consider abundance-weighted bp_match_at_rank in tax

[bluegenes]

bp_match_at_rank as reported by sourmash tax metagenome is the number of unique base pairs matched at a given rank.

I've received a few questions about this that have made me realize that it's a bit counterintuitive for this to not be abundance-weighted.

Perhaps we should consider reporting both a unique and abundance-weighted version of this value.

Also, our documentation seems to be a little out of date in terms of the included columns. Perhaps a column index (as we now have for sourmash gather) would be helpful?