Function printRunStats

samblaster.cpp:1420–1541

Output the runtine stats. Used both for normal and premature exit.

// Output the runtine stats. Used both for normal and premature exit.
void printRunStats(state_t * state)
{
    // This check should be redundant.
    if (idCount == 0)
    {
        fprintf(stderr, "samblaster: No reads processed.\n");
        return;
    }
    // Output stats on auxiliary files.
    if (!state->quiet)
    {
        if (state->discordantFile != NULL)
            fprintf(stderr, "samblaster: Output  %10" PRIu64 " discordant read pairs to %s\n", discCount/2, state->discordantFileName);
        if (state->splitterFile != NULL)
            fprintf(stderr, "samblaster: Output  %10" PRIu64 " split reads to %s\n", splitCount, state->splitterFileName);
        if (state->unmappedClippedFile != NULL)
            fprintf(stderr, "samblaster: Output  %10" PRIu64 " unmapped/clipped reads to %s\n", unmapClipCount, state->unmappedClippedFileName);
    }

    // Now the stats that might indicate a problem that therefore deserve reporting even if quiet set.
    if (readTooLongCount > 0)
    {
        fprintf(stderr, "samblaster: Found   %10" PRIu64 " of %10" PRIu64 " (%5.3f) total read ids longer than the --maxReadLength(%d)\n."
                        "samblaster: The longest of which is %d bases long.\n",
                readTooLongCount, idCount, ((double)100)*readTooLongCount/idCount, state->maxReadLength, readTooLongMax);
        fprintf(stderr, "samblaster: Consider rerunning samblaster with a larger --maxReadLength.\n");
    }
    if (state->ignoreUnmated)
    {
        fprintf(stderr, "samblaster: Found   %10" PRIu64 " of %10" PRIu64 " (%5.3f%%) total read ids are marked paired yet are unmated.\n",
                unmatedCount, idCount, ((double)100)*unmatedCount/idCount);
        if (unmatedCount > 0) fprintf(stderr, "samblaster: Please double check that input file is read-id (QNAME) grouped.\n");
    }
    if (noPrimaryIdCount > 0)
    {
        fprintf(stderr, "samblaster: Found   %10" PRIu64 " of %10" PRIu64 " (%5.3f%%) total read ids with no primary alignment.\n",
                noPrimaryIdCount, idCount, ((double)100)*noPrimaryIdCount/idCount);
        if (unmatedCount > 0) fprintf(stderr, "samblaster: Please double check that input file is read-id (QNAME) grouped.\n");
    }

    // Now the main output stats.
    // We need no unmated reads in order to ensure that the various id types form a partition of all read ids.
    // This is because unmated reads could indicate an incorrectly sorted file which could cause double counting of ids.
    // Therefore, to be conservative, we should only output the table with both-unmapped, orphan and both-mapped stats in that case.
    // NOTE: the correct test is the unmated reads instead of --ignoreUnmated, as the latter doesn't mean we actually FIND any unmated.
    // However, I have decided to output them even if there are unmated reads because we do warn above in that case.

    // We need to have at least two id types to make sense to output the table.
    int idCountTypes = 0;
    if (bothUnmappedIdCount > 0)   idCountTypes += 1;
    if (unmappedOrphanIdCount > 0) idCountTypes += 1;
    if (mappedOrphanIdCount > 0)   idCountTypes += 1;
    if (bothMappedIdCount > 0)     idCountTypes += 1;
    // We need dups for the denominators to avoid div by zero.
    if (!state->quiet && dupCount > 0 && idCountTypes > 1)
    {
        // We vary the names and spacing of the type names depending on whether or not there are any unmapped orphans.
        // Use shorter strings with less blank space if there are none, as should be common for properly grouped paired-end data.