Skip to main content

Release Notes

Release Notes

Version 3.2.0

  • EasyRPA Automation Process Development Framework (ADF) is realeased and documented with sample
  • Third-party services version levels up
  • Various platform bugfixes

API/Engine

  • Performane improovment for AP with big number of task. Split 1000x8 tasks 3h40m → 25m
  • Screen driver updated with ScreenSearch.text(String text) method
  • JavaDriver various fixes. Now it supports JavaFX

ML

  • LLM HOCR IE model with customizeable promts config added

OCR

  • Performance degradation on 4CPU workaround provided

Samples

  • Java Sample update
  • Datastore/DocumentSet indexes are aplied to all samples
  • LLM model DocumentSet sample

Documentation

  • Academy update
  • ML/OCR documentation update

Version 3.1.0

  • EasyRPA Automation Process Development Framework (ADF) PoC
  • Third-party services version levels up
  • Various platform bugfixes

API/Engine

  • WebBrowser Driver implementation has been switched from Selenium 3 to Selenium 4
  • Page elements support added to page objects.
  • Datastore field encryption support added
  • Datastore engine various improovment and bug fixes
  • Split mechanizm has been reworked, now it supports maximum number of paralell threads from all split within AP.

Human Tasks

  • Various bugfixing

Samples

  • Text IE Samples updated to use tabular like text invoices
  • EasyRPA Automation Process Development framework demo application 

ML

  • Various bugfixes with HTML IE model, now it support Text tabular Information Extraction

OCR

  • OCR pipline has been update to support the target images format JPG, PNG, an so on. It could improoves OCR results on diffical images.
  • OCR pipline has been update, now it support ImageMagick 6 or 7 and dynamic command line arguments

Version 3.0.0

  • Third party libriary and services versions are upgrated to decrease number of vulnearabilities.
  • Platform Java switched to version 17 (JNLP is not supported anymore)
  • Login user attemts record added into Audit logs
  • Reduced platfrom package installation size
  • Platforms third party services that used additional ssl ports: 8443 (Kinana), 8445 (Minio UI), 8446 (InfluxDB) are now proxied by nginx from 443 port
  • All CS opertions with minio using external 8444 port are internalized inside docker network, the 8444 port is optional now, and could be used for external s3 clients and additional OCR container on another machine.
  • The datastore versioning mechanizme anounced using usecase.properties

API/Engine

  • Various bugfixing

Human Tasks

  • Various bugfixing

Samples

  • New "Accounting Information Extraction from Multi-Page Reports" sample added. It demostrates ML/OCR post processing posibilities.
  • New "Information extraction from TEXT" sample added. It demostrates posibility to trocess pure text document using HTML IE Document Processing
  • IE HTML Sample reworked to use more meanefull documents
  • JNLP samples removed becaue it is not supported by java 17 anymore
  • OCR Sample added. It demostrates posibility how to use OCR scripts.
  • Debtor Analysis sample removed

ML

  • various bugfixes with HTML IE model, now it can process TEXT documents
  • ml_ie_spacy2_model updated, now it uses spacy-lookups-data, that improoves entity recognitions, but increase (~50%) training time.

OCR

  • Tesseract switched on version 5
  • Shell scripts pipeline support added

Version 2.9.1

  • APS Document Processor project reworked to support new PostProcesors ideology
  • Platforms OOTB PreProcessors, PostProcesors , Validators and Parsers for Document Processors added for all Human Task Types
  • Various performance improvement and bugfixing in API/Engine, Control Server, Node Agent, Human task

Samples

  • ML document samples updated by post processors and validators
  • Signature sample updated by signature verification

ML

  • All Models has been re-released with fixed 3d party dependency versions
  • IE HTML model bugfixing - now it correctly handles pre tag
  • IE HTML Auto-Training supports MATCHED and CONTAINS tags

Documentation

  • Document Processor articles updated with PostProcessing guides

Version 2.9.0

  • Worked dashboard added
  • Dedicated node page reworked
  • ML tools project reworked to have all resources in one place. Now easy-rpa-aps AP package import brings all ML resources
  • Model import reworked, now it is possible to create model from python package existing in platform nexus
  • A custom attribute could be defined in document type settings entities

API/Engine

  • Split task mechanism reworked. Now it takes less thread resources. Now the splited flow is not blocked by Human Task (the new remoteExecutionService.maxBatchSize parameter configured is added).
  • Standalone platform package generation added
  • Indexes and columns could be defined for document set in the Automation Process package
  • Fixes in Error Handling - now the same error code figurates in AP an CS logs

Node Agent

  • Node Agent refactored to use less thread resources

Human Tasks

  • Multi document human task sample added

Samples

  • Signature detection sample added
  • Invoice Plane sample reworked
  • Dedicated Automation Project sample added
  • Google Translate RPA sample added
  • HT sample updated with multi document human task sample

ML

  • Signature detection custom model added
  • IE ML post processors configured by JSON PoC added
  • All model re-released to fix libraries version

Infrastructure

  • public samples code repository is added
  • public platform documentation site added
  • ML Document Processing developer guide improved

Version 2.8.0

  • Additional OCR installation package is generated during installation procedure. Now is is possible to add more OCR containers to another machines.
  • Authorization system reworked - cookies is using in additional with tokens
  • CS API Context subsystem, an improvement that optimizes CS file import/export requests, big data upload is now more stable.
  • Add/Remove Datastore/DocumentSet indexes 
  • Add/Remove Datastore/DocumentSet columns
  • Node Health status icon in Nodes list
  • CS search filter UX update
  • RPA browser selenium driver auto-update on node machine
  • All CS download switched on browser native downloads, file upload % is displayed during file uploading
  • Task history - input/output view now shows hierarchical JSONs
  • Document Processors reworked to have more usable code to reuse in customers APs
  • Logs obtaining performance improvement

API/Engine

  • Datastore/DocumentSet indexes 
  • New sugar methods UIElement
  • Expected Conditions reworked and now are the same for all driver
  • Driver API methods reworked
  • All waiting methods in drivers and Expected Conditions are reworked
  • Page objects switched on @Wait for all drivers
  • Page API method reworked to support wait and expected conditions elements search
  • SELENIUM_DIRECT node feature released
  • @Driver(exclusiveDesktopAccess) anounced

Human Tasks

  • Human Task security update - now Human Task access to CS API using current user. As a result - the no public bucket on minio for Human Task
  • HTML Human Task input json reworked to have long context stored on minio
  • Fixes in HTT Form document type validation
  • IE Human Task performance optimization

Samples

  • Al ML samples has been reworked to use document processors code as a base

Version 2.7.0

  • Model autotraining has been added: clicking the Autotraining icon divides the documents in the document set automatically into training and test sets, based on the settings in the document set. The new model is trained on the training set and then tested on the test set.
  • A new button, Create Model, has been added, which allows users to create models in which versions will be stored.
  • Models module has been redesigned: models and their versions are shown in a hierarchy.
  • Active model selection functionality has been added: the user can choose the best active version of the model, or the best model can be determined automatically by comparing the autotrained model with the active model.

  • It is now possible to download resources for the model version. Reports for autotrained models can now be downloaded.

  • Added the ability to exclude documents from autotraining processes using the Invalid Document checkbox.

  • The Document Set module now includes a Metrics tab. Autotrained models can be evaluated using the graphs on this page.

  • UX improvements has been made to Document Set record view to increase readability of document fields.
  • Notification channels have been updated. Multiple recipients can now be added on the same line with delimiters set in the control server configuration.
  • An action confirmation message has been added when selecting an overwrite strategy for a data store or document set.

  • Advanced filters are now saved between sessions.

  • The cross icon can now be clicked to close informational messages.

  • The history of informational messages can be accessed via a modal window.

  • The advanced text filter can now be configured to include multiple conditions.

  • A migration was made to Grafana version 9.3.2.

  • All entity modules have been updated with the Created By/Creation Date columns.

Human Tasks

  • The ability to combine fields into blocks has been added to Forms Human Task.

API/Engine

  • Improvement in data store API - added custom queries.

  • As a result of an update to the Engine API, it is now possible to save the results of human tasks to a document set.

  • The parallel processing of tasks during an automated process launch has been improved.

Samples

  • A new version of the IDP Sample has been released - human tasks results are now saved to the document set.

Version 2.6.0

  • API Secrets for the Control Server API has been announced.
  • UX improvements for group management have been made. Multiple contexts/context names can be added in one role creation operation using checkboxes.
  • Platform Health Check service has been announced.
  • New Puddle OCR engine has been added to platform, so users have more options for document recognition.
  • VirtualBox images for developers has been announced.
  • Admin users now have access to ActiveMQ's Admin UI.
  • UX improvements has been made to Data Store record view to increase readability of records.
  • There have been various improvements to Document Processors in order to reduce Task Input JSON size.
  • Migration from 2.5.0 to 2.6.0 has been announced.
  • In order to improve the stability of the platform, a number of bug fixes and changes have been made.

Human Tasks

  • File attachments are now possible with Forms Human Tasks.
  • In order to make document tagging more convenient, an icon has been added to human tasks that allows switching to full-screen.
  • There have been several fixes and improvements to Classification Human Task.
  • The Information Extraction/Classification human tasks have been reworked in order to reduce the size of Task Input JSON. 

API/Engine

  • Implementation of storage service was reworked - there are no more timeframe restrictions for tasks.
  • The possibility of setting a dynamic subject in notifications has been announced.
  • API for FindBy annotation has been reworked - now all drivers can fully use it.
  • The Query annotation has been reworked so that it now supports different types of returns.
  • There are new search attributes available in DesktopSearch.
  • The task error timeouts have been announced.
  • Task local run feature has been announced.

Samples

  • Java Oracle Form sample process has been implemented and added to Nexus.
  • Updated APIs have been applied to all samples.
  • Java samples have been updated to run with minimum configuration work on the node.

Version 2.5.0

  • Metrics representations are reworked, now they are competed Grafana dashboard and it layout can be managed
  • Custom dashboards view added
  • Migration support added

Engine

  • Various changes in API
  • Split engine API and engine implementation
  • Merge task rule reworked, API changed
  • Parallel task execution performance improved
  • Various bug fixing

Version 2.4.0

  • Multiple dedicated nodes can now run on the same machine.
  • Option to use lemmatizer in ML model training configuration was added.
  • Automation Process history representation was reworked from paths to graphs. 
  • Dynamic language loading in spacy model trainers was added.
  • New HTML Information Extraction and HTML Classification samples were implemented in order to demonstrate automatic HTML document processing.

Engine

  • Parallel task execution is now really simultaneous.
  • Business process development API is separated from CS core allowing to run the same processes on different versions of CS (assuming the API are compatible).

Human Tasks

  • Dynamic (defined by business process per document) values for checkboxes, listboxes and radiobuttons in Human Task.
  • Human tasks can now process HTML documents (in addition to PDF/image).
  • A timer was added to the Human Tasks.

Control Server UI

  • Document types now include context help when creating or editing them. Clicking the question icon will redirect the user to the appropriate document type settings section depending on the selected Human Task Type.
  • Depending on the Human Task Type selected for creating and editing Document Type, a default template appears in the settings window.
  • A default template for OCR settings was added to the settings window when creating and editing a Document Set.
  • Validation of Document Type settings has been added. If the user makes a mistake when editing or creating Document Type, the corresponding error message will be displayed and the system will not allow updating/creating Document Type with incorrect settings.
  • The JSON editor has been redesigned with the following features: syntax validation, minimap, JSON search, word wrap, data formatting.
  • The number of tasks that were executed during the Automation Process run is now displayed.
  • Now you can view the details of errors that occurred during dependency scanning for Automation Process.
  • A reworked model training dialog simplifies the training process. Default training configuration is generated automatically now in accordance with selected Training Document Set.
  • Special labels were added to development runs.
  • New additional columns were added to CS Logs, Node Logs, AP Run Logs, Run Task Logs and Node Feature Logs.
  • Multiple independent groups/departments can now be organized using the Execution groups.
  • Users now have the ability to adjust logging levels for specified packages directly from the UI.
  • Instead of separate graphics, metrics are now linked to the entire dashboard.
  • Now that the mechanism for downloading logs has been improved, it is possible to download filtered logs as well.
  • Search in task details now allows searching by Task Input and Output JSON data by following the search syntax.
  • The log search has been improved, you can now filter the required logs using the Log Search Syntax rules.
  • Now, the user has access to the logs for all containers through the Logs module

Infrastructure

  • The pipeline has been modified to create an image that comes with EasyRPA pre-installed, allowing local development.
  • Improved Export plugin that allows easier development and streamlines CI/CD. New possibility to include storage content in AP package.
  • Virtual Box image created for local development.
  • Data Purge functionality was implemented. The interval for automatic data cleanup can now be controlled via CS Configuration: Runs, Logs and Activity Logs.

Version 2.3.1

Fixed vault initialization issue causing login impossibility.

Version 2.3.0

OAuth

  • The OAuth server has been added into platform. As part of the migration to OAuth, the authentication process for nodes, swagger, Minio and Grafana has also been modified.
  • The Minio UI has been updated in conjunction with Oauth-related updates to enhance the user experience.
  • For the Group Management module, a new Third Party Roles tab has been added that allows you to configure user access to 3rd party applications. Currently, Third Party Roles include roles to manage MinIO and Grafana.

Metrics

  • EasyRPA metrics are announced. InfluxDB, Grafana and Telegraf containers have been added into platform.
  • Grafana preconfigured with EasyRPA dashboards.
  • A metrics tab was added for Node, Automation Process, Automation Process Run, and Document Type. 
  • A monitoring page was has been added that allows you to view the actual state of the Control Server, ML, and OCR containers.

Control Server UI

  • Context help system has been reworked. Now it contains search.
  • All modules of EasyRPA now have feature advanced filters that allow you to retrieve data according to a variety of detailed criteria.
  • There is now a dropdown called "Table settings" available on the platform for all tables. You can now turn on/off the display of advanced filters for columns and select which columns will appear in the table by using this dropdown. 
  • The login page UI was redesigned so that it is more attractive for a user.  

Human Tasks

  • The "Start Available Task" button has been added to the Workspace module. The user can now start any available task from either the general queue or the queue for a specific Document type, depending on where the button was clicked: at the top of the page or in the row next to the Document type name.
  • Skip Human Task task logic was implemented in Workspace.
  • The following Human Task display option has been added to the Workspace in order to reduce the information displayed by default:
    • Using the Show Empty Document Types toggle, you can exclude Document Types that have no tasks from the Workspace list.
    • Using the Hide Skipped toggle, you can exclude Skipped Human Tasks from the Workspace list.

Document Sets

  • For Document Sets module, fields filled in the More tab are now supported. The new "Synchronize Metadata Columns" option adds columns from the Metadata section and populates the fields based on the data entered by the user. 

Human Task Type and Document Type

  • In order to make it easier to understand the hierarchy and logic of the different Document Types, Human Task Types, and Human Tasks, the Human Task Types module was reworked and split into two modules: Human Task Types and Document Types 
    •  Document Types view was added to Navigation.
    •  Tab Document Types was removed from Human Task Type details.

Framework

  • Automation Process Capability assignment logic was reworked and improved.
  • Fixed bug in delayed in next execution in dedicated node.
  • Parallel Task execution logic reworked and improved.
  • Retry with parallel task threads execution issue is fixed

Version 2.2.4

Fixed another vault initialization issue causing login impossibility.

Version 2.2.3

Fixed vault initialization issue causing login impossibility. 

Version 2.2.2

Log4j vulnerability fix for Control Server, Node Agent,  and EasyRPA Engine API.

Version 2.2.1

Control Server UI

  • The document set data export/import UI has been updated to make it more visually understandable for the user. In this version, it is clarified that the document set export/import only supports data migration, but not all artifacts related to Document Set.
  • The initial automation process package can now be exported for automation processes that have been uploaded to Control Server through the package import.

Human Tasks

  • Added the ability to display and fill a configured metadata form in ML Tasks (Information Extraction and Classification) in the More tab of the Document Set and Workspace modules.
  • The feature to display the metadata form in ML Tasks (IE and CS) in the More tab of the Document Set and Workspace modules has been added.
    • There is an option to check the "Invalid" checkbox, which disables fields validation when saving the document and excludes the document from the following steps of the automation process. 
    • IDP Sample have been updated to include and transfer data generated by filling out the metadata form into Workspace
  • HT Classification and Information Extraction interface was improved to make it more clear and easy to use. 
  • The following Human Task display option has been added to the Workspace in order to reduce the information displayed by default:
    • Show Completed Tasks toggle allows excluding documents in Completed status from the list in Workspace.
  • A new dialog box for selecting a model when creating a new Document Set or editing an existing one has been added.
  • It is now possible to display messages in Human Task in three different statuses (info, warning, and error).

API/Framework

  • Split/merge functionality has been implemented and added to the API/framework: the new "split" and "merge" methods significantly reduce the complexity of asynchronous processing of task/data groups.
    • IDP Sample was updated in accordance with the new split/merge functionality.