Identifying documents that have been processed in the thin client

I have configured a robot that ingests documents in a Desktop Automation step using one of the built-in document transformation projects (invoices_vat). Documents that pass all validity checks get released to a database and the others are routed to RPA Document Transformation Thin Client Validation. I'm trying to figure out how to identify those documents that have been processed through the thin client so that the robot can then release then to the database. How can I accomplish this?

Thanks!

Hi,

The best approach for this is to have your master robot check the status of the transformation result after processing.

After transformation runs, the Transformation browser should return a status for the document (dts > application > document > valid attribute). You can check this within your DA workflow and then push all the successful documents to the database. For those that return invalid, they will be automatically sent to the thin client for validation but you can extract the validation URL from the status bar to send users directly to that document for validation.

Any documents that get sent to validation will need to handled by your call back robot (this is set when you make the call through the Open DT browser step). After validation in the thin client this robot will be automatically called and can then handle the submission of the validated documents into the DB.

If you're running version 10.4 onwards there should be a detailed step by step guide on setting up document transformation and some samples that should give you everything you need

Thanks for the response. I previously setup a robot using the 10.5 guide as an example as well as in my own robot. 

In both cases, in the step after validation, the callback robot appears to not be automatically called as it does not handle the submission of the validated docs to the DB. In the \Document Transformation\Batch folder the documents get stuck in the Processing folder and in some instances go to the Error folder.

A couple of things to check..

1. Make sure that the call back robot is actually being called once the document has been processed in the thin client. The call back robot should be executed through the MC via a web service so any interaction should be tracked in the logs. If there's nothing in the logs there's a possibility that the call back isn't even being executed.
2. Test the call back robot. I think it's possible to zip your error folder (DTS install folder\Batch\Error\Error Folder XYZ) and provide this to the call back robot as your document variable. This would allow you to step through your call back robot in design studio and see if this is generating an error.

Transformation tends to write error messages out to the event viewer, so worth checking here for any error messages along with the usual log files.

Failing that the folks at Kofax support are usually pretty good at resolving these kind of issues

I verified that the callback robot is being executed. However, documents that are routed to the thin client keep getting routed back to the thin client (you can see the files go from Document Transformation\Batch\Validation to Document Transformation\Batch\Processing and then back to the Validation folder after a few minutes).

The issue according to the logs seems to be the missing validation.properties file ("validation.properties could not be loaded by any means. fail"). I'll follow-up with Kofax unless you see an obvious "gotcha."

Thanks!

So the call back robot is executed and then the batch goes back to the thin client?

If the call back robot is being executed, the only reason I can think it would go back to the thin client is if the document returns invalid when transformation runs the second time round. This shouldn't be the case if you have validated it already.

A couple more ideas..
Make sure that the document is fully validated in the thin client (don't force anything through), if it's the invoice project make sure amounts add up, dates are valid etc.
In your call back robot make sure the document variable is set as an input to the robot. If it's not an input variable it might just be using your default document that you've been using for testing.

Another idea might be to use a more straightforward project than the invoice one. This contains quite a lot of index data so something with less fields and validation rules might be easier to get through.

If none of these help, probably best running it through Kofax support. The validation.properties message could be the problem but I think I've also seen that messages in installs where it works without any issues.

Hope this helps