Integrate results when cohort id's dont match but cohort json match

[gowthamrao]

Use case:

• a site cherry picked cohorts used in a study and imported to their atlas
• created a new study package
• executed the study and shared results
• we would like to integrate the results to main study package

Problems:

• the cohort id's do not match, because usually the cohort id are the same as atlas id.

Solution:

• If we really want to integrate the results then
• Check and confirm that cohort json from donating atlas matches receiving cohort jsons in study package
• If cohort json match, check if the corresponding cohort id’s match.
• If they don’t match - we need a way to override the donating atlas cohort id so that it matches the cohort id in the study package

[azimov]

@chrisknoll on the atlas side is there a standard hashing approach? I think having a global solution for cohorts is a good idea. But tracking changes between a cohort in Cohort DIagnostics is more complicated than it first sounds as there is no history stored within a cohort. If the hash implementation is not consistent then 2 identical cohorts could be different because line endings are change, whitespace changes etc.

[gowthamrao]

@chrisknoll any insight on this would be great! i ran into another scenario this week, where a collaborator imported the cohort definitions, thereby reassigning the cohort ids, executing the results and contributing.

Its tedious to manually compare cohorts - and a process to some how compare, deduplicate, reassign id would make it very useful.