Ontotext has released GraphDB 9.8, which offers text mining integration, notifications over Kafka, Helm charts, and performance improvements.
The GraphDB text mining plugin integrates text mining algorithms and allows for generating new relationships between entities without coding or developing bespoke solutions and integration code.
The plugin comes with out-of-the-box support for popular text analytic services such as Ontotext’s Tag API, GATE Cloud, and spaCy server, as well as an expressive mapping language, to register new services without coding. The extracted text annotations can be manipulated with SPARQL and either returned to the caller for further processing or stored directly into the repository where they will enrich the existing knowledge graph. This functionality covers a number of use-cases that rely on both RDF and text analytics.
Synchronize Downstream Systems with the Kafka Connector
The Kafka connector provides a means to synchronize changes to the RDF model to any downstream system via the Apache Kafka framework. This enables easy processing of RDF updates in any external system and covers a variety of use-cases where a reliable synchronization mechanism is needed.
Each Kafka connector instance will stay automatically up-to-date with the GraphDB repository data. The implementation is built on the same rock-solid framework as the existing Elasticsearch, Solr and Lucene connectors end employs multiple features that allow for precise mapping from RDF to JSON, such as defining fields based on property chains, nested document support as well as advanced filtering by type, literal language or a complex expression.
Standard Deployment on Kubernetes Using Helm Charts
GraphDB 9.8 comes with standard Helm charts and instructions that can help you get started with GraphDB Enterprise Edition on Kubernetes. Every component is configured with sensible defaults and it is very easy to customize it to meet your needs. The Helm charts are open source and provided as a reference for setting up complex GraphDB deployments.
Performance and Database Throughput Improvements
The new version of the GraphDB engine implements a major performance improvement in query evaluation due to optimized memory usage and algorithm optimizations. The LDBC SPB-256 benchmark showed around 50% increase in read performance, while the BSBM benchmark showed a 15% increase in read/explore operations. All changes are backward compatible and would positively affect large repositories under moderate to heavy query load.
Upgrade to RDF4J 3.6.3
GraphDB comes with the latest RDF4J 3.6.3, which includes many SHACL improvements and a SPARQL parser optimization.