[prev in list] [next in list] [prev in thread] [next in thread] 

List:       spambayes
Subject:    [Spambayes] Spambayes for MAC
From:       skip () pobox ! com (skip at pobox ! com)
Date:       2007-09-15 1:48:22
Message-ID: 18155.14822.390236.404555 () montanaro ! dyndns ! org
[Download RAW message or body]


(warning: long response... also, please keep spambayes at python.org in your
replies so you get the benefit of the other users' expertise...)

    Jessica> I may be a bit incompetant here, but I can?t find a link to
    Jessica> download either one of the things you mentioned.  DO you know
    Jessica> where I can find them?

Yes, everything comes in one bundle.  Just go here:

    http://sourceforge.net/project/showfiles.php?group_id=61702&package_id=58141&release_id=518556

and download the "spambayes-1.1a4.tar.gz" package.  You may need to expand
this manually, though on my Mac using Firefox it just prompted me to open it
with Stuffit Expander.  That left a folder on my desktop named
"spambayes-1.1a4".

Now it gets a bit interesting because we don't have an official Mac
installer (maybe someday).  I don't know how facile you are with the Unix
command line.  Sorry if this is too basic or too advanced.  You need to
launch your Terminal app (in Applications>Utilities).  When you run it a
window will be launched with a prompt ending in a "$".  At that prompt type

    cd ~/Desktop/spambayes-1.1a4

and hit return.  This will take you to the expanded SpamBayes distribution.
At the next prompt type

    ls

and hit return again.  You'll see a list of files in that folder:

    CHANGELOG.txt           README-DEVEL.txt        scripts
    LICENSE.txt             README.txt              setup.py
    MANIFEST.in             TESTING.txt             spambayes
    NEWTRICKS.txt           WHAT_IS_NEW.txt         src
    Outlook2000             contrib                 testtools
    PKG-INFO                pspam                   utilities
    POP3PROXY.txt           runtest.sh              windows

The file setup.py is key here.  Type

    /usr/bin/python setup.py install

and hit return.  That will install the SpamBayes tools into

    /System/Library/Frameworks/Python.framework/Versions/2.3/bin

Let's shorten up the rest of the commands by changing directory to the above
folder:

    cd /System/Library/Frameworks/Python.framework/Versions/2.3/bin

Now type

    ./sb_server.py --help

You will be rewarded with a hopefully not-to-cryptic help message.  Key is
the summary of the tool's usage:

    sb_server.py [options] [<server> [<server port>]]

It also lists the possible options.  The important one is

    -b      : Launch a web browser showing the user interface.

A likely command to get things going would be

    ./sb_server.py -b

Your web browser will open up looking at this page:

    http://localhost:8880/

Click the "Configuration Page" link.  Let's assume you normally get your
email via POP from pop.comcast.net.  Enter pop.comcast.net in the Remote
Servers field.  Leave the rest of the options alone for now.  Next save your
configuration (the Save button is at the bottom of the page).

Now, return to Entourage.  In your account setup instead of telling
Entourage to fetch mail from pop.comcast.net you want to tell it to fetch
mail from "localhost" or "127.0.0.1" (both without the quotes).  Either
should work.  What you have done is interpose the SpamBayes server between
Entourage and your regular POP server:

    +---------------+	    +---------+	     +---------+
    |pop.comcast.net|  <--> |SpamBayes| <--> |Entourage|
    +---------------+	    +---------+	     +---------+

SpamBayes will score every message that arrives from your POP server.
Initially it won't know anything about your mail preferences, so you have to
start classifying those it's seen.  Back on the SpamBayes start page
(http://localhost:8880/) click the Review messages link.  If it has seen
some messages you'll be presented with a tabular display which allows you to
mark messages individually or in groups as spam or ham.  Do that for a few
messages.  Now we're to the point where I can't help much more since I don't
actually use this particular SpamBayes application.  I don't actually have
any messages to classify.  I believe you should be able to rescore
everything after classifying just a few spam and ham messages.

Hopefully that's enough to get you started.  If you wade in but find
yourself up to your neck in quicksand, send us another note and someone here
can probably guide you along.

-- 
Skip Montanaro - skip at pobox.com - http://www.webfast.com/~skip/

[prev in list] [next in list] [prev in thread] [next in thread]