[Spambayes] Speeding up Outlook plugin

[Tony Meyer]

> Is there any way to speed up processing of messages by SB? 

If you use a pickle for storage rather than the default bsddb opening
Outlook and training will be slower, but classifying should be faster.  (You
can only do this with the source version or the upcoming 1.1).

Scoring from a pst file is probably faster than from Exchange (because
SpamBayes has to pull down the message content, which is probably going to
be quicker with a local file than via the network), if that is possible.

Do you have any experimental options enabled?  Some of those may slow
scoring down.

I was offered a pyrex version of the main classifier code, which would speed
things up a bit (in 1.1), but that hasn't arrived yet.  It's possible that
we might work on speed, but I suspect that this would be after 1.1.

> By any chance, you haven't heard of any SB implementation 
> (barring the Outlook plugin) that is able to scan an Exchange 
> mailbox using its native MAPI/whatever protocol even when the 
> user is not interactively logged on (kind of a service/daemon thing)?

Is this for just one user?  In other words do the Outlook plug-in's scoring
job without Outlook being open?  Or is this running on the server?  There's
some information about doing the latter on
<http://spambayes.org/server_side.html>.  We could create a script for the
former, I suppose, but you would still be logged in really, just not through
Outlook.  It would be much simpler to just have Outlook open.

=Tony.Meyer

-- 
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.