Research VectorChord
Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.
Background
Here's the announcement about a new PG based addon https://blog.pgvecto.rs/vectorchord-store-400k-vectors-for-1-in-postgresql
Proposal
We should research whether it could be adopted in the future as an additional adapter.
The main concern at the moment is filtering capabilities.
Comments from @DylanGriffith:
No mention of composite index on https://github.com/tensorchord/VectorChord nor could I find any mentions in issues/discussions. So I doubt they have a focus on filtered searches right now. I found one question about this at https://github.com/tensorchord/VectorChord/issues/77#issuecomment-2475555276 but no responses to that question.
So many of these academic efforts for vector search seem to completely gloss over the fact that real world enterprise applications need to combine filtering and search to meet business requirements. And the reality is that this almost always invalidates the indexing algorithms they are using.
I opened this issue on VectorChord to start the discussion https://github.com/tensorchord/VectorChord/issues/121