Re: processing of mail considered spam

2009-07-12 Thread Michelle Konzack 
Hi Noah,

Am 2009-07-12 11:58:23, schrieb Admin:
> okay how do I script sa-learn to learn the contents of a particular
> file.

I use courier (Maildir) with procmail and if I have spams, then  I  move
it to the folder "INBOX.Learn_IS_Spam" and if I have ham  found  in  the
spamfolders, I move it to "INBOX.Learn_NO_Spam".

[ '/home/michelle.konzack/bin/check_sa_learn' ]-
#!/bin/bash

VAL_USERS=$(cut -d ':' -f1,3,6 /etc/passwd |tr ':' ' ')
HN=$(hostname -s)
LOG=/var/log/check_sa_learn.$(date +%Y-%m-%d).log

touch ${LOG}
chown root:adm ${LOG}
chmod 660 ${LOG}

(echo "${VAL_USERS}" ; echo) |
while read VAL_USER VAL_UID VAL_HOME
do
  if [ -n "${VAL_USER}${VAL_UID}${VAL_HOME}" ] ; then
if [ ${VAL_UID} -ge 1000 ] && [ ${VAL_UID} -lt 65000 ] ; then


  NUM_SPAM=$(ls ${VAL_HOME}/Maildir/.Learn_IS_Spam/{cur,new}/* 2>/dev/null 
|wc -l)
  NUM_HAM=$(ls  ${VAL_HOME}/Maildir/.Learn_NO_Spam/{cur,new}/* 2>/dev/null 
|wc -l)

  NUM=$((${NUM_SPAM}+${NUM_HAM}))

  if [ ${NUM} -gt 0  ] ; then

echo "Working on user:  ${VAL_USER}" 1>&2
DATE=$(date +"%Y-%m-%d %H:%M:%S")
echo -n "${DATE} [${HN}] USER=${VAL_USER}, NUM_SPAM=${NUM_SPAM}, 
NUM_HAM=${NUM_HAM}" >>${LOG}



if [ -d ${VAL_HOME}/Maildir/.Learn_IS_Spam ] ; then
  LIST_SPAM=$(find ${VAL_HOME}/Maildir/.Learn_IS_Spam/cur -type f -mmin 
+2 ; find ${VAL_HOME}/Maildir/.Learn_IS_Spam/new -type f -mmin +2)
  echo -n ", SPAM=" >>${LOG}
  if [ -n "${LIST_SPAM}" ] ; then
for MSG in ${LIST_SPAM} ; do
  su - ${VAL_USER} -c "sa-learn --spam ${MSG} ; if [ $? -eq 0 ] ; 
then procmail <${MSG} ; if [ $? -eq 0 ] ; then rm -f ${MSG} ; fi ; fi"
done
echo -n "scaned" >>${LOG}
  else
echo -n "none" >>${LOG}
  fi
fi



if [ -d ${VAL_HOME}/Maildir/.Learn_NO_Spam ] ; then
  LIST_HAM=$(find ${VAL_HOME}/Maildir/.Learn_NO_Spam/cur -type f -mmin 
+2 ; find ${VAL_HOME}/Maildir/.Learn_NO_Spam/new -type f -mmin +2)
  echo -n ", HAM=" >>${LOG}
  if [ -n "${LIST_HAM}" ] ; then
for MSG in ${LIST_HAM} ; do
  DE=$(date +%s)
  for CNT in $(seq --equal-width 1 100) ; do
if [ ! -f ${VAL_HOME}/Maildir/new/${DE}.${CNT}.${HN} ] ; then
  su - ${VAL_USER} -c "sa-learn --ham ${MSG} ; if [ $? -eq 0 ] 
; then mv -f ${MSG} ${VAL_HOME}/Maildir/new/${DE}.${CNT}.${HN} ; fi"
  break
fi
  done
done
echo "scaned" >>${LOG}
  else
echo "none" >>${LOG}
  fi
fi



  fi
fi
  fi
done


and a cronjob:

[ '/etc/cron.d/check_sa_learn' ]
MAILTO=cron

*/10 * * * *root/home/michelle.konzack/bin/check_sa_learn


and works perfectly...  on more then 2000 user!

Thanks, Greetings and nice Day/Evening
Michelle Konzack
Systemadministrator
Tamay Dogan Network
Debian GNU/Linux Consultant


-- 
Linux-User #280138 with the Linux Counter, http://counter.li.org/
# Debian GNU/Linux Consultant #
Michelle Konzack   c/o Shared Office KabelBW  ICQ #328449886
+49/177/9351947Blumenstasse 2 MSN LinuxMichi
+33/6/61925193 77694 Kehl/Germany IRC #Debian (irc.icq.com)


signature.pgp
Description: Digital signature

Re: processing of mail considered spam

2009-07-12 Thread Jari Fredriksson 
> Jari Fredriksson wrote:
>>> Hi there,
>>> Some spam is getting past the spamassassin.  So I;d like to devise a
scheme where I manually place the emails not caught by spamassass in
my
>>> 'spam-mail' folder.  Is there any way to get spamassassin to process the
>>> contents of the folder so I those accepted messages are considered
spam
>>> in the future?
>> It depends. If the "spam-mail folder" is a folder in a POP3 client, it is
>> not easy to automate it. if the "spam-mail folder" is a folder in a
IMAP-server, and the server uses Maildir format under Linux or other
Unix,
>> it is easy. The mails are plain text files in the server file system, and
>> a cron job can easily run sa-learn --spam with that folder.
>> If the server is some MS Exchange I have no idea. Maybe the folder
could
>> be downloaded periodically with fetchmail or such, and fed to sa-learn.
Anyway, SpamAssassin itself does not help besides sa-learn, so this
needs
>> scripting on your part.
>
> okay how do I script sa-learn to learn the contents of a particular
file.
>
> Cheers,
>
> Noah

If the user used to call spamc is noah, then this might do it. The last
parameter is file name or a folder name, in which case all files in the
folder will be learnt.

man sa-learn will tell more..



#!/bin/sh

/usr/bin/sa-learn -u noah --spam /home/noah/Maildir/spam-mail/cur

Re: processing of mail considered spam

2009-07-12 Thread Admin 

Jari Fredriksson wrote:

Hi there,

Some spam is getting past the spamassassin.  So I;d like to devise a
scheme where I manually place the emails not caught by spamassass in my
'spam-mail' folder.  Is there any way to get spamassassin to process the
contents of the folder so I those accepted messages are considered spam
in the future?



It depends. If the "spam-mail folder" is a folder in a POP3 client, it is
not easy to automate it. if the "spam-mail folder" is a folder in a
IMAP-server, and the server uses Maildir format under Linux or other Unix,
it is easy. The mails are plain text files in the server file system, and
a cron job can easily run sa-learn --spam with that folder.

If the server is some MS Exchange I have no idea. Maybe the folder could
be downloaded periodically with fetchmail or such, and fed to sa-learn.

Anyway, SpamAssassin itself does not help besides sa-learn, so this needs
scripting on your part.




okay how do I script sa-learn to learn the contents of a particular file.

Cheers,

Noah

Re: processing of mail considered spam

2009-07-12 Thread Jari Fredriksson 
> Hi there,
>
> Some spam is getting past the spamassassin.  So I;d like to devise a
> scheme where I manually place the emails not caught by spamassass in my
> 'spam-mail' folder.  Is there any way to get spamassassin to process the
> contents of the folder so I those accepted messages are considered spam
> in the future?
>

It depends. If the "spam-mail folder" is a folder in a POP3 client, it is
not easy to automate it. if the "spam-mail folder" is a folder in a
IMAP-server, and the server uses Maildir format under Linux or other Unix,
it is easy. The mails are plain text files in the server file system, and
a cron job can easily run sa-learn --spam with that folder.

If the server is some MS Exchange I have no idea. Maybe the folder could
be downloaded periodically with fetchmail or such, and fed to sa-learn.

Anyway, SpamAssassin itself does not help besides sa-learn, so this needs
scripting on your part.

processing of mail considered spam

2009-07-12 Thread Admin 

Hi there,

Some spam is getting past the spamassassin.  So I;d like to devise a 
scheme where I manually place the emails not caught by spamassass in my 
'spam-mail' folder.  Is there any way to get spamassassin to process the 
contents of the folder so I those accepted messages are considered spam 
in the future?


Cheers,
Noah