Write a PDF parser which extracts position and text

I need to index a lot of pdf files and extract various parts of the documents as specific fields. For this I have previously used pdf2text and used regex to extract part of the text. In unstructured text this is a tedious process and I would much rather base this process on the actual position of the text. Therefor I need a commandline software to be run on linux which will parse PDF files and extract not only the text, but also the text position. I need exact location of every word and or phrase in the text. There are some libraries in C++ here: [login to view URL] And JAVA libraries: [login to view URL] Also Python: [login to view URL] Output should be in plain text to stout, either as csv (x-position, y-position, word/phrase) or JSON. The tool should be able to run on a linux terminal with the following setup: Linux x86_64 x86_64 x86_64 GNU/Linux PHP 5.3.29 (cli) (built: Aug 20 2014 16:41:34) Python 2.6.9 OpenJDK Runtime Environment (amzn-2.5.4.0.53.amzn1-x86_64 u75-b13) I have attached a PDF-file, typical for my scenario.