Squirro 3.4.7 is a long-term support (LTS) release and will receive updates for security issues and important bug fixes for the next two years. See the Squirro Release Process document for details on Squirro’s versioning. |
These release notes cover all the new features and improvements since the previous LTS release 3.3.9. Most of these were already introduced in the intermediate releases of the 3.4.x series and have been covered in those release notes as well.
This 3.4.7 LTS release comes with an upgrade of Elasticsearch to version 7.16.3. Upgrading Squirro to the LTS release requires to update Elasticsearch before Squirro is updated. See the upgrade instructions for details. |
On new deployments, Squirro creates now a monitoring project for user activity tracking upon first admin-user sign up. The monitoring project is setup to collect data on the user's interactions with Squirro through the new Squirro Activity Log data loader plugin and a pre-defined pipeline containing the Activity Tracking pipelet. The recorded user logins, dashboard usage and submitted queries provide valuable insights on the user behaviour.
Particularly interesting are the recorded queries. These can be used to improve the users search experience by suggesting popular queries to the user.
Squirro can now suggest queries to the user that have been used in a previous search. While typing in a query, the matching query suggestions are proposed to the user as autocompletion. Popular query suggestions require that activity tracking is setup.
Newly created projects are now equipped with a pre-configured query processing workflow. Query processing improves the search experience by providing more relevant results. The submitted query gets processed and enriched before the actual search (document retrieval) happens.
The individual steps of the query processing pipeline are configurable in the UI. It is also possible to customize the entire pipeline by uploading a custom workflow and use it as the default query processing.
Besides the functionality for rerunning an entire pipeline, Squirro now also supports rerunning individual pipeline steps in the pipeline editor. Submitting a query together with the rerun allows to filter the items on which the enrichment step is (re-) applied.
The AI Studio supports now the creation of proximity filters with multiple labels. The proximity filter is Squirro’s rule-based classification approach. Proximity filters are created using the same workflow and functionality as is used for creating ML classifiers.
The detail view of items has been improved and redesigned. Particularly the user experience for viewing PDF documents has been improved. The item detail view for PDFs now features an in-document search.
Squirro introduces the reference screen, a split screen view of two documents, that helps the user identify entities in one document and retrieve relevant information about that entity in a second (reference) document.
The reworked Data Sources page in the Setup space is now paginated and has filtering and search options.
The Salesforce connector now supports SOQL queries to retrieve arbitrary Salesforce objects.
Standalone external Tika Service that is used during Content-Extraction-Step (instead of spawning a new JVM every time)
Ability to adjust batch size for reruns in the pipeline
Squirro supports now OCR (optical character recognition) with an out-of-the box PDF OCR pipeline step. The step converts images of text embedded in PDF files (for example, scanned documents) into machine-encoded text.
Optimization: Process consecutive ML Enrichment steps in the pipeline-workflow automatically as a group.
Rules in the AI Studio are first sorted by label, then by query
Improved validation of proximity filters.
The optimization for execution of multiple ML workflows increases the performance by grouping similar workflows in the pipeline and deduplicate the execution of identical steps.
Squirro’s libNLP is now available for download to our partners.
Publishing of ML workflow directly to the ingestion pipeline using the Squirro client.
Communities are now automatically synchronized from facets and updating KEE.
The project configuration is now part of the configuration service. Note, the former project configuration is no longer supported (see breaking changes).
Self-service instances support now the management of the upgrade instance behaviour. It is, for example, now possible to having an instance automatically updated to the latest stable release.
Newly created self-service instances now come with a pre-configured monitoring project for activity tracking.
Handle user rejection in SAML / SSO authentication if Squirro side enforces role-based access (give a proper error instead of a generic 500 error message)
Projects can now hold project pictures that show up in the grid view of the project selection.
This LTS release comes with an upgrade of Elasticsearch to 7.16.3. Upgrading Squirro to the LTS release requires to update Elasticsearch before Squirro is updated. See the Upgrades for Squirro 3.4.x and later instruction for details.
[3.4.0] Validation of the SAML redirect URLs could break login flows if the hostname setup has not been done correctly. Before upgrading, check that the SAML configuration reflects the correct hostname. After upgrading, ensure the redirects still work, and if they don't, edit the SAML configuration to update the host name.
[3.4.0] The API of the MLPublishMixin.ml_publish_model and MLPublishMixin.modify_published_model has changed.
[3.4.4] Deleted the old project configuration plugin from the assets and introduced a new plugin that uses the configuration service. This implies that the old plugin is not supported anymore and cannot be used after upgrading. Any existing configurations would need to be manually transferred to the new plugin.
[3.4.5] Rename "Natural-Language-Query-Parser" to "Query Understanding" and move it to the AI Studio section.
[3.4.7] The Calendar dashboard widget has been removed and will no longer work on dashboards that make use of it.
See the intermediate release notes for a list of all the bugs fixed since the last LTS release.
For new installations please follow the Setup on Linux instructions.
To upgrade an existing installation, please consult the Upgrades for Squirro 3.4.x and later guide.