I made available a JSON Schema draft-4 for the Atlas Cohort Definition at Github. The objective is to have a tool to validate JSON cohort definitions as a way to help spread the use of OHDSI outside Atlas. With a JSON Schema, a cohort JSON output can be exchanged between tools with confidence. I hope that someone finds useful, feedback welcomed.
@pmadril - it is great to see that someone is also thinking about standardizing on ATLAS JSON definitions! In OHDSI Community, we have had quite a lot of discussions on that and we are totally in agreement it is needed. The question is when and what it should look like. I believe we said we will target those in ATLAS v3
How about this - I would propose you come to our ATLAS/WebAPI Workgroup and share your ideas and proposal in that session with others? The group meets on Wednesdays at 10am ET.
@gregk - Sorry by the delay to respond, I was traveling during these days.
Thank you for your interest about the schema, and yes, for me is totally needed. I will love to participate on the next meeting.
Answering the question about when, can be just now.
I followed the Atlas version 2.7.4 interface for cohort definition and implemented the
full extension of it using JSON Schema draft-04 .
There are two versions, one suited for use with JSON Editor, a project which I also contribute,
and one without extensions and in just one file, best suited for use in validation of JSON Output of Atlas.
So, if anybody had ever asked how it will look like, it is ready to be tested.
I do not expect it to be final, but just a starting point to better develop ideas, specially on setting a cohort definition standard.
Now it is possible to experiment with a JSON cohort definition with confidence and start thinking about other tools that exchange cohorts beside Atlas.
During the developing of the schema, I also get insights about the actual implementation of the cohort definition and will like to share on next posts.
Any feedback is welcomed.
This is a really great initiative! Any updates on this post from about 4 years ago? I think it would be a great benefit to have a standard around cohort definitions created. Thanks!