sa-learn for per-user sql bayes
Hi, I'm looking for how to tell sa-learn to learn against a per-user Bayes database (in MySQL) instead of learning sitewide. I swear I saw this was going to be in 3.0.2; some kind of change to sa-learn... but a couple hours searching around didn't turn up anything helpful. If I missed something obvious, a tip on where to look would be great! TIA! __ Celebrate Yahoo!'s 10th Birthday! Yahoo! Netrospective: 100 Moments of the Web http://birthday.yahoo.com/netrospective/
Re: sa-learn for per-user sql bayes
email builder wrote: Hi, I'm looking for how to tell sa-learn to learn against a per-user Bayes database (in MySQL) instead of learning sitewide. I swear I saw this was going to be in 3.0.2; some kind of change to sa-learn... but a couple hours searching around didn't turn up anything helpful. If I missed something obvious, a tip on where to look would be great! TIA! As in manually training the database? sa-learn -u username
Re: sa-learn for per-user sql bayes
I'm looking for how to tell sa-learn to learn against a per-user Bayes database (in MySQL) instead of learning sitewide. As in manually training the database? sa-learn -u username I don't know... as long as this will do the same thing spamd does with the username to work against that user's sql-bayes stuff. Looking at the POD docs on the SA website does not show this option, but I do see it in the tarball. I thought the docs on the site would be the most up to date guess not. Thanks much! __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Re: sa-learn for per-user sql bayes
I'm looking for how to tell sa-learn to learn against a per-user Bayes database (in MySQL) instead of learning sitewide. As in manually training the database? sa-learn -u username Can anyone tell me how to verify that a message was learnt correctly for the right user? Running with -D helps a little bit: /usr/bin/sa-learn -D --spam [EMAIL PROTECTED] message Had this line in the output: debug: bayes: Using username: [EMAIL PROTECTED] But I would like to either dig in the Bayes database and verify or send myself another message with the same content and watch it get tagged as spam next time (which doesn't seem to work with my first tests). Any pointers on how to verify learning? Thanks! __ Celebrate Yahoo!'s 10th Birthday! Yahoo! Netrospective: 100 Moments of the Web http://birthday.yahoo.com/netrospective/