Closed nbokulich closed 4 years ago
This looks really great @nbokulich! My only concerns are really the help text, of "super". Otherwise everything seems to run well locally, all tests pass.
Thanks @mikerobeson ! I fixed the description text with your recommendation.
Ultimately, I think usage will be best demonstrated with a detailed tutorial, on the to-do list. For now hopefully it is clear enough to guide users!
We may also want to consider what mode to use as the default for this method. For now I have kept as len
(choose the taxonomy with the most elements) but super
sort of has the best of all worlds (hence the name) so we may want to promote that.
No problem @nbokulich, I am leaning towards the super
approach as the default. They way I look at it, we are making use of the most information available. I agree about clarifying details more within a tutorial.
merge_taxa
has two new modes:majority
performs LCA merging, with priority to majority labelssuper
performs LCA merging, with priority to majority labels. Labels that are substrings of other labels (including empty rank handles) are collapsed prior to majority/LCA.In general also includes a decent code refactor and addition of several more tests.