Closed wasade closed 9 years ago
Just a few comments
Thanks, @wasade! Do we not have Travis CI for this repo? I'm going to run the tests manually, but it would be nice to have Travis, too
I don't think we do actually... this repo predates skbio if i remember, so it is potentially still a bit lagging
On Wed, Mar 25, 2015 at 11:00 AM, adamrp notifications@github.com wrote:
Thanks, @wasade https://github.com/wasade! Do we not have Travis CI for this repo? I'm going to run the tests manually, but it would be nice to have Travis, too
— Reply to this email directly or view it on GitHub https://github.com/biocore/American-Gut/pull/137#issuecomment-86121622.
I'll add an issue about it
Tests pass on my local system, merging
This produces JSON-based taxonomy summaries using nested objects. Each taxon is represented, as well as its count and relative frequency. The intent here is to provide a reasonably easily consumable format for web applications for representing per-sample taxonomy results. Newick is not ideal here as you have to hack it in order to embed metadata. JSON is not great for trees due to the nesting, however taxonomy in bacteria and archaea top out at 7 levels (8 including root) and eukarya top out at I believe 14 (15 with root), so the nesting is not extreme.
Note, this is not using scikit-bio's
TreeNode
. The taxtree code was put in place prior to scikit-bio, and we needed code that was BSD (hence not using cogent). It was easier to just adapt the existing code here than to do a larger refactoring.@madprime, just to keep you in the loop as it may be cool to provide these in the Open Humans pulldown. What I think we'll end up doing is hosting these via www.microbio.me as public downloads but I haven't resolved that specifically. What this then lets us do is provide relative abundance tab-delimited taxonomy tables and well as a computable format for the full taxonomy easily.