/**
* Generates a BAM index file, either textual or binary, from an input BAI file. Only used for
* testing, but located here for visibility into CachingBAMFileIndex.
*
* @param output BAM Index (.bai) file (or bai.txt file when text)
* @param textOutput Whether to create text output or binary
*/
public static void createAndWriteIndex(
final File input, final File output, final boolean textOutput) {
// content is from an existing bai file.
final CachingBAMFileIndex existingIndex = new CachingBAMFileIndex(input, null);
final int n_ref = existingIndex.getNumberOfReferences();
final BAMIndexWriter outputWriter;
if (textOutput) {
outputWriter = new TextualBAMIndexWriter(n_ref, output);
} else {
outputWriter = new BinaryBAMIndexWriter(n_ref, output);
}
// write the content one reference at a time
try {
for (int i = 0; i < n_ref; i++) {
outputWriter.writeReference(existingIndex.getQueryResults(i));
}
outputWriter.writeNoCoordinateRecordCount(existingIndex.getNoCoordinateCount());
outputWriter.close();
} catch (final Exception e) {
throw new SAMException("Exception creating BAM index", e);
}
}
@Test(enabled = false, dataProvider = "indexComparisonData")
/** Test linear index at all references and windows, comparing with existing index */
public void compareLinearIndex(String testName, String bamFile, String bamIndexFile)
throws IOException {
// compare index generated from bamFile with existing bamIndex file
// by testing all the references' windows and comparing the counts
// 1. generate bai file
// 2. count its references
// 3. count bamIndex references comparing counts
// 1. generate bai file
File bam = new File(bamFile);
assertTrue(bam.exists(), testName + " input bam file doesn't exist: " + bamFile);
File indexFile1 = createIndexFile(bam);
assertTrue(
indexFile1.exists(), testName + " generated bam file's index doesn't exist: " + indexFile1);
// 2. count its references
File indexFile2 = new File(bamIndexFile);
assertTrue(indexFile2.exists(), testName + " input index file doesn't exist: " + indexFile2);
final CachingBAMFileIndex existingIndex1 =
new CachingBAMFileIndex(indexFile1, null); // todo null sequence dictionary?
final CachingBAMFileIndex existingIndex2 = new CachingBAMFileIndex(indexFile2, null);
final int n_ref = existingIndex1.getNumberOfReferences();
assertEquals(n_ref, existingIndex2.getNumberOfReferences());
final SAMFileReader reader1 = new SAMFileReader(bam, indexFile1, false);
final SAMFileReader reader2 = new SAMFileReader(bam, indexFile2, false);
System.out.println(
"Comparing " + n_ref + " references in " + indexFile1 + " and " + indexFile2);
for (int i = 0; i < n_ref; i++) {
final BAMIndexContent content1 = existingIndex1.getQueryResults(i);
final BAMIndexContent content2 = existingIndex2.getQueryResults(i);
if (content1 == null) {
assertTrue(
content2 == null,
"No content for 1st bam index, but content for second at reference" + i);
continue;
}
int[] counts1 = new int[LinearIndex.MAX_LINEAR_INDEX_SIZE];
int[] counts2 = new int[LinearIndex.MAX_LINEAR_INDEX_SIZE];
LinearIndex li1 = content1.getLinearIndex();
LinearIndex li2 = content2.getLinearIndex();
// todo not li1 and li2 sizes may differ. Implies 0's in the smaller index windows
// 3. count bamIndex references comparing counts
int baiSize = Math.max(li1.size(), li2.size());
for (int win = 0; win < baiSize; win++) {
counts1[win] = countAlignmentsInWindow(i, win, reader1, 0);
counts2[win] = countAlignmentsInWindow(i, win, reader2, counts1[win]);
assertEquals(
counts2[win], counts1[win], "Counts don't match for reference " + i + " window " + win);
}
}
indexFile1.deleteOnExit();
}
