MNI PD data part of public federation

[surchs]

Context

We have PD data in nipoppy and neurobagel but have taken down the search interface / demo for them in favour of the public OpenNeuro query endpoint. The PD data are cool, and our users want to search them. We should find ways to make the PD data queryable in a way that respects their specific characteristics.

It is important to clearly communicate and show that we are providing a cross-dataset view at the participant level, and not recreate the work of a local database.

Why

• Part of goal: Grow the amount and usefulness of data #11
• So MNI users (and possibly others) can query cohorts across more than one PD dataset
• To enhance the quality and amount of findable data
• To understand the data governance constraints of non-public data in a semi-public search context (e.g. aggregate only, yes/no only, full results ...)
• To communicate to MNI users what it is that neurobagel does, how it is different from a database, and why it is useful to them

Outcomes

• We know what the external constraints are for making PD data publicly queryable
• If possible, there is now a web interface I can go to search cohorts across QPN and PPMI
• If possible, I can get phenotypic information of my cohort if I have the necessary permissions (e.g. encryption)
• I can find out the sample size across QPN and PPMI that fit specific cohort criteria

What

• Clarify what constraints exist to make local PD data queryable with a public endpoint (yes/no result, aggregate, full, ...) -> JB
• Update the MNI-PD graph-ready data and make it federation-accessible #39
• Create separate nodes for the PPMI and QPN #44
• Understand if node constraints are interoperable with the rest of the neurobagel network, and connect the PD nodes if possible
• provide a query interface that can be constrained to only the PD data API endpoints
• Bump @types/react from 18.3.3 to 18.3.4 query-tool#262
• Make a list of API results returned in aggregate mode so that we can individually OK them
• Decide whether to return paths [for PD data] in aggregate mode api#208
• Remove session path values from aggregate results api#210

[alyssadai]

@surchs, the only subtask left is:

provide a query interface that can be constrained to only the PD data API endpoints

Currently, there is no query tool running that federates over only the PPMI and QPN, but our public federation API is pointing to OpenNeuro as well as the two PD databases. Do we still want to have the former?

Otherwise, I think this milestone is complete.