#!/usr/bin/perl
# ---------------------------------------------------------------------------------------------
#
# classify_eudora_mbx.pl --- Process Eudora MBX files using Popfile's
#                            Classifier::Bayes and emit pf_[bucketname].mbx
#                            files for use by Eudora.
#
# This program authored by Scott W Leighton (helphand@pacbell.net)
# based upon bayes.pl, a Popfile component, which is Copyrighted
# by John Graham-Cumming. The author hereby contributes this code
# to the Popfile project under the terms of the Popfile License
# Agreement.    /Scott W Leighton/  February 15, 2003
#
# Modified May 24, 2003 - added support for v 0.19.x
#
# Popfile and Classifier::Bayes 
# Copyright (c) 2001-2003 John Graham-Cumming
#
# ---------------------------------------------------------------------------------------------

use strict;
use Classifier::Bayes;
use POPFile::Configuration;

# version check

my $version;

if ( -s "Popfile/Module.pm" ) {
  $version = 19;
} else {
  $version = 18;
}

# main



if ( $#ARGV == 0 ) 
{

    my $start_time = time;

    my $b;
    my $c;



    if ($version == 18) {

        $b = new Classifier::Bayes;
    
        if ( $b->initialize() == 0 ) {
           die "Failed to start while initializing the classifier module";
        }
    
        $b->{debug} = 0; 
        $b->{parser}->{debug} = 0;
        $b->load_word_matrix();

    } else {

        $b = new Classifier::Bayes;
        $c = new POPFile::Configuration;

        $b->configuration($c);
        $c->configuration($c);

        $c->initialize();
        $b->initialize();

        $c->load_configuration();

        $b->{unclassified__} = $c->parameter("bayes_inclassified_probablity") || 0.5;

        $b->{debug} = 0;
        $b->{parser__}->{debug} = 0;
        $b->start();
    

#        my @dirs = glob $c->parameter("html_archive_dir") . "/*";

        # setup dummy buckets

#        foreach my $dir (@dirs) {
#            my ($x, $bucket) = split /\//,$dir;
#            $b->{total__}{$bucket} = 100;
#            $b->{colors__}{$bucket} = 'black';
#        }
    }


    my @files   = glob $ARGV[0];
    foreach my $file (@files) {
        open FILE, "<$file" or die "Unable to open $file :$!";
        my $nexthdr;
        while (<FILE>) {
            my $hdr = "";
            if (/^From \?{3}\@\?{3} / or defined $nexthdr) {
                $hdr = defined $nexthdr ? $nexthdr : $_;
                my @body = ();
                undef $nexthdr;
LINE:           while  (<FILE>) {
                    if (/From \?{3}\@\?{3} /) {
                        $nexthdr = $_;
                        last LINE;
                    }
                    push @body,$_;
                }
                my $fname = "pf_message.tmp";
                open MSG, ">$fname" or die "Unable to create $fname :$!\n";
                binmode MSG;
                foreach (@body) {
                    print MSG $_;
                }
                close MSG;
                my $d = $b->classify_file($fname);
                if (open(FROM,"<$fname")) {
                    if (open (TO, ">>pf_${d}.mbx")) {
                        binmode FROM;
                        print TO $hdr;
                        while (<FROM>) {
                            print TO $_;
                        }
                        close TO;
                    }
                    close FROM;
                }
                unlink($fname);
            }
        }
        close FILE;
    }

    my $end_time = time;
    print "Done, process took " . ($end_time - $start_time) .  " seconds\n";

    
}
else
{
    print "classify_eudora_mbx.pl - determine bucket for mails in mbx file and output pf_[bucketname].mbx files for classified mail. Original input Eudora mbx file left untouched.\n\n";
    print "Usage: classify_eudora_mbx.pl <eudora mbx file>\n";
    print "       <eudora mbx files>     filename of Eudora *.mbx file to classify\n";
}