Generalizing variant normalization beyond VCF/BGEN representations about glow HOT 2 OPEN

Would providing a simple conversion between CPRA string ids and our current id representation be sufficient? Or is there something else that is necessary.

from glow.

Yes, conversion to and from CPRA strings or chromosome, position, ref, alt as separate columns should be sufficient. The only other immediate caveats I can think of are:

1. this also relates to #108 since normalize is intended for VCF/BGEN schema data frames and therefore drops supplemental columns. For external data set site lists, they are rarely coming in from VCF/BGEN so should not be expected to fit this schema.
2. external data is more likely to have different chromosome representations as compared to the reference genome file. For example, contigNames "1" and "chr1" mean the same thing, or "X" and "23". We might want to include some basic consistency checking/reporting, even if it just entails listing how many sites could not be mapped to the reference.

from glow.