We need to refactor current integ tests that were created for aggregations (https://github.com/opensearch-project/neural-search/pull/632). Currently there are 20+ aggregations that are execute with and without concurrent segment search each. They are split into 3 classes based on type of aggregations that are there.
At least we need to optimize tests for cases when concurrent segment search, so we don't need to flip that setting for every test. We can split each class into 2 classes, or we can create two separate indexes one with concurrent search and one without.
We need to refactor current integ tests that were created for aggregations (https://github.com/opensearch-project/neural-search/pull/632). Currently there are 20+ aggregations that are execute with and without concurrent segment search each. They are split into 3 classes based on type of aggregations that are there.
At least we need to optimize tests for cases when concurrent segment search, so we don't need to flip that setting for every test. We can split each class into 2 classes, or we can create two separate indexes one with concurrent search and one without.