[Rspamd-Users] Disable autolearn per user for spam-reporting mailbox

[Stefan Bauer]

Hi Mario,

it depends on how the forwarding is done in the application. The
approach you mentioned requires specific IMAP backends, access to user
mailboxes and strong trust in the users. None of this i have.

Another thing is that i do not want to let users train our filers.
Most of them do not understand the difference between spam and
unwanted messages. The latter are mostly serious mails, the user just
now dont like anymore. Others dont understand the differene between
junk and deleted items. As we have a user helpdesk, we want humans
make the final decision / train the filters.

Stefan

Am Mi., 19. Feb. 2020 um 15:35 Uhr schrieb Mario Thies via Users
<users at lists.rspamd.com>:
>
> Hi,
>
> forwarding is a bad idea, because the relevant header will be stripped.
> What do you think about this ided? The users move the spam mails to
> their junk folder and a cron job is running every day, purge old junk
> mails and train the bayes filter?
>
> You can use rspamc learn_spam /home/..../Maildir/.Junk/cur for that.
>
>
> Am 19.02.2020 um 13:57 schrieb Stefan Bauer:
> > Hi,
> >
> > our users should be able to forward spam-mails to our helpdesk, so we
> > can train the filters better.
> > How can we disable bayes auto-learning for this recipient-mailbox?
> >
> > We have
> >
> > classifier "bayes" {
> >     tokenizer {
> >     name = "osb";
> >     }
> >
> >     backend = "redis";
> >     #servers = "127.0.0.1:6379";
> >     min_tokens = 11;
> >     min_learns = 200;
> >     autolearn = [-0.5, 15]
> >
> > in place. Thank you.
> >
> > PS: Can we maybe just let the users forward mails to our
> > spamtrap-addresses or would that be bad, as coming from a legitimate
> > email?
> --
> Mit freundlichem Gruß
> Mario Thies
>
>
> --
> Users mailing list
> Users at lists.rspamd.com
> https://lists.rspamd.com/mailman/listinfo/users