[
https://issues.apache.org/jira/browse/SOLR-12890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16948876#comment-16948876
]
Trey Grainger commented on SOLR-12890:
--------------------------------------
I'd like to start working on this issue. I've been doing some work with
indexing and scoring on vectors in Solr recently, and am hearing requests for
this feature more all the time. It's already possible to index vectors in Solr
and do vector scoring leveraging streaming expressions (has been for a while),
but it's not currently easy to do it with the search handler / query component
(i.e. standard search), so we need to get that implemented.
I'm traveling next week (talking about vector search and semantic search at a
meetup, in fact), but plan to get started when I get back. Obviously would
welcome input into the design and approach we should take here, and review and
help with the implementation is always welcome, too. I'll post some comments
here about current approaches.
> Vector Search in Solr (Umbrella Issue)
> --------------------------------------
>
> Key: SOLR-12890
> URL: https://issues.apache.org/jira/browse/SOLR-12890
> Project: Solr
> Issue Type: New Feature
> Reporter: mosh
> Priority: Major
>
> We have recently come across a need to index documents containing vectors
> using solr, and have even worked on a small POC. We used an URP to calculate
> the LSH(we chose to use the superbit algorithm, but the code is designed in a
> way the algorithm picked can be easily chagned), and stored the vector in
> either sparse or dense forms, in a binary field.
> Perhaps an addition of an LSH URP in conjunction with a query parser that
> uses the same properties to calculate LSH(or maybe ktree, or some other
> algorithm all together) should be considered as a Solr feature?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
