Expose _vectorize_counts_and_tree

[mortonjt]

As shown in the PhyloRPCA paper, leveraging internal node counts in the tree (in addition to tip abundances) can double the accuracy in downstream ML pipelines. It makes a lot of sense to directly expose this functionality, which is leveraged internally within fast unifrac

https://github.com/scikit-bio/scikit-bio/blob/main/skbio/diversity/_util.py#L122