#! /bin/bash -f

function copyright()
{
    echo "Copyright (C) 2002 David Kreil <D.Kreil@IEEE.Org>"
    echo "Modified by Barak A. Pearlmutter <bap@debian.org>"
    echo "         and Yaroslav Halchenko <debian@onerussian.com>"
    echo "Released under the GNU GPL v2, 21-Oct-2002."
}

function warranty()
{
    echo "This program is distributed in the hope that it will be useful,"
    echo "but WITHOUT ANY WARRANTY; without even the implied warranty of"
    echo "MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the"
    echo "GNU General Public License for more details."
}

function disclaimer()
{
    echo "By using this tool you accept the following disclaimer:"
    echo "  Because the any2djvu service is free of charge, there is no"
    echo "  warranty of any kind.  In particular, no effort is made to"
    echo "  prevent anyone from downloading the files submitted to or"
    echo "  produced by the any2djvu server."
}

# TO DO:
#  - error handling

rurl="http://any2djvu.djvuzone.org"
rcgi="any2djvu.php"
res=400
ocr=1
docformat=2

function warn()
{
    echo "Notes:"
    echo " - Internet connection is required."
    echo " - Filenames are assumed to require no URL-encoding."
    echo " - Documents must be PostScript (.ps, .ps.gz) or PDF (.pdf)."
    echo " - Defaults are: conversion is at 400dpi, with English OCR enabled."
    echo " - This script should not be used for large scale conversions of"
    echo "   documents, as it may badly affect and hence endanger the free web"
    echo "   service to the community."
    echo " - This software comes with NO WARRANTY."
}

function format_help()
{
    echo "Codes for the formats of the input documents to use with -f"
    echo "  1 - DjVu Document (for verification or OCR)"
    echo "  2 - PS/PS.GZ/PDF Document (default)"
    echo "  3 - Photo/Picture/Icon"
    echo "  4 - Scanned Document - B&W - <200 dpi"
    echo "  5 - Scanned Document - B&W - 200-400 dpi"
    echo "  6 - Scanned Document - B&W - >400 dpi"
    echo "  7 - Scanned Document - Color/Mixed - <200 dpi"
    echo "  8 - Scanned Document - Color/Mixed - 200-400 dpi"
    echo "  9 - Scanned Document - Color/Mixed - >400 dpi"
}

function usage()
{
    echo "Convert files from .ps/.ps.gz/.pdf to .djvu"
    echo "Usage:	$0 [options] [url] {filename(s)}"
    echo
    echo "Options:"
    echo "  -q          quiet mode: reduce chatter on the screen"
    echo "  -a          acknowledge internet transmission"
    echo "  -c          clean after work: remove log file"
    echo "  -r dpi      resolution in dpi (200/300/400, default: $res)"
    echo "  -o ocr      perform OCR (0 - no, 1 - yes, default: $ocr)"
    echo "  -f format   format of input document; invoke -f help for list"
    echo "  -u url      base URL of server (default: $rurl)"
    echo "  -p s        CGI script name (default: $rcgi)"
    echo
    echo "If no local file is named by the 1st argument the 1st"
    echo "argument is taken to be an external web-accessible directory and,"
    echo "the specified files are uploaded from that location."
    echo
    echo "Examples:"
    echo "  cd ~bap/public_html/foo"
    echo " # uploads from web-accessible directory"
    echo "  any2djvu http://www.bcl.hamilton.ie/~barak/papers mesh-preprint.ps.gz"
    echo "  any2djvu http://www.inference.phy.cam.ac.uk/mackay *.ps.gz bar.pdf"
    echo " # uploads from current directory"
    echo "  any2djvu b*.pdf"
    echo "  any2djvu -s -c -r 300 -o 0 bp.ps"
}


while getopts "hacqsr:o:f:u:p:" opt
  do
  case "$opt" in
      h) usage; exit 0 ;;
      a) DJVU_ONLINE_ACK=1 ;;
      c) doclean=1 ;;
      s) echo "option -s is deprecated, please use -q"; silent=1  ;;
      q) silent=1  ;;
      r) res="$OPTARG" ;;
      o) ocr="$OPTARG" ;;
      f) docformat="$OPTARG" ;;
      u) rurl="$OPTARG" ;;
      p) rcgi="$OPTARG" ;;
      *) usage; exit 2 ;;
  esac
done

shift $((OPTIND - 1))

# check resolution
if [ ${res#[234]} != '00' ]; then
    echo 'error: -r resolution must be one of: 200, 300, 400'
    exit 2
fi

# check OCR option
if [ ! "x$ocr" == x0 ] && [ ! "x$ocr" == x1 ]; then
    echo 'error: -o OCR must be 0 or 1'
    exit 2
fi

# if help is requested or docformat is not specified right - show help
if [ "x$docformat" == x'help' ]; then
    format_help
    exit 0
fi

if [ -z "$docformat" ] || [ ! -z "${docformat#[0-9]}" ]; then
    echo "error: requested document format unknown: $docformat".
    format_help
    exit 2
fi

lurl=''
if [ ! -e "$1" ]; then	     # if 1st argument is not an existing file
    lurl="$1"			# assume that it is a URL
    shift
fi

#log=`date -u`": "`whoami`'@'`hostname`", pid $$: $0 (cwd "`pwd`")"
log="`date -u`: `whoami`@`hostname`, pid $$: $0 (cwd `pwd`)"
[ -z $silent ]  && echo '/-- Started' "$log" >&2

in="$1"
shift
if [ -z "$in" ]; then
    echo "error: no files to convert"
    usage
    exit 1
fi

if [ ! -z $s ]; then
    copyright
    disclaimer
    warn
fi

if [ -z "$DJVU_ONLINE_ACK" ]; then
    cat <<EOF
WARNING!

any2djvu uses an external server which is willing to perform the
conversion and requires the document transfer over to that server.
There is a security issue in operating on documents not intended for
widespread distribution, which could be partially although not
completely ameliorated by using a secure web connection.

Do you acknowledge and allow the transmission of the document?
(Type 'yes' to acknowledge. You can define non-empty environment
 variable DJVU_ONLINE_ACK to avoid seeing this dialog, or use -a
 command line parameter to any2djvu).
EOF
    echo -n "[yes/no]:"
    read ack
    if [[ ! "$ack" =~ ^[yY][eE][sS]$ ]]; then
        warn
        exit 1
    fi
fi

wgetopts=''
curlopts=''
shellopts=''
if [ ! -z $silent ]; then
    curlopts=' -s '
    wgetopts=' -q '
    shellopts=' 1>/dev/null ' # 2>&1
fi

cgiopts="&docformat=$docformat&resolution=$res&ocr=$ocr&legal=1"
while [ -n "$in" ]; do
    b="$in"
    b="${b%.ps.gz}"
    b="${b%.ps}"
    b="${b%.pdf}"
    b="${b%.PDF}"
    echo `date -u` Processing $b ...
    log="$b-any2djvu.log"
    if [ -z "$lurl" ]; then
        if [ ! -e "$in" ]; then
            echo "File $in wasn't found. Conversion stopped"
            exit 2
        fi
        curl $curlopts -F "fupload=@$in" -F submit=Submit               \
            -F docformat=$docformat -F resolution=$res                  \
            -F ocr=$ocr -F legal=1 "$rurl/$rcgi"                          \
            | eval tee "'$log'" $shellopts
    else
        wget $wgetopts -O - "$rurl/$rcgi?urlupload=$lurl/$in$cgiopts"   \
            | eval tee "'$log'" $shellopts
    fi
    l=`egrep 'href=djvu/.*\.djvu' "$log"`
    l="${l##*href=}"
    l="${l%%>*}"
    if [ -z "$l" ]; then
        echo "error: something got wrong. check log file"
        exit 1
    fi
    wget $wgetopts -O "$b.djvu" "$rurl/$l"

    [ -z "$doclean" ] || rm "$log"
    [ -z $silent ]  && ls -l "$b.djvu"

    in="$1"
    shift
done

log=`date -u`": "`whoami`'@'`hostname`", pid $$: $0 (cwd "`pwd`")"

[ -z $silent ]  && echo '\-- Done' $log >&2