Got back most of the performance lost when I fixed the dropped reads problem.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2835 348d0f76-0448-11de-a6fe-93d51630548a

java/src/net/sf/samtools/BAMFileReader2.java

@@ -547,7 +547,7 @@ class BAMFileReader2
                 mRegionEnd = (end <= 0) ? Integer.MAX_VALUE : end;
             }
             mQueryType = queryType;
-            advance();
+            nextRead = advance();
         }
 
         /**
@@ -565,7 +565,7 @@ class BAMFileReader2
             if(!hasNext())
                 throw new NoSuchElementException("BAMQueryFilteringIterator: no next element available");
             final SAMRecord currentRead = nextRead;
-            advance();
+            nextRead = advance();
             return currentRead;
         }
 

java/src/org/broadinstitute/sting/gatk/datasources/shards/BAMFormatAwareShard.java

@@ -6,6 +6,8 @@ import net.sf.samtools.SAMFileReader2;
 import java.util.List;
 import java.util.Map;
 
+import org.broadinstitute.sting.utils.GenomeLoc;
+
 /**
  * A common interface for shards that natively understand the BAM format.
  *
@@ -17,5 +19,13 @@ public interface BAMFormatAwareShard extends Shard {
      * Get the list of chunks delimiting this shard.
      * @return a list of chunks that contain data for this shard.
      */
-    public Map<SAMFileReader2,List<Chunk>> getChunks();    
+    public Map<SAMFileReader2,List<Chunk>> getChunks();
+
+    /**
+     * Get the bounds of the current shard.  Current bounds
+     * will be the unfiltered extents of the current shard, from
+     * the start of the first interval to the end of the last interval.
+     * @return The bounds of the shard.
+     */
+    public GenomeLoc getBounds();
 }

java/src/org/broadinstitute/sting/gatk/datasources/shards/BlockDelimitedReadShard.java

@@ -7,6 +7,8 @@ import java.util.List;
 import java.util.Map;
 import java.util.Collections;
 
+import org.broadinstitute.sting.utils.GenomeLoc;
+
 /**
  * Expresses a shard of read data in block format.
  *
@@ -37,6 +39,16 @@ public class BlockDelimitedReadShard extends ReadShard implements BAMFormatAware
         return Collections.singletonMap(reader,chunks);
     }
 
+    /**
+     * Get the bounds of the current shard.  Current bounds
+     * will be the unfiltered extents of the current shard, from
+     * the start of the first interval to the end of the last interval.
+     * @return The bounds of the shard.
+     */
+    public GenomeLoc getBounds() {
+        return null;
+    }
+
     /**
      * String representation of this shard.
      * @return A string representation of the boundaries of this shard.

java/src/org/broadinstitute/sting/gatk/datasources/shards/IndexDelimitedLocusShard.java

@@ -2,6 +2,7 @@ package org.broadinstitute.sting.gatk.datasources.shards;
 
 import org.broadinstitute.sting.utils.GenomeLoc;
 import org.broadinstitute.sting.utils.StingException;
+import org.broadinstitute.sting.utils.GenomeLocParser;
 import net.sf.samtools.Chunk;
 import net.sf.samtools.SAMFileReader2;
 
@@ -71,6 +72,27 @@ public class IndexDelimitedLocusShard extends LocusShard implements BAMFormatAwa
         return chunks;
     }
 
+    /**
+     * Get the bounds of the current shard.  Current bounds
+     * will be the unfiltered extents of the current shard, from
+     * the start of the first interval to the end of the last interval.
+     * @return The bounds of the shard.
+     */    
+    public GenomeLoc getBounds() {
+        if(loci == null)
+            return null;
+
+        String contig = null;
+        long start = Long.MAX_VALUE, stop = 0;
+        for(GenomeLoc locus: loci) {
+            if(contig == null) contig = locus.getContig();
+            start = Math.min(locus.getStart(),start);
+            stop = Math.max(locus.getStop(),stop);
+        }
+
+        return GenomeLocParser.createGenomeLoc(contig,start,stop);
+    }
+
     /**
      * returns the type of shard.
      */

java/src/org/broadinstitute/sting/gatk/datasources/shards/LocusShard.java

@@ -3,40 +3,19 @@ package org.broadinstitute.sting.gatk.datasources.shards;
 import org.broadinstitute.sting.utils.GenomeLoc;
 import org.broadinstitute.sting.utils.Utils;
 
-import java.util.Collections;
 import java.util.List;
 
 /**
- *
- * User: aaron
- * Date: Apr 7, 2009
- * Time: 1:19:49 PM
- *
- * The Broad Institute
- * SOFTWARE COPYRIGHT NOTICE AGREEMENT 
- * This software and its documentation are copyright 2009 by the
- * Broad Institute/Massachusetts Institute of Technology. All rights are reserved.
- *
- * This software is supplied without any warranty or guaranteed support whatsoever. Neither
- * the Broad Institute nor MIT can be responsible for its use, misuse, or functionality.
- *
- */
-
-
-/**
- * @author aaron
- * @version 1.0
- * @date Apr 7, 2009
- * <p/>
- * Class Shard
- * <p/>
  * This is the base class for locus shards.  Right now it does little more then
  * wrap GenomeLoc (actually nothing more), but it's good to have the class
  * in place so it's easier to change guts later.
+ * @author aaron
+ * @version 1.0
+ * @date Apr 7, 2009
  */
 public class LocusShard implements Shard {
     // currently our location
-    final List<GenomeLoc> loci;
+    protected final List<GenomeLoc> loci;
 
     public LocusShard(List<GenomeLoc> loci) {
         this.loci = loci;

java/src/org/broadinstitute/sting/gatk/datasources/simpleDataSources/BlockDrivenSAMDataSource.java

@@ -154,8 +154,8 @@ public class BlockDrivenSAMDataSource extends SAMDataSource {
         Map<SAMFileReader,CloseableIterator<SAMRecord>> readerToIteratorMap = new HashMap<SAMFileReader,CloseableIterator<SAMRecord>>();
         for(Map.Entry<SAMFileReader2,List<Chunk>> chunksByReader: bamAwareShard.getChunks().entrySet()) {
             SAMFileReader2 reader = chunksByReader.getKey();
-            List<Chunk> chunks = chunksByReader.getValue();
-            readerToIteratorMap.put(reader,reader.iterator(chunks));
+            GenomeLoc bounds = bamAwareShard.getBounds();
+            readerToIteratorMap.put(reader,reader.queryOverlapping(bounds.getContig(),(int)bounds.getStart(),(int)bounds.getStop()));
         }
 
         // Set up merging and filtering to dynamically merge together multiple BAMs and filter out records not in the shard set.