Here is my quandry. I use Spamassassin to filter the mail on my mail server, a mail server that hosts more that just my mail. It works quite well. But those wiley spammers are good at hiding their spam from the likes of SA. I want to use the sa-learn to start teaching my bayesian filter what gets missed. Using outlook, I can automatically mark items as junk, which does it's own learning in outlook. Wanting to share this education with everyone else using the mail server, I would like to create a PERL script that reads my 'Junk Mail' folder in the Outlook PST file and runs those mails against the sa-learn binary.
Has anyone done that before, a quick search on CPAN didn't find anything, likewise my searching skills found minimal solutions here within the monestary.