[Spambayes] Feature Request

Tony Meyer tameyer at ihug.co.nz
Mon May 2 03:50:38 CEST 2005

[Bob Coe]
>> I don't recall that anyone has explained how to trim the database
>> with the Outlook plugin.

[Skip Montanaro]
> Yeah, I sort of glossed over that bit.  I'm not sure there's 
> really a convenient way to do it either (not being an Outlook 
> user I can't explore the idea in that context).
> In any case, this is all very un-automatic.  It might be nice 
> of the Outlook plugin could do the grunt work for you.

There really isn't any convenient way to trim the database in Outlook.
AFAIK the only way is as Skip described, where you manually maintain (adding
and pruning) ham/spam training folders and do a "rebuild entire database"
retrain every so often.  A fair bit of work.

The situation with sb_server isn't really that much better, FWIW.  I'd
really like to change this (I have fading hopes of finishing up my work with
this in time to put it in 1.1), but just don't have the time to work on code
at this level at the moment.

1.1 does keep track of mail that is trained, by date, so, assuming that the
mail can be found again with the Outlook ids that we use to identify the
mail, it would be theoretically possible to automatically purge old mail.
There's also IIRC a fairly old patch (by Skip, I think) that adds date
information to tokens in the database that would allow doing this in a
different way.

> My training regimen 
> (using the train-to-exhaustion script, tte.py) also has an 
> option to delete messages that train correctly on each pass, 
> thus limiting growth of those special mailboxes.

I'd really like to make some sort of tte (optionally) available to
Outlook/sb_server users, too.  One of those 'if I have time' things, unless
someone beats me to it.


Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes.
http://www.massey.ac.nz/~tameyer/writing/reply_all.html explains this. 

More information about the Spambayes mailing list