PrintFileDiagnostics for cram, crai and bai. (#8577)

* New experimental tool to print out human readable file diagnostics for cram/crai/bai files.

src/main/java/org/broadinstitute/hellbender/tools/PrintFileDiagnostics.java

@@ -0,0 +1,79 @@
+package org.broadinstitute.hellbender.tools;
+
+import org.broadinstitute.barclay.argparser.*;
+import org.broadinstitute.hellbender.cmdline.CommandLineProgram;
+import org.broadinstitute.hellbender.cmdline.StandardArgumentDefinitions;
+import org.broadinstitute.hellbender.engine.GATKPath;
+import org.broadinstitute.hellbender.tools.filediagnostics.HTSAnalyzer;
+import org.broadinstitute.hellbender.tools.filediagnostics.HTSAnalyzerFactory;
+import picard.cmdline.programgroups.OtherProgramGroup;
+
+import java.io.File;
+
+/**
+ * A diagnostic tool that prints meta information about a GATK input file.
+ *
+ * Works on files ending in .cram, .crai, and .bai.
+ *
+ * Sample Usage:
+ *
+ * gatk PrintFileDiagnostics \
+ *   -I input.cram \
+ *   -count-limit 10
+ */
+@ExperimentalFeature
+@WorkflowProperties
+@CommandLineProgramProperties(
+        summary = "Print diagnostic information about a genomics file to stdout",
+        oneLineSummary = "Print diagnostic information about a genomics file to stdout",
+        programGroup = OtherProgramGroup.class
+)
+public class PrintFileDiagnostics extends CommandLineProgram {
+
+    @Argument(fullName = StandardArgumentDefinitions.INPUT_LONG_NAME,
+            shortName = StandardArgumentDefinitions.INPUT_SHORT_NAME,
+            doc = "Input path for diagnostics",
+            optional = false,
+            common = true)
+    @WorkflowInput
+    public GATKPath inputPath;
+
+    @Argument(fullName = StandardArgumentDefinitions.OUTPUT_LONG_NAME,
+            shortName = StandardArgumentDefinitions.OUTPUT_SHORT_NAME,
+            doc = "Outut file for diagnostics (must be a local file)",
+            optional = false,
+            common = true)
+    @WorkflowInput
+    public File outputFile;
+
+    @Argument(shortName="count-limit",
+            fullName="count-limit",
+            doc="Limit on how much output to emit (.cram only)")
+    private long countLimit = 1000;
+
+    private HTSAnalyzer htsAnalyzer;
+
+    @Override
+    protected void onStartup() {
+        super.onStartup();
+        htsAnalyzer = HTSAnalyzerFactory.getFileAnalyzer(inputPath, outputFile, countLimit);
+    }
+
+    @Override
+    protected Object doWork() {
+        htsAnalyzer.analyze();
+        return 0;
+    }
+
+    @Override
+    protected void onShutdown() {
+        if ( htsAnalyzer != null ) {
+            try {
+                htsAnalyzer.close();
+            } catch (Exception e) {
+                throw new RuntimeException(e);
+            }
+        }
+    }
+
+}

src/main/java/org/broadinstitute/hellbender/tools/filediagnostics/BAIAnalyzer.java

@@ -0,0 +1,31 @@
+package org.broadinstitute.hellbender.tools.filediagnostics;
+
+import htsjdk.samtools.BAMIndexer;
+import org.broadinstitute.hellbender.engine.GATKPath;
+
+import java.io.File;
+import java.io.IOException;
+
+/**
+ * Analyzer for BAI files.
+ */
+public class BAIAnalyzer extends HTSAnalyzer {
+
+    public BAIAnalyzer(final GATKPath inputPath, final File outputFile) {
+        super(inputPath, outputFile);
+    }
+
+    /**
+     * Run the analyzer for the file.
+     */
+    protected void doAnalysis() {
+        System.out.println(String.format("\nOutput written to %s\n", outputFile));
+        BAMIndexer.createAndWriteIndex(inputPath.toPath().toFile(), outputFile, true);
+    }
+
+    @Override
+    public void close() throws IOException {
+    }
+
+}
+

src/main/java/org/broadinstitute/hellbender/tools/filediagnostics/CRAIAnalyzer.java

@@ -0,0 +1,59 @@
+package org.broadinstitute.hellbender.tools.filediagnostics;
+
+import htsjdk.samtools.CRAMCRAIIndexer;
+import htsjdk.samtools.cram.CRAIIndex;
+import htsjdk.samtools.util.RuntimeIOException;
+import org.broadinstitute.hellbender.engine.GATKPath;
+
+import java.io.File;
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+
+/**
+ * Analyzer for CRAM (.crai) index files.
+ */
+public class CRAIAnalyzer extends HTSAnalyzer {
+
+    final FileOutputStream fos;
+
+    public CRAIAnalyzer(final GATKPath inputPath, final File outputFile) {
+        super(inputPath, outputFile);
+        try {
+            fos = new FileOutputStream(outputFile);
+        } catch (final IOException e) {
+            throw new RuntimeIOException(e);
+        }
+    }
+
+    protected void emitln(final String s) {
+        try {
+            fos.write(s.getBytes());
+            fos.write('\n');
+        } catch (IOException e) {
+            throw new RuntimeException(e);
+        }
+    }
+
+    /**
+     * Run the analyzer for the file.
+     */
+    protected void doAnalysis() {
+        try (final InputStream is = inputPath.getInputStream()) {
+            final CRAIIndex craiIndex = CRAMCRAIIndexer.readIndex(is);
+            emitln("\nSeqId AlignmentStart AlignmentSpan ContainerOffset SliceOffset SliceSize\n");
+            craiIndex.getCRAIEntries().stream().forEach(e -> emitln(e.toString()));
+        } catch (IOException e) {
+            throw new RuntimeException(e);
+        }
+    }
+
+    @Override
+    public void close() throws IOException {
+        if (fos != null) {
+            fos.close();
+        }
+    }
+
+}
+