felix at compsoc.nuigalway.ie
Wed Nov 3 11:37:21 GMT 2004
Quoting Kenn Humborg <kenn at bluetree.ie>:
> > sa-learn just took about 10mins of cpu time to process a 30Mb/1000
> > message mbox file of ham.
> > Is this normal or am I doing it wrong?
> Did you use sa-learn --mbox? If you forget the --mbox, sa-learn
> will parse it as one very large message.
It sounds to me as though there is a load of attachments with the mail. I have
about 1000 ham and spam messages in mboxes totaling sizes of 6.8MB and 8MB
Perhaps you might want to consider deleting some of the non text attachments
from the mail. As the size of the mbox sounds just a little large for 1000
I'm using mbox format as well and the time taken to perform
sa-learn --spam --no-rebuild --mbox ~/mail/spam/spam && sa-learn --ham
--no-rebuild --mbox ~/mail/spam/ham && sa-learn --rebuild
is about 1 minute.
"Nothing's foolproof to a sufficently talented fool"
More information about the ILUG