Release Notes 05.09.2025 - v3.18.0

Release Notes

📅 Date: 05.09.2025
📦 Version: 3.18.0

✨ Features

• Add Case Model Positive/Negative time handling to the new JSON structure
  Introduced better support for handling positive and negative time within case models, fully integrated into the updated JSON format.
• Add Hypothetical and Omitted Activities to the new JSON structure (Case Model)
  Expanded case model flexibility by including both hypothetical and omitted activities, improving overall data completeness.
• Fix parallelism with StopWatch in AI Center
  Enhanced AI Center with optimized parallel processing through the StopWatch, resulting in faster and more consistent performance.
• Better Visual Distinction for Datatypes
  Improved the UI by making datatypes more visually distinguishable, helping users navigate data more easily.
• Minerva - Include recent N messages as short-term memory
  Extended Minerva’s capabilities with short-term memory by allowing it to recall recent N messages, supporting more context-aware responses.
• Create new Agent for User Settings Data Retrieval
  Built a new dedicated agent for retrieving user settings data, streamlining access and reliability of user configurations.
• Explore and Test the Spring AI Evaluation Library
  Conducted an exploration and validation of the Spring AI Evaluation Library, preparing the ground for future AI evaluation improvements.
• Add Oracle Connector
  Added a new Oracle database connector, enabling broader integration possibilities for enterprise systems.
• AI-Center - Integrate Graph-Storage-Reader Tools (Case Instance)
  Enhanced AI-Center with Graph-Storage-Reader integration for case instances, improving graph-based process insights.
• Graph-Storage-Reader - Integrate MCP Server (Case Instance)
  Strengthened graph-storage capabilities by integrating with the MCP Server, further advancing process intelligence.
• Integrate Tools for "Case Model View" access
  Enabled tool integration for direct access to the Case Model View, simplifying workflows for end-users.
• Investigate Creation of Dimension Model from Entity Graph
  Carried out an initial investigation into deriving dimension models from the entity graph, laying groundwork for advanced reporting.

🐛 Bug Fixes

• Embedding model not mounted correctly
  Improved the setup process for embedding models to ensure they are always mounted correctly and available for use.
• Hypothetical relationships don’t show simplifications
  Enhanced visualization so that hypothetical relationships now consistently display simplifications as intended.
• User-Settings error when executing regeneration of embeddings in the reset-vm-script
  Fixed inconsistencies with user settings during embedding regeneration, ensuring smoother execution of the reset script.
• Where clause case sensitivity leads to problems
  Improved handling of case sensitivity in “where” clauses to avoid mismatches and unintended query issues.
• Filtering process pattern and variant don’t work
  Refined filtering logic so that both process patterns and variants are correctly identified and processed.
• Changing Language changes detail-table data
  Fixed an issue where switching languages unintentionally altered detail-table data, ensuring consistent display across languages.
• Title and summary are German but conversation English
  Improved alignment between titles, summaries, and conversation language for better consistency.
• Importer - Deadlock during dimension load when removing case instance properties in Insurance Dimension
  Resolved deadlock issues in the importer when handling dimension loads, making insurance dimension processing more reliable.