1. extract images from one or more  pdf file/files converting them into tiff/jpg images;

2.  autocrop them (to remove a white margin), then split them, since images are two-pages scans that need to be splitted into single-page images;

3a . load the resulting images into a zip file or

3b. wrap them again into a pdf file with a lossless compression. 

Presently I'm doing the job using a local routine into my PC, and I use nconvert for step 1, and python (with PIL) for steps 2-3a.

I better would like to to this job into Labs; the best being, to addo routines to grab files from an URL and to load them into Internet Archive. 

Just to avoid "rediscovering the wheel", is someone doing something similar into Labs? 

If the wheel has to be built ;-) , which tools have I to (painfully) study and test?

Can I install nconvert for linux into my account for personal use? 

Alex (from it.wikisource)