Partner Register Search
How partner register search is split between a standard adapter, the core search engine, and the domain extension.
Partner register search is a read path on the Partner API. Staff search is a separate implementation and is not covered here.
The design keeps the database search independent of any one interoperability standard.
DCI controller, query decoders, envelopes, signing, consent calls
Partner API
Turn a DCI body into a core search, then build the DCI response.
PartnerRegisterSearch
openg2p-registry-core
Compile an allowlisted clause, page the root register, and load related rows in batches.
Register classes and Jinja templates
Domain extension and object storage
Supply the tables and the external payload shape.
DCI is the only adapter shipped today. POST /dci/registry/sync/search and POST /dci/registry/async/search both decode into the same core search. The difference is delivery: the synchronous call returns the envelope, and the asynchronous call publishes it to WebSub after an acknowledgement.
A future adapter would own its own routes and body format. It would still call PartnerRegisterSearch and render through the outgoing template bound to its data model. UNDP-style search is an example of that extension point. It is not implemented.
The feature overview is in Partner Register Search. WebSub hub behaviour is in WebSub. Register change fan-out, which shares outgoing topics but not this search path, is in Outgestion Pipeline.
Last updated
Was this helpful?