openSNP
Create new json endpoint
A list of URLs for user's genotype files would be helpful for Open Humans integration.
That sounds good and shouldn't be too hard to implement - something living at https://opensnp.org/genotypes.json and just printing all genotypes with date uploaded, associated user ID, genotype ID, and URL link!
If it has to happen now, the users JSON already includes all genotypings too: http://opensnp.org/users.json You'd just have to iterate over each user, via https://github.com/openSNP/snpr/wiki/JSON-API
We already have this at the https://opensnp.org/users.json endpoint. The problem is that we have too many users by now! As it's not paginated and just returns all users it's too big to be useful:
- the JSON blob is large and takes long to process
- it takes us by far too long to produce the JSON before delivering it
What @arvkevi and I had discussed was an endpoint that lives at https://opensnp.org/users/1.json and returns for the user_id=1 all files we have.
This would also play nicely with the Open Humans integration we already have in place, as we upload a custom JSON to Open Humans that contains a link to the openSNP user-profile. So it would be easy for someone to grab the JSON for that user instead of having to parse our profile HTML.
I should probably give some more context: @arvkevi and I discussed this earlier today (yesterday for @philippbayer :joy:) as he's building http://openimpute.com/. That project takes genotyping files that are stored in Open Humans, imputes them and then uploads an imputed version back to Open Humans.
So far he can support VCF files from other Open Humans projects as he can get the download-URLs easily from Open Humans. But the openSNP integration is special as we don't deposit any actual data into a person's Open Humans account, but rather just a file that links to the openSNP user account.