Enhancement/bam2fastq

[gongyixiao]

This might also be useful since we added #1008 and planning to just store BAMs instead of FASTQs.

[anoronh4]

As discussed earlier today, we should consider accommodating cases in which the bam has the OQ tag, so that we can get something as close to the original fastq as possible. This can be done with gatk's RevertSam tool. We might have to first parse the input bam file to see if the OQ tag is present, before running the tool. We should consider adding OQ to our aligned bam outputs as well, which was brought up in #947

[gongyixiao]

As discussed earlier today, we should consider accommodating cases in which the bam has the OQ tag, so that we can get something as close to the original fastq as possible. This can be done with gatk's RevertSam tool. We might have to first parse the input bam file to see if the OQ tag is present, before running the tool. We should consider adding OQ to our aligned bam outputs as well, which was brought up in #947

Yes. I think this can be done by first check if OQ is present, and and then pipe RevertSam output into SamToFastq as input like this https://www.biostars.org/p/56566/ to save disk space.

[gongyixiao]

We need to bring this back to our radar.

[johnoooh]

This fails with large bams that exceed the max int value. The type of files_pe1.size() would then be 'long' which is not able to be used by groupKey.