It’s annoying and time consuming having to map DocumentTypeId back to the names when the names are already in the taxonomy. Why isn’t the name exposed in the results?
Hello @postwick ,
Can you please give more details on what the issue is? is it that we don’t expose Document Type Name in the ClassificationResult?
Yes that is the issue. To get the name I have to parse the taxonomy into a dictionary and then lookup the name by the DocumentTypeID
Hate to break it to you Paul.
But that data has always been in the extraction result…
Assuming you have the datatype of UiPath.DocumentProcessing.Contracts.Results.ExtractionResult you have the following.
If you use the modern activities you can still get this information even though the data types are different, but I assume you are on the older Windows Compatible ones and not the cross platform ones.
I think we don’t yet have it in a ClassificationResult and that would be @postwick 's issue
Please confirm so we can analyze correctly and potentially plan.
Could be. I dont ever use the classic activities these days so don’t have a sandbox to try them out. Pretty sure the cross platform ones do it.
Perhaps @postwick can confirm if, in the classic activities, those properties are populated or not by the extraction activities. If not its understandably annoying to need to cross reference the taxonomy if the underlying objects even have the properties but dont use them.
Correct, I’m not talking about extraction result I’m talking about classification result.
Ah, I apologize there Paul. I am always focused on the extraction part of didnt read clearly enough.
It seems to be present on the Modern Cross Platform activities, can you switch to them?
I think you are using on premise right? I am not sure what version you’d need to make these work or if they are fully compatible with any version.
Next project I’ll try that. Could be useful.
got it, will try to plan adding it.
if you need any sort of advanced DU functionality, please be careful when using the DocumentUnderstanding.Activities pack intead of the the IntelligentOCR.Activities one. The first is much simpler to use but limited in what it allows you to do.