Optimizations for HC infrastructure

-- outgoingVerticesOf and incomingVerticesOf return a list not a set now, as the corresponding values must be unique since our super directed graph doesn't allow multiple edges between vertices
-- Make DeBruijnGraph, SeqGraph, SeqVertex, and DeBruijnVertex all final
-- Cache HashCode calculation in BaseVertex
-- Better docs before the pruneGraph call

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/DeBruijnAssembler.java

@@ -185,6 +185,14 @@ public class DeBruijnAssembler extends LocalAssemblyEngine {
         final SeqGraph seqGraph = deBruijnGraph.convertToSequenceGraph();
         if ( debugGraphTransformations ) seqGraph.printGraph(new File("sequenceGraph.1.dot"), pruneFactor);
 
+        // TODO -- we need to come up with a consistent pruning algorithm.  The current pruning algorithm
+        // TODO -- works well but it doesn't differentiate between an isolated chain that doesn't connect
+        // TODO -- to anything from one that's actuall has good support along the chain but just happens
+        // TODO -- to have a connection in the middle that has weight of < pruneFactor.  Ultimately
+        // TODO -- the pruning algorithm really should be an error correction algorithm that knows more
+        // TODO -- about the structure of the data and can differeniate between an infrequent path but
+        // TODO -- without evidence against it (such as occurs when a region is hard to get any reads through)
+        // TODO -- from a error with lots of weight going along another similar path
         // the very first thing we need to do is zip up the graph, or pruneGraph will be too aggressive
         seqGraph.zipLinearChains();
         if ( debugGraphTransformations ) seqGraph.printGraph(new File("sequenceGraph.2.zipped.dot"), pruneFactor);

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/graphs/BaseVertex.java

@@ -58,6 +58,7 @@ import java.util.Arrays;
  */
 public class BaseVertex {
     final byte[] sequence;
+    int cachedHashCode = -1;
 
     /**
      * Create a new sequence vertex with sequence
@@ -128,8 +129,10 @@ public class BaseVertex {
      */
     @Override
     public int hashCode() {
-        // TODO -- optimization, could compute upfront once and cached in debruijn graph
-        return Arrays.hashCode(sequence);
+        if ( cachedHashCode == -1 ) {
+            cachedHashCode = Arrays.hashCode(sequence);
+        }
+        return cachedHashCode;
     }
 
     @Override

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/graphs/DeBruijnGraph.java

@@ -59,7 +59,7 @@ import java.util.Map;
  * User: rpoplin
  * Date: 2/6/13
  */
-public class DeBruijnGraph extends BaseGraph<DeBruijnVertex> {
+public final class DeBruijnGraph extends BaseGraph<DeBruijnVertex> {
     /**
      * Create an empty DeBruijnGraph with default kmer size
      */

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/graphs/DeBruijnVertex.java

@@ -54,7 +54,7 @@ import com.google.java.contract.Ensures;
  * User: ebanks, mdepristo
  * Date: Mar 23, 2011
  */
-public class DeBruijnVertex extends BaseVertex {
+public final class DeBruijnVertex extends BaseVertex {
     private final static byte[][] sufficesAsByteArray = new byte[256][];
     static {
         for ( int i = 0; i < sufficesAsByteArray.length; i++ )

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/graphs/SeqVertex.java

@@ -70,7 +70,7 @@ import java.util.Arrays;
  * @author: depristo
  * @since 03/2013
  */
-public class SeqVertex extends BaseVertex {
+public final class SeqVertex extends BaseVertex {
     private static int idCounter = 0;
     public final int id;
 

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/graphs/SharedSequenceMerger.java

@@ -75,7 +75,7 @@ public class SharedSequenceMerger {
         if ( graph == null ) throw new IllegalArgumentException("graph cannot be null");
         if ( ! graph.vertexSet().contains(v) ) throw new IllegalArgumentException("graph doesn't contain vertex " + v);
 
-        final Set<SeqVertex> prevs = graph.incomingVerticesOf(v);
+        final List<SeqVertex> prevs = graph.incomingVerticesOf(v);
         if ( ! canMerge(graph, v, prevs) )
             return false;
         else {

protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/graphs/BaseGraphUnitTest.java

@@ -241,9 +241,11 @@ public class BaseGraphUnitTest extends BaseTest {
         graph.printGraph(tmp, 10);
     }
 
-    private void assertVertexSetEquals(final Set<SeqVertex> actual, final SeqVertex ... expected) {
+    private void assertVertexSetEquals(final Collection<SeqVertex> actual, final SeqVertex ... expected) {
+        final Set<SeqVertex> actualSet = new HashSet<SeqVertex>(actual);
+        Assert.assertEquals(actualSet.size(), actual.size(), "Duplicate elements found in vertex list");
         final Set<SeqVertex> expectedSet = expected == null ? Collections.<SeqVertex>emptySet() : new HashSet<SeqVertex>(Arrays.asList(expected));
-        Assert.assertEquals(actual, expectedSet);
+        Assert.assertEquals(actualSet, expectedSet);
     }
 
     @Test(enabled = true)

protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/graphs/SeqGraphUnitTest.java

@@ -58,7 +58,7 @@ import java.util.LinkedList;
 import java.util.List;
 
 public class SeqGraphUnitTest extends BaseTest {
-    private final static boolean DEBUG = true;
+    private final static boolean DEBUG = false;
 
     private class MergeNodesWithNoVariationTestProvider extends TestDataProvider {
         public byte[] sequence;