Provide validation against current "schema" for image 3

[yarikoptic]

NDA does change the details for the image 3 type of submissions quite regularly: see https://ndar.nih.gov/data_structure_history.html?short_name=image03 for the "changelog". Unfortunately they nohow reflect having introduced changes within the version -- I would have expected it to become "3.1", "3.2" etc as long as it stays compatible (no new required field introduced) and jump to "4" whenever a new field become required or some field gets removed. So, in principle, prepared submission file might become "incompatible" even when it was compatible for the version "3" and there is no other version we could specify to signal the level of compatibility.
I have not looked into details of changes, i.e. if anything is incompatible already -- at least on our quick trial (without actually uploading anything) for a sample study and it was all "kosher" according to the NDA's validator.

But I wonder if the schema could be fetched from https://ndar.nih.gov/data_structure.html?short_name=image03 once in a while (e.g. in the tests? or upon bids2nda call and cached locally) and internal assumptions of bids2nda could be validated (that all required fields are "covered" and whatnot). May be there is some way to get that schema in a standardized form (json schema) and validate against internal structure?

what do you think?

references

• support ticket with NDA asking about versioning of changes: http://ndar.zendesk.com/hc/requests/49134 -- unfortunately seems to be not public, requires login

[bpoldrack]

May be there is some way to get that schema in a standardized form (json schema) and validate against internal structure?

There is:

import requests
import pprint

pprint.pprint(requests.get("https://nda.nih.gov/api/datadictionary/datastructure/image03").json())