[
https://issues.apache.org/jira/browse/OAK-9339?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17286296#comment-17286296
]
Amrit Verma commented on OAK-9339:
----------------------------------
[~mreutegg] Thanks for pointing out. I will add license to
elasticstartscript.sh. Do we also need license in the imagedata.txt file? I
would have to then modify the test class code to skip those license lines. I
didn't add license in the text file because i didn't see license in one other
simiar file file -
[https://github.com/oak-indexing/jackrabbit-oak/blob/trunk/oak-lucene/src/test/resources/org/apache/jackrabbit/oak/query/fvs.csv]
Regarding size, we have reduced it quite a lot by truncating the decimal points
used in double values. Do we need more compression? Then I think we need to
save it as zip and then unzip it during the tests.
Reagrding, if we are allowed to use it or not, I think we should be but let me
check again.
> Image Similarity: LSH based search
> ----------------------------------
>
> Key: OAK-9339
> URL: https://issues.apache.org/jira/browse/OAK-9339
> Project: Jackrabbit Oak
> Issue Type: New Feature
> Components: indexing
> Reporter: Amrit Verma
> Assignee: Amrit Verma
> Priority: Major
> Fix For: 1.40.0
>
>
> Image Similarity introduced in OAK-9213 won't scale well with millions of
> images because it does brute force comparisons. We need to explore
> alternative approaches such as using LSH for similarity search.
> One of the plugins which provides such functionality is elastiknn -
> [https://elastiknn.com/api/]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
