import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Override
public List<String> getChromosomeNames(){
	
	if(this.getTrackFormat().equals(TrackFormat.TDF)){

		ResourceLocator resourceLocator= new ResourceLocator(this.getWorkFilename());
		TDFReader reader= new TDFReader(resourceLocator);
		List<String> chroms= new ArrayList<String>(reader.getChromosomeNames());
		if(chroms.get(0).equals("All")){
			chroms.remove(0);
		}
		return chroms;
		// chroms.addAll();
	}
	if(this.getTrackFormat().equals(TrackFormat.BEDGRAPH)){
		TabixIndex tbi= (TabixIndex) IndexFactory.loadIndex(this.getWorkFilename() + TabixUtils.STANDARD_INDEX_EXTENSION);
		return tbi.getSequenceNames();
	}
	if(this.getTrackFormat().equals(TrackFormat.BIGWIG)){
		return this.bigWigReader.getChromosomeNames();
	}
	return null;
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexHeaderlessVCF() throws ClassNotFoundException, IOException, InvalidRecordException, SQLException{

	String infile= "test_data/noheader.vcf";
	File outfile= new File("test_data/noheader.vcf.gz");
	outfile.deleteOnExit();
	
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();

	new MakeTabixIndex(infile, outfile, TabixFormat.VCF);
	
	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 200);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 100);

	TabixReader tbx = new TabixReader(outfile.getAbsolutePath());
	Iterator x = tbx.query("1", 1, 10000000);
	assertTrue(x.next().startsWith("1"));

}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void testRealFileSizeVCF() throws ClassNotFoundException, IOException, InvalidRecordException, SQLException{
	
	// See test_data/README.md for this file. This is fairly large and we want to check it is processed
	// in a reasonable amount of time.
	String infile= "test_data/ALL.wex.union_illumina_wcmc_bcm_bc_bi.20110521.snps.exome.sites.vcf";
	
	File outfile= new File("deleteme.vcf.gz");
	outfile.deleteOnExit();
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	long t0= System.currentTimeMillis();
	new MakeTabixIndex(infile, outfile, TabixFormat.VCF);
	long t1= System.currentTimeMillis();
	
	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 1000);
	assertTrue((t1 - t0) < 20000); // Should be << than 20 sec, ~2 sec

	// Check you can read ok
	this.vcfTester(outfile.getAbsolutePath());
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexVCF_CEU() throws ClassNotFoundException, IOException, InvalidRecordException, SQLException{

	String infile= "test_data/CEU.exon.2010_06.genotypes.vcf";
	
	File outfile= new File("deleteme.vcf.gz");
	outfile.deleteOnExit();
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(infile, outfile, TabixFormat.VCF);

	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 1000);
	
	// Check you can read ok
	this.vcfTester(outfile.getAbsolutePath());
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test 
public void canHandleEmptyFile() throws ClassNotFoundException, IOException, InvalidRecordException, SQLException, InvalidGenomicCoordsException{
	String infile= "test_data/empty.bed";
	File outfile= new File("test_data/empty.tmp.bed.gz");
	outfile.deleteOnExit();
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(infile, outfile, TabixFormat.BED);

	assertTrue(outfile.exists());
	assertTrue(expectedTbi.exists());
	
	assertEquals(28, outfile.length()); // Checked with `bgzip empty.bed && ls -l empty.bed.gz`
	assertTrue(expectedTbi.length() > 0);

}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexSortedFile() throws IOException, InvalidRecordException, ClassNotFoundException, SQLException {
	
	String infile= "test_data/overlapped.bed";
	File outfile= new File("test_data/tmp.bed.gz");
	outfile.deleteOnExit();
	
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(infile, outfile, TabixFormat.BED);
	
	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 80);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 80);
	
	TabixReader tbx = new TabixReader(outfile.getAbsolutePath());
	Iterator x = tbx.query("chr1", 1, 1000000);
	assertTrue(x.next().startsWith("chr1"));
	
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexSortedGzipFile() throws IOException, InvalidRecordException, ClassNotFoundException, SQLException {
	
	String infile= "test_data/hg19_genes.gtf.gz";
	File outfile= new File("test_data/tmp2.bed.gz");
	outfile.deleteOnExit();
	
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(infile, outfile, TabixFormat.GFF);

	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 7000000);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 500000);
	
	TabixReader tbx = new TabixReader(outfile.getAbsolutePath());
	Iterator x = tbx.query("chr1", 1, 1000000);
	assertTrue(x.next().startsWith("chr1"));
	
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexUnsortedFile() throws IOException, InvalidRecordException, ClassNotFoundException, SQLException {
	
	String infile= "test_data/refSeq.hg19.bed.gz";
	File outfile= new File("test_data/tmp3.bed.gz");
	outfile.deleteOnExit();
	
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(infile, outfile, TabixFormat.BED);
	
	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 200000);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 100000);
	
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexSortedURL() throws IOException, InvalidRecordException, ClassNotFoundException, SQLException {
	
	String infile= "http://hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/wgEncodeMapability/wgEncodeDacMapabilityConsensusExcludable.bed.gz";
	File outfile= new File("test_data/tmp4.bed.gz");
	outfile.deleteOnExit();
	
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(infile, outfile, TabixFormat.BED);
	
	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 1000);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 1000);
	
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexUnsortedURL() throws IOException, InvalidRecordException, ClassNotFoundException, SQLException {
	
	String infile= "http://hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/wgEncodeSydhTfbs/wgEncodeSydhTfbsGm12878P300bStdPk.narrowPeak.gz";
	File outfile= new File("test_data/tmp5.bed.gz");
	outfile.deleteOnExit();
	
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(infile, outfile, TabixFormat.BED);
	
	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 1000);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 1000);
	
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void canCompressAndIndexVCF() throws ClassNotFoundException, IOException, InvalidRecordException, SQLException{

	String infile= "test_data/CHD.exon.2010_03.sites.unsorted.vcf";
	File outfile= new File("test_data/tmp6.bed.gz");
	outfile.deleteOnExit();
	
	File expectedTbi= new File(outfile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();

	new MakeTabixIndex(infile, outfile, TabixFormat.VCF);
	
	assertTrue(outfile.exists());
	assertTrue(outfile.length() > 1000);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 1000);

	TabixReader tbx = new TabixReader(outfile.getAbsolutePath());
	Iterator x = tbx.query("1", 20000000, 30000000);
	assertTrue(x.next().startsWith("1"));

	// Check you can read ok
	this.vcfTester(outfile.getAbsolutePath());
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void blockCompressFileInPlace() throws IOException, ClassNotFoundException, InvalidRecordException, SQLException{
	// Test we can compress and index a file and overwrite the original file. 
	
	File testFile= new File("deleteme.bed.gz");
	testFile.deleteOnExit();
	Files.copy(new File("test_data/refSeq.hg19.bed.gz"), testFile);
	
	File expectedTbi= new File(testFile.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION); 
	expectedTbi.deleteOnExit();
	
	new MakeTabixIndex(testFile.getAbsolutePath(), testFile, TabixFormat.BED);
	
	assertTrue(testFile.exists());
	assertTrue(testFile.length() > 200000);
	assertTrue(expectedTbi.exists());
	assertTrue(expectedTbi.length() > 100000);

}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
/**
 * Try to load the tabix index from disk, checking for out of date indexes and old versions
 * @return an Index, or null if we're unable to load
 */
public static Index loadTabixIndex(final File featureFile) {
    Utils.nonNull(featureFile);
    try {
        final String path = featureFile.getAbsolutePath();
        final boolean isTabix = new AbstractFeatureReader.ComponentMethods().isTabix(path, null);
        if (! isTabix){
            return null;
        }
        final String indexPath = ParsingUtils.appendToPath(path, TabixUtils.STANDARD_INDEX_EXTENSION);
        logger.debug("Loading tabix index from disk for file " + featureFile);
        final Index index = IndexFactory.loadIndex(indexPath);
        final File indexFile = new File(indexPath);
        checkIndexVersionAndModificationTime(featureFile, indexFile, index);
        return index;
    } catch (final IOException | RuntimeException e) {
        return null;
    }
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
/**
 * Creates a temp file that will be deleted on exit
 *
 * This will also mark the corresponding Tribble/Tabix/BAM indices matching the temp file for deletion.
 * @param name Prefix of the file.
 * @param extension Extension to concat to the end of the file.
 * @return A file in the temporary directory starting with name, ending with extension, which will be deleted after the program exits.
 */
public static File createTempFile(String name, String extension) {
    try {

        if ( !extension.startsWith(".") ) {
            extension = "." + extension;
        }

        final File file = File.createTempFile(name, extension);
        file.deleteOnExit();

        // Mark corresponding indices for deletion on exit as well just in case an index is created for the temp file:
        new File(file.getAbsolutePath() + Tribble.STANDARD_INDEX_EXTENSION).deleteOnExit();
        new File(file.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION).deleteOnExit();
        new File(file.getAbsolutePath() + ".bai").deleteOnExit();
        new File(file.getAbsolutePath() + ".md5").deleteOnExit();
        new File(file.getAbsolutePath().replaceAll(extension + "$", ".bai")).deleteOnExit();

        return file;
    } catch (IOException ex) {
        throw new GATKException("Cannot create temp file: " + ex.getMessage(), ex);
    }
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void testVCFGZIndex_tabix() {
    final File ORIG_FILE = getTestFile("test_variants_for_index.vcf.blockgz.gz"); //made by bgzip
    final File outName = createTempFile("test_variants_for_index.blockgz.gz.", TabixUtils.STANDARD_INDEX_EXTENSION);

    final String[] args = {
            "--feature-file" ,  ORIG_FILE.getAbsolutePath(),
            "-O" ,  outName.getAbsolutePath()
    };
    final Object res = this.runCommandLine(args);
    Assert.assertEquals(res, outName.getAbsolutePath());

    final Index index = IndexFactory.loadIndex(res.toString());
    Assert.assertTrue(index instanceof TabixIndex);

    Assert.assertEquals(index.getSequenceNames(), Arrays.asList("1", "2", "3", "4"));
    checkIndex(index, Arrays.asList("1", "2", "3", "4"));
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
@Test
public void testVCFGZIndex_inferredName(){
    final File ORIG_FILE = getTestFile("test_variants_for_index.vcf.blockgz.gz"); //made by bgzip
    final String[] args = {
            "--feature-file" ,  ORIG_FILE.getAbsolutePath(),
    };
    final Object res = this.runCommandLine(args);
    final File tabixIndex = new File(ORIG_FILE.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION);
    Assert.assertEquals(res, tabixIndex.getAbsolutePath());
    tabixIndex.deleteOnExit();

    Assert.assertTrue(tabixIndex.exists(), tabixIndex + " does not exists");
    final Index index = IndexFactory.loadIndex(tabixIndex.toString());
    Assert.assertTrue(index instanceof TabixIndex);

    Assert.assertEquals(index.getSequenceNames(), Arrays.asList("1", "2", "3", "4"));
    checkIndex(index, Arrays.asList("1", "2", "3", "4"));
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
public void setDbIndex() {
	if(!SeekableStreamFactory.isFilePath(db)) {
		if(this.indexFormat == null) {
			throw new IllegalArgumentException("Please set up index format for remote file.");
		} else if(this.indexFormat == IndexFormat.TBI) {
			this.dbIndex = this.db + TabixUtils.STANDARD_INDEX_EXTENSION;
			if(!isExist(this.dbIndex))
				throw new IllegalArgumentException("Index file: " + this.dbIndex + " is not find.");
		} else {
			this.dbIndex = this.db + BinIndexWriter.PLUS_INDEX_EXTENSION;
			this.plusFile = this.db + BinIndexWriter.PLUS_EXTENSION;
			if(!isExist(this.dbIndex)) throw new IllegalArgumentException("Index file: " + this.dbIndex + " is not find.");
			if(!isExist(this.plusFile)) throw new IllegalArgumentException("Plus file: " + this.plusFile + " is not find.");
		}
	} else {
		if(this.indexFormat == IndexFormat.PLUS) {
			checkPlus();
		} else if(this.indexFormat == IndexFormat.TBI){
			checkTbi();
		} else {
			this.dbIndex = this.db + BinIndexWriter.PLUS_INDEX_EXTENSION;
			if(!isExist(this.dbIndex)) {
				this.dbIndex = this.db + TabixUtils.STANDARD_INDEX_EXTENSION;
				if(!isExist(this.dbIndex)) {
					throw new IllegalArgumentException("We support two types of index file. The first is the tabix index file, the second type needs plus file and plus index file both. Please put the index in the same folder with the database file!");
				} else {
					this.indexFormat = IndexFormat.TBI;
				}
			} else {
				this.plusFile = this.db + BinIndexWriter.PLUS_EXTENSION;
				if(!isExist(this.plusFile)) {
					throw new IllegalArgumentException("We cannot find plus file: "+ this.plusFile + ", you must put plus file and plus index file in the same folder!");
				} else {
					this.indexFormat = IndexFormat.PLUS;
				}
			}
		}
	}
	idx = IndexFactory.readIndex(dbIndex);
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
/**
 * Get a temporary file path based on the prefix and extension provided.
 * This file (and possible indexes associated with it) will be scheduled for deletion on shutdown
 *
 * @param prefix a prefix for the file name
 *               for remote paths this should be a valid URI to root the temporary file in (ie. gcs://hellbender/staging/)
 *               there is no guarantee that this will be used as the root of the tmp file name, a local prefix may be placed in the tmp folder for example
 * @param extension and extension for the temporary file path, the resulting path will end in this
 * @return a path to use as a temporary file, on remote file systems which don't support an atomic tmp file reservation a path is chosen with a long randomized name
 *
 */
public static String getTempFilePath(String prefix, String extension){
    if (isCloudStorageUrl(prefix) || (isHadoopUrl(prefix))){
        final String path = randomRemotePath(prefix, "", extension);
        deleteOnExit(path);
        deleteOnExit(path + Tribble.STANDARD_INDEX_EXTENSION);
        deleteOnExit(path + TabixUtils.STANDARD_INDEX_EXTENSION);
        deleteOnExit(path + ".bai");
        deleteOnExit(path + ".md5");
        deleteOnExit(path.replaceAll(extension + "$", ".bai")); //if path ends with extension, replace it with .bai
        return path;
    } else {
    	try {
            final File file = File.createTempFile(prefix, extension);
            file.deleteOnExit();

            // Mark corresponding indices for deletion on exit as well just in case an index is created for the temp file:
            new File(file.getAbsolutePath() + Tribble.STANDARD_INDEX_EXTENSION).deleteOnExit();
            new File(file.getAbsolutePath() + TabixUtils.STANDARD_INDEX_EXTENSION).deleteOnExit();
            new File(file.getAbsolutePath() + ".bai").deleteOnExit();
            new File(file.getAbsolutePath() + ".md5").deleteOnExit();
            new File(file.getAbsolutePath().replaceAll(extension + "$", ".bai")).deleteOnExit();

            return file.getAbsolutePath();
        } catch (IOException ex) {
            throw new UserException("Cannot create temp file: " + ex.getMessage(), ex);
        }
    }
}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
private VcfChunk initChunk(String chr, int chunkStart, int chunkEnd, boolean phased, List<String> header)
		throws IOException {
	overallChunks++;

	String chunkName = null;

	chunkName = FileUtil.path(chunksDir, "chunk_" + chr + "_" + chunkStart + "_" + chunkEnd + ".vcf.gz");

	// init chunk
	VcfChunk chunk = new VcfChunk();
	chunk.setChromosome(chr);
	chunk.setStart(chunkStart);
	chunk.setEnd(chunkEnd);
	chunk.setVcfFilename(chunkName);
	chunk.setIndexFilename(chunkName + TabixUtils.STANDARD_INDEX_EXTENSION);
	chunk.setPhased(phased);

	BGzipLineWriter writer = new BGzipLineWriter(chunk.getVcfFilename());
	for (String headerLine : header) {
		writer.write(headerLine);
	}

	chunk.vcfChunkWriter = writer;

	return chunk;

}
 

import htsjdk.tribble.util.TabixUtils; //导入依赖的package包/类
public void testTabixIndexCreationChr20() throws IOException {

		String configFolder = "test-data/configs/hapmap-chr1";
		// input folder contains no vcf or vcf.gz files
		String inputFolder = "test-data/data/chr20-phased";

		// create workflow context
		WorkflowTestContext context = buildContext(inputFolder, "hapmap2", "eagle");

		// create step instance
		InputValidation inputValidation = new InputValidationMock(configFolder);

		// run and test
		boolean result = run(context, inputValidation);

		// check if step is failed
		assertEquals(true, result);
		assertTrue(context.hasInMemory("[OK] 1 valid VCF file(s) found."));

		
		// test tabix index and count snps
		String vcfFilename = inputFolder + "/chr20.R50.merged.1.330k.recode.small.vcf.gz";
		VCFFileReader vcfReader = new VCFFileReader(new File(vcfFilename),
				new File(vcfFilename + TabixUtils.STANDARD_INDEX_EXTENSION), true);
		CloseableIterator<VariantContext> snps = vcfReader.query("20", 1, 1000000000);
		int count = 0;
		while (snps.hasNext()) {
			snps.next();
			count++;
		}
		snps.close();
		vcfReader.close();
		
		//check snps
		assertEquals(7824, count);

	}