feat: Vector Search #2526
Closed
feat: Vector Search #2526
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
hinthornw
force-pushed
the
wfh/_/pgsearch
branch
from
cd3d10f
to
694e2ab
Compare
hinthornw
commented
Nov 25, 2024
| - Cohere embed-multilingual-light-v3.0: 384 | ||
| """ | ||
|
|
||
| embed: Embeddings |
There was a problem hiding this comment.
Should probably accept a raw function.
We also don't (rn) support multimodal well?
hinthornw
commented
Nov 25, 2024
hinthornw
commented
Nov 25, 2024
| value: The stored value. | ||
| created_at: When the item was first created. | ||
| updated_at: When the item was last updated. | ||
| response_metadata: Optional metadata about the response/result. |
There was a problem hiding this comment.
Rn contains the similarity score.
hinthornw
force-pushed
the
wfh/_/pgsearch
branch
from
694e2ab
to
585a194
Compare
hinthornw
force-pushed
the
wfh/_/pgsearch
branch
from
6cc4662
to
9c25bd1
Compare
|
Transfered to #2535 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Right now you enable it by:
Each document has 1 or more vectors associated with it for each json path in the embedding config.
Would welcome critique and requests!
Especially around handling migrations, whether we want to support multimedia (multimodal embeddings is old school already so kinda bad langchain embeddings classes don't support well), querying (do we need mmr or other strategies...?) and anything else that feels off.
Also happy to shelve the use of json path for saving embeddings if we think that's too complicated.
Also the default
__root__embedding of the whole JSON object retains the keys. I'm guessing there's a bit of an impact here but probably better to have it remain self descriptive than to drop the keys before embedding the object.