Skip to main content

Release Notes - PoolParty 2024 Release 2 (9.4.0-9.4.3)

13/11/2024

Our latest PoolParty release significantly improves user experience with an enhanced Taxonomy Advisor, redesigned interface, and improved performance while maintaining our commitment to high security standards.

Highlights

Introducing the enhanced Taxonomy Advisor

  • PoolParty’s Taxonomy Advisor, the cutting-edge LLM-based suggestion tool, harnesses the capabilities of Generative AI to empower subject matter experts in creating and refining knowledge graphs and taxonomies with unprecedented ease, efficiency and precision.

    • The auto-generated definitions enrich your knowledge graphs and taxonomies and provide crucial context for concepts hence enhancing interpretability and supporting advanced AI applications including RAG (Retrieval Augmented Generation) implementations.

    • The enhanced Taxonomy Advisor streamlines bulk operations allowing generation of complete concept suggestions containing preferred labels, alternative labels, and definitions - all within a single seamless process. This feature not only saves time but allows you to accomplish lots more with less clicks.

    • The improved Taxonomy Advisor feature offers an intuitive single-screen interface - all data is clearly and conveniently presented on a single user-friendly screen allowing you to effortlessly edit and refine data, remove unnecessary elements, and efficiently review and store results within a single operation.

UI redesign: a fresh, modern look for enhanced usability 

  • A new sleek interface and improved color schemes simplifies navigation and enhance visual comfort

    • We have introduced a contemporary new design that goes beyond mere aesthetics. This comprehensive overhaul makes the UI even more intuitive and user-friendly. The redesigned interface incorporates ergonomic design principles, significantly reducing eye strain and enhancing visual comfort during extended use.

    • Our comprehensive redesign includes improvements of the overall visual experience with particular attention to styling, color schemes and contrast ratios. Users can even more intuitively understand how different elements are interrelated within the application. 

    • The updated visuals delineate various sections and functionalities within the interface. This enhanced visual hierarchy not only allows for quicker comprehension of the information structure but also facilitates more intuitive navigation throughout the application. 

Significant improvement of GraphSearch performance and scalability 

  • GraphSearch functionality has been improved with considerably enhanced overall performance and scalability. GraphSearch scales to much larger search spaces while maintaining high levels of search and filtering performance.

Improvements

Improvements implemented in the current release include:

  • The PoolParty Workbench now supports selection of projects and search spaces. Additionally, text, file and URL are supported as input sources with correspondingly color-coded results further enhancing user experience. It also offers the Expansion Query Builder functionality for subject matter experts less familiar with SPARQL queries.

  • Significant reduction of the time required to open or close a project in the Thesaurus Manager.

  • Corpus Analysis has received a face-lift and functional overhaul.

    • Overall usability and design of the feature have been improved.

    • Candidate concepts are now displayed in a separate tree alongside the taxonomy tree, improving interaction between the two. Users can create custom nodes in the candidate concepts tree to organize concepts and use drag-and-drop to move nodes within and between trees.

    • Extracted terms can now be directly added to the taxonomy tree via drag and drop, bypassing the candidate concepts section. This streamlined process improves efficiency when quickly incorporating extracted terms into the taxonomy without further editing or structuring.

Security Fixes

This section lists remedied vulnerabilities attributable to third party software.

  • The 9.4.2 version incorporates Keycloak 25.0.6 addressing the vulnerability (CVE-2024-8698) identified in the previous Keycloak version.

  • The 9.4.3 version addresses the OAuth vulnerability related to JOSE+JWT Token (CVE-2023-52428)

Bug Fixes

This section lists bug fixes grouped by the respective PoolParty Suite component.

Thesaurus Manager 

  • Assigning custom classes to ordered and non-ordered collections works as intended.

  • PoolParty now prevents deletion of preferred labels of concepts when the project default language uses a locale like en-US or de-DE.

  • The SPARQL and the SPARQL Update endpoints are compliant with the specification (https://www.w3.org/TR/sparql11-protocol/).

  • The options for the SPARQL endpoint are: 

    • Content type is application/sparql-query and the query format is raw. 

    • Content type is application/x-www-form-urlencoded and the query is passed in the parameter query.

  • The options for the SPARQL Update endpoint are: 

    • Content type is application/sparql-update and the query format is raw.

    • Content type is application/x-www-form-urlencoded and the query is passed in the parameter update.

  • Advanced SKOS relations (such as skos:exactMatch or skos:narrowerMatch) are now shown in the concept details of the Thesaurus Manager after being added by the user or imported using tabular import.

Data Validator 

  • Selecting the data validation option during a sub-tree import using tabular data no longer leads to an error.

  • Data Validator scalability has been enhanced to support loading of larger lists of errors allowing users to repair the reported issues.

  • Data Validator can now deal with quotes present in labels and does not fail validation.

  • A Data Validator issue was fixed where sometimes after running the Data Validator the reported domain and range errors were not being shown and the user could not repair them.

Extractor 

  • Concept annotation positions returned by the Extractor for PDF documents are now correct and correspond to the plain text representation of the document.

GraphSearch 

  • Deleting documents in a GraphSearch search space now works correctly for index based settings (Solr or Elasticsearch setups).

  • The facet view now shows all concepts when the user switches from list to tree view.

LinkedData Frontend 

  • The Linked Data frontend can now handle concepts that have a URL present as a value (either as an attribute or a relation).

PoolParty Workbench 

  • Duplicate concepts are no longer shown in the recommendations list.

  • Expanded and manually added concepts are now correctly marked.

  • Queries without extracted concepts now show labels for manually added concepts.

  • Expansion queries are functional even if the prefix SKOS has not been selected.

  • Configurations can be saved even if no search space has been selected.

  • Scores are now correctly calculated for multiplied concepts.

  • Previously removed manually added concepts can be added again.

User Management 

  • Large number of users are now reliably listed in User Management and the loading time of the user list has also been improved.

Component Versions

  • Tomcat: 9.0.91

  • Java 21 (The current version deprecates Java 17.)

  • Keycloak: 25.0.6

  • rdf4j: 4.3.6

  • Spring Core: 5.3.32

  • Spring Security: 5.8.10

  • Tika: 2.9.1

  • Elasticsearch: 8.12.1

  • Solr: 9.4.0 (The current version deprecates Solr.)

Deprecation Notices

  • Support for Java 17 has been deprecated with this release. The PoolParty platform has been tested on both Java 17 and 21 and you can safely switch over to Java 21. Starting from the next PoolParty version - i.e. PoolParty 2024 Release 3 - only Java 21 will be supported.

  • The use of Solr on the PoolParty platform has been deprecated with this release. This affects PoolParty Extractor, GraphSearch, Recommender and the features Corpus Management and Semantic Classifier. Before upgrading to the next PoolParty version - i.e. PoolParty 2024 Release 3 - you will need to execute the corresponding index data migration procedure and configure Elasticsearch as an index.

In this section: