The automation before I came consist on reading a PDF and extracting values using Machine Learning Skills. Then sending to action center so users can approve them or edit them.
But there are too many fields and now I need to hide or delete some of those fields.
How can I hide/delete now this extra fields so when it is passed to action center, it is cleaner?
Solution Open and save the data extraction scope window
*I want to delete yellow fields and leave green ones
I believe you have option to show all fields only as of now…unless the field is removed directly from extraction i dont think youc an restrict the display
I tried just deleting those fields on the taxonomy, but when it came to the extracting activity, when using the ML Skill, it failed.
Saying that those fields were missing.
Yes, I need them to disapear, and users will be thankful. They will get used with a cleaner action center
It doesn’t necessarily have to be deleted from the labelling and you would not have to re-do the work by labelling or re-training the model (if the model does give out results properly for the other fields which are required accurately).
You would just have to Delete/Remove the fields in the Taxonomy and that would make also the other parts (Data Extraction Scope) to get updated. But if you are receiving errors elsewhere it would mean the value/Key is being used in further checks. You would need to modify the code to adjust to the fields which are required.
If you could provide us with screenshot of the Activities where the errors take place, we can take a further look on what needs to be done.
actually, when I edited the taxonomy and deleted the fields. They appear no more on the extraction scope (exactly the step it was failing), but good to make sure.
Now waiting 36 hours to finish the training. Last pipeline took that time.
I’ll let you know about my results. Thanks again
I did run a test myself using DU template. I did also encountered the same kind of error, in my case there were two, as Train Extractors module had the Train Extractor Scope, where we had to check the Configuration of it as well.