working on fasta.gz support in gatk

broadinstitute · Aug 22, 2018 · c6bf892 · c6bf892
1 parent 401d217
commit c6bf892
Show file tree

Hide file tree

Showing 26 changed files with 318 additions and 281 deletions.
diff --git a/src/main/java/org/broadinstitute/hellbender/engine/AssemblyRegion.java b/src/main/java/org/broadinstitute/hellbender/engine/AssemblyRegion.java
@@ -410,14 +410,14 @@ public List<ActivityProfileState> getSupportingStates() {
     /**
      * See #getActiveRegionReference but using the span including regions not the extended span
      */
-    public byte[] getFullReference( final IndexedFastaSequenceFile referenceReader ) {
+    public byte[] getFullReference( final ReferenceSequenceFile referenceReader ) {
         return getFullReference(referenceReader, 0);
     }
 
     /**
      * See #getActiveRegionReference but using the span including regions not the extended span
      */
-    public byte[] getFullReference( final IndexedFastaSequenceFile referenceReader, final int padding ) {
+    public byte[] getFullReference( final ReferenceSequenceFile referenceReader, final int padding ) {
         return getReference(referenceReader, padding, spanIncludingReads);
     }
 

diff --git a/...org/broadinstitute/hellbender/tools/walkers/haplotypecaller/AssemblyBasedCallerUtils.java b/...org/broadinstitute/hellbender/tools/walkers/haplotypecaller/AssemblyBasedCallerUtils.java
@@ -20,12 +20,14 @@
 import org.broadinstitute.hellbender.utils.haplotype.Haplotype;
 import org.broadinstitute.hellbender.utils.haplotype.HaplotypeBAMWriter;
 import org.broadinstitute.hellbender.utils.io.IOUtils;
-import org.broadinstitute.hellbender.utils.read.*;
+import org.broadinstitute.hellbender.utils.read.AlignmentUtils;
+import org.broadinstitute.hellbender.utils.read.GATKRead;
+import org.broadinstitute.hellbender.utils.read.ReadCoordinateComparator;
+import org.broadinstitute.hellbender.utils.read.ReadUtils;
 import org.broadinstitute.hellbender.utils.smithwaterman.SmithWatermanAligner;
 import org.broadinstitute.hellbender.utils.variant.GATKVariantContextUtils;
 
 import java.io.File;
-import java.io.FileNotFoundException;
 import java.util.*;
 import java.util.stream.Collectors;
 
@@ -147,12 +149,8 @@ public static SimpleInterval getPaddedReferenceLoc(final AssemblyRegion region,
     }
 
     public static CachingIndexedFastaSequenceFile createReferenceReader(final String reference) {
-        try {
             // fasta reference reader to supplement the edges of the reference sequence
-            return new CachingIndexedFastaSequenceFile(IOUtils.getPath(reference));
-        } catch( FileNotFoundException e ) {
-            throw new UserException.CouldNotReadInputFile(IOUtils.getPath(reference), e);
-        }
+            return CachingIndexedFastaSequenceFile.checkAndCreate(IOUtils.getPath(reference));
     }
 
     /**

diff --git a/...ain/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/HaplotypeCaller.java b/...ain/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/HaplotypeCaller.java
@@ -2,7 +2,6 @@
 
 import htsjdk.samtools.SAMSequenceDictionary;
 import htsjdk.samtools.reference.ReferenceSequenceFile;
-import htsjdk.variant.variantcontext.VariantContext;
 import htsjdk.variant.variantcontext.writer.VariantContextWriter;
 import org.broadinstitute.barclay.argparser.Argument;
 import org.broadinstitute.barclay.argparser.ArgumentCollection;
@@ -14,18 +13,14 @@
 import org.broadinstitute.hellbender.cmdline.programgroups.ShortVariantDiscoveryProgramGroup;
 import org.broadinstitute.hellbender.engine.*;
 import org.broadinstitute.hellbender.engine.filters.ReadFilter;
-import org.broadinstitute.hellbender.exceptions.UserException;
-import org.broadinstitute.hellbender.tools.walkers.annotator.*;
+import org.broadinstitute.hellbender.tools.walkers.annotator.Annotation;
+import org.broadinstitute.hellbender.tools.walkers.annotator.VariantAnnotatorEngine;
 import org.broadinstitute.hellbender.utils.fasta.CachingIndexedFastaSequenceFile;
 import org.broadinstitute.hellbender.utils.io.IOUtils;
 
-import java.io.FileNotFoundException;
-import java.util.*;
-import java.util.ArrayList;
-import java.util.List;
-import org.broadinstitute.hellbender.utils.io.IOUtils;
-import org.broadinstitute.hellbender.utils.variant.HomoSapiensConstants;
 import java.nio.file.Path;
+import java.util.Collection;
+import java.util.List;
 
 
 /**
@@ -233,14 +228,8 @@ public void onTraversalStart() {
     }
 
     private static CachingIndexedFastaSequenceFile getReferenceReader(ReferenceInputArgumentCollection referenceArguments) {
-        final CachingIndexedFastaSequenceFile referenceReader;
         final Path reference = IOUtils.getPath(referenceArguments.getReferenceFileName());
-        try {
-            referenceReader = new CachingIndexedFastaSequenceFile(reference);
-        } catch (FileNotFoundException e) {
-            throw new UserException.CouldNotReadInputFile(reference, e);
-        }
-        return referenceReader;
+        return CachingIndexedFastaSequenceFile.checkAndCreate(reference);
     }
 
     @Override

diff --git a/src/main/java/org/broadinstitute/hellbender/tools/walkers/rnaseq/OverhangFixingManager.java b/src/main/java/org/broadinstitute/hellbender/tools/walkers/rnaseq/OverhangFixingManager.java
@@ -4,6 +4,7 @@
 import htsjdk.samtools.SAMFileHeader;
 import htsjdk.samtools.TextCigarCodec;
 import htsjdk.samtools.reference.IndexedFastaSequenceFile;
+import htsjdk.samtools.reference.ReferenceSequenceFile;
 import htsjdk.samtools.util.Tuple;
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
@@ -60,7 +61,7 @@ public class OverhangFixingManager {
     private final GATKReadWriter writer;
 
     // fasta reference reader to check overhanging edges in the exome reference sequence
-    private final IndexedFastaSequenceFile referenceReader;
+    private final ReferenceSequenceFile referenceReader;
 
     // the genome unclippedLoc parser
     private final GenomeLocParser genomeLocParser;
@@ -92,7 +93,7 @@ public class OverhangFixingManager {
     public OverhangFixingManager(final SAMFileHeader header,
                                  final GATKReadWriter writer,
                                  final GenomeLocParser genomeLocParser,
-                                 final IndexedFastaSequenceFile referenceReader,
+                                 final ReferenceSequenceFile referenceReader,
                                  final int maxRecordsInMemory,
                                  final int maxMismatchesInOverhangs,
                                  final int maxBasesInOverhangs,
@@ -469,7 +470,7 @@ public Splice(final String contig, final int start, final int end) {
             loc = genomeLocParser.createGenomeLoc(contig, start, end);
         }
 
-        public void initialize(final IndexedFastaSequenceFile referenceReader) {
+        public void initialize(final ReferenceSequenceFile referenceReader) {
             reference = referenceReader.getSubsequenceAt(loc.getContig(), loc.getStart(), loc.getStop()).getBases();
         }
 

diff --git a/src/main/java/org/broadinstitute/hellbender/tools/walkers/rnaseq/SplitNCigarReads.java b/src/main/java/org/broadinstitute/hellbender/tools/walkers/rnaseq/SplitNCigarReads.java
@@ -1,13 +1,14 @@
 package org.broadinstitute.hellbender.tools.walkers.rnaseq;
 
-import htsjdk.samtools.*;
-import htsjdk.samtools.reference.IndexedFastaSequenceFile;
+import htsjdk.samtools.CigarElement;
+import htsjdk.samtools.CigarOperator;
+import htsjdk.samtools.SAMFileHeader;
+import htsjdk.samtools.TextCigarCodec;
+import htsjdk.samtools.reference.ReferenceSequenceFile;
 import org.broadinstitute.barclay.argparser.Advanced;
 import org.broadinstitute.barclay.argparser.Argument;
 import org.broadinstitute.barclay.argparser.CommandLineProgramProperties;
 import org.broadinstitute.barclay.help.DocumentedFeature;
-import org.broadinstitute.hellbender.utils.io.IOUtils;
-import picard.cmdline.programgroups.ReadDataManipulationProgramGroup;
 import org.broadinstitute.hellbender.engine.FeatureContext;
 import org.broadinstitute.hellbender.engine.ReferenceContext;
 import org.broadinstitute.hellbender.engine.TwoPassReadWalker;
@@ -21,10 +22,13 @@
 import org.broadinstitute.hellbender.utils.SATagBuilder;
 import org.broadinstitute.hellbender.utils.clipping.ReadClipper;
 import org.broadinstitute.hellbender.utils.fasta.CachingIndexedFastaSequenceFile;
-import org.broadinstitute.hellbender.utils.read.*;
+import org.broadinstitute.hellbender.utils.io.IOUtils;
+import org.broadinstitute.hellbender.utils.read.ArtificialReadUtils;
+import org.broadinstitute.hellbender.utils.read.CigarUtils;
+import org.broadinstitute.hellbender.utils.read.GATKRead;
+import org.broadinstitute.hellbender.utils.read.SAMFileGATKReadWriter;
+import picard.cmdline.programgroups.ReadDataManipulationProgramGroup;
 
-import java.io.File;
-import java.io.FileNotFoundException;
 import java.io.IOException;
 import java.util.ArrayList;
 import java.util.Collections;
@@ -128,7 +132,7 @@ public boolean requiresReference() {
 
     private SAMFileGATKReadWriter outputWriter;
     private OverhangFixingManager overhangManager;
-    private IndexedFastaSequenceFile referenceReader;
+    private ReferenceSequenceFile referenceReader;
     SAMFileHeader header;
 
     @Override
@@ -164,15 +168,10 @@ public ReadTransformer makePostReadFilterTransformer(){
     @Override
     public void onTraversalStart() {
         header = getHeaderForSAMWriter();
-        try {
-            referenceReader = new CachingIndexedFastaSequenceFile(referenceArguments.getReferencePath());
-            GenomeLocParser genomeLocParser = new GenomeLocParser(getBestAvailableSequenceDictionary());
-            outputWriter = createSAMWriter(IOUtils.getPath(OUTPUT), false);
-            overhangManager = new OverhangFixingManager(header, outputWriter, genomeLocParser, referenceReader, MAX_RECORDS_IN_MEMORY, MAX_MISMATCHES_IN_OVERHANG, MAX_BASES_TO_CLIP, doNotFixOverhangs, processSecondaryAlignments);
-
-        } catch (FileNotFoundException ex) {
-            throw new UserException.CouldNotReadInputFile(referenceArguments.getReferencePath(), ex);
-        }
+        referenceReader = CachingIndexedFastaSequenceFile.checkAndCreate(referenceArguments.getReferencePath());
+        GenomeLocParser genomeLocParser = new GenomeLocParser(getBestAvailableSequenceDictionary());
+        outputWriter = createSAMWriter(IOUtils.getPath(OUTPUT), false);
+        overhangManager = new OverhangFixingManager(header, outputWriter, genomeLocParser, referenceReader, MAX_RECORDS_IN_MEMORY, MAX_MISMATCHES_IN_OVERHANG, MAX_BASES_TO_CLIP, doNotFixOverhangs, processSecondaryAlignments);
     }
 
     @Override