Haplotype matching code needs to use Hartigan parsimony

[jeromekelleher]

Not strictly a bug as correctness doesn't depend on how parsimonious the compression of likelihoods is, but the code should be updated. See #1030 for background.

[jeromekelleher]

This isn't part of any released API yet, so not tagging with any particular milestones, etc.

[hyanwong]

I assume that @astheeggeggs has been using the Hartigan parsimony algorithm. Is this issue therefore fixed in tskit @jeromekelleher ?

[jeromekelleher]

Yes, this was fixed in #1053