Crop random regions from a series of PDF documents (https://askubuntu.com/q/932193/175814)

crop-calc.py

import sys, random

if len(sys.argv) < 3:
	sys.argv.append(input())

cropsize = float(sys.argv[1])
llx, lly, urx, ury = map(float, sys.argv[2].split(None, 3))
width = urx - llx
height = ury - lly

if width < cropsize or height < cropsize:
	print('Crop size too small for bounds', llx, lly, urx, ury, file=sys.stderr)
	sys.exit(1)

cropllx = random.uniform(0, width - cropsize) + llx
croplly = random.uniform(0, height - cropsize) + lly
cropurx = cropllx + cropsize
cropury = croplly + cropsize

print(*map(lambda x: format(x, '.6f'), (cropllx, croplly, cropurx, cropury)))

crop.sh

#!/bin/bash
set -e -u -o pipefail

# default values:
size=256
cropsize=50
density=600
quality=75%
dstext=jpg


get_cropbox()
{
	gs -q -dSAFER -dBATCH -dNOPAUSE -dNOPROMPT -sDEVICE=bbox -f "$1" 2>&1 >&- |
	sed -ne 's/^%%HiResBoundingBox:\s*//p' |
	python3 "$exedir/crop-calc.py" "$2"
}


cat_pdf_page()
{
	gs -q -dSAFER -dBATCH -dNOPAUSE -dNOPROMPT -sDEVICE=pdfwrite \
		-dFirstPage="$3" -dLastPage="$3" -sOutputFile="$2" -f "$1"
}


crop_pdf()
{
	gs -q -dSAFER -dBATCH -dNOPAUSE -dNOPROMPT -sDEVICE=pdfwrite \
		-sOutputFile=%stdout% \
		-c "[/CropBox [$(get_cropbox "$1" "$cropsize")] /PAGES pdfmark" \
		-f "$1" |
	convert -define pdf:use-cropbox=true -colorspace sRGB -density "$density" pdf:- \
		-flatten -resize "${size}x${size}^" -crop "${size}x${size}+0+0!" \
		-quality "$quality" "$2"
}


get_random_area()
{
	local -i pagecount="$(pdfinfo "$1" | sed -ne 's/^Pages:\s*//p')" &&
	local -i page=$(($RANDOM % $pagecount + 1)) &&
	local pagefile="${1%.*}.p${page}.${1##*.}" &&
	[ "$1" -ot "$pagefile" ] || cat_pdf_page "$1" "$pagefile" "$page" &&
	crop_pdf "$pagefile" "$2"
}


get_random_areas()
{
	local -i i
	for ((i=1; i <= "$2"; i++)); do
		get_random_area "$1" "${1%.*}.$i.$dstext" || return
	done
}


exe="$(exec readlink -f -- "$0")"
exedir="${exe%/*}"
args="$(exec getopt -s bash -n "${0##*/}" -o 's:c:d:q:' -l 'size:,cropsize:,density:,quality:,dstext:' -- "$@")" || exit 64
eval set -- "$args"
unset args

while :; do
	case "$1" in
		--size|--cropsize|--density|--quality|--dstext)
			eval "${1:2}=\$2"; shift 2;;
		--)
			shift; break;;
		-*)
			echo 'Internal error!' "$1" >&2; exit 127;;
		*)
			break;;
	esac
done

declare -i areacount="$1"
shift
for f; do
	get_random_areas "$f" "$areacount" || exit
done