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

Unified byLoci and byLociByInterval traversals. It now figures out what to do for you based on the presence of an index and set of required locations to process. 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@191 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
depristo 2009-03-25 16:01:58 +00:00
parent c18f8fbf5f
commit d457778283
2 changed files with 27 additions and 42 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
java/src/org/broadinstitute/sting/gatk/GenomeAnalysisTK.java
View File

@ -185,10 +185,7 @@ public class GenomeAnalysisTK extends CommandLineProgram {
// Try to get the walker specified
try {
LocusWalker<?, ?> walker = (LocusWalker<?, ?>) my_walker;
if ( INTERVALS_FILE == null )
engine.traverseByLoci(walker);
else
engine.traverseByLociByInterval(walker);
engine.traverseByLoci(walker);
}
catch (java.lang.ClassCastException e) {
// I guess we're a read walker LOL

64
java/src/org/broadinstitute/sting/gatk/TraversalEngine.java
View File

@ -244,6 +244,10 @@ public class TraversalEngine {
}
}
public boolean hasLocations() {
return this.locs != null;
}
/**
* Returns true iff we have a specified series of locations to process AND we are past the last
* location in the list. It means that, in a serial processing of the genome, that we are done.
@ -551,15 +555,33 @@ public class TraversalEngine {
* @return 0 on success
*/
protected <M, T> int traverseByLoci(LocusWalker<M, T> walker) {
samReadIter = initializeReads();
samReader = initializeSAMFile(readsFile);
verifySortOrder(true);
// initialize the walker object
walker.initialize();
// We aren't locus oriented
T sum = carryWalkerOverInterval(walker, samReadIter, walker.reduceInit(), null);
T sum = walker.reduceInit();
if ( samReader.hasIndex() && hasLocations() ) {
// we are doing interval-based traversals
for ( GenomeLoc interval : locs ) {
logger.debug(String.format("Processing locus %s", interval.toString()));
CloseableIterator<SAMRecord> readIter = samReader.queryOverlapping( interval.getContig(),
(int)interval.getStart(),
(int)interval.getStop() );
Iterator<SAMRecord> wrappedIter = WrapReadsIterator( readIter, false );
sum = carryWalkerOverInterval(walker, wrappedIter, sum, interval);
readIter.close();
}
}
else {
// We aren't locus oriented
samReadIter = WrapReadsIterator(getReadsIterator(samReader), true);
sum = carryWalkerOverInterval(walker, samReadIter, sum, null);
}
printOnTraversalDone("loci", sum);
walker.onTraversalDone();
@ -617,40 +639,6 @@ public class TraversalEngine {
return sum;
}
/**
* Same as the normal locus traverser, but oriented by locus, rather than implicitly
*
* @param walker A locus walker object
* @param <M> MapType -- the result of calling map() on walker
* @param <T> ReduceType -- the result of calling reduce() on the walker
* @return 0 on success
*/
protected <M, T> int traverseByLociByInterval(LocusWalker<M, T> walker) {
//verifySortOrder(true);
// initialize the walker object
walker.initialize();
// Initialize the T sum using the walker
T sum = walker.reduceInit();
samReader = initializeSAMFile(readsFile);
for ( GenomeLoc interval : locs ) {
logger.debug(String.format("Processing locus %s", interval.toString()));
CloseableIterator<SAMRecord> readIter = samReader.queryOverlapping( interval.getContig(),
(int)interval.getStart(),
(int)interval.getStop() );
Iterator<SAMRecord> wrappedIter = WrapReadsIterator( readIter, false );
sum = carryWalkerOverInterval(walker, wrappedIter, sum, interval);
readIter.close();
}
printOnTraversalDone("loci", sum);
walker.onTraversalDone();
return 0;
}
/**
* 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