What's new in version 1.7

New

Guides

Imagery and raster analysis

• Orthomapping (part 1) - creating image collections
• Orthomapping (part 2) - generating elevation models
• Orthomapping (part 3) - managing image collections

Deep Learning with ArcGIS

• How does feature categorization work
• Categorizing features using arcgis.learn

Samples

• Pawnee Fire analysis
• Finding grazing allotments
• Extracting building footprints from drone data
• Detecting super blooms using image classification
• Assessing building damage using feature classification
• Analyzing New York City growth factors

Module Changes

• Adds ability to run Geoprocessing tools asynchronously with new GPJob class. See future parameter in:
• arcgis.features submodules
• arcgis.raster.analytics module
• arcgis.network.analysis module
• arcgis.geoanalytics submodules

arcgis.gis

• Adds manager classes:
• ProfileManager
• CertificateManager
  • provides missing certificate management in ArcGIS Online
• InvitatationManager
• Adds support to add a geopackage file
• Adds folder parameter to import_data()
• Adds reset_by_email parameter to user.reset() to allow user to reset their accounts by email
• Adds replace_metadata parameter to replace_service()
• Adds support for hosted tile layers and hosted scene layers to replace_service() operation
• Adds delete_thumbnail() to user class

arcgis.gis.server

• Adds hardware property to Machine class
• Adds ItemInformationManager to for administration
• Adds new federate_data_item. See Add To Server
• Adds change_provider() method to Service
• Adds synchronize() method to Machine class

arcgis.features

• Adds analysis tools
• generate_tesselation analysis tool
• create_route_layers
• Adds spatial_filter parameter on Spatially Enabled Dataframe.from_featureclass() method
• Adds missing in_sr parameter to extract_changes documentation
• Adds return_delete_results parameter to delete_features method

arcgis.raster

• ImageryLayer.save() function
• Adds a cache_manager to ImageryLayer

arcgis.raster.analytics

• New multidimensional functions:
• generate_multidimensional_anomaly()
• aggregate_multidimensional_raster()
• build_multidimensional_transpose()
• find_argument_statistics()
• generate_trend_raster()
• predict_using_trend_raster()
• linear_spectral_unmixing()
• subset_multidimensional_raster()
• costpath_as_polyline()
• New parameters:
• Adds statistics_type parameter to flow_distance function
• Adds and context parameter to raster analytics functions to enable setting of function-specific processing
• Adds destination_field parameter to determine_travel_costpath_as_polyline
• Adds process_as_multidimensional and build_transpose parameters to generate_raster and copy_raster
• Documentation improvements:
• add_image
• delete_image_collection
• create_image_collection
• Adds ability to input a list to list_datastore_content()
• Fixes error when using keyboard to cancel a raster analytics geoprocessing job
• Adds support for item as input to aggregate_multidimensional_raster()

arcgis.raster.functions

• Adds the WNDWI predefined algorithm to band_arithmetic function
• Adds ability to accept list of integers or string as value for band_indexes parameter in band_arithmetic function
• Adds the process_as_multiband option to:
• local()
• majority()
• max()
• mean()
• mid()
• min()
• minority()
• cellstats_range
• std()
• sum()
• variety()
• New functions:
• constant_raster()
• random_raster()
• Enhances focal_statistics() and focatl_stats() documentation

arcgis.raster.functions.gbl

• New functions:
• euclidian_back_direction
• flow_length
• expand
• distance_allocation
• distance_accumulation
• shrink
• sink
• snap_pour_point
• Adds in_barrier_data parameter to:
• euclidean_allocation
• euclidean_distance
• euclidean_direction
• Adds generate_out_back_direction_raster, in_barrier_data, and distance_method parameters to calculate_distance
• Adds default value of MEAN to zonal_statistics() statistics_type parameter
• Adds process_as_multidimensional capability to zonal_statistics()

arcgis.network.analysis

• Adds save_route_data parameter to find-routes and find-closest-facilities tasks

and to_featurelayer methods

• Improves pixel_size parameter documentation for appropriate ImageLayer operations
• Adds zoom_to_layer function
• Adds new network analysis functions
• Adds edit_vehicle_routing_problem

arcgis.geoanalytics

• Documentation improvements:

  • Adds

  • find-point-clusters

  • Enhances

  • calculate_fields

  • aggregate_points

  • build_multivariable_grid

  • find_hot_spots

  • create_buffers

  • dissolve_boundaries

  • detect_incidents

  • calculate_density()

  • aggregate_points()

  • forest

  • glr

  • run_python_script

  • copy_to_data_store

  • overlay_data

  • find_similiar_locations

  • enrich_from_grid

  • merge_layers

  • join_features

  • summarize_within

  • summarize_attributes

  • reconstruct_tracks

  • geocode_locations

arcgis.apps

• Adds new parameters to build_collector_url. See Deploy your map for parameter details.

arcgis.learn

• New models:
• RetinaNet
• EntityRecognizer
• PSPNetClassifier
• MaskRCNN
• Documentation improvements:
• export_training_data
• Adds option to UnetClassifier to not predict the background class
• Adds predict() and predict_video() methods to SingleShotDetector
• Adds plot_hard_examples() function
• Adds padding support to SingleShotDetector
• Adds ability to train and inference on attachments in a Feature Layer using FeatureClassifier
• Adds ability to train detect_objects model with any dataset
• Adds categorize_features function and enhances to work with local rasters and feature classes
• Adds Mobilenet backbone support to FeatureClassifier and SingleShotDetector
• Adds SingleShotDetector.from_model() method
• Adds resize_to paramter to prepare_data() function
• Adds support for full range of multispectral imagery
• Adds support for visualizing training progress
• Changes lr parameter to optional in fit() methods of models
• Adds from_model() method to UnetClassifier
• Adds FeatureClassifier inferencing when using ArcGIS Pro
• Adds support for multispectral imagery to Unet type models for inferencing when using Classify Pixels using Deep Learning

Fixes

Module Changes

arcgis.gis

• Corrects max_results parameter documentation for the search() method of UserManager
• Fixes issue in ArcGIS Online that a user who does not own public data could not share() an Item to group
• Fixes issue where from arcgis import * fails with AttributeError regarding find_centroids or summarize_center_and_dispersion
• Fixes issue with setting group.protected = False on a group whose protected property is already False
• Fixes error when returning a Pandas dataframe with item.usage(date_range='6m'
• Improves messaging when trying to read a File Geodatabase with environment that does not contain ArcPy and fiona packages
• Fixes issue in ContentManager.add() if data contains multiple files
• Fixes invite() method on ArcGIS Online UserManager instance to accept email parameter
• Fixes known issues when using PKI
• Fixes publising error with items of type Compact Tile Package
• Fixes issue with item.usage failing on non-hosted layers in ArcGIS Online
• Fixes advanced_search() method to properly return results based on sort_field

arcgis.gis.server

• Fixes issue with ItemInformationManager to retrieve service metadata
• Fixes BUG-000112855 AttributeError thrown when trying to rename service

arcgis.features

• Fixes error with Spatially Enabled DataFrame to_featureset() method failing if DataFrame contains non-unique index
• Improves ouput for Spatially Enabled DataFrame.to_featurecollection()
• Fixes issue with HeatMapRenderer not honoring color breaks (See Spatially Enabled DataFrame.plot()
• Fixes TypeError when using FeatureLayerCollection.query_domains()
• Improves FeatureLayer.query() performance with large result sets

arcgis.features.analysis

• Fixes multiple functions to write output types
• Fixes issue where create_route_layers() only returns a single layer from multiple result layers
• Fixes error with find_nearest() when input is a Feature Collection
• Fixes handling of the travel_mode input parameter for:
• find_nearest() improperly
• choose_best_facilities()
• connect_origins_to_destinations()
• create_drive_time_areas()
• Fixes aggregate_points failing when output is a FeatureCollection
• New parameters:
• route_shape in connect_origins_to_destinations()
• travel_direction, show_holes, and include_reachable_streets in create_drive_time_areas()

arcgis.raster

• Fixes export_image failure on images with a raster function template containing a constant function that requires input RasterInfo applied
• Fixes issue where a Raster Function Template fails to dynamically render output when federated servers serve multiple server functions
• Fixes issue when service_url parameter used to ImageLayer.rasters.add() rasters to ImageryLayer
• Fixes issue with get_samples() failing with large payloads
• Fixes to how convert_raster_to_feature() reads input properties
• Fixes to how train_classifier() parses input layer

arcgis.raster.analytics

• Fixes issue with classify and segment not honoring extent
• Fixes error with optimum_travel_cost_network() when using output_neighbor_network_name parameter

arcgis.raster.functions

• Fixes issue where raster functions errored out when displaying results of raster functions with private image services from different Web GIS portals
• Fixes issue in raster functions where jobs did not honor rendering_rule setting
• Fixes curvature function outputs
• Fixes gvitm documentation. (See Green Vegetation Index at raster function objects for reference)

arcgis.geoanalytics

• Fixes issue with geoanalytics tools failing to write output to Big Data File Shares

arcgis.geoprocessing

• Improves geoprocessing.import_toolbox() load time
• Fixes geoprocessing functions to accept item of type Image Collection

arcgis.mapping

• Fixes issue when WebMap.save() would not reproject data to correct spatial reference if the view when saved was outside the range

arcgis.widgets

• Fixes issue with map widget.save() not saving layer input as a dictionary with map widget.add_layer()
• Fixes map to add Raster from a datastore and display output from a raster rendering service
• Fixes issue with map widget not displaying in JupyterLab

arcgis.learn

• Fixes issue when prepare_data() would now work with output from LabelImg tool
• Fixes SingleShotDetector to work with all combinations of grid sizes
• Improves performance for arcgis.learn by making use of all GPUs present on a device
• Improves SingleShotDetector to create grid size, aspect ratios, and zoom by reading dataset
• Fixes UnetClassifier to accept 0 for background value
• Fixes SingleShotDetector training error when resizing empty tensor in batch
• Fixes known issue with object detection padding
• Fixes issue with show_results() improperly visualizing results with multispectral data using the UnetClassifier