[Tony Meyer]
>> No user data is stored in that directory, it's all stored under the
>> Windows Application Data directory. With XP this is usually
>> 'C:\Documents and Settings\{username}\Application
>> Data\Spambayes\Proxy'.
> When looking in there at that ..\Proxy' location, I find
> that there are
> three folders...
> \pop3proxy-ham-cache
> \pop3proxy-spam-cache
> \pop3proxy-unknown-cache
> ...the first two are full of files comprising solely of
> digits and the last,
> the 'unknown' is empty.
These are the message caches. SpamBayes caches a copy of messages when they
are retrieved so that the cached copy can be used for training if necessary.
These messages auto-expire after a certain number of days (7 by default), or
are removed when you use the 'discard' option on the review page. Untrained
messages are in the unknown-cache folder, and trained messages get moved to
the ham- or spam- cache folders.
*Training data*, however, is stored in a single, persistent, database file.
It's probably called hammie.db and will be in the Proxy directory that the
cache directories are in. (There will also be a spambayes.messageinfo.db
that keeps track of messages that have been trained/classified).
> I wondered if simply deleting all these files would have
> the same effect as starting training anew?
No. Deleting these messages will have no effect on training at all.
(Deleting the hammie.db file will remove all training information). The
only effect that deleting a message in the ham- or spam- caches will have is
that you won't be able to find it via the web interface (to correct a
training mistake, for example). That's why these messages auto-expire.
Deleting a message in the unknown cache will remove it from the review page
in the web interface (next time you start sb_server), which means you won't
be able to train on it. It won't effect any existing training, though.
> I wonder, and I realise I'm being a bit cheeky not to mention
> nosy in not just doing what I'm told, If the actual file name
> hammie.db was changed 'in place' so to speak to say 2hammie.db
> so 'hiding' it from SpamBAyes, would the application regenerate
> a new and of course empty hammie.db and so start training anew?
Yes. This is a safe way to do that, because if something goes wrong, you
can always just delete the newly created hammie.db and rename the 2hammie.db
one back to hammie.db to go back to how things were. All of this should
really be done when SpamBayes isn't running, of course.
(NZ stuff answered off-list)
=Tony.Meyer
_______________________________________________
[email protected]
http://mail.python.org/mailman/listinfo/spambayes
Check the FAQ before asking: http://spambayes.sf.net/faq.html
