an easy and consistent way of installing software ported to FreeBSD
bmf is an extremely efficient Bayesian mail filter based on
Paul Graham's article, "A Plan for Spam." It's fast, small,
efficient, and more versatile than other Bayesian filters.
-- bmf-user@lists.sourceforge.net
WWW: http://sourceforge.net/projects/bmf
============================
=== How to integrate bmf ===
============================
For maildrop(1), use only one of the following examples. The first
sample invokes bmf in its normal mode of operation and the second
invokes bmf as a filter:
### Sample One
# Invoke bmf and use return code to filter spam in one step
exception {
`bmf`
if ( $RETURNCODE == 0 )
to $MAILDIR/spam
}
### Sample Two
# Invoke bmf as a filter
exception {
xfilter "bmf -p"
if (/^X-Spam-Status: Yes/)
to $MAILDIR/spam
}
For procmail(1), use only one of the following samples. SPAM will be
filtered into $MAILDIR/spam. The first sample invokes bmf in its
normal mode of operation and the second invokes bmf as a filter.
### begin sample one ###
# Invoke bmf and use return code to filter spam in one step
:0HB
* ? bmf
| formail -A"X-Spam-Status: Yes, tests=bmf" >>$MAILDIR/spam
### begin sample two ###
# Invoke bmf as a filter
:0 fw
| bmf -p
# Filter spam
:0:
^X-Spam-Status: Yes
$MAILDIR/spam
If you put bmf in your maildrop or procmail scripts as suggested
above, it will always register an email as either spam or non-spam.
To reverse this registration and train bmf, the following mutt macros
may be useful:
macro index \ed "
macro index \et "
macro index \eu "
These will override these commands:
Please see the ${PREFIX}/share/bmf/README for further details.
bin/bmf
bin/bmfconv
%%DATADIR%%/README
@dirrm %%DATADIR%%