Bogofilter training on KMail mailboxes?

[Todd Slater]

On 1/16/06, paul cooke <paul.cooke100 at blueyonder.co.uk> wrote:
> On Sunday 15 January 2006 17:13, Rob Blomquist wrote:
> > I am attempting to use bogofilter after a couple of years of running
> > SpamAssassin.
> >
> > Under SA, it was easy to teach it about spam: every night I ran a cron job
> > on my spamfolder, which has "maildir" style directories.
> >
> > according to what I have read on the bogofilter page, all I need to do is
> > run "bogofilter -s ~/.Mail/SpamPile/cur/, and the filtering will be done.
> >
> > Last night, while in ~/.Mail/SpamPile/cur/ I tried to run bogofilter -s,
> > and it hung all night.
> >
> > Today I tried while in ~ to run "bogofilter -s .Mail/SpamPile/cur/" and I
> > got "Extra arguments given, first: .Mail/SpamPile/cur/. Aborting."
> >
> > It appears to want a command format I don't understand. Thoughts?
>
> at least you're getting bogofilter to run,
>
> all I get is
>
> paulc at broken-drum:~$ bogofilter -B < /home/paulc/Mail/spamtrap/cur/
> Can't open file 'wordlist.db' in directory '/home/paulc/.bogofilter'.
> error #2 - No such file or directory.
>
> so how do I create this wordlist.db? the faq on the homepage is seriously
> geeky and unhelpful and so's the Man page

You have to first train bogofilter to know what is spam and what is
ham. You do this by registering mails as such. This means you need a
corpus of spam and ham for initial training. You use the -n switch for
ham and -s for spam.

When you do the initial registration the wordlist.db gets created.
wordlist.db lives wherever your bogofilter config file tells it to
live, probably in ~/.bogofilter. Check /etc/bogofilter.cf for other
settings.

HTH,
Todd