Agree with Saikatc8 above - design a suitable 'eFrom' if possible - this will be easier to map in the long run.
if you have no control over the PDFs though (say you are processing existing PDFs which are not eForms): you can use 'pdfBox' (which ships with PRPC) to extract the text - you'll still need to do some additional work (regex ?) to extract out the specific bits of text you need to map to properties.
There is some example code on this post that illustrates how to strip text from a PDF using PDFBox:
Shows a way to extract the (raw text) from a PDF (and maps it to a single Clipboard Property).
In my understanding: there is no easy way of creating a mapping here - because a PDF document (non eform type at least) isn't structured in 'fields' that can be mapped in a generic way; you are basically working with blocks of text (and other entities like images) when you look at a PDF.