Bogofilter and reclassifying

[Stroller]

On Dec 5, 2003, at 9:44 am, Nathaniel wrote:
>
> Is this correct?  Most howtos I've seen either recreate a wordlist or 
> just
> mark as spam/ham the entire corpus, but I didn't want to maintain large
> corpuses and wanted something fairly efficient...

Please find attached a shell script I'm currently working on - it scans 
Maildir folders for new messages & calls bogofilter to add their 
contents to the wordlist based on whether they're in a spam 
(~/.Maildir/.Junk.Definite) or ham folder. I hope you might find this 
useful - you could easily add a line to delete old messages once 
they've been read - I intend to tar up the spam as part of the process, 
and to run this as part of a cron job. I'll have mailfilter set to drop 
messages with a high bogosity into a ~/.Maildir/.Junk.Probable folder, 
and so all I'll need to do is peruse them myself & drop them into 
~/.Maildir/.Junk.Definite to confirm their spamicity & train Bogofilter 
further based on their contents.

I think you should find this script self-explanatory, but please feel 
free to ask any questions, or since I'm new to shell programming, any 
suggestions. This script works here, but the usual disclaimers 
(provided as is, no warranty, back-up your data, yadayadayada) apply.

Stroller.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: bogo-update.sh
Type: application/octet-stream
Size: 2634 bytes
Desc: not available
URL: <http://www.bogofilter.org/pipermail/bogofilter/attachments/20031205/5f67dff9/attachment.obj>