docs: Sample to use AlloyDB interface for vector search in PGVector database #204
Conversation
product-auto-label
bot
added
the
api: alloydb
| " \"cats\", k=5, filter=f\"collection_id='{uuid}' and cmetadata->>'topic' = 'animals'\"\n", | ||
| ")" | ||
| ] | ||
| } |
There was a problem hiding this comment.
Do we want to also provide guidance on migrate collections to new tables? This would reduce the need to add a filter on each similarity search. If so can you give more background on why we think they should migrate there tables to individual tables. We will also need to show the users how to create a new table, load the old data and add the documents.
There was a problem hiding this comment.
I have a Colab Notebook with code that users can reference to migrate to the new tables. The notebook contains several code blocks and SQL commands for data migration.
However, we're also currently developing a migration method within our engine class. It might be better to wait for this implementation and then provide users with a concise code snippet to facilitate their migration process.
What do you think?
Co-authored-by: Averi Kitsch <[email protected]>
Co-authored-by: Averi Kitsch <[email protected]>
Co-authored-by: Averi Kitsch <[email protected]>
twishabansal
force-pushed
the
pg_to_alloy
branch
from
9f7fd1b
to
1c0f6de
Compare
|
@twishabansal This PR can be close in favor of #249 , correct? |
This code samples shows users how to use AlloyDB interface for vector search in any DB created using the PGVector interface.
This would allow you to migrate from using PGVector to AlloyDB Vector Store search methods.