ENH: Output reference tree attachment node into tsv/csv

[corneliusroemer]

I'm interested in seeing which regions of the tree are hot spots for attachment and which are not.

To that end, it would be super nice if we output the name of the nearest neighbour node on the reference tree for each sample.

This is not very important, but it could be interesting information for others, too.

[corneliusroemer]

Particularly interesting would be to know if there are other nodes with the same "distance" metric, this could indicate uncertainty in placement.

UShER, for example, outputs the number of "equally parsimonious placements". In this case, instead of outputting a single node, one could output a list of nodes - the first one being the one on which private mutations etc are based on.