Examples with SortingCollection htsjdk.samtools.util.SortingCollection used on opensource projects

Example 1 with SortingCollection

use of htsjdk.samtools.util.SortingCollection in project gatk by broadinstitute.

the class FixMateInformation method doWork.

@Override
protected Object doWork() {
    // Open up the input
    boolean allQueryNameSorted = true;
    final List<SamReader> readers = new ArrayList<>();
    for (final File f : INPUT) {
        IOUtil.assertFileIsReadable(f);
        final SamReader reader = SamReaderFactory.makeDefault().referenceSequence(REFERENCE_SEQUENCE).open(f);
        readers.add(reader);
        if (reader.getFileHeader().getSortOrder() != SAMFileHeader.SortOrder.queryname)
            allQueryNameSorted = false;
    }
    // or into a temporary file that will overwrite the INPUT file eventually
    if (OUTPUT != null)
        OUTPUT = OUTPUT.getAbsoluteFile();
    final boolean differentOutputSpecified = OUTPUT != null;
    if (differentOutputSpecified) {
        IOUtil.assertFileIsWritable(OUTPUT);
    } else if (INPUT.size() != 1) {
        throw new UserException("Must specify either an explicit OUTPUT file or a single INPUT file to be overridden.");
    } else {
        final File soleInput = INPUT.get(0).getAbsoluteFile();
        final File dir = soleInput.getParentFile().getAbsoluteFile();
        try {
            IOUtil.assertFileIsWritable(soleInput);
            IOUtil.assertDirectoryIsWritable(dir);
            OUTPUT = File.createTempFile(soleInput.getName() + ".being_fixed.", BamFileIoUtils.BAM_FILE_EXTENSION, dir);
        } catch (final IOException ioe) {
            throw new RuntimeIOException("Could not create tmp file in " + dir.getAbsolutePath());
        }
    }
    // Get the input records merged and sorted by query name as needed
    final PeekableIterator<SAMRecord> iterator;
    final SAMFileHeader header;
    {
        // Deal with merging if necessary
        final Iterator<SAMRecord> tmp;
        if (INPUT.size() > 1) {
            final List<SAMFileHeader> headers = new ArrayList<>(readers.size());
            for (final SamReader reader : readers) {
                headers.add(reader.getFileHeader());
            }
            final SAMFileHeader.SortOrder sortOrder = (allQueryNameSorted ? SAMFileHeader.SortOrder.queryname : SAMFileHeader.SortOrder.unsorted);
            final SamFileHeaderMerger merger = new SamFileHeaderMerger(sortOrder, headers, false);
            tmp = new MergingSamRecordIterator(merger, readers, false);
            header = merger.getMergedHeader();
        } else {
            tmp = readers.get(0).iterator();
            header = readers.get(0).getFileHeader();
        }
        // And now deal with re-sorting if necessary
        if (ASSUME_SORTED || allQueryNameSorted) {
            iterator = new SamPairUtil.SetMateInfoIterator(new PeekableIterator<>(tmp), ADD_MATE_CIGAR);
        } else {
            logger.info("Sorting input into queryname order.");
            final SortingCollection<SAMRecord> sorter = SortingCollection.newInstance(SAMRecord.class, new BAMRecordCodec(header), new SAMRecordQueryNameComparator(), MAX_RECORDS_IN_RAM, TMP_DIR);
            while (tmp.hasNext()) {
                sorter.add(tmp.next());
            }
            iterator = new SamPairUtil.SetMateInfoIterator(new PeekableIterator<SAMRecord>(sorter.iterator()) {

                @Override
                public void close() {
                    super.close();
                    sorter.cleanup();
                }
            }, ADD_MATE_CIGAR);
            logger.info("Sorting by queryname complete.");
        }
        // Deal with the various sorting complications
        final SAMFileHeader.SortOrder outputSortOrder = SORT_ORDER == null ? readers.get(0).getFileHeader().getSortOrder() : SORT_ORDER;
        logger.info("Output will be sorted by " + outputSortOrder);
        header.setSortOrder(outputSortOrder);
    }
    if (CREATE_INDEX && header.getSortOrder() != SAMFileHeader.SortOrder.coordinate) {
        throw new UserException("Can't CREATE_INDEX unless sort order is coordinate");
    }
    try (final SAMFileWriter out = createSAMWriter(OUTPUT, REFERENCE_SEQUENCE, header, header.getSortOrder() == SAMFileHeader.SortOrder.queryname)) {
        logger.info("Traversing query name sorted records and fixing up mate pair information.");
        final ProgressLogger progress = new ProgressLogger(logger);
        while (iterator.hasNext()) {
            final SAMRecord record = iterator.next();
            out.addAlignment(record);
            progress.record(record);
        }
        iterator.close();
        if (header.getSortOrder() == SAMFileHeader.SortOrder.queryname) {
            logger.info("Closing output file.");
        } else {
            logger.info("Finished processing reads; re-sorting output file.");
        }
    }
    // TODO throw appropriate exceptions instead of writing to log.error and returning
    if (!differentOutputSpecified) {
        logger.info("Replacing input file with fixed file.");
        final File soleInput = INPUT.get(0).getAbsoluteFile();
        final File old = new File(soleInput.getParentFile(), soleInput.getName() + ".old");
        if (!old.exists() && soleInput.renameTo(old)) {
            if (OUTPUT.renameTo(soleInput)) {
                if (!old.delete()) {
                    logger.warn("Could not delete old file: " + old.getAbsolutePath());
                    return null;
                }
                if (CREATE_INDEX) {
                    final File newIndex = new File(OUTPUT.getParent(), OUTPUT.getName().substring(0, OUTPUT.getName().length() - 4) + ".bai");
                    final File oldIndex = new File(soleInput.getParent(), soleInput.getName().substring(0, soleInput.getName().length() - 4) + ".bai");
                    if (!newIndex.renameTo(oldIndex)) {
                        logger.warn("Could not overwrite index file: " + oldIndex.getAbsolutePath());
                    }
                }
            } else {
                logger.error("Could not move new file to " + soleInput.getAbsolutePath());
                logger.error("Input file preserved as: " + old.getAbsolutePath());
                logger.error("New file preserved as: " + OUTPUT.getAbsolutePath());
                return null;
            }
        } else {
            logger.error("Could not move input file out of the way: " + soleInput.getAbsolutePath());
            if (!OUTPUT.delete()) {
                logger.error("Could not delete temporary file: " + OUTPUT.getAbsolutePath());
            }
            return null;
        }
    }
    CloserUtil.close(readers);
    return null;
}