Configurable Tika ocr strategies for PDFs

[wombat94]

OCR of PDFs in Tika can take a long time. This is unnecessary if the PDF has already been ORCed.

I would like to see an option to define the OCR strategy used by Tika in the lodestone front end.

Ideally, this would be multi-pass with a first pass being no_ocr and if the size of returned data is below a threshold (perhaps 500 bytes of text) then re-process with text_and_ocr to recognize the document.