One of the fastest ways to compile an index would be to get a high resolution scan of the text, run the image through a good quality optical character resolution program (OCR) and from there search the text for terms and create an index. The slowest part of that fastest way is the OCR part, in which an individual might have to key in text where the program doesn't recognize it if the font is an old one or the paper has foxing and such that offer random unrecognized marks. (I've done this work - it can be slow.)

The other way would be to have someone read the book and take notes regarding keywords and pages. The resulting PDF from the first method would give you access to any word anywhere in the book, the second would be the index alone.