The GA4GH API needs a reference implementation

ga4gh / ga4gh-schemas

Models and APIs for Genomic data. RETIRED 2018-01-24

http://ga4gh.org

Apache License 2.0

214 stars 114 forks source link

The GA4GH API needs a reference implementation #114

Closed cassiedoll closed 10 years ago

cassiedoll commented 10 years ago

A new repository should be created that serves up a minimum viable reference implementation of the API.

As a possible implementation detail - it could serve data from the usual file formats (BAM, VCF, etc), but we could choose another route.
The compliance suite should be able to use this reference implementation as an example of how to get all API methods passing.
It should be easy for casual users to read the code and get a sense of how they might write a real API implementation of their own

cassiedoll commented 10 years ago

Another goal: it should be easy to run the implementation on various cloud providers in addition to a local machine. (in other words - it should be as portable as possible)

pgrosu commented 10 years ago

+1 I agreed on the original post, comprehensive and consistent with a less-is-more approach which I like :)

+1 On the second post, which goes back to the command-line driven pipeline approach I was suggesting a while back :) Though most of these would be consolidated into scripts, the trend will most likely be on the cloud by utilizing pipelines with cached data/results, such as comparing HapMap with 1000 Genomes or having an online pipelines that keep updating a cloud storage collection of buckets as new data becomes available.

jeromekelleher commented 10 years ago

Excellent idea! Since performance isn't a key issue and clarity, readability, etc are, can I suggest Python as the implementation language? This should also tick all the boxes for portability, and a PyPI package would make it very easy for users to install and try out.

cassiedoll commented 10 years ago

Sounds good to me!

cassiedoll commented 10 years ago

@jeromekelleher - would you like to take a stab at a skeleton project? I feel like once we can get a rough outline of the code together it will be easy to pull in contributions to flush out all the API methods.

There is a java version of v0.1 in this repo if it helps at all!

jeromekelleher commented 10 years ago

Sounds like fun @cassiedoll, sign me up! How about I make a start on a repo in my personal github account, which we can then move into ga4gh when it's ready for input?

cassiedoll commented 10 years ago

Great! I'll do ya one better and create a new repo that you can fork. ga4gh/server will be up in a sec. (we can always rename it if we want later :)

jeromekelleher commented 10 years ago

Even better, thanks! I'll report back here once there's a working skeleton.

richarddurbin commented 10 years ago

On 6 Aug 2014, at 20:54, cassiedoll notifications@github.com wrote:

Another goal: it should be easy to run the implementation on various cloud providers in addition to a local machine. (in other words - it should be as portable as possible)

— Reply to this email directly or view it on GitHub.

The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE.

jeromekelleher commented 10 years ago

I've made a PR for the initial proposal: https://github.com/ga4gh/server/pull/1

The original repo is at https://github.com/jeromekelleher/server if you would like to browse the code there. The README should explain what has been implemented and the approach that we've taken.

Any comments would be much appreciated!

cassiedoll commented 10 years ago

Marking this closed as the first commit is in. All other reference impl discussion should now happen in that repo: https://github.com/ga4gh/server