Consistent API for file retrieval

[lynaghk]

I'm looking into hooking up the filtering functionality with the VCF frontend. Right now you have a nicely exposed plot-ready-metrics function in the metrics API that will grab the file in the background. The do-analysis multimethods have no such convenience---are you working on some unified code/backend (datomic?) to deal with these concerns? We should sync up on this before the VCF starts reaching out into analyses in bcbio if the latter are going to change up in the next few weeks.

[chapmanb]

Kevin; The do-analysis calls should do things the same way. It starts by using the same file API call:

https://github.com/chapmanb/bcbio.variation/blob/master/src/bcbio/variation/api/run.clj#L20

This will bring GenomeSpace files locally (into a cache dir) and just return files that are already local. So if you pass an ID string like gs:/Home/your_gs_user/whatever it should do the right thing. Let me know if you're running into problems with it and happy to refactor to make it consistent.

The backend still needs work but hopefully that shouldn't mess with calling the API.

[lynaghk]

I tried

(def files (get-files :vcf creds))
(do-analysis :filter
             {:filename (:id (first files))
              :metrics {:dp [0 100]}}
             creds)

and got The fasta file you specified (/Users/kevin/work/harvard/vcf) does not exist. [Thrown class org.broadinstitute.sting.utils.exceptions.UserException]

[chapmanb]

Kevin; That's a GATK exception complaining about the FASTA input file. That's specified in the web config YAML:

https://github.com/chapmanb/bcbio.variation/blob/master/config/web-processing.yaml#L8

and should be a local file since that's part of the server configuration. Hope that fixes it.

[lynaghk]

Ah, got it. I've added a similar config file to the VCF repo and the default FASTA genome and things appear to be running fine now:

https://github.com/lynaghk/vcf/commit/5cce7995acc402d2ab9d2044b65607f9f0930e3a