FileDataSource is in SAMRecord

[LeeTL1220]

Bug Report

Affected tool(s) or class(es)

SAMRecord from GATKRead

Affected version(s)

• Latest master branch as of January 30, 2024

Description

When I run a tool with a bam file as input, the following code will give me a null:

    @Override
    public void apply(GATKRead read, ReferenceContext referenceContext, FeatureContext featureContext) {

        // Build sets of read IDs for each file.
        final SAMRecord samRecord = read.convertToSAMRecord(getHeaderForReads());
        final SAMFileSource fileSource = samRecord.getFileSource();
        System.out.println(fileSource);

Output:
(a long list of null)

Steps to reproduce

Create a ReadWalker that takes in a bam file. Here is an integration test that will replicate the issue:

public class ReadConcordanceIntegrationTest extends CommandLineProgramTest {

    @Test
    public void testTwoCrams() throws IOException {
        final File output = createTempFile("testReadConcordanceOutputFile", ".txt");
        final File input = new File(GATKBaseTest.largeFileTestDir, "expected.K-562.splitNCigarReads.chr20.bam");

        final ArgumentsBuilder args = new ArgumentsBuilder();

        args.addInput(input);
        this.runCommandLine(args.getArgsArray());
    }
}

Expected behavior

Output should be the file used in the read data source (bam file) for each read.

Actual behavior

I get nulls instead

[droazen]

We should expose a getSource() method at the GATKRead level, and have GATK delegate to samRecord.getFileSource() in the SAMRecord case. We might need to do something extra to get HTSJDK to populate this field for us.

Use case is a ReadsDataSource backed by multiple bam/cram inputs, with the reads merged into a single sorted stream, and the tool needing to be able to tell where each read came from. Can't always use read groups / sample names to accomplish this.