More gracefully handle missing read groups

[dpryan79]

Thanks for the nice tool. I installed it via Bioconda and noticed that if you input a BAM file that lacks read groups that you get a segfault in line 75 here: https://github.com/zzhofict/FastDup/blob/v1.0.0/src/markdup/markdup.cpp#L75 This is easy enough to fix by adding read groups, but I wonder if you might instead want to (A) catch this and print an informative error message and (B) document that read groups are required (I missed that if you wrote it somewhere).

[zzhofict]

Thank you for reporting this issue and for the detailed information. This is very helpful. Actually, in Picard, the duplicate status of reads is determined based on read groups, so I assumed the input BAM files would contain read group information by default. This approach is not rigorous, and I will fix this issue as soon as possible.