dm-scripts/pdfcompress

#!/bin/bash

set -euo pipefail

help_message="$(cat <<END
usage: $(basename "$0") [OPTIONS...] INPUT_PDF OUTPUT_PDF

Converts a PDF file to a series of JPEG images to reduce the quality 
and then merges them back in a single pdf. 

Options:
    -q, --quality    Set the JPEG quality (default: 75)
    -h, --help       Display this help message

END
)"

function show_usage() {
    echo "$help_message"
    exit 1
}

quality=75

while [[ $# -gt 0 ]]
do
    key="$1"

    case $key in
        -q|--quality)
            quality="$2"
            shift # past argument
            shift # past value
            ;;
        -h|--help)
            show_usage
            ;;
        -*)
            echo "Unknown option: $key"
            show_usage
            ;;
        *)
            break
            ;;
    esac
done

if [ -z "$1" ] || [ -z "$2" ]; then
    show_usage
fi

input_file="$1"
output_file="$2"

temp_pages_dir="$(mktemp -d)"

function logf {
    printf "$@"
}

logf "Rendering each page to a jpeg image (quality = $quality)\n"
pdftoppm -r 300 -jpeg -jpegopt "quality=$quality,optimize=y" -progress $input_file "$temp_pages_dir/page"

for f in "$temp_pages_dir"/page-*.jpg; do
    logf "Wrapping $f into ${f/.jpg/.pdf}\n"
    gs -dNOSAFER  -sDEVICE=pdfwrite -o "${f/.jpg/.pdf}" $(find -iname viewjpeg.ps /usr/share | head -1) -c "($f)" viewJPEG
done

logf "Merging compressed pages to a single PDF\n"
pdfunite "$temp_pages_dir"/page-*.pdf $output_file

rm -r "$temp_pages_dir"
logf "Done\n"