OBI node deployment

[surchs]

Context

The OBI is ready to start federating with the MNI, for now with synthetic data. They have already done a (partially?) successful deployment of their own node. The goal is to demonstrate a query that finds results from both an OBI and a MNI dataset together. We will for now not list the OBI node in our list of public nodes.

Why

• To improve our integration with the OBI
• To demonstrate the inter-institute federation capability
• To show to other interested partner institutes that we are successfully federating
• To learn more about the OBI specific needs, e.g. how can we link from cohort results to the BrainCode platform, etc

Outcomes

• OBI is running a publicly routable full-stack node
• The node has SSL/HTTPS (probably by using our newly public nginx-acme recipe)
• The node contains 2 or 3 (synthetic datasets), for now without imaging data or derivatives
• The node runs in aggregated mode
• The node has a mimimum cell size of 5 participants. Any query that would result in a smaller number of subjects included will not have a response for that dataset (i.e. if 4 participants match, it looks as if 0 participants match)
• There is one public node that we consider to be the MNI node (e.g. the one that holds QPN). May or may not have to rename the node in our public list (TBD)

What

• Provide OBI with link to the nginx-acme recipe (assume they use separate subdomains for each service)
• (Optional) Give the OBI assistance as needed to get the Node to run on their Azure VM (e.g. zoom call)
• Create or re-use one of our f-APIs (other than federate.neurobagel.org) and add to its local node list the n-API URL for the OBI. Make sure there is a query tool (no auth needed) that is publicly routable and talks to this f-API, so I can demonstrate a query that hits both the OBI and a "MNI node"