What's new in version 2.4.0

Note: The ArcGIS API for Python 2.4.0 release is supported with ArcGIS Pro 3.4 and later cloned environments for use with arcpy. The 2.4.0 arcgis and arcgis-mapping packages are incompatible with ArcGIS Pro 3.3.x and earlier cloned environments. The default ArcGIS Pro 3.3.x environment can be cloned and upgraded to arcgis packages of 2.3.x.

New

Reorganized modules and refactored functionality into new classes to incorporate the new Jupyter Lab and Notebook 7 architectures.

Guides

Samples

Module Changes

arcgis.gis

  • GIS
    • Adds documentation:
      • to class description for retrieving a token
      • to token keyword for retrieving a token
  • Item
  • ContentManager
    • clone_items()
      • Enhances processing to export items in cases where it applies to prevent excess server processing and 504 errors
  • Folder
    • add()
      • Adds deprecation notice when using the overwrite parameter as it has been deprecated in the REST API
  • UserManager
    • create()
      • BUG-000168602 - Adds documentation note for using the org_admin value to assign built-in Administrator role to a user
  • GroupMigrationManager
    • load()
      • New parameter:
        • keep_epk_item

arcgis.gis Dataclasses

arcgis.gis.admin

arcgis.gis.server

arcgis.gis.agonb

arcgis.gis.workflowmanager

arcgis.features

arcgis.features.managers

arcgis.raster

arcgis.raster.functions

  • speckle()
    • Adds support for RefinedLee and GammaMap as filter_type arguments

arcgis.raster.utils

arcgis.raster.orthomapping

arcgis.notebook

arcgis.geoanalytics

  • Removed from the 2.4.0 code base. Supported only in 2.3.x releases.

arcgis.geocoding

  • suggest()
    • Deprecates parameters:
      • distance - use search_extent
    • Adds parameters:
      • preferred_label_values
      • return_collections

arcgis.geoenrichment

  • Enhances error messaging when attempted in environments without ArcPy or Shapely

arcgis.graph

arcgis.layers

arcgis.map

arcgis.widgets

arcgis.apps

arcgis.learn

arcgis.datapipelines

Fixes

Guides

  • Fixes display issue in documentation where Item Type icon appeared incorrectly when displaying Item cards
  • Fixes display of internal information on User cards
  • Fixes display issue with Group card in notebooks

Module Changes

arcgis.gis

  • GIS
    • Fixes issue where session not created if certificate Common Name does not match Server Name
  • Item
    • Fixes inconsistency with return values for layers and tables property for Map Service items
    • Fixes issue with improperly item objects from CodeType items based on Python version
  • SharingGroupManager
    • add()
      • Fixes BUG-000168928 where method retuns True only for first Group added in a sequence of calls
  • ContentManager
    • clone_items()
      • Fixes issue when cloning StoryMap items containing web maps on Windows not properly cloning all resoures
      • Fixes BUG-000168118 where cloning Web Experiences created through the Create Web App button on Web Map item details page
      • Fixes Public Repo Issue #1761 where survey content was causing failures because of character encoding
      • Fixes BUG-000167100 where a TypeError raised when cloning Dashboard items of latest version
      • Fixes BUG-000166690 where read-only fields are not properly populated
      • Fixes issue where editor tracking information was not cloned in feature services
    • import_data()
      • Fixes optimization issues when using overwrite
      • Fixes Analyze Service Error in environments where ArcPy is present and service source name generated from file is not unique
  • Folders
    • list()
      • Fixes performance issue by reducing number of calls to endpoints
  • Folder
    • add()
      • Fixes TypeError when file input is a io.StringIO or io.BytesIO object
      • Fixes ValueError if file, text, url, or data_url argument not provided
  • UserManager
    • create()
      • Fixes BUG-000168590 where user_type value from New Member Defaults settings not honored when user_type not provided

arcgis.gis.nb

  • Runtime
    • update()
      • Fixes FileNotFoundError because of incorrectly formatted post request

arcgis.features

  • FeatureLayer
    • edit_features()
      • Fixes issue where no response returned when run asynchronously using the deletes argument
  • Spatially Enabled DataFrame - GeoAccessor
    • centroid
      • Fixes BUG-000164784 where ValueError returned in an ArcPy environment on dataframe containing geometries with z values
    • from_featureclass()
      • Fixes AttributeError when accessing wkid property on geometries
    • to_featureclass()
      • Fixes BUG-000162383 where Utility Network layers failed to export because numeric column values of 0 were getting set to Null
    • to_featurelayer()
      • Fixes optimization issues when overwrite=True
      • Fixes BUG-000169830 where m-values are not retained
    • from_layer()
      • Fixes issue where date values were read incorrectly and returning all values at beginning of Unix Epoch

arcgis.features.managers

arcgis.raster

  • ImageLayer
    • save()
      • Fixes BUG-000167021 where misleading error message returned in service property of allow copy is set to False
  • Raster

arcgis.raster.utils

arcgis.network.analysis

  • get_tool_info()
    • Fixes TypeError because of passing an additional parameter

arcgis.notebook()

arcgis.geoanalytics

  • Removed from the 2.4.0 code base. Supported only in 2.3.x releases.

arcgis.geocoding

arcgis.geometry

  • Updates how GeometryService is accessed within the API for increased flexibility in different deployments
  • Geometry
    • has_z
      • Fixes BUG-000169189 where property returns incorrect results in environment with Shapely only
      • Fixes issue where Polygon objects were missing property if initialized from a feature class

arcgis.geometry.functions

  • Fixes issues multipart polygon processing by spatial operations performed inconsistently between environments with Shapely and ArcPy

arcgis.layers

arcgis.apps

arcgis.learn

  • Fixes out of memory issues by providing backbone support during model initialization when batch size is being estimated
  • Fixes error in format of default value for TTA scales used in inferencing
  • Fixes error with show_batch() when training data had fewer elements than the batch_size argument
  • Fixes ValueError failures on most models with predict_video()
  • Data Preparation Methods
    • prepare_data()
      • Fixes error initializing PSETAE model from output when class doesn't contain any files
      • Fixes incorrect recursion error raised if working_dir argument does not exist
    • prepare_textdata()
      • Fixes issue where similar entities are extracted multiple times using Mistral
  • Automated Machine Learning
    • Fixes issue where early stopping was triggered before adequate processing time elapsed on time calculations
    • AutoML
      • Fixes crash when training using prepare_tabulardata object created using explanatory or predict variables
      • fit()
        • Fixes KeyError: 0 error
      • report()
        • Fixes FileNotFound error when run after a call to save in Jupyter notebook
  • Object Classification Models
    • FeatureClassifier
      • from_model()
        • Fixes failure to load model because of incompatibility between supported backbone and saved backbone names using torchgeo
      • save()
        • Fixes NotADirectoryError using torchgeo
  • Object Detection Models
    • Fixes warning messages received during initialization
    • SingleShotDetector
      • Fixes issue when initializing model from data bunch without specifying backbone
      • from_model()
        • Fixes compatible backbone exception using torchgeo
      • Fixes performance issue for objects of varying sizes related to ordering of grids
    • MMDetection
      • fit()
        • Fixes info messages printing during training of vfnet models
        • Fixes issue resulting from incorrect available_metrics for dino detr models
  • Pixel Classification Models
  • Image Translation Models
    • Pix2Pix
    • Fixes backbone names for backbone parameter description
  • 3d Models
    • Fixes RuntimeError and memory leak when using compute_precision_recall() on large datasets:
  • Scanned Maps
  • Unstructured Text Models
    • arcgis.learn.text module
      • Fixes issue with path of the emd file not properly transferred to the extension file
      • Fixes performance issues for models using the Mistral backbone
      • EntityRecognizer
        • extract_entities()
          • Fixes issue where fails to pull out information from text while inferencing using Mistral
        • from_model()
          • Fixes EntityRecognizer not present issue on model created from ner_json dataset type
      • TextClassifier
        • from_model()
          • Fixes UnboundLocalError resulting from incorrect passing of parameters to extension model

Your browser is no longer supported. Please upgrade your browser for the best experience. See our browser deprecation post for more details.