I work in a small company that will start dealing with some data processing, as an additional service to its core business. We need to accomplish two things for a large number of scanned PDF files (good quality) of bank transactions: 1. Sort the PDF files in folders according to the account number; 2. Extract data from some of them, depending on the transaction type.
The transaction files are very variegated: 35 different banks, about 20 different transactions for each bank and 4 languages. More or less, we are talking about 2,800 templates...
The first step would be to recognize the bank, then the account number and then the transaction type (and so forth). If transactions are of type X, simply move the files somewhere, otherwise read on and extract a few fields.
We checked some commercial solutions but they are way to expensive for our budget, so I'm wondering I might find a solution by putting together several pieces (either Windows or Linux).
I'm used to hack my way through problems, so any crazy idea is welcome!