datasetindex_v2 perhaps?

my best guess is that when a user submits a search request, its sent via graphql and is resolved by using this method: https://github.com/linkedin/datahub/blob/18dd5b6c13c47c0dfccf6a00927472fd23fa95c9/metadata-io/src/main/java/com/linkedin/metadata/recommendation/RecommendationsService.java#L54 and it looks like it actually takes a collection of "CandidateSources": https://github.com/linkedin/datahub/tree/18dd5b6c13c47c0dfccf6a00927472fd23fa95c9/metadata-io/src/main/java/com/linkedin/metadata/recommendation/candidatesource Which I assume each have their own index that they query, and then they are all merged together and ranked using this: https://github.com/linkedin/datahub/blob/2f03ad854b77acf38234a012211a7252ab2f54f5/metadata-io/src/main/java/com/linkedin/metadata/recommendation/ranker/SimpleRecommendationRanker.java Am I on the right path here?

however for this thread to be appropriate in chatter, you need to replace index with

kibble

and search with

dog

😄

Essentially, we have one index per entity type (so datasetindex_v2 is one of those indices for dataset entities)

so anything *index_v2 are indices used by our search system

hope that helps

is that right?