zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Removing the old reads traversal. 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@842 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
aaron 2009-05-27 18:36:11 +00:00
parent c7b032cc88
commit 5c6163ecbf
1 changed files with 0 additions and 134 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

134
java/src/org/broadinstitute/sting/gatk/traversals/TraverseByReads.java
View File

@ -1,134 +0,0 @@
package org.broadinstitute.sting.gatk.traversals;
import org.broadinstitute.sting.gatk.walkers.ReadWalker;
import org.broadinstitute.sting.gatk.walkers.LocusWalker;
import org.broadinstitute.sting.gatk.walkers.Walker;
import org.broadinstitute.sting.gatk.LocusContext;
import org.broadinstitute.sting.gatk.refdata.ReferenceOrderedData;
import org.broadinstitute.sting.gatk.refdata.ReferenceOrderedDatum;
import org.broadinstitute.sting.gatk.iterators.ReferenceIterator;
import org.broadinstitute.sting.utils.GenomeLoc;
import org.broadinstitute.sting.utils.fasta.FastaSequenceFile2;
import java.util.List;
import java.util.Arrays;
import java.util.ArrayList;
import java.util.LinkedList;
import java.io.File;
import net.sf.samtools.SAMRecord;
import net.sf.samtools.util.StringUtil;
/**
* Created by IntelliJ IDEA.
* User: mdepristo
* Date: Mar 27, 2009
* Time: 10:26:03 AM
* To change this template use File | Settings | File Templates.
*/
public class TraverseByReads extends TraversalEngine {
public TraverseByReads(List<File> reads, File ref, List<ReferenceOrderedData<? extends ReferenceOrderedDatum>> rods) {
super(reads, ref, rods);
}
@Override
public <M,T> T traverse(Walker<M,T> walker, List<GenomeLoc> locations) {
if ( walker instanceof ReadWalker ) {
Walker x = walker;
ReadWalker<?, ?> readWalker = (ReadWalker<?, ?>)x;
return (T)this.traverseByRead(readWalker, locations);
} else {
throw new IllegalArgumentException("Walker isn't a read walker!");
}
}
/**
* Traverse by read -- the key driver of linearly ordered traversal of reads. Provides a single read to
* the walker object, in coordinate order. Supports all of the
* interaction contract implied by the read walker
* sor
*
* @param walker A read walker object
* @param <M> MapType -- the result of calling map() on walker
* @param <> ReduceType -- the result of calling reduce() on the walker
* @return 0 on success
*/
public <M, T> Object traverseByRead(ReadWalker<M, T> walker, List<GenomeLoc> locations) {
samReadIter = initializeReads();
if ( refFileName != null && !locations.isEmpty() )
GenomeLoc.setupRefContigOrdering(new FastaSequenceFile2(refFileName));
if (refFileName == null && !walker.requiresOrderedReads() && verifyingSamReadIter != null) {
logger.warn(String.format("STATUS: No reference file provided and unordered reads are tolerated, enabling out of order read processing."));
if (verifyingSamReadIter != null)
verifyingSamReadIter.setCheckOrderP(false);
}
if ( samReader != null )
verifySortOrder(refFileName != null || walker.requiresOrderedReads());
// Initialize the walker
walker.initialize();
// Initialize the sum
T sum = walker.reduceInit();
List<Integer> offsets = Arrays.asList(0); // Offset of a single read is always 0
boolean done = false;
// copy the locations here in case we ever want to use the full list again later and so that we can remove efficiently
LinkedList notYetTraversedLocations = new LinkedList(locations);
while (samReadIter.hasNext() && !done) {
TraversalStatistics.nRecords++;
// get the next read
final SAMRecord read = samReadIter.next();
final List<SAMRecord> reads = Arrays.asList(read);
GenomeLoc loc = new GenomeLoc(read);
char[] refBases = null;
// Jump forward in the reference to this locus location
LocusContext locus = new LocusContext(loc, reads, offsets);
if (!loc.isUnmapped() && refIter != null) {
final ReferenceIterator refSite = refIter.seekForward(loc);
locus.setReferenceContig(refSite.getCurrentContig());
// MAQ alignments sometimes go off the end of the reference. Take this
// into account by only returning a shortened version of the reference.
int refStart = read.getAlignmentStart()-1;
int refLength = Math.min( read.getAlignmentEnd() - read.getAlignmentStart() + 1,
refSite.getCurrentContig().length() - read.getAlignmentStart() );
refBases = StringUtil.bytesToString( refSite.getCurrentContig().getBases(),refStart,refLength ).toCharArray();
}
GenomeLoc.removePastLocs(loc, notYetTraversedLocations);
if (GenomeLoc.overlapswithSortedLocsP(loc, notYetTraversedLocations, locations.isEmpty())) {
//
// execute the walker contact
//
final boolean keepMeP = walker.filter(refBases, read);
if (keepMeP) {
M x = walker.map(refBases, read);
sum = walker.reduce(x, sum);
}
if (this.maxReads > 0 && TraversalStatistics.nRecords > this.maxReads) {
logger.warn(String.format(("Maximum number of reads encountered, terminating traversal " + TraversalStatistics.nRecords)));
done = true;
}
}
printProgress("reads", loc);
if (GenomeLoc.pastFinalLocation(loc, locations))
done = true;
//System.out.printf("Done? %b%n", done);
}
//printOnTraversalDone("reads", sum);
walker.onTraversalDone(sum);
return sum;
}
}