Amazon Textract

[OslienMesa]

Hello

Our team is experimenting with several APIs that perform document reading using OCR techniques, we are not able to use the Amazon Textract API. Anyone have any idea how it would be?

Att

------------------------------
Oslien Mesa
Developer
Elumini
America/Sao_Paulo
------------------------------

[ChrisHogan]

I am working on a VBO for Textract that will be published to  the DX along with the existing AWS VBOs  for Rekognition and Comprehend. This will be available soon.

------------------------------
Chris Hogan
Senior Ecosystem Architect
Blue Prism
------------------------------

[OslienMesa]

Hello Chris:

Thank you very much, we will wait for your VBO, although we would like to have some clue on how to work with the AWS APIs to get ahead of our work. So when your VBO arrives, we would only do an update.

Thank you again.
Att

------------------------------
Oslien Mesa
Developer
Elumini
America/Sao_Paulo
------------------------------

[ChrisHogan]

I am not sure what you mean by work with the API. The VBO does all the API connectivity for you. You can see what the raw JSON from the Textract service looks like by visiting the Textract portal and submitting an image to analyze: https://aws.amazon.com/textract/

The VBO will convert the raw JSON to a nested collection of pages, forms, key/value pairs and tables for easier use inside Blue Prism.


------------------------------
Chris Hogan
Senior Ecosystem Architect
Blue Prism
------------------------------

[EdGarabedian1]

Hi Oslien ,

If you mean working with APIs without a specific VBO you could refer to this document on the portal: https://portal.blueprism.com/documents/v64-user-guide-web-apis
I also wrote a blog post going over the basics of how to call a rest API in this blog post: https://rpasuccess.com/blog/2019/1/1/how-to-call-a-rest-api-from-blue-prism
Blue Prism is building out an ecosystem of connectedRPA VBOs, as Chris mentioned, that do the API connection for you but there are also base level, out-of-the box VBOs that enable you to build the API call yourself. Hope this helps.
Thanks,



------------------------------
Ed Garabedian
Sr. Solutions Consultant
Blue Prism
------------------------------

[BenedictDeroose]

​Hi Chris, we are also looking at using AWS Textract API to perform image to digital conversions. Is there an expected date when the AWS Textract VBO will be available in the DX?

------------------------------
Benedict Deroose
Configuration Process Analyst
Schroders
Asia/Singapore
------------------------------

[MattPalmer]

We were already using the AWS Rekognition VBO and we found that if we copied and edited that VBO to point to the Textract Host / Target and update the actions we could get the Textract service to work.  You might find success doing the same...

------------------------------
Matt Palmer
Business Process Specialist
Chesapeake Energy
America/Chicago
------------------------------

[BenedictDeroose]

​Thanks for the information Matt! Unfortunately, we are currently on Blue Prism 6.2.1 (no plans to upgrade till next major version) and the AWS Rekognition VBO is only compatible with v6.4 onwards.

------------------------------
Benedict Deroose
Configuration Process Analyst
Schroders
Asia/Singapore
------------------------------

[MelanieGiuliani]

Hi Chris - thanks for sharing! Once your VBO is published to the DX, could you please link it here? I don't see it up yet but looks like there's quite a bit of interest! 🙂

------------------------------
Melanie Giuliani
Online Community Manager
Blue Prism
------------------------------

[VinothkumarC_N]

Hi @Matt Palmer - i tried to use same rekognition for Textract by changing the service name and actions. However i could not establish connection. I receive unknow operation exception.
Can you confirm on the service name to refer. I was using Textract.DetectDocumentText action,but could not establish connection. Kindly help me

------------------------------
Vinothkumar C N
Developer
Accenture
Asia/Kolkata
------------------------------