2.4.2

Summary

See the Release Notes for 2.4.2 for a comprehensive list of changes since the 2.4.1 release.

Dependency Managment changes

The 2.4.2 Release upgraded some major dependencies within the arcgis package:

Python 3.13
NumPy 2.3.3
timm 1.0.14

It also removed the TensorFlow library as a dependency.

These upgrades provide substantial performance improvements in backend processing. Depending upon the environment, Python and Numpy can leverage different frameworks that can significantly improve speed on fundamental operations such as sorting. These changes are transparent to Python developers and scripters - just continue writing your code as is, but alterations behind the scenes will contribute to better overall performance.

Improvements to GIS Administration and Content Management

Upgrades and new classes within the arcgis.gis module and the arcgis.gis.admin submodule streamline administrative tasks for ArcGIS Enterprise and ArcGIS Enterprise on Kubernetes deployments. The introduction of the AboutManager for detailed inspection of Enterprise deployment components allows for keen oversight into aspects of usage and licensing. New HealthCheckManager, SuitesManager, and ReportManager classes provide a suite of tools for functional and availability checks to assess overall organization health in Kubernetes deployments. Enhanced performance for dashboards and improved handling of various item types like Survey123 Forms and Web Experience items in the clone_items() method continues to advance the ability to migrate complex content. New parameters to import_content() while offline augment the range and control of moving content to a target deployment.

Enhancements in Deep Learning and Machine Learning Capabilities

Introduction of the new model Hyperspectral3DRCNet and support for new backbones (clay, dofa, satlas) for vision tasks in object classification, detection, and pixel classification improve advanced geospatial analysis significantly. Added support for GRAD-CAM for multi-label and multipsectral imagery in object classification and the classify_objects() inferencing method increase understanding of model results. The timm upgrade supports models signifcantly better. It also modernizes access to intermediate feature maps from backbones. This upgrade also allows the Python API to take advantage of dynamic image sizing for other tasks besides segmentation, which should reveal better overall performance for complex tasks like object detection.

Updates to Core GIS Functionality

Focused peformance and data handling improvements for essential GIS data, particularly with geometry and mapping, make for a better analysis experience. Enhancements to FeatureLayer.query() and initializing spatially enabled dataframes with FeatureSet.sdf reduce processing times and ensure correct results. New BasemapStylesService and BasemapStyle classes introduce modern controls for managing basemap styles, languages, and worldviews. The new arcgis.raster.realitymapping provides automation capabilities for full resolution meshes and advanced products to expand on orthomapping capabilities. Numerous bug fixes address issues in cloning hosted feature layers and dashboards and attribute handling for spatially enabled data frames to make the API even more impactful for GIS workflows.

Additional Details

Processing option: Spatially Enabled DataFrame

We've updated the previously titled ARCGIS_GEOMETRY_ENGINE to ARCGIS_IO_ENGINE. It refines the ability to explicitly set the library used for certain data file reading and/or writing operations of shapefiles and file geodatabases. See the API Reference for the GeoAccessor for details. Methods to which this processing applies are noted in the Reference documentation.

You can set the processing environment at the top of a document with:

import os
os.environ["ARCGIS_IO_ENGINE"] = "Engine_of_Choice"

Developer Credentials

New classes:

DeveloperCredentialManager
DeveloperCredential classes

New dataclass:

TokenPrivilege dataclass

These objects work together to control the access and permissions granted to users and applications by defining the specific scope of privileges for an OAuth authentication token. Among other capabilities, this provides SAML-authenticated users full access to all content and administrative tasks. See the platform developer credentials documentation for full conceptual information on the types of items these classes can help manage.

Notebook Workspace Management for Online and Notebook support for Kubernetes

New class:

NotebookFolder

Updated class implementations for:

The NotebookDataAccess and NotebookFile classes have been updated with implementations for ArcGIS Online and ArcGIS Enterprise on Kubernetes notebook workspaces in addition to the previous Notebook Server for ArcGIS Enterprise implementation. The new NotebookFolder class manages capabilities for folders in workspaces. Together these classes provide workspace functionality such as uploading and downloading files, creating folders, and deleting or renaming file content. In addition, these classes allow for transferring of workspaces as well as folders and files to help administrators manager user content prior to removal from an organization.

New `arcgis.raster.realitymapping` Submodule

A new submodule dedicated to automating capabilities and functionality for working with Reality Mapping Project items. See the Introduction to the ArcGIS Reality for ArcGIS Pro extension documentation for full details on creating items that this submodule can manage.

Modernized capabilities for styling Basemaps