shrinkpdf

#!/usr/bin/env bash
# vim: set shiftwidth=4 softtabstop=4 expandtab foldlevel=0 foldmethod=indent: # vim-modline, don't delete
#############################################################################
# $HGFile: shrinkpdf $
# $HGLocation: /home/mb/src/tools $

# 1st Author: Massimo
# Created: 2012-10-26

# $HGChecked in by: massimo $
# $HGDate: 2018-09-21 14:08 +0200 $
version='$HGRevision: 10 $'
# $HGBranch:  $
# $HGDesc: Fixed du calculation $
#############################################################################

# Checking Bash version
(( ${BASH_VERSINFO[0]} >= 4 )) || die "This script needs at least bash version 4"

# Libraries
libs=( output )
for lib in "${libs[@]}"; do
    lib_path="${BASH_PREFIX:-/}usr/local/lib/lib_$lib"
    if [[ lib_$lib ]]; then
        source "lib_$lib"
        continue
    elif [[ -e "$lib_path" ]]; then
        source "$lib_path"
        continue
    else
        echo "$lib_path not found" 1>&2; exit 1;
    fi
done

usage() {
    cat <<EOF
This is $0 $version

Usage:
    $0 [OPTIONS] <target pdf files>

Description:
    Compressing pdf files by 2 step process through qpdf and pdftk.
    Automatically keeps original if compression is zero or even worse.
    
    File with subsctring "noshrink" as part of their filename are skipped, such as manual_noshrink.pdf .

Options:
    -k, --keep              keep new files and don't overwrite old ones
    -p, --pdfsetting        PDFSETTINGS, with decreasing quality: default, prepress, printer, ebook, screen
    -v, --verbose           Verbose
    -h, --help              Show this help

Dependencies:

    * pdftk
    * qpdf

EOF
}

while [[ $1 == -* ]]; do
    case "$1" in
        -k|--keep)          opt_keep=1
                            shift;;
        -v|--verbose)       opt_verbose=1
                            shift;;
        -p|--pdfsetting)    shift
                            case $1 in
                                default|prepress|printer|ebook|screen)  true;;
                                *) die "invalid pdfsetting: $1";
                            esac
                            opt_pdfsetting=$1
                            shift;;
        -h|--help|-\?)      usage
                            exit 0;;
        -*) echo "invalid option: $1" 1>&2; usage; exit 1;;
    esac
done

#gs	-q -dNOPAUSE -dBATCH -dSAFER \
#	-sDEVICE=pdfwrite \
#	-dCompatibilityLevel=1.3 \
#	-dPDFSETTINGS=/screen \
#	-dEmbedAllFonts=true \
#	-dSubsetFonts=true \
#	-dColorImageDownsampleType=/Bicubic \
#	-dColorImageResolution=72 \
#	-dGrayImageDownsampleType=/Bicubic \
#	-dGrayImageResolution=72 \
#	-dMonoImageDownsampleType=/Bicubic \
#	-dMonoImageResolution=72 \
#	-sOutputFile="${1%.*}_gs.pdf" \
#	 "$1"

shopt -s nullglob
export LC_NUMERIC=C


# Dependency checks
type -ap pdftk &>/dev/null || die "pdftk not found"
type -ap qpdf &>/dev/null || die "qpdf not found"
type -ap gs &>/dev/null || die "gs not found"

for file in "$@"; do
    [[ $file == *"noshrink"* ]] && continue
    (( ${#file} > ${maxlength_filename:-0} )) && maxlength_filename="${#file}"
done

for file in "$@"; do
    [[ $file == *"noshrink"* ]] && {
        echo "Skipped $file"
        continue
    }

    printf "Processing %*s: " "$maxlength_filename" "$file"

    if [[ ! -r $file ]]; then
        warn "File $file does not exist, skipped"
        continue
    fi

    this_type="$(file -b "$file")"
    if [[ $this_type != PDF* ]]; then
        warn "File $file is not of PDF type but $this_type, skipped"
        continue
    fi

    if [[ $file != *.pdf && $file != *.PDF ]]; then
        warn "File $file does not fit naming convention: *.pdf or *.PDF, skipped"
        continue
    fi

    original_size=( $(du -b "$file" 2>/dev/null) )
    printf " original: %-5s ->" "$(du -bh "$file" 2>/dev/null |awk '{print $1}')" 


    current_file="${file}"

    # 0. step: gs compression
    if [[ -n $opt_pdfsetting ]]; then
        gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dPDFSETTINGS=/${opt_pdfsetting} -dNOPAUSE -dQUIET -dBATCH -sOutputFile="${current_file}_${opt_pdfsetting}" "${current_file}"
        gs_status="$?"
        if (( $gs_status != 0 )); then
            rm -f ${opt_verbose:+-v} "${current_file}_${opt_pdfsetting}"
            error "gs failed, ${current_file}_${opt_pdfsetting} dropped."
        fi
        # 1.2: gs size
        if [[ -s ${current_file}_${opt_pdfsetting} ]]; then
            gs_size=( $(du -b "${current_file}_${opt_pdfsetting}" 2>/dev/null) )
        else
            gs_size="0"
        fi
        gs_ratio="$(bc <<< "scale=4; $gs_size / $original_size * 100")"
        printf " 0.gs: %-5s " \
            "$(du -bh "${current_file}_${opt_pdfsetting}" 2>/dev/null |awk '{print $1}')"

        current_file="${current_file}_${opt_pdfsetting}"
    fi

    # 1.step: qpdf
    #         decrypting useless passwords for stupid pdftk
    qpdf \
        --decrypt\
        --linearize\
        --object-streams=generate\
        --stream-data=compress\
        "${current_file}" "${current_file}_qpdf" 2>/dev/null
    qpdf_status="$?"
    if (( $qpdf_status != 0 )); then
        if (( $qpdf_status == 3 )); then
            true
            # qpdf returned warnings, resulting file may have some problems 
        else
            rm -f ${opt_verbose:+-v} "${curent_file}_qpdf"
            error "qpdf failed, ${current_file}_qpdf dropped."
        fi
    fi

    # 1.2: qpdf size
    if [[ -s ${current_file}_qpdf ]]; then
        qpdf_size=( $(du -b "${current_file}_qpdf" 2>/dev/null) )
    else
        qpdf_size="0"
    fi
    qpdf_ratio="$(bc <<< "scale=4; $qpdf_size / $original_size * 100")"
    printf " 1.qpdf: %-5s " \
        "$(du -bh "${current_file}_qpdf" 2>/dev/null |awk '{print $1}')"

    # 2.step: pdftk
    pdftk "${current_file}_qpdf" output "${current_file}_pdftk" compress ${opt_verbose:+-verbose} 2>/dev/null
    if (( $? != 0 )); then
        rm -f ${opt_verbose:+-v} "${current_file}_pdftk"
        #error "pdftk failed."
        error "pdftk failed, ${current_file}_pdftk dropped."
        continue
    fi

    # 2.2: pdftk size
    if [[ -s ${current_file}_pdftk ]]; then
        pdftk_size=( $(du -b "${current_file}_pdftk" 2>/dev/null) )
    else
        pdftk_size="0"
    fi
    pdftk_ratio="$(bc <<< "scale=4; $pdftk_size / $original_size * 100")"
    printf " 2.pdftk: %-5s " \
        "$(du -bh "${current_file}_pdftk" 2>/dev/null |awk '{print $1}')"

    # 3.1: Decide wether to keep original or to overwrite, first checking pdftk size

    if [[ -n $opt_keep ]]; then
            mv ${opt_verbose:+-v} "${file}" "${file}_orig"

        if (( $? != 0 )); then
            warn_n "Option --keep enabled, but failed to backup $file, skipped."
            continue
        fi
    fi
    if (( $pdftk_size > 0 && $pdftk_size < $original_size && $pdftk_size < $qpdf_size )); then
        mv -f "${current_file}_pdftk" "${file}" || {
            warn "Moving ${current_file}_pdftk over original failed, skipped."
            continue
        }
        rm -f ${opt_verbose:+-v} "${current_file}_qpdf" || {
            warn "Removing ${file}_qpdf failed, skipped."
            continue
        }
        success_n "pdftk $(printf "[%5.1f%%]" "$pdftk_ratio") and kept. "
    else
        rm -f ${opt_verbose:+-v} "${current_file}_pdftk"
        warn_n "pdftk $(printf "[%5.1f%%]" "$pdftk_ratio") and dropped. "

        # 3.2: Checking qpdf size

        if (( $qpdf_size > 0 && $qpdf_size < $original_size )); then
            mv -f "${current_file}_qpdf" "$file" || {
                warn "Moving ${file}_qpdf over original failed, skipped."
                continue
            }
            success_n "qpdf $(printf "[%5.1f%%]" "$qpdf_ratio") and kept."
        else
            rm -f ${opt_verbose:+-v} "${current_file}_qpdf" || {
                warn "Removing ${current_file}_qpdf failed, skipped."
                continue
            }
            [[ -z $opt_keep ]] && rm -f ${opt_verbose:+-v} "${file}_orig"
            warn_n "qpdf $(printf "[%5.1f%%]" "$qpdf_ratio") dropped and reverted."
        fi
    fi
    [[ -z $opt_keep ]] && [[ -n $opt_pdfsetting ]] && [[ -f ${file}_${opt_pdfsetting} ]] && rm -f ${opt_verbose:+-v} "${file}_${opt_pdfsetting}"
    printf "\n"

    if (( $qpdf_status == 3 )); then
        warn "qpdf returned warnings, resulting file may have some problems "
    fi

done