Phenotype sort

[jmcmurry]

We'd like an algorithm that orders a list of phenotypes in a way that minimizes the sum of the sim distances between adjacent terms, with an option to co-sort two similar lists as well (retaining their membership in the original lists).

Note that ideally, any runner-up matches would be noted (and sorted) as such; see https://github.com/monarch-initiative/monarch-app/issues/1525 which replaces https://github.com/genome-one/monarch/issues/35