1. Reading PDF

Jiffy has an inbuilt node available for reading PDF documents. The node is available under “Miscellaneous” section. PDF Reader node is used to read the contents from a text PDF. Jiffy supports integration with OCR technologies to read image PDFs. The image PDF would get converted to text PDF and the same would be used by the PDF reader too extract the data.

How to Start?
The below steps need to be followed:

• Drag and drop a PDF node from “Miscellaneous” section
• Configuration: Select a PDF configuration from the drop down. Refer - Task Design -> Configuration for setting a Document configuration.
• Mappings: Provide the Location where PDF is saved in the input mapping section. The PDF files should be placed in the Windows server where Bot (Jiffy client) is running. Refer - Mapping section for details on providing input mappings.
• Template: Template flag need to be turned on/off. In case an XML Template has been created, turn “Template” flag to ON.
• Template Tag: Provide the Template tag that is saved by creating the template Refer - Repository -> Document Templates for more information.

Data Extraction Approaches
Jiffy supports the following different approaches for extracting data from the PDF files

• Rule-based Data Extraction
  This approach is mostly used in scenarios where there are limited number of PDF formats available and the frequency in which a new format being introduced is also low. An XML template need to be defined for each PDF format.
  

• Intelligent PDF Processing
  This approach is mostly used in scenarios where the number of PDF formats are huge and the frequency in which a new format being introduced is also high. The IQ automation works on the principle of having a dynamic repository to store all the data dictionary and a dynamic model (XML template) for extract information from the PDF.
  

• Cognitive Automation
  This approach is also used in scenarios where the number of PDF formats are huge and the frequency in which a new format being introduced is also high. The major difference between IQ Automation and Cognitive automation is that, the training data (approximately 1 year past data) need to be fed into the system to train the BOT and an ML algorithm need to be devised.