Treat .vcf.bgz as a valid vcf extension in FileExtension (#1727)

* Fix so that .vcf.bgz is treated as a valid compressed vcf extension.
samtools · Oct 29, 2024 · 4ea3285 · 4ea3285
1 parent fd56496
commit 4ea3285
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 8 deletions.
diff --git a/src/main/java/htsjdk/beta/codecs/variants/vcf/VCFCodec.java b/src/main/java/htsjdk/beta/codecs/variants/vcf/VCFCodec.java
@@ -32,7 +32,7 @@ public abstract class VCFCodec implements VariantsCodec {
         {
             add(FileExtensions.VCF);
             add(FileExtensions.COMPRESSED_VCF);
-            add(".vcf.bgz");
+            add(FileExtensions.COMPRESSED_VCF_BGZ);
         }
     };
 

diff --git a/src/main/java/htsjdk/beta/plugin/variants/VariantsBundle.java b/src/main/java/htsjdk/beta/plugin/variants/VariantsBundle.java
@@ -206,7 +206,7 @@ private static <T extends IOPath> Optional<Tuple<String, String>> getInferredCon
             final String ext = extension.get();
             if (ext.equals(FileExtensions.VCF)) {
                 return Optional.of(new Tuple<>(BundleResourceType.CT_VARIANT_CONTEXTS, BundleResourceType.FMT_VARIANTS_VCF));
-            } else if (ext.equals(FileExtensions.COMPRESSED_VCF)) {
+            } else if (ext.equals(FileExtensions.COMPRESSED_VCF) || ext.equals(FileExtensions.COMPRESSED_VCF_BGZ)) {
                 return Optional.of(new Tuple<>(BundleResourceType.CT_VARIANT_CONTEXTS, BundleResourceType.FMT_VARIANTS_VCF));
             }
         }

diff --git a/src/main/java/htsjdk/samtools/util/FileExtensions.java b/src/main/java/htsjdk/samtools/util/FileExtensions.java
@@ -65,9 +65,10 @@ public final class FileExtensions {
     public static final String VCF = ".vcf";
     public static final String VCF_INDEX = TRIBBLE_INDEX;
     public static final String BCF = ".bcf";
+    public static final String COMPRESSED_VCF_BGZ = ".vcf.bgz"; // suffix used by gnomad see https://gnomad.broadinstitute.org/data#v4
     public static final String COMPRESSED_VCF = ".vcf.gz";
     public static final String COMPRESSED_VCF_INDEX = ".tbi";
-    public static final List<String> VCF_LIST = Collections.unmodifiableList(Arrays.asList(VCF, COMPRESSED_VCF, BCF));
+    public static final List<String> VCF_LIST = Collections.unmodifiableList(Arrays.asList(VCF, COMPRESSED_VCF, COMPRESSED_VCF_BGZ, BCF));
 
     public static final String INTERVAL_LIST = ".interval_list";
     public static final String COMPRESSED_INTERVAL_LIST = ".interval_list.gz";

diff --git a/src/test/java/htsjdk/variant/variantcontext/writer/VCFWriterUnitTest.java b/src/test/java/htsjdk/variant/variantcontext/writer/VCFWriterUnitTest.java
@@ -90,7 +90,7 @@ private void createTemporaryDirectory() {
     public void testBasicWriteAndRead(final String extension) throws IOException {
         final File fakeVCFFile = File.createTempFile("testBasicWriteAndRead.", extension, tempDir);
         fakeVCFFile.deleteOnExit();
-        if (FileExtensions.COMPRESSED_VCF.equals(extension)) {
+        if (FileExtensions.COMPRESSED_VCF.equals(extension) || FileExtensions.COMPRESSED_VCF_BGZ.equals(extension)) {
             new File(fakeVCFFile.getAbsolutePath() + FileExtensions.VCF_INDEX);
         } else {
             Tribble.indexFile(fakeVCFFile).deleteOnExit();
@@ -136,7 +136,7 @@ public void testBasicWriteAndRead(final String extension) throws IOException {
     public void testWriteAndReadVCFHeaderless(final String extension) throws IOException {
         final File fakeVCFFile = File.createTempFile("testWriteAndReadVCFHeaderless.", extension, tempDir);
         fakeVCFFile.deleteOnExit();
-        if (FileExtensions.COMPRESSED_VCF.equals(extension)) {
+        if (FileExtensions.COMPRESSED_VCF.equals(extension) || FileExtensions.COMPRESSED_VCF_BGZ.equals(extension)) {
             new File(fakeVCFFile.getAbsolutePath() + ".tbi");
         } else {
             Tribble.indexFile(fakeVCFFile).deleteOnExit();
@@ -159,7 +159,7 @@ public void testWriteAndReadVCFHeaderless(final String extension) throws IOExcep
         try (BlockCompressedInputStream bcis = new BlockCompressedInputStream(fakeVCFFile);
                 FileInputStream fis = new FileInputStream(fakeVCFFile)) {
             AsciiLineReaderIterator iterator =
-                    new AsciiLineReaderIterator(new AsciiLineReader(".vcf.gz".equals(extension) ? bcis : fis));
+                    new AsciiLineReaderIterator(new AsciiLineReader(FileExtensions.COMPRESSED_VCF.equals(extension) || FileExtensions.COMPRESSED_VCF_BGZ.equals(extension) ? bcis : fis));
             int counter = 0;
             while (iterator.hasNext()) {
                 VariantContext context = codec.decode(iterator.next());
@@ -300,7 +300,7 @@ public void TestWritingLargeVCF(final String extension) throws FileNotFoundExcep
 
         final File vcf = new File(tempDir, "test" + extension);
         final String indexExtension;
-        if (extension.equals(FileExtensions.COMPRESSED_VCF)) {
+        if (extension.equals(FileExtensions.COMPRESSED_VCF) || extension.equals(FileExtensions.COMPRESSED_VCF_BGZ)) {
             indexExtension = FileExtensions.TABIX_INDEX;
         } else {
             indexExtension = FileExtensions.TRIBBLE_INDEX;
@@ -333,7 +333,8 @@ public void TestWritingLargeVCF(final String extension) throws FileNotFoundExcep
                 // TODO: BCF doesn't work because header is not properly constructed.
                 // {".bcf"},
                 {FileExtensions.VCF},
-                {FileExtensions.COMPRESSED_VCF}
+                {FileExtensions.COMPRESSED_VCF},
+                {FileExtensions.COMPRESSED_VCF_BGZ}
         };
     }