potential additional functionality for lca gather

[ctb]

@brooksph and I had a great conversation on Friday about what to do with the "equivalent matches" output in lca gather. Right now sourmash lca gather will simply say "there's 5 matches to this; here's the first one, alphabetically sorted." But this is confusing.

One idea was to "pull" the whole genome matches back to the LCA of all of the equivalent matches.

I am not sure how this would interact with the p_query or p_match numbers output by sourmash lca gather, however. Maybe we'd take the intersection of all of the equivalent matches, or something, and calculate with respect to that? Seems like this could be really good in the context of unknown organisms; I can implement it quickly if we find a good benchmark data set.

[ctb]

lca gather has been removed in 4.0, #1307.

however, this idea lives on in thumper as well as #817