Issues

Upgrade to Hibernate ORM 6.4.2.Final

Add compatiblity with Elasticsearch 8.12.0 and upgrade to Elasticsearch client 8.12.0

Upgrade to AWS SDK 2.23.3

Rename VectorField#beamWidth / VectorField#maxConnections to align with Elasticsearch/OpenSearch naming

Upgrade to Elasticsearch client 8.11.4 and test against Elasticsearch 8.11.4

knn predicate: filter out results based on similarity

Upgrade JBeret to 2.2.0.Final

Upgrade to AWS SDK 2.22.1

Upgrade to Lucene 9.9.1

Upgrade to Hibernate ORM 6.4.1.Final

Test against latest OpenSearch 1.3.14

Use Maven 3.9.6 in CI builds and as a required minimum version for the build

Align vector similarity naming between backends

Test against latest Elasticsearch 7.17.16

Upgrade to Elasticsearch client 8.11.3 and test against Elasticsearch 8.11.3

Allow passing BeanReference<? extends T> when registering beans to BeanConfigurationContext

Upgrade to Elasticsearch client 8.11.2 and test against Elasticsearch 8.11.2

Mass indexing ID loading queries with discriminator multi-tenancy may cross tenant boundaries in some cases

Simplify entity cleanup in ORM-related tests

Test against latest OpenSearch 2.11.1

Add container extraction support to @VectorField

Simplify the format for indexNullAs on float[]/byte[] bridges

Improve support for custom value bridges for vector fields

Upgrade to Lucene 9.9

Default to SqlTypes.UUID instead of CHAR for the OutboxEvent/Agent ID in databases supporting it

Vector search using Elasticsearch/OpenSearch

Allow looking up the capabilities of each field in the metamodel