Make VCFHeader not throw exception if contig header lines lack length…

… field (#1418)

* Make VCFHeader not throw exception if contig header lines lack length field

src/main/java/htsjdk/variant/vcf/VCFContigHeaderLine.java

@@ -76,10 +76,23 @@ public Integer getContigIndex() {
         return contigIndex;
     }
 
+    /**
+     * Get the SAMSequenceRecord that corresponds to this VCF header line.
+     * If the VCF header line does not have a length tag, the SAMSequenceRecord returned will be set to have a length of
+     * SAMSequenceRecord.UNKNOWN_SEQUENCE_LENGTH. Records with unknown length will match any record with the same name
+     * when evaluated by SAMSequenceRecord.isSameSequence.
+     * @return The SAMSequenceRecord containing the ID, length, assembly, and index of this contig. Returns null if the
+     * contig header line does not have a length.
+     */
 	public SAMSequenceRecord getSAMSequenceRecord() {
 		final String lengthString = this.getGenericFieldValue("length");
-		if (lengthString == null) throw new TribbleException("Contig " + this.getID() + " does not have a length field.");
-		final SAMSequenceRecord record = new SAMSequenceRecord(this.getID(), Integer.parseInt(lengthString));
+		final int length;
+		if (lengthString == null) {
+		    length = SAMSequenceRecord.UNKNOWN_SEQUENCE_LENGTH;
+        } else {
+		    length = Integer.parseInt(lengthString);
+        }
+		final SAMSequenceRecord record = new SAMSequenceRecord(this.getID(), length);
         record.setAssembly(this.getGenericFieldValue("assembly"));
 		record.setSequenceIndex(this.contigIndex);
 		return record;

src/main/java/htsjdk/variant/vcf/VCFHeader.java

@@ -261,16 +261,18 @@ public List<VCFContigHeaderLine> getContigLines() {
 
     /**
      * Returns the contigs in this VCF file as a SAMSequenceDictionary. Returns null if contigs lines are
-     * not present in the header. Throws SAMException if one or more contig lines do not have length
-     * information.
+     * not present in the header. If contig lines are missing length tags, they will be created with
+     * length set to SAMSequenceRecord.UNKNOWN_SEQUENCE_LENGTH. Records with unknown length will match any record with
+     * the same name when evaluated by SAMSequenceRecord.isSameSequence.
      */
     public SAMSequenceDictionary getSequenceDictionary() {
         final List<VCFContigHeaderLine> contigHeaderLines = this.getContigLines();
         if (contigHeaderLines.isEmpty()) return null;
 
         final List<SAMSequenceRecord> sequenceRecords = new ArrayList<SAMSequenceRecord>(contigHeaderLines.size());
         for (final VCFContigHeaderLine contigHeaderLine : contigHeaderLines) {
-            sequenceRecords.add(contigHeaderLine.getSAMSequenceRecord());
+            final SAMSequenceRecord samSequenceRecord = contigHeaderLine.getSAMSequenceRecord();
+            sequenceRecords.add(samSequenceRecord);
         }
 
         return new SAMSequenceDictionary(sequenceRecords);

src/test/java/htsjdk/variant/vcf/VCFHeaderUnitTest.java

@@ -25,6 +25,8 @@
 
 package htsjdk.variant.vcf;
 
+import htsjdk.samtools.SAMSequenceDictionary;
+import htsjdk.samtools.SAMSequenceRecord;
 import htsjdk.samtools.util.CloseableIterator;
 import htsjdk.samtools.util.FileExtensions;
 import htsjdk.samtools.util.IOUtil;
@@ -239,6 +241,21 @@ public void testVCFHeaderAddContigLine() {
     }
 
     @Test
+    public void testVCFHeaderContigLineMissingLength() {
+        final VCFHeader header = getHiSeqVCFHeader();
+        final VCFContigHeaderLine contigLine = new VCFContigHeaderLine(
+                "<ID=chr1>", VCFHeaderVersion.VCF4_0, VCFHeader.CONTIG_KEY, 0);
+        header.addMetaDataLine(contigLine);
+        Assert.assertTrue(header.getContigLines().contains(contigLine), "Test contig line not found in contig header lines");
+        Assert.assertTrue(header.getMetaDataInInputOrder().contains(contigLine), "Test contig line not found in set of all header lines");
+
+        final SAMSequenceDictionary sequenceDictionary = header.getSequenceDictionary();
+        Assert.assertNotNull(sequenceDictionary);
+        Assert.assertEquals(sequenceDictionary.getSequence("chr1").getSequenceLength(), SAMSequenceRecord.UNKNOWN_SEQUENCE_LENGTH);
+
+    }
+
+        @Test
     public void testVCFHeaderHonorContigLineOrder() throws IOException {
         try (final VCFFileReader vcfReader = new VCFFileReader(new File(variantTestDataRoot + "dbsnp_135.b37.1000.vcf"), false)) {
             // start with a header with a bunch of contig lines