Skip to content

Conversation

oaganesh
Copy link
Contributor

@oaganesh oaganesh commented Apr 30, 2025

Description

Adding implementation for profiling and statistical analysis of KNN vector segments in OpenSearch with compression. Provides functionality to collect, process, and store statistical information about vector dimensions across different shards and segments.

Related Issues

Check List

  • [ ✔️ ] New functionality includes testing.
  • [ ✔️] New functionality has been documented.
  • [ ✔️] API changes companion pull request created.
  • [ ✔️] Commits are signed per the DCO using --signoff.
  • [ ✔️] Public documentation issue/PR created.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check ✔️.

oaganesh and others added 30 commits April 2, 2025 13:55
Signed-off-by: Arun Ganesh <[email protected]>
Signed-off-by: Arun Ganesh <[email protected]>
For derived source, we need to locate the first child document for a
given parent document. This positions the iterators in the correct place
when injecting the derived source. This change adds an optimization to
start this look up from different starting points before shorting out to
starting at the beginning.

Along with this, added some null checks in the test case.

Signed-off-by: John Mazanec <[email protected]>
…-project#2620)

* add github action to run ITs against remote index builder

Signed-off-by: AnnTian Shao <[email protected]>

* fixes and refactored build.gradle to reuse integ test components

Signed-off-by: AnnTian Shao <[email protected]>

* fixes and exclude RemoteBuildIT from default integTest

Signed-off-by: AnnTian Shao <[email protected]>

* fixes to github action

Signed-off-by: AnnTian Shao <[email protected]>

* fix to build.gradle

Signed-off-by: AnnTian Shao <[email protected]>

* fix to github action

Signed-off-by: AnnTian Shao <[email protected]>

* fixes to test and build.gradle

Signed-off-by: AnnTian Shao <[email protected]>

* parameterize spaceType for tests and add filtering test

Signed-off-by: AnnTian Shao <[email protected]>

* fixes and add KNN unfiltered search test

Signed-off-by: AnnTian Shao <[email protected]>

---------

Signed-off-by: AnnTian Shao <[email protected]>
Signed-off-by: Tommy Shao <[email protected]>
Co-authored-by: AnnTian Shao <[email protected]>
Co-authored-by: Navneet Verma <[email protected]>
…ensearch-project#2647)

* Add multi-vector-support faiss patch to IndexHNSW::search_level_0

Signed-off-by: AnnTian Shao <[email protected]>

* Add tests to JNI and KNN

Signed-off-by: AnnTian Shao <[email protected]>

* Update tests by adding hnsw cagra index binary and remove JNI layer method updateIndexSettings

Signed-off-by: AnnTian Shao <[email protected]>

* test fixes

Signed-off-by: AnnTian Shao <[email protected]>

---------

Signed-off-by: AnnTian Shao <[email protected]>
Co-authored-by: AnnTian Shao <[email protected]>
…oject#2646)

* Combine method and lucene mappers to EngineFieldMapper

Signed-off-by: Kunal Kotwani <[email protected]>

* Change the default doc values to false, retain old value for flat field

Signed-off-by: Kunal Kotwani <[email protected]>

* Update flat field mapper checks

Signed-off-by: Kunal Kotwani <[email protected]>

* Fix the default doc value logic

Signed-off-by: Kunal Kotwani <[email protected]>

---------

Signed-off-by: Kunal Kotwani <[email protected]>
…ject#2652)

Fixes a bug that was already fixed in
opensearch-project#2494 but was then
reverted by accident in a refactor. It makes it so that instead of
opening up readers for each transform request, it opens up once per
reader.

Signed-off-by: John Mazanec <[email protected]>
Signed-off-by: Dooyong Kim <[email protected]>
Co-authored-by: Dooyong Kim <[email protected]>
Signed-off-by: Arun Ganesh <[email protected]>
"Updating serialization changes."
oaganesh added 15 commits April 29, 2025 16:49
Signed-off-by: Arun Ganesh <[email protected]>
@oaganesh oaganesh closed this Apr 30, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.