SWIRL v3.9 Community Release
The SWIRL team is thrilled to release the SWIRL AI Connect 3.9, Community Edition, a significant upgrade that enhances functionality, user experience, and compatibility.
This release is especially relevant to AI search and retrieval-augmented generation (RAG) users, with critical new features, refined improvements, and extended support for Python 3.12.6. Below is an overview of the key updates in this version.
New Features in SWIRL 3.9
1. Enhanced Usability with “Unselect All” for RAG:
A new “Unselect All” button has been introduced in the “Select Items” menu for RAG, providing users greater control over their selections. This feature simplifies bulk operations by allowing users to deselect items at once, streamlining their interactions and enhancing ease of use.
2. Database Migrations Using Django:
This release introduces Django-based database migrations for handling schema changes. Now, any update to the SWIRL database schema will automatically generate a corresponding migration file, ensuring a smooth, version-controlled update process. This change improves the maintainability and scalability of SWIRL deployments, especially for developers managing data integrations.
3. Python 3.12.6 Validation:
To keep up with Python’s advancements, SWIRL 3.9 has been tested and validated on Python 3.12.6. This ensures that the platform leverages the latest optimizations and security patches in Python, offering a stable environment for users looking to deploy SWIRL on modern Python versions.
Improvements in SWIRL 3.9
1. Galaxy UI Enhancements:
The Galaxy UI, known for its visualization capabilities, now better handles longer result titles and adjusts seamlessly with browser window resizing. These improvements aim to deliver a more responsive and accessible interface, enhancing readability and user interaction across different screen sizes.
2. Developer Tag for SearchProviders:
A new “Dev” tag for SearchProviders has been added, making it easier for developers to search specifically across development-centric providers. This helps segment search results based on developer relevance, streamlining workflows for users working on technical and code-centric queries.
3. Elasticsearch Result Configuration Fix:
For users relying on Elasticsearch, SWIRL 3.9 now ensures that the search results align accurately with the configured result count. This addresses issues where the number of results did not match settings, enhancing predictability and precision in search outcomes.
4. Optimized Outlook and OneDrive Integrations:
The FILE_SYSTEM directive has been removed from the default OneDrive configuration, reducing unnecessary data requests. Additionally, the order of result_processors for Outlook email has been restructured, improving the display of threaded conversations and making email retrieval more intuitive.
Known Issues
While SWIRL v3.9 brings many enhancements, there are known issues that the team is actively working to resolve. Notably, Microsoft Teams results sometimes fail to redirect users correctly if they aren’t currently in the corresponding chat. Users are advised to have the Teams app open before clicking on SWIRL-generated links to mitigate this issue. Additionally, using predictive services in browsers may create duplicate search objects, which can be resolved by disabling browser prediction features.
Upgrading and Documentation
Upgrading to SWIRL v3.9 requires database migration due to the new Django-based migration framework. Users are encouraged to follow the upgrade instructions available on the official documentation site. The documentation has been updated and expanded to support this release, offering comprehensive resources for all users. See the upgrading instructions here.
Community and Support
The SWIRL team welcomes feedback from users to improve the platform continually. For any questions, issues, or contributions, the SWIRL community on Slack and direct support channels are available. Join our community on Slack.
Overall, SWIRL AI Connect 3.9 Community Edition marks a substantial evolution in SWIRL’s AI-powered search capabilities, with a focus on usability, development efficiency, and seamless upgrades. View the release on GitHub.
