Releases: smclab/openk9
Releases · smclab/openk9
2.0.0
Complete changelog
- Refactored sorting
- Improved handling for rejecting messages during enrichment
- Migration from Elasticsearch to Opensearch
- Added vector and hybrid query parser using Approximate k-NN search and Hybrid search
- Schedule statuses revised. Added new statuses: STALE, FAILURE.
- Added new pipeline for vectorization
- New graphql endpoints to handle new admin ui experience
- New sections to configure generative AI elements
- Added score to search response
- Integrated generative AI api to search frontend
- Modified Search Query to support vector indices
- Added file vault config source
- Modifica richiesta Search Query per scelta indici
- Various fixes
- Docker images update using
registry.access.redhat.com/ubi8/openjdk-17-runtime:1.20-3.1727147550
to fix vulnerabilities; - Bug fixes to better handle UI embedding
- Fix types oracle
Docker images are available on Docker Hub at https://hub.docker.com/repositories/smclab
1.7.1
Complete changelog
- Better handling for I18n support
- Added support to create ingresses through applicative manner
- Handling better trigger schedulation
- Added support to deploy connectors through applicative manner
- Added extra configurations for datasource: schedulation workers and scheduling timeout
- Resolved some bugs for better handling errors during schedulation
- Added new frontend components for sorting and calendar
- Resolved problems for handling buffer memory Akka
- Fixed parsing for Tika
- Added supporto to deploy connectors using ArgoCD
- Tenant Manager: initialization schema with default data
- New endpoint to create datasource in multi step way
- Fix query analysis to better handling autocomplete
- Docker images update using
registry.access.redhat.com/ubi8/openjdk-17-runtime:1.19-1
to fix vulnerabilities; - Acl profiling for Query Analysis
Docker images are available on Docker Hub at https://hub.docker.com/repositories/smclab
1.6.0
Complete changelog
- Integrated the hibernate-reactive fork to implement the Multi-Tenancy Scheme
- Refactored Datasource
*K9Service
and replaced use ofTransactionInvoker
(removed) withMutiny.SessionFactory
- Quarkus Cache implementation for rest and grpc api, used in search phase. Cache synchronized with Akka cluster
- Updated Quarkus version
2.16.12.Final
to fix vulnerabilities - Retry Policy of non-indexed messages, managed via RabbitMQ queues
- Starting Job consumers in progress when nodes start
- Ingestion scalability, now multiple messages from the same queue can be processed in parallel (default: 2 * num-nodes)
- Various fixes
- Added translations for configuration from the administration panel
- Hugging Face integration (supported on Kubernetes and OpenShift only)
- Accessibility improving
- Tika optimizing
- Docker images update using
registry.access.redhat.com/ubi8/openjdk-17-runtime:1.18-2
to fix vulnerabilities; - Query Analysis new UI
- Bug fixes to better handle UI embedding
- Tenant UI refactoring
Docker images are available on Docker Hub at https://hub.docker.com/repositories/smclab