should we default to ANI based on average containment for `compare`?

[ctb]

per conversation with @bluegenes,

our similarity, avg, and max-containment based ANI values are symmetric, which seems like a strong advantage for matrix generation and analysis.

this is not true for straight containment-based ANI.

so should we use avg containment-based ANI for compare output?

I think this is probably a v5 change.