guard against shell pipefails in align_and_count in instances where empty inputs would result in div-by-zero conditions
This is also somewhat dependent on https://github.com/broadinstitute/viral-core/pull/102, which helps ensure samtools idxstats emits useful outputs in the case of empty read groups in the input (i.e. zero counts are reported for the sequences in the reference fasta used by align_and_count in the event the bam is empty, rather than a single zero for the catchall "*" read category—though with the changes in this PR the latter should now be tolerated)
guard against shell pipefails in align_and_count in instances where empty inputs would result in div-by-zero conditions
This is also somewhat dependent on https://github.com/broadinstitute/viral-core/pull/102, which helps ensure
samtools idxstatsemits useful outputs in the case of empty read groups in the input (i.e. zero counts are reported for the sequences in the reference fasta used by align_and_count in the event the bam is empty, rather than a single zero for the catchall "*" read category—though with the changes in this PR the latter should now be tolerated)