On Apr 14, 2004, at 1:22 AM, Bob Stia wrote:
> Hello Sluggers,
>
> Somewhat confused with sa-learn. I have a script to run sa-learn and
> it
> seems to work but not to work. Originally built a pretty good data
> base using sa-learn manually. Caught most of the spam and put it into
> a folder in KDE called "Spam" (using KDE filters) Also made a folder
> called "MissedSpam" and when something slipped through I would put it
> in there to run sa-learn on. Over the past few months when I ran the
> script sa-learn it did not learn anything. "Lear
You shouldn't be running "sa-learn" on missed spam messages, you should
be running "spamasassin -r" on them instead. It combines both training
the built-in bayesian system like sa-learn, and also reporting to any
DNS-based things like razor or pyzor that spamassassin might be using.
Plus, since it's, um, well, part of the spamassassin package, it will
definitely be calling sa-learn correctly for you. :)
Your script looks ok, so I don't know why sa-learn wouldn't be working,
unless it just can't find your spamassassin configs or something.
(Which is maybe another reason to use "spamassassin -r" and not call
sa-learn directly.)
I usually do this against my "misses" folder every couple of days:
cat misses | formail -s spamassassin -r
Since the "spamassassin" command-line wants one message at a time, use
"formail" to split apart each individual message in the folder and send
to the "spamassassin -r" tool one at a time. You might just consider
whacking your spamassassin bayes databases entirely and starting over
using the different syntax.
Also, I would recommend NOT deleting spam once it's been processed, but
saving it somewhere away from your "misses" folders, if only so you can
"start over" with your Bayesian training if you need to, whether you
want to try some new filtering tool or an upgrade invalidates your
existing training or just because. I have a corpus of probably
hundreds of thousands of spams over the last year and a half or so,
which helps immensely in training new filters I might want to try out
and for "seeding" filters if I want to set up spamassassin for someone
else.
(I'm also keeping my fingers crossed that one day a law will be passed
demanding a payment of $1 per spam received, retroactively...)
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
"We all enter this world in the | Support Electronic Freedom
same way: naked; screaming; soaked | http://www.eff.org/
in blood. But if you live your | http://www.anti-dmca.org/
life right, that kind of thing |---------------------------
doesn't have to stop there." -- Dana Gould
-----------------------------------------------------------------------
This list is provided as an unmoderated internet service by Networked
Knowledge Systems (NKS). Views and opinions expressed in messages
posted are those of the author and do not necessarily reflect the
official policy or position of NKS or any of its employees.
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 20:36:47 EDT