> > Dino wrote: > > I'm using standard DB based Bayes > > >>See how many spam (nspam) and ham (nham) are in the database (under > >>the amavis user): > > >>su amavis -c 'sa-learn --dump magic' > >>or: > >>sudo -H -u amavis sa-learn --dump magic > > So I ran sudo -H -u amavis sa-learn --dump magic and here's > the result. > Am I to assume that there are 1429 nspam messages and 2039 > nham messages? > > 0.000 0 3 0 non-token data: bayes > db version > 0.000 0 1429 0 non-token data: nspam > 0.000 0 2039 0 non-token data: nham > 0.000 0 140491 0 non-token data: ntokens > 0.000 0 1170889169 0 non-token data: oldest atime > 0.000 0 1172327394 0 non-token data: newest atime > 0.000 0 1172311175 0 non-token data: last journal > sync atime > 0.000 0 1172271573 0 non-token data: last expiry > atime > 0.000 0 1382400 0 non-token data: last expire > atime delta > 0.000 0 11750 0 non-token data: last expire > reduction count
Yes, these are the message counts. And one can see that your DB is getting updated by the timestamps, which are roughly: Wed Feb 7 14:59:29 2007 Sat Feb 24 06:29:54 2007 Sat Feb 24 01:59:35 2007 You don't have a terribly large database yet. > > >>Debug bayes as the amavis user: > >> (get a sample from me if you need one): > >>wget http://www200.pair.com/mecham/spam/sample-spam.txt > > >>su amavis -c 'spamassassin -D bayes < sample-spam.txt' > >>or: > >>sudo -H -u amavis spamassassin -D bayes < sample-spam.txt > > I ran both commands. The first one amavis -c 'spamassassin -D > bayes < sample-spam.txt' returned nothing it just gave me a > new line, and the second one sudo -H -u amavis spamassassin > -D bayes < sample-spam.txt gave me the following output: > > debug: SpamAssassin version 3.0.4 > debug: Score set 0 chosen. > debug: running in taint mode? no > bad areas in --debug option (bayes)! > Usage: > spamassassin [options] [ < *mailmessage* | *path* ... ] ... Hmmm. You certainly have an older SA. Amavis suggests: Mail::SpamAssassin for doing spam scanning (2.64 or 3.0.5 or >=3.1) Your output should look like: # sudo -H -u amavis spamassassin -D bayes < spam-zjmpTf6PYEGc [12117] dbg: bayes: tie-ing to DB file R/O /var/spool/amavis/.spamassassin/bayes_toks [12117] dbg: bayes: tie-ing to DB file R/O /var/spool/amavis/.spamassassin/bayes_seen [12117] dbg: bayes: found bayes db version 3 [12117] dbg: bayes: DB journal sync: last sync: 0 [12117] dbg: bayes: DB journal sync: last sync: 0 [12117] dbg: bayes: corpus size: nspam = 39697, nham = 86142 [12117] dbg: bayes: header tokens for *p = "" [12117] dbg: bayes: header tokens for X-Envelope-From = " <[EMAIL PROTECTED]>" ... pts rule name description ---- ---------------------- -------------------------------------------------- -0.0 SPF_HELO_PASS SPF: HELO matches SPF record 0.0 HTML_MESSAGE BODY: HTML included in message 4.0 BAYES_99 BODY: Bayesian spam probability is 99 to 100% [score: 1.0000] 0.0 MIME_HTML_ONLY BODY: Message only has text/html MIME parts 3.0 URIBL_OB_SURBL Contains an URL listed in the OB SURBL blocklist [URIs: tipstoolstricks.com] -0.7 AWL AWL: From: address is in the auto white-list ... Note the high BAYES_99 score. > > I've seen message headers with a BAYES hit, but yet they > never score high enough. For instance, here's part of the > header of a message that I know the BAYES filter has been > trained with. As you can see, this message has been processed > by two mail servers both running amavisd-new with SA yet > still it's not tagged as spam. > Can you update your SA ? MrC ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ AMaViS-user mailing list AMaViS-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amavis-user AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3 AMaViS-HowTos:http://www.amavis.org/howto/