fixing a bug where reads in overlapping interval based locus traversals could get assigned to only one of two the regions

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@1407 348d0f76-0448-11de-a6fe-93d51630548a

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

@@ -40,8 +40,12 @@ public class IntervalShard implements Shard {
 
     /** a collection of genomic locations to interate over */
     private GenomeLoc mSet;
+    private Shard.ShardType mType = Shard.ShardType.LOCUS_INTERVAL;
 
-    IntervalShard(GenomeLoc myLocation) {
+    IntervalShard(GenomeLoc myLocation, Shard.ShardType intervalType) {
+        if (intervalType != Shard.ShardType.LOCUS_INTERVAL && intervalType != Shard.ShardType.READ_INTERVAL)
+            throw new IllegalArgumentException("The specified interval type must be either LOCUS_INTERVAL or READ_INTERVAL");
+        mType = intervalType;
         mSet = myLocation.clone();
     }
 
@@ -56,6 +60,6 @@ public class IntervalShard implements Shard {
      * @return READ, indicating the shard type
      */
     public Shard.ShardType getShardType() {
-        return ShardType.INTERVAL;
+        return mType;
     }
 }

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

@@ -49,7 +49,7 @@ public class IntervalShardStrategy implements ShardStrategy {
 
     /** their prefered size of the shard, we can modify this based on what we see in the shards */
     private long size;
-
+    private Shard.ShardType type;
     /**
      * change the recommended shard size for the next shard we generate.  The code will do it's
      * best to respect this value, but there are no guarantees.
@@ -66,10 +66,11 @@ public class IntervalShardStrategy implements ShardStrategy {
      * @param size
      * @param locations
      */
-    IntervalShardStrategy( long size, GenomeLocSortedSet locations ) {
+    IntervalShardStrategy( long size, GenomeLocSortedSet locations, Shard.ShardType shardType ) {
         if (locations == null || locations.isEmpty()) {
             throw new StingException("IntervalShardStrategy: genomic regions list is empty.");
         }
+        type = shardType;
         this.regions = locations.clone();
         this.size = size;
     }
@@ -110,7 +111,7 @@ public class IntervalShardStrategy implements ShardStrategy {
         GenomeLoc loc = regions.iterator().next();
 
         regions.removeRegion(loc);
-        return new IntervalShard(loc);
+        return new IntervalShard(loc,type);
 
     }
 

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

@@ -173,11 +173,11 @@ public abstract class LocusShardStrategy implements ShardStrategy {
 
         if (loc.getStop() - loc.getStart() <= proposedSize) {
             intervals.removeRegion(loc);
-            return new IntervalShard(loc);
+            return new IntervalShard(loc,Shard.ShardType.LOCUS_INTERVAL);
         } else {
             GenomeLoc subLoc = GenomeLocParser.createGenomeLoc(loc.getContigIndex(), loc.getStart(), loc.getStart() + proposedSize - 1);
             intervals.removeRegion(subLoc);
-            return new IntervalShard(subLoc);
+            return new IntervalShard(subLoc,Shard.ShardType.LOCUS_INTERVAL);
         }
     }
 

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

@@ -30,7 +30,7 @@ import java.io.Serializable;
  */
 public interface Shard extends Serializable {
     enum ShardType {
-        READ, LOCUS, INTERVAL
+        READ, LOCUS, READ_INTERVAL, LOCUS_INTERVAL
     }
 
     /** @return the genome location represented by this shard */

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

@@ -106,8 +106,9 @@ public class ShardStrategyFactory {
             case EXPONENTIAL:
                 return new ExpGrowthLocusShardStrategy(dic, startingSize, lst, limitDataCount);
             case INTERVAL:
+                return new IntervalShardStrategy(startingSize, lst, Shard.ShardType.LOCUS_INTERVAL);
             case READS:
-                return new IntervalShardStrategy(startingSize, lst);
+                return new IntervalShardStrategy(startingSize, lst, Shard.ShardType.READ_INTERVAL);
             default:
                 throw new StingException("Strategy: " + strat + " isn't implemented");
         }

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

@@ -76,7 +76,7 @@ public class SAMDataSource implements SimpleDataSource {
     /**
      * A histogram of exactly what reads were removed from the input stream and why.
      */
-    private SAMReadViolationHistogram violations = new SAMReadViolationHistogram();    
+    private SAMReadViolationHistogram violations = new SAMReadViolationHistogram();
 
     // A pool of SAM iterators.
     private SAMResourcePool resourcePool = null;
@@ -120,24 +120,24 @@ public class SAMDataSource implements SimpleDataSource {
         return resourcePool.getHeader();
     }
 
-    
+
     /**
      * Returns Reads data structure containing information about the reads data sources placed in this pool as well as
      * information about how they are downsampled, sorted, and filtered
      * @return
      */
     public Reads getReadsInfo() { return reads; }
-    
+
-    /** 
+    /**
      * Returns header merger: a class that keeps the mapping between original read groups and read groups
      * of the merged stream; merger also provides access to the individual file readers (and hence headers
-     * prior to the merging too) maintained by the system. 
+     * prior to the merging too) maintained by the system.
      * @return
      */
     public SamFileHeaderMerger getHeaderMerger() { return resourcePool.getHeaderMerger(); }
 
     /**
-     * 
+     *
      * @param shard the shard to get data for
      *
      * @return an iterator for that region
@@ -162,7 +162,8 @@ public class SAMDataSource implements SimpleDataSource {
                     reads.getDownsamplingFraction(),
                     reads.getSafetyChecking(),
                     reads.getSupplementalFilters());
-        } else if (shard.getShardType() == Shard.ShardType.INTERVAL) {
+        } else if ((shard.getShardType() == Shard.ShardType.LOCUS_INTERVAL) ||
+                   (shard.getShardType() == Shard.ShardType.READ_INTERVAL)) {
             iterator = seekLocus(shard.getGenomeLoc());
             iterator = applyDecoratingIterators(false,
                     iterator,
@@ -170,8 +171,8 @@ public class SAMDataSource implements SimpleDataSource {
                     reads.getSafetyChecking(),
                     reads.getSupplementalFilters());
 
-            // add the new overlapping detection iterator, if we have a last interval
+            // add the new overlapping detection iterator, if we have a last interval and we're a read based shard
-            if (mLastInterval != null && queryOverlapping) iterator = new IntervalOverlapIterator(iterator,mLastInterval,false);
+            if (mLastInterval != null && shard.getShardType() == Shard.ShardType.READ_INTERVAL ) iterator = new IntervalOverlapIterator(iterator,mLastInterval,false);
             mLastInterval = shard.getGenomeLoc();
         } else {
 
@@ -258,7 +259,7 @@ public class SAMDataSource implements SimpleDataSource {
      */
     void setResourcePool( SAMResourcePool resourcePool ) {
         this.resourcePool = resourcePool;
-    }    
+    }
 
     /**
      * Retrieve unmapped reads.