This solution must be followed only if your documents have more than 1000 different templates.
Currently, this approach doesnot support the predictions for tabular data. Example:- Line item, Line item amounts etc.
Every installation of docube will have a folder named TrainingCognitive. This can be used as reference for any semi structured data processing solution. Duplicate this folder and work on it. PDF to CSV Convertor and Model Application are the reference tasks in JIFFY. These can be imported into your project
For any document processing soltion, we will have a one time process of creating a model and an ongoing process of applying these models on new documents.
Please refer the create model section in order to create a model.