AutoClassifier 7.2.0.0 Release Notes
This page outlines all of the updates and enhancements in this release. For more information on upgrading, see How to Upgrade AutoClassifier.
September 2025- v7.2.0.0
The following features and updates are new in this release of AutoClassifier:
System Enhancements
|
Feature |
Description |
|---|---|
| Support for Python scripting |
This release of AutoClassifier adds a new Python Scripting component, allowing you to utilize the capability of a Python script of your choosing. For example, you can use Langchain text splitters to break down content from Confluence and HTML pages, improving the relevancy and accuracy of retrieved results from your indexed content, while also enhancing AI-driven functionality such as content summaries and contextual answers. For more information, see Using Python script capabilities. To be able to integrate Python scripting functionality in your AutoClassifier environment, you must install the AutoClassifier Python Script Runner Rest API when you install the AutoClassifier engine. For more information, see Install AutoClassifier Engine Components. |
| Improved event logging | The AutoClassifier logs now provide in depth relevant information on administrative events, such as configuration updates, setting modifications, and logging modifications, to enhance the auditing and tracking experience for AutoClassifier. For more information, see View detailed information on admin actions. |
| Support for on-demand annotation testing | In the AutoClassifier SharePoint add-in, you can now test the annotation of a single document in your SharePoint environment for debugging, configuration issues, or pipeline validation on-demand. For more information, see Annotating your documents on-demand for testing. |
| Document Chunking enhancements | In the Document Chunker component, the default value for the Max Text To Extract field is now set to 2MB. For more information, see Configure the document chunker component. |
| Video Indexer enhancements | In addition to local storage, the Microsoft Video Indexer component now supports network paths for storing and processing videos. For more information, see how to classify videos. |
Bug fixes
-
When a metadata property was overridden using the AutoClassifier script component, existing metadata from parent properties were missing from the response.
-
When using the document chunker component on a PowerPoint presentation, the component was not inheriting the section title from the previous slide if the current slide did not explicitly define a section.
-
In some scenarios, AutoClassifier was failing to process requests from Connectivity Hub when it was deployed on a server that did not use the en-US regional settings.
-
The Azure Open AI Embeddings Generator component was generating metadata for metadata properties that did not exist in the pipeline.
-
In some scenarios, pipeline stages would perpetually run when specific documents were being tested. (1715804)