I removed that script because people expected it to do too much and recognize arbitrary pages at any resolution. Writing a script that reliably deals with all the odd cases people might throw at it is a lot of work. We will be revisiting that in OCRopus 0.8, when preprocessing will be a lot more automated as well.
Generally, you should be able to run ocropus-nlbin, ocropus-gpageseg, and ocropus-rpred in sequence.
Tom