Replies: 1 comment
|
I think this is essential. Since MCP capabilities currently rely on pgvector, baking this functionality into the database would enable out-of-the-box usage for users, delivering much better convenience. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Proposers
Dianjin Wang (@tuhaihe)
Proposal Status
Under Discussion
Abstract
Bundle pgvector as a first-class, built-in extension in Apache Cloudberry, following the existing
gpcontrib/pattern (the same wayorafceandpg_hint_planare vendored). Beyond just vendoring, the proposal includes the MPP integration work required to make vector search correct and well-documented.Motivation
Vector search has become a table-stakes capability in the AI era, and PostgreSQL users expect pgvector to just work.
Today Cloudberry users must install it separately from a maintained fork (cloudberry-contrib/pgvector). Bundling it gives us:
There is already strong precedent:
orafce,pg_hint_plan, and more cases are all third-party extensions vendored undergpcontrib/.Implementation
gpcontrib/pgvectorfollowing the existing pattern:PGAC_ARG_BOOL(enable, pgvector, ...)inconfigure.ac(mirroringenable_orafce).pgvectorintogpcontrib/Makefile(recurse_targets,install,installcheck).--enable-pgvectorto the releaseCONFIGFLAGSingpAux/Makefileand the devops configure script.cloudberry-contrib/pgvectorRollout/Adoption Plan
--enable-pgvector), add CI.Are you willing to submit a PR?
All reactions