Release-1.10

Release 1.10.126 (Maintenance release) - April 18, 2023

1. Bugfix in adectl command line tool where on certain Ubuntu Linux distributions, the custom memory/CPU resource configuration settings for local mode setup was not accepted.

Release 1.10.125 (Maintenance release) - April 14, 2023

1. A link to the support desk for raising tickets has been added to the left navigation panel below.
   
2. On the 'Organization registration' page, a link to the Data Explorer application endpoint is added for already registered users.
   
3. Bug fixes
   1. In some cases, the search space reduction filters applied for similarity search ran the search on incorrect reduced space.
   2. A corner case UI issue when resultset creation was cancelled before saving.
   3. A corner case issue with importing the catalog through adectl when the CSV had spaces in a float type of column.

Release 1.10.120 (Maintenance release) - March 30, 2023

1. Video tutorials are available within the app for a quick introduction to the product features.
   
   
2. Attaching pipelines to datasets filters out the applicable pipelines based on the data type.
   
3. Bugfixes
   1. Fix for catalog-based filtering of points producing incorrect results when multiple conditions are used on the same column.
   2. Additional field validations in weight filter, confidence filter and sample size fields on the job visualization page.
4. Notes
   1. The default landing page is changed from the 'Jobs' page to the 'Datasets' page.

Release 1.10.115 - March 24, 2023

New features

1. Multiple bounding boxes for patch search: You can draw multiple bounding boxes(up to 5) to specify multiple objects to look for in a patch search. The example below has 3 objects selected for search.
   
2. Grouped thumbnails: When viewing thumbnails in the selection panel and the search results panel, similar images are grouped into a folder. The folder can be browsed into to view individual images that were assigned to the group. This feature is available only on new datasets created with this or later releases.
   
3. Support for Video Sequence (BETA): A video dataset is ingested as a sequence of frames. For certain use cases like searching for a specific action, the features of a sequence of frames need to be viewed as an aggregate unit. Towards this objective, the following capabilities on a sequence of frames are now available.
   1. Search catalog with a sequence specification. A sequence specification consists of
      1. Sequence length: Duration of the sequence.
      2. Sequence stride: Gap between successive sequences.
      3. Sequence offset: Time before the anchor frame(the frame chosen by the catalog query) that defines the start of the sequence. This is intended for use cases where the sequence must start a few frames before the frame with a specific attribute in the catalog.
         
   2. View a video sequence against a thumbnail - A sequence-explore type of job can be created from a catalog query with sequence parameters. This type of job provides a control on the thumbnail to view a video snippet for the sequence the thumbnail represents. The sequence-explore type of job will evolve to support clustering, embedding and searches based on sequence aggregate features.
      
      This feature is available only on new datasets created with this or later releases.
4. Increase in size of resultset operations: Support for up to 50000 images/frames to be added/updated into a resultset. This allows certain bulk operations like sampling a 20% core set sample of the points in the job and saving this into a resultset.
5. Fine-grained access control on external catalogs: An external catalog endpoint may have many schemas and tables, and all of these schemas and tables may not be relevant for integration with Data Explorer. This feature allows an OrganizationAdmin role user to whitelist a specific set of schemas and tables to be visible to other users in the organization.
   
6. Delegated role-based cross-account access to AWS resources: Support for recommended best practice of using delegated role-based access to AWS hosted resources like S3 buckets, ECR repositories, Athena endpoints etc., that are integrated with Data Explorer. The delegated role details are configured as a new type of secret.
   
   
7. Support for Github as a single sign-on provider.

Enhancements

1. Job page enhancements
   
2. New design for full-resolution image view with easier navigation through different catalog attributes.