Rasterize and upload CUPS filter

#!/usr/bin/env perl
# -----------------------------------------------------------------------------------------
# rasterize, parse and upload print filter for CUPS
# -----------------------------------------------------------------------------------------
# if it's fed a PDF file, it queries a web service with the MD5 hash to see if it's been
# converted already, and if not:
# 1. it rasterizes the files and uploads the raster
# 2. it extracts text from it, and uploads the text
# 3. it extracts the color histogram from the raster and uploads it
# -----------------------------------------------------------------------------------------
# TODO:
# - error handling
# - histogams may be best extracted from a non-aliased raster
# - probably more
# -----------------------------------------------------------------------------------------



use strict;

$\ = "\n";

my $tempdir = '/var/tmp';           # we need a world writeable temp dir

my $PERL_BADLANG; $PERL_BADLANG = 0 if ($PERL_BADLANG);     # Suppress warnings about locale problems
my $PDFINFO  = '/opt/local/bin/pdfinfo';
my $PDFTOPPM = '/opt/local/bin/pdftoppm -r 150 -png';
my $PDFTOTXT = '/opt/local/bin/pdftotext -layout';
my $CONVERT  = '/opt/local/bin/convert';
my $MD5      = '/sbin/md5 -q';
my $UPLURL   = 'http://127.0.0.1:3000/upload';


my $CUPS_BACKEND_OK = 0;
my $CUPS_BACKEND_FAILED = 1;

my %args;
my $arg_count = scalar @ARGV;

@args{qw/jobid username title copies options file_to_print/} = @ARGV;
$args{file_to_print} = '-' if (! defined $args{file_to_print});

for ($arg_count) {
    /^0$/ && do {
    print qq/file rasterize "Unknown" "rasterize pdf and upload it"/;
    last;
    };
    {
    my $file = $args{file_to_print};
    my $title = $args{title};

    #-------------------------
    # get file type and exit
    # if it isn't a PDF
    #-------------------------
    my $type = qx/file -b "$file"/;

    unless ($type =~ /^PDF/) {
        unlink $file;
        exit 0
    }

    #-----------------------
    # get md5 hash of file
    #-----------------------
    my $md5 = qx/$MD5 "$file"/; $md5 =~ s/\s*$//;
    my $pct = qx|curl $UPLURL/f/$md5|;

    #------------------
    # get file info
    #------------------
    my $info = qx/$PDFINFO "$file"/;

    #---------------------
    # extract page count
    #---------------------
    $info =~ /Pages:\s*(\d+)/; my $pages = $1;

    #---------------------
    # print stuff to log
    #---------------------
    open my $fh, '>>', '/var/tmp/foo.txt';
    exit 0 if $pct == $pages;

    print $fh join "\n", (@ARGV);
    print $fh '-' x 80;

    #---------------------------
    # convert pages, make them
    # readable and upload
    #---------------------------
    for my $p (1..$pages) {
        #---------------------
        # convert to png
        #---------------------
        my $png = sprintf "$tempdir/%s-%03d.png", $md5, $p;
        system(qq/$PDFTOPPM -f $p -l $p "$file" > "$png"/);
        system(qq|curl -F file=\@$png $UPLURL|);
        chmod 0666, $png;

        #---------------------
        # convert to text
        #---------------------
        my $txt = sprintf "$tempdir/%s-%03d.txt", $md5, $p;
        system(qq/$PDFTOTXT -f $p -l $p "$file" - > "$txt"/);
        system(qq|curl -F file=\@$txt $UPLURL|);
        chmod 0666, $txt;

        #---------------------
        # create histogram
        #---------------------
        my $hst = sprintf "$tempdir/%s-%03d.hst", $md5, $p;
        # print $fh qq(convert "$png" -format \%c histogram:info:"$hst");
        system qq/$CONVERT "$png" -format \%c histogram:info:"$hst"/;
        system(qq|curl -F file=\@$hst $UPLURL|);
        chmod 0666, $hst;

        unlink $png, $txt, $hst;
        chmod 0666, '/var/tmp/foo.txt'
    }
    close $fh;
    unlink $file;
    last;
    };
}

exit 0;