Optionally error if not all clades defined in clades.tsv are found in tree

[corneliusroemer]

Context

On my SARS-CoV-2 Pango tree, I didn't notice at first that clade 24H had disappeared because augur clades only warns but doesn't error when a clade present in the clades.tsv is missing from the tree. I don't read all warnings. It would be good if there was an optional --error-on-missing-clade argument that would make such issues become apparent more quickly.

In the case mentioned above the clade definition of 24H wasn't specific enough - it also matched 25C. That was fine, until 25C became bigger than 24H. At this point, 24H was put on the 25C branch and then overwritten by the further-down-specified 25C.

Description

Add an option --error-on-missing-clade that if set makes augur clades fail if not all clades in the clades tsv are found in the tree.

[huddlej]

That sounds reasonable!