On Thu, 27 May 2010 21:16:33 +0200 Stevan Bajić <ste...@bajic.ch> wrote:
> On Thu, 27 May 2010 08:11:46 -0700 > Bradley Giesbrecht <bradley.giesbre...@gmail.com> wrote: > > > I do not have dspam signature in message body, only in header. > > > > Using Apple Mail.app I set my preferences to show "X-*" headers. > > > > If I forward normally to dspam for retraining the headers and body are > > quoted in the forward and sum(dspam_stats.spam_misclassified) does not > > change. > > If I forward as an attachment to sum(dspam_stats.spam_misclassified) > > is increased by 1. > > If I select 10 messages and forward as attachments > > sum(dspam_stats.spam_misclassified) is increased by 1. > > Could you forward me a bunch of messages as attachment so that I can see how they arrive at my end? > > Does anyone know if it works to forward multiple email as attachments > > to dspam for retraining? > > > Some time ago I have coded for someone on the DSPAM mailing list a code in > Perl that is able to extract multiple signatures and retrain all of them. If > you want I can quickly search for the script or quickly write you one. It > should be no big time investment creating such a retraining script. Let me > know if you want one. > I found my old script. Boy is that script old. I coded it in 2006 after some one asked me how to make the script from the old DSPAM wiki (http://dspamwiki.expass.de/DspamRetrainScript) to be able to process more then just one signature at once. Here a quick test with a message just containing: =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From: test To: test Subject: test X-DSPAM-Signature: 0,12345678901234567890 this is a test !DSPAM:1,12345678901234567890! !DSPAM:2,12345678901234567890! !DSPAM:3,12345678901234567890! !DSPAM:4,12345678901234567890! =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Test with standard parameters: =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:2,12345678901234567890!\n!DSPAM:3,12345678901234567890!\n!DSPAM:4,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=error --mode=teft --client --user t...@localhost.localdomain DEBUG: Found DSPAM signature '0,12345678901234567890' in header. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=0\,12345678901234567890 --mode=teft --user te...@localhost\.localdomain nyx ~ # =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Now the same test telling the retraining script to parse the body and search for signatures: =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:2,12345678901234567890!\n!DSPAM:3,12345678901234567890!\n!DSPAM:4,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=error --header-only=no --mode=teft --client --user t...@localhost.localdomain DEBUG: Found DSPAM signature '1,12345678901234567890' in body. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=1\,12345678901234567890 --mode=teft --user te...@localhost\.localdomain DEBUG: Found DSPAM signature '2,12345678901234567890' in body. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=2\,12345678901234567890 --mode=teft --user te...@localhost\.localdomain DEBUG: Found DSPAM signature '3,12345678901234567890' in body. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=3\,12345678901234567890 --mode=teft --user te...@localhost\.localdomain DEBUG: Found DSPAM signature '4,12345678901234567890' in body. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=4\,12345678901234567890 --mode=teft --user te...@localhost\.localdomain nyx ~ # =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The script detects if it already has processed a signature and does not call DSPAM again. A small test with 4 times the same signature in the body: =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=error --header-only=no --mode=teft --client --user t...@localhost.localdomain DEBUG: Found DSPAM signature '1,12345678901234567890' in body. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=1\,12345678901234567890 --mode=teft --user te...@localhost\.localdomain nyx ~ # =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= And the script is checking (if needed) for valid --class, --source and --mode values: =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=not_valid_source --header-only=no --client --user t...@localhost.localdomain ERROR: Source 'not_valid_source' is not known by DSPAM. nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=not_valid_class --source=error --header-only=no --client --user t...@localhost.localdomain ERROR: Class 'not_valid_class' is not known by DSPAM. nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=error --mode=not_valid_mode --header-only=no --client --user t...@localhost.localdomain ERROR: Mode 'not_valid_mode' is not known by DSPAM. nyx ~ # =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The script is using the current user if no --user switch is used: =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=error --header-only=no --client DEBUG: Found DSPAM signature '1,12345678901234567890' in body. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=1\,12345678901234567890 --user root nyx ~ # =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= And the script can be used for inoculation / corpus fed as well: =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=corpus --client DEBUG: Piping message to: /usr/bin/dspamc --source=corpus --class=spam --user root nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=inoculation --client DEBUG: Piping message to: /usr/bin/dspamc --source=inoculation --class=spam --user root nyx ~ # =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= And the script is able to process spam/ham aliases if no --user switch is used and change the class to spam/innocent even if the script is called with a (wrong) --class value. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=spam --source=error --client --rcpt-to notspam-t...@my.tld DEBUG: Found DSPAM signature '0,12345678901234567890' in header. DEBUG: /usr/bin/dspamc --source=error --class=innocent --signature=0\,12345678901234567890 --user te...@my\.tld nyx ~ # echo -ne 'From: test\nTo: test\nSubject: test\nX-DSPAM-Signature: 0,12345678901234567890\n\nthis is a test\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n!DSPAM:1,12345678901234567890!\n'|./dspam-retrain-forward.pl --debug --class=innocent --source=error --client --rcpt-to spam-t...@my.tld DEBUG: Found DSPAM signature '0,12345678901234567890' in header. DEBUG: /usr/bin/dspamc --source=error --class=spam --signature=0\,12345678901234567890 --user te...@my\.tld nyx ~ # =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= etc, etc, etc... the script however does not try to decode base64 attachments and such. When I coded that script I did not wanted to invest to much time in making it über intelligent. > > Thank you, > > Bradley Giesbrecht > > -- Kind Regards from Switzerland, Stevan Bajić ------------------------------------------------------------------------------ _______________________________________________ Dspam-user mailing list Dspam-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dspam-user
