
 <h3><a name="tcf">Theta Choosing Function (TCF)</a></h3>
-The Theta Choosing Function (TCF) and the theory behind it is fully described in this
-<a href="ThetaSketchFramework.pdf">paper</a>.
+The Theta Choosing Function (TCF) and the theory behind it is fully described in the
+<a href="https://github.com/DataSketches/DataSketches.github.io/blob/master/docs/ThetaSketchFramework.pdf">Theta Sketch Framework</a> paper.
 
 <h3><a name="theta">Theta, &theta;</a></h3>
 Refers to the mathematical random variable &theta; that represents the current probability
@@ -209,7 +209,7 @@ <h3><a name="thetaSketch">Theta Sketch Framework</a></h3>
 This framework also enables the sketches to share estimation, upper and lower bounds algorithms and 
 a common serialization data structure.
 The Theta Sketch Framework, Theta Choosing Functions and the theory behind them is fully described 
-in this <a href="ThetaSketchFramework.pdf">paper</a>.
+in the <a href="https://github.com/DataSketches/DataSketches.github.io/blob/master/docs/ThetaSketchFramework.pdf">Theta Sketch Framework</a> paper.
 
 <h3><a name="updateReturnState">Update Return State</a></h3>
 Provides useful detail for sketch characterization and debugging.  It is not required that any of

src/test/java/com/yahoo/sketches/theta/DirectUnionTest.java

@@ -612,6 +612,45 @@ public void printlnTest() {
     println("PRINTING: "+this.getClass().getName());
   }
 
+  @Test
+  public void checkPrimitiveUpdates() {
+    int k = 32;
+    Memory uMem = new NativeMemory(new byte[getMaxUnionBytes(k)]);
+    Union union = SetOperation.builder().initMemory(uMem).buildUnion(k);
+    
+    union.update(1L);
+    union.update(1.5); //#1 double
+    union.update(0.0);
+    union.update(-0.0);
+    String s = null;
+    union.update(s); //null string
+    s = "";
+    union.update(s); //empty string
+    s = "String";
+    union.update(s); //#2 actual string
+    byte[] byteArr = null;
+    union.update(byteArr); //null byte[]
+    byteArr = new byte[0];
+    union.update(byteArr); //empty byte[]
+    byteArr = "Byte Array".getBytes();
+    union.update(byteArr); //#3 actual byte[]
+    int[] intArr = null;
+    union.update(intArr); //null int[]
+    intArr = new int[0];
+    union.update(intArr); //empty int[]
+    int[] intArr2 = { 1, 2, 3, 4, 5 };
+    union.update(intArr2); //#4 actual int[]
+    long[] longArr = null;
+    union.update(longArr); //null long[]
+    longArr = new long[0];
+    union.update(longArr); //empty long[]
+    long[] longArr2 = { 6, 7, 8, 9 };
+    union.update(longArr2); //#5 actual long[]
+    CompactSketch comp = union.getResult();
+    double est = comp.getEstimate();
+    assertEquals(est, 7.0, 0.0);
+  }
+  
   /**
    * @param s value to print
    */

src/test/java/com/yahoo/sketches/theta/HeapUnionTest.java

@@ -491,6 +491,44 @@ public void checkGetResult() {
     assertEquals(csk.getCurrentBytes(true), 8);
   }
 
+  @Test
+  public void checkPrimitiveUpdates() {
+    int k = 32;
+    Union union = Sketches.setOperationBuilder().buildUnion(k);
+    
+    union.update(1L);
+    union.update(1.5); //#1 double
+    union.update(0.0);
+    union.update(-0.0);
+    String s = null;
+    union.update(s); //null string
+    s = "";
+    union.update(s); //empty string
+    s = "String";
+    union.update(s); //#2 actual string
+    byte[] byteArr = null;
+    union.update(byteArr); //null byte[]
+    byteArr = new byte[0];
+    union.update(byteArr); //empty byte[]
+    byteArr = "Byte Array".getBytes();
+    union.update(byteArr); //#3 actual byte[]
+    int[] intArr = null;
+    union.update(intArr); //null int[]
+    intArr = new int[0];
+    union.update(intArr); //empty int[]
+    int[] intArr2 = { 1, 2, 3, 4, 5 };
+    union.update(intArr2); //#4 actual int[]
+    long[] longArr = null;
+    union.update(longArr); //null long[]
+    longArr = new long[0];
+    union.update(longArr); //empty long[]
+    long[] longArr2 = { 6, 7, 8, 9 };
+    union.update(longArr2); //#5 actual long[]
+    CompactSketch comp = union.getResult();
+    double est = comp.getEstimate();
+    assertEquals(est, 7.0, 0.0);
+  }
+  
   //used by DirectUnionTest as well
   public static void testAllCompactForms(Union union, double expected, double toll) {
     double compEst1, compEst2;