Ontotext are pleased to announce the release of KIM 3.7 which brings significant performance and stability improvements. This release has been coordinated with the latest release of the OWLIM family and most notably allows smooth integration with OWLIM's replication cluster. The combination achieves resilience and linearly scalable parallel query performance and also makes KIM easily deployable in industrial environments such as The National Archives.
The most important changes and improvements cover the following areas:
- Indexing and search
- Fully compliant SPARQL 1.1 support and full-text RDF indexing through the updated OWLIM 5.2.
- Support for the N-Quads RDF format
- Handles millions of documents and hundreds of millions of annotations
- Increased customization flexibility - in KIM 3.7 we give users the option to replace some of the pre-defined KIM modules with their own algorithms and behavior. This provides a higher level of customization and flexibility for the KIM server. Four different types of modules can be user-implemented:
- Custom Document Repository module - integrate your preferred indexing/storage engine or change document repository behavior.
- Custom CORE (shorthand for Co-Occurence and Ranking of Entities) module - customize document CRUD operations and search mechanics.
- Custom KIM Document Handlers - facilities which can be executed before/after a document is added in KIM, or when a document is deleted.
- Custom KIM Extensions - write your own extensions for KIM(in Java), which will run in the same virtual machine as the server, for the sake of lowest communication overhead.
- Text analysis
- Improved text analysis capabilities based on GATE 7.0.
- Default pipeline performance improvements.
- User interface (UI)
- Optimized AJAX requests and DOM manipulation in the Facets search screen.
- Improved smart auto-complete in the semantic Patterns and Timelines search interface.
- Added a configuration option which allows you to start the UI and the KIM server with a single command in the same Java Virtual Machine(using Jetty embedded), which is very useful when developing showcases or small demos. Still, running the KIM UI in a separate web application container such as Tomcat 6.x/7.x is recommended for production environments.
- Dropped support of the Web Services API
- Custom web services can still be developed, all you need to get going is the KIM API.
- For additional information ask on the KIM discussion mailing list.
- Documentation
- More detailed and easy to follow online documentation publicly available at https://confluence.ontotext.com/display/Kim36Docs.
- Notes on upgrading
- Although the Java API of the KIM Server has been altered significantly, the overall structure remains the same. If your client software uses a removed method for some interface, please consult the interface Javadocs to find a replacement method.
- Upgrading from an existing KIM 3.0 installation is possible, but complicated. Upgrading the OWLIM image from 3.x to 5.x is not straightforward, as migration tools exist only for the purpose of upgrading from OWLIM 4.x to 5.x. We will post instructions as soon as the OWLIM team posts instructions for upgrading from 3.x to 5.x. Re-populating your data is always an option, though not very efficient for large amounts of data.
- KIM 3.7 requires Java Development Kit 6 and is not compatible with Java 7. Although we are keen to provide Java 7 support, there are no such plans yet.
This is not the complete changelog as the full list of improvements is too long. Please contact us if you are interested in any more details.
Labels:
None