cancel
Showing results for 
Search instead for 
Did you mean: 

Amazon Textract VBO (TextractVBORelease_v1.0-062020)

PrathyushaMelap
Level 3
Hello,

I am using Amazon Textract VBO and trying to see if the tables in a document are read.
But I don't see the table data being pulled in the collection. While debugging into the AWS Textract Basic VBO, End action on Get tables page has output "Table Collection" . But the data item is not present, I tried adding it and running but still the output is empty.

Can someone help on this?


Thank you in advance!
1 BEST ANSWER

Best Answers

Unfortunately the logic to convert the raw json table data to collection is not yet there in the vbo, so you will have to build it. If you look at the "Response blocks" data item, each line with a block type of "LINE" denotes content of a tabular data which you can use to build your own collection. Alternatively you may convert the Full text to collection using code stage or use the raw response directly to format your output.
8097.png


------------------------------
Shashank Kumar
DX Integrations Partner Consultant
Blue Prism
Singapore
+6581326707
------------------------------

View answer in original post

5 REPLIES 5

Hi,

Would you be able to share your pdf file?

------------------------------
Shashank Kumar
DX Integrations Partner Consultant
Blue Prism
Singapore
+6581326707
------------------------------

Hello Shashank,

I tried attaching the file here but shows "Failed to Attach " Error. I have downloaded a sample Invoice image from the link below which has form and table.
The VBO reads raw text but not form or table. I tried the same document directly on AWS service and it did read the data as Key value and tables.

https://www.invoicesimple.com/invoice-template/sample-invoice-template

Please let me know if more details are required. Thank you for your time on this

------------------------------
Prathyusha Melapindi
------------------------------

Hi Prathyusha,

I have just tried it out using the attached invoice. Note that supported document types are jpg and png. I got the output response in Full Text variable. As you also pointed out, I did notice that the output parameter may not have been declared and logic not built out to create a output collection. That would need to be updated. 


------------------------------
Shashank Kumar
DX Integrations Partner Consultant
Blue Prism
Singapore
+6581326707
------------------------------

Yes, right. I have the image in .png format. I tried adding the data item too but it doesn't get any values and is still empty.
I am able to get full text too but over other ocr's textract standsout with Table and form feature which doesn't seem working.

------------------------------
Prathyusha Melapindi
------------------------------

Unfortunately the logic to convert the raw json table data to collection is not yet there in the vbo, so you will have to build it. If you look at the "Response blocks" data item, each line with a block type of "LINE" denotes content of a tabular data which you can use to build your own collection. Alternatively you may convert the Full text to collection using code stage or use the raw response directly to format your output.
8097.png


------------------------------
Shashank Kumar
DX Integrations Partner Consultant
Blue Prism
Singapore
+6581326707
------------------------------