PoolParty 10.1 Release Notes

19/01/2025

Taxonomy Builder
Building a good taxonomy is a notorious bottleneck for knowledge graph initiatives because it relies on scarce specialists and manual curation, which inevitably drives up costs and delays progress.
With our new Taxonomy Builder taxonomists and domain experts can create and expand taxonomies faster and more efficiently. This feature leverages the power of Large Language Models (LLMs) to streamline the creation and enrichment of taxonomies and delivers a significant reduction of expertise and effort required to build a foundational knowledge model. Instead of creating concepts one by one, users can now generate
- a hierarchical taxonomy skeleton with a top-down LLM-powered approach
- labels and definitions with AI assistance to improve clarity and discoverability
With the Taxonomy Builder users can describe a domain, generate structure, and refine suggestions while maintaing control through review and acceptance workflows.
For Graphwise customers, this is a meaningful step towards making semantic backbone systems easier to start, faster to iterate, and stronger as a foundation for reliable AI.
Custom user roles
PoolParty previously offered only fixed user roles with predetermined permissions. Now, customers can create custom user roles and assign specific permissions as needed. This enhancement provides fine-grained access control, enabling organizations to tailor user privileges precisely to their security requirements and operational workflows.
Key features:
- API-based role management using Keycloak integration
- Automated policy creation
- Flexible resource control
- Comprehensive scope operations
This enables granular role creation, such as an “API read-only user” for secure data queries or an "Ontology manager" who can modify schemas without requiring full administrator-level access.
Extractor API cleanup
To simplify the integration process and reduce API complexity, we have introduced a new, unified endpoint: /extractor/api/tag. This new endpoint endpoint consolidates the core functionalities previously spread across /extractor/api/extract, /extractor/api/annotate, /extractor/api/categorization.
Benefits for developers:
- Simplified integration: Replaces the legacy /extract method and its over 30 parameters with a cleaner, more intuitive design that is significantly easier to navigate and implement
- Unified input handling: A single entry point now handles various data types (including plain text, URLs, and files) eliminates endpoint switching
- Consolidated functionality: The /tag endpoint merges extraction, categorization, and annotation logic into one streamlined workflow supporting both JSON and RDF outputs
- Efficient batch & asynchronous processing: Integrated support for ZIP archives and asynchronous requests makes large-scale processing more predictable and maintainable

Improved GraphSearch ranking with annotated concept relevance
GraphSearch ranking has been overhauled to prioritize results based on extracted concept relevance scores. Now, users see the most pertinent documents first, ranked by precision scores assigned during concept extraction. This approach surfaces the most valuable information immediately, even for complex, domain-specific searches, eliminating the need to sift through pages of loosely related results.
Dockerization of PoolParty apps
Application Development Framework (ADF) and the Semantic Workbench can now be installed using the same Docker installation as PoolParty 10.1 simplifying and unifying the installation process.

jQuery UI 1.12.1 and 1.13.1 were updated to jQuery 1.13.2 to address the following vulnerabilities CVE-2022-31160, CVE-2021-41184, CVE-2021-41183, and CVE-2021-41182.
tomcat-embed-jasper was upgraded to version 11.0.14 addressing CVE-2025-55752.
tomcat-embed-core was upgraded to version 11.0.14 to address CVE-2025-55752, CVE-2025-55754, and CVE-2025-61795.
lz4-java was upgraded to version 1.8.1 addressing the CVE-2025-8941 vulnerability.
logback-core 1.5.18 was upgraded to version 1.5.19 to address the CVE-2025-11226 vulnerability.
lz4-java 1.8.1 was updated to lz4-java 1.10.1 addressing the CVE-2025-66566 vulnerability.

Warning

If Keycloak is to be deployed behind a reverse proxy, it is necessary to set up a reverse proxy in line with the official Keycloak documentation.

Custom labels dialog for the tree now works as intended.
A message is now displayed informing the user of completion of snapshot creation and restoration.
Automatic generation of API documentation for Thesaurus, Extractor and GraphSearch has been unified and streamlined improving user experience.
The Extractor/api/annotate POST endpoint is now working as intended.
Workflows and SKOS-XL options have been removed from the project creation dialog when not included in the license.
The issue where SPARQL endpoint UI sometimes omitted file extension from downloads has been remedied.
Long ontology labels are now truncated with a trailing (...) on the ontology synchronization dialog.
An error is now shown on the UI informing why user login is not possible due to missing a mandatory user's attribute.
The /PoolParty/api/projects/create endpoint now accepts region language identifiers (e.g., en_US) as values for the defaultLanguage parameter.
The Ontology Manager now correctly selects the display language for multilingual ontologies.
Concept can now be added to a custom relation with a unique property on the UI.
Concept history is now correctly displayed.

In this section: