Zarafa Bayesian Learning

From SME Server
Jump to: navigation, search

Contents

Zarafa Bayesian learning

This howto enables SpamAssasin Bayesian learning for Zarafa

The DMZS script (LGPL) works over IMAP. It reads the mail from two folders (LearnAsSpam and LearnAsHam) and feeds it to SpamAssasin's sa-learn. This script is implemented here in a way that it makes use of public folders in Zarafa.

Installation

Bayes

yum install perl-Mail-IMAPClient --enablerepo=smecontribs

Create a new script-file:

nano -w /usr/bin/DMZS-sa-learn.pl

Paste the code below in this script-file and change the 'SpamAdminPassword' into a proper (strong) password:

#!/usr/bin/perl
#
# Process mail from imap server shared folder 'Public folders/LearnAsSpam' & 'Public folders/LearnAsHam' through spamassassin sa-learn
# dmz@dmzs.com - March 19, 2004
# http://www.dmzs.com/tools/files/spam.phtml
# http://www.dmzs.com/tools/files/spam/DMZS-sa-learn.pl [modified for SMEServer]
# LGPL

use Mail::IMAPClient;

my $debug=0;
my $salearn;

# # # # # # # # # # EDIT USER AND PASSWORD  # # # # # # # # # #

my $imap = Mail::IMAPClient->new( Server=> '127.0.0.1:8143',
                                  User => 'SpamAdmin',
                                  Password => 'SpamAdminPassword',
                                  Debug => $debug);

if (!defined($imap)) { die "IMAP Login Failed"; }

# If debugging, print out the total counts for each mailbox
if ($debug) {
 my $spamcount = $imap->message_count('Public folders/LearnAsSpam');
 print $spamcount, " Spam to process\n";

 my $nonspamcount = $imap->message_count('Public folders/LearnAsHam');
 print $nonspamcount, " Notspam to process\n" if $debug;
}

# Process the spam mailbox
$imap->select('Public folders/LearnAsSpam');
my @msgs = $imap->search("ALL");
for (my $i=0;$i <= $#msgs; $i++)
{
 # I put it into a file for processing, doing it into a perl var & piping through sa-learn just didn't seem to work
 $imap->message_to_file("/tmp/salearn",$msgs[$i]);

 # execute sa-learn w/data
 if ($debug) { $salearn = `/usr/bin/sa-learn -D --no-sync  --spam /tmp/salearn`; } 
 else { $salearn = `/usr/bin/sa-learn --no-sync  --spam /tmp/salearn`; }
 print "-------\nSpam: ",$salearn,"\n-------\n" if $debug;

 # delete processed message
 $imap->delete_message($msgs[$i]);
 unlink("/tmp/salearn");
}
$imap->expunge();
$imap->close();

# Process the not-spam mailbox
$imap->select('Public folders/LearnAsHam');
my @msgs = $imap->search("ALL");
for (my $i=0;$i <= $#msgs; $i++)
{
 $imap->message_to_file("/tmp/salearn",$msgs[$i]);
 # execute sa-learn w/data
 if ($debug) { $salearn = `/usr/bin/sa-learn -D --no-sync  --ham /tmp/salearn`; }
 else { $salearn = `/usr/bin/sa-learn --no-sync  --ham /tmp/salearn`; }
 print "-------\nNotSpam: ",$salearn,"\n-------\n" if $debug; 

 # delete processed message
 $imap->delete_message($msgs[$i]);
 unlink("/tmp/salearn");
}
$imap->expunge();
$imap->close();

$imap->logout();

# integrate learned stuff
my $sarebuild = `/usr/bin/sa-learn --sync`;
print "-------\nRebuild: ",$sarebuild,"\n-------\n" if $debug;

Set proper permissions on the script:

 chmod 555 /usr/bin/DMZS-sa-learn.pl

Zarafa

Create a user-account in Zarafa for reading the public spam-folders.

db method, Replace the <MyPassword> with a proper strong password.

 zarafa-admin -c 'SpamAdmin' -p '<MyPassword>' -f 'Spam Administration Account' -e root@localhost

If you have configured Zarafa to use the unix method and if you enable Zarafa usage on a per user base:

 db accounts setprop SpamAdmin zarafa enabled
 /etc/e-smith/events/actions/qmail-update-user

Login to Zarafa with an account that has admin rights and make two new folders LearnAsSpam and LearnAsHam under: Public folder > Public folders. Set the permissions (right-click folder > Properties > Permission-tab) on both these new folders to:

 Spam administration account
 * Folder visible
 * Read items
 * Edit items: all
 * Delete items: all
 
 Everyone (and/or other users/groups you've added at least need:)
 * Folder visible
 * Create items
 * Edit items: none
 * Delete items: none
Important.png Note:
Dropping mail in the public 'LearnAsHam' folder may pose a privacy problem if permissions are set less restrictive as shown above!

Cron

Create a new crontab fragment:

 nano -w /etc/e-smith/templates/etc/crontab/91_SpamAssasinLearn

Add the following to the template (change the execution times to your own likings -- Wikipedia on Cron):

 # Running the Spamassasin Bayesian SPAM learning script every hour from 8:00 to 22:00 during weekdays
 0 8-22 * * 1-5 root /usr/bin/DMZS-sa-learn.pl

Make the new fragment active by expanding the template:

 expand-template /etc/crontab

Configuration

Spamassassin has to be enabled in the Email Panel

Bayesian learning has to be enabled and configured in SME with

config setprop spamassassin UseBayes 1 
config setprop spamassassin BayesAutoLearnThresholdSpam 6.00 
config setprop spamassassin BayesAutoLearnThresholdNonspam 0.10 
expand-template /etc/mail/spamassassin/local.cf 
sa-learn --sync --dbpath /var/spool/spamd/.spamassassin -u spamd 
chown spamd.spamd /var/spool/spamd/.spamassassin/bayes_* 
chown spamd.spamd /var/spool/spamd/.spamassassin/bayes.mutex 
chmod 640 /var/spool/spamd/.spamassassin/bayes_* 
signal-event email-update

These commands will:

  • enable bayesian filter
  • 'autolearn' as SPAM any email with a score above 6.00
Note: SpamAssassin requires at least 3 points from the header, and 3 points from the body
to auto-learn as spam.
Therefore, the minimum working value for this option is 6, to be changed in increments of 3,
12 considered to be a good working value..
  • 'autolearn' as HAM any email with a score below 0.10

Usage

Warning.png Warning:
All mail dropped in the LearnAsSpam and LearnAsHam folders will be automatically deleted !!

  • Move/copy spam messages that are delivered to your Inbox to the public LearnAsSpam folder.
  • COPY regular messages that end up in your Junk E-mail folder to the public LearnAsHam folder.

After the messages have been processed they will be deleted to save your valuable space.

Personal tools
Koozali SME Server wiki