IQ Bot Version 11.3.5.3 Release Notes

Review the changed features, fixed features, and known limitations in Version 11.3.5.3 of Automation Anywhere IQ Bot. There are no new features or security fixes in this release.

Changed features

View confidence level of extracted fields (Service Cloud case ID: 00648941)

You can now view the confidence level of the fields extracted using the Tesseract OCR and Tegaki API OCR engines in IQ Bot. This information is available in a JSON file in the following output directory: C:\Users\Public\Documents\Automation Anywhere IQBot Platform\Output\<Learning Instance>\.

IQ Bot 11.x: Improve output quality using OCR confidence

Automatically detect check box

IQ Bot can now automatically detect check boxes and radio buttons in a document. You can enable this feature when creating a learning instance.

Tegaki API OCR improvements
IQ Bot now uses enhanced Tegaki API OCR segmentation to improve accuracy in the following scenarios:
  • Multiple fields are merged into a single SIR
  • Segmentation on empty spaces
  • Overlapping SIRs
  • Incorrectly located SIRs
  • Partial or incorrectly captured values from SIRs

IQ Bot 11.x: Use Tegaki API OCR engine

Custom encryption key (Service Cloud case ID: 00656823)

During a new installation of IQ Bot, you can now specify a custom encryption key.

IQ Bot 11.x: Installing IQ Bot in Custom mode

Performance improvement (Service Cloud case ID: 00629302)

IQ Bot now takes significantly less time to create a bot for groups containing only production documents.

Timeout limit for extracting data from tables (Service Cloud case ID: 00674110, 00662968, 00679587)

Data extraction from a table sometimes stops working indefinitely if the training is not done correctly. To avoid such a scenario, the timeout limit for table data extraction is set to 30 minutes. After the timeout (30 minutes), the document is moved to Validator with empty table rows, where you can manually add the data.

Fixed features

Service Cloud case ID Description
00653929 With segmentation improvements, you can now train Korean language documents using the Tegaki API OCR engine without any system-identified region (SIR) errors.

IQ Bot 11.x: Use Tegaki API OCR engine

00430923, 00618234, 00636422, 00662137, 00478986, 00414921

All uploaded documents are now rendered correctly, irrespective of their content orientation while uploading, which enables you to train them without encountering any issues. Previously, when you created a learning instance, documents that did not have the correct content orientation (they were upside down or tilted) could not be trained.

00639117 Numeric format in the Polish language is now accepted in Designer and Validator.

IQ Bot 11.x Designer validation patterns

00590165 Fixed an issue that caused the following error: Cannot read property 'startPx' of undefined. You can now see the extraction results and save your documents without any errors.

IQ Bot 11.x: Preview extracted data

00500914 The TesseractLog.log file is now available in the following directory: C:\Users\Public\Public Documents\Automation Anywhere IQBot Platform\Logs\Engine\

The log level is set to error by default. You can change this setting in the CognitiveServiceConfiguration.json file.

00445378 You can now create and edit learning instances even if the role name consists of Unicode characters in Chinese, Japanese, and Korean languages.
00459301, 00677204, 00532325 Custom Python logic now displays extraction results in the Designer without any errors.
00619139 When you now run a bot and try to edit it while it is still running, the following correct message is displayed: Staging Documents for this Vision bot is in progress. Please try after some time. Previously, the following incorrect message was displayed: Learning instance not found
00668625 When you upload documents containing digital signatures to learning instances, Designer now displays the correct success or error message.
-- When you now select Hide successful fields, successful table fields are accurately hidden.
00663790 Performance is now improved for the projects listing page.
00672470, 00650232, 00542411, 00605607, 00595731 After validating the fields of production documents, you can now use the Save current document option to save all the validated document, without any error messages.
-- Fixed an issue that caused a higher API response time when more than 20 users tried to access IQ Bot.

You must add the ignite.security.disable=true property in the cluster.properties file.

Create or update cluster.properties file

00672702

You can now export and import backup (.iqba) files successfully even if the data entered in the Validator contains spaces or tabs.

00669197, 00681599 Node.js is now upgraded to version v10.22.1 to fix a security vulnerability.
00662968 Sometimes, when you click See Extraction Results and the preview fails, IQ Bot now correctly displays the following error message after a timeout of 2 minutes: Error occurred while extracting the document. Please re-check the training/document. Previously, when the preview failed, the page kept loading indefinitely.
00644948 When you now upload a document in production, deadlocks no longer occur in the system and documents are correctly sent to the Validator.

Known limitations

In a cluster setup, sometimes after training, when you click See Extraction Results, the preview page does not load and the system displays an Internal Server Error message.

IQ Bot Version 11.3.5.3 is a full installer and not a patch release. Therefore, to import learning instance backups to Version 11.3.5.3, the backup (.iqba) file must be exported only from Version 11.3.5.3.

Processing time for bulk documents (4000 approximately) takes longer than before because IQ Bot now fixes the orientation of documents first before processing them.

When the value entered for GroupDocumentThreshold in the settings.txt file is zero (0) or negative, the Threshold to create new groups field does not display the default value of 10 in the Learning Instances > Summary page.

The settings.txt file is available in C:\Program Files (x86)\Automation Anywhere IQ Bot\Configurations folder.

If you create a learning instance with Checkbox auto-detection enabled, check box SIRs cannot be selected.

Workaround: Draw on top of the check box SIR instead of selecting the identified SIR field.

When you install IQ Bot with a custom key, documents might not be classified if they are used with Microsoft Azure or Google Vision API.

Sometimes, when you train a bot and verify the extracted data, and then if you retrain that bot and view the extraction results again, the preview might fail but the cached extraction results are still displayed.

When you upload a document to production immediately after deleting a bot it was associated with, extraction results are not displayed.

Workaround: Use one of the following options:
  • Select the Disable cache option in the Network tab.
  • Ensure you wait at least 3 minutes between deleting the bot and uploading the document.
  • Empty the cache and reload the page.

Sometimes, if a bot is not trained correctly and the preview fails, IQ Bot does not display an error and the page keeps loading.

When two users try to access a bot for a learning instance at the same time, IQ Bot redirects the user to the learning instances listing page and displays an error message there instead of the learning instance details page.

When a user with a custom role with all permission except View learning instance from the same role and Launch Validator clicks Save and go to next group, IQ Bot displays an Access Denied error. However, the group information is saved.

In the Designer, if a segment drawn over a field maps less than 50% of the total area of that field, the field value is displayed in the Designer.

If you are using Internet Explorer, you cannot save a document or mark it as invalid in the learning instances validator screen.

Workaround: Use the Google Chrome browser to access the learning instances validator screen.

Sometimes, IQ Bot does not display an error when you enter a date in an invalid format in the Validator.

When IQ Bot is installed with Microsoft SQL Azure (RTM) database, and if the IQ Bot application and database servers are in different sub-networks, the Healthcheck API returns NOT_OK for the alias service due to network latency.

Workaround:
  1. Open the C:\Program Files (x86)\Automation Anywhere IQ Bot\Services\Application.yaml file.
  2. Update the following attributes to increase the healthcheck timeout period to 10000 from the current 5000:

    connect-timeout: 1000

    read-timeout: 10000

During a new installation, IQ Bot does not restrict a user from including spaces in the custom encryption key.

When you export a learning instance, the status is not displayed on the Migration page until you manually refresh the page data.

When table fields are validated in the Validator, the user is navigated to the associated page for that field but the SIRs are sometimes not displayed.

Workaround: Navigate to the page by entering the page number in the pagination bar.