Added a basic test implementation using the latest htsjdk

This uses htsjdk 2.23.0 but currently hardcodes the reference path. It fails for the test case because of versioning issues. Refers to #54
s-andrews · Aug 18, 2020 · 36d48a4 · 36d48a4
1 parent 3e35053
commit 36d48a4
Show file tree

Hide file tree

Showing 5 changed files with 23 additions and 14 deletions.
diff --git a/.classpath b/.classpath
@@ -3,7 +3,8 @@
 	<classpathentry kind="src" path=""/>
 	<classpathentry kind="con" path="org.eclipse.jdt.launching.JRE_CONTAINER"/>
 	<classpathentry kind="lib" path="jbzip2-0.9.jar"/>
-	<classpathentry kind="lib" path="sam-1.103.jar"/>
 	<classpathentry kind="lib" path="cisd-jhdf5.jar"/>
+	<classpathentry kind="lib" path="htsjdk.jar"/>
+	<classpathentry kind="lib" path="commons-compress.jar"/>
 	<classpathentry kind="output" path="bin"/>
 </classpath>
diff --git a/sam-1.103.jar b/sam-1.103.jar
diff --git a/uk/ac/babraham/FastQC/FileFilters/SequenceFileFilter.java b/uk/ac/babraham/FastQC/FileFilters/SequenceFileFilter.java
@@ -37,6 +37,7 @@ public boolean accept(File f) {
 				|| f.getName().toLowerCase().endsWith(".fastq") 
 				|| f.getName().toLowerCase().endsWith(".bam") 
 				|| f.getName().toLowerCase().endsWith(".sam")
+				|| f.getName().toLowerCase().endsWith(".cram")
 				|| f.getName().toLowerCase().endsWith(".compact-reads")
 				|| f.getName().toLowerCase().endsWith(".goby")
 

diff --git a/uk/ac/babraham/FastQC/Sequence/BAMFile.java b/uk/ac/babraham/FastQC/Sequence/BAMFile.java
@@ -25,11 +25,15 @@
 import java.util.Iterator;
 import java.util.List;
 
-import net.sf.samtools.CigarElement;
-import net.sf.samtools.CigarOperator;
-import net.sf.samtools.SAMFileReader;
-import net.sf.samtools.SAMFormatException;
-import net.sf.samtools.SAMRecord;
+import htsjdk.samtools.CigarElement;
+import htsjdk.samtools.CigarOperator;
+import htsjdk.samtools.SAMFormatException;
+import htsjdk.samtools.SAMRecord;
+import htsjdk.samtools.SamInputResource;
+import htsjdk.samtools.SamReader;
+import htsjdk.samtools.SamReaderFactory;
+import htsjdk.samtools.ValidationStringency;
+
 
 public class BAMFile implements SequenceFile {
 
@@ -43,7 +47,7 @@ public class BAMFile implements SequenceFile {
 	// only way to access the file pointer.
 	private FileInputStream fis;
 
-	private SAMFileReader br;
+	private SamReader br;
 	private String name;
 	private Sequence nextSequence = null;
 	Iterator<SAMRecord> it;
@@ -55,11 +59,11 @@ protected BAMFile (File file, boolean onlyMapped) throws SequenceFormatException
 		name = file.getName();
 		this.onlyMapped = onlyMapped;
 
-		SAMFileReader.setDefaultValidationStringency(SAMFileReader.ValidationStringency.SILENT);
-
 		fis = new FileInputStream(file);
 
-		br = new SAMFileReader(fis);
+		SamInputResource sir = SamInputResource.of(fis);
+
+		br = SamReaderFactory.makeDefault().validationStringency(ValidationStringency.SILENT).referenceSequence(new File("c:/Users/andrewss/Desktop/phix-illumina.fa")).open(sir);
 
 		it = br.iterator();
 		readNext();
@@ -134,7 +138,10 @@ record = it.next();
 		// through the file.
 		if (recordSize == 0) {
 			recordSize = (record.getReadLength()*2)+150;
-			if (br.isBinary()) {
+
+			// TODO: Work out if this is actually needed or if we can pull
+			// what we need from the fis
+			if (true) {
 				recordSize /= 4;
 			}
 		}

diff --git a/uk/ac/babraham/FastQC/Sequence/SequenceFactory.java b/uk/ac/babraham/FastQC/Sequence/SequenceFactory.java
@@ -77,10 +77,10 @@ public static SequenceFile getSequenceFile(File file) throws SequenceFormatExcep
 		if (config.sequence_format != null) {
 			// We're not autodetecting the format, but taking whatever they said
 
-			if (config.sequence_format.equals("bam") || config.sequence_format.equals("sam")) {
+			if (config.sequence_format.equals("bam") || config.sequence_format.equals("sam") || config.sequence_format.equals("cram")) {
 				return new BAMFile(file,false);				
 			}
-			else if (config.sequence_format.equals("bam_mapped") || config.sequence_format.equals("sam_mapped")) {
+			else if (config.sequence_format.equals("bam_mapped") || config.sequence_format.equals("sam_mapped") || config.sequence_format.equals("cram_mapped")) {
 				return new BAMFile(file,true);				
 			}
 			else if (config.sequence_format.equals("fastq")) {
@@ -95,7 +95,7 @@ else if (config.sequence_format.equals("fastq")) {
 
 		// Otherwise we just use the extension on the end of the file name to try to determine
 		// the type
-		if (file.getName().toLowerCase().endsWith(".bam") || file.getName().toLowerCase().endsWith(".sam")) {
+		if (file.getName().toLowerCase().endsWith(".bam") || file.getName().toLowerCase().endsWith(".sam") || file.getName().toLowerCase().endsWith(".cram")) {
 			// We default to using all reads
 			return new BAMFile(file,false);
 		}