########################################################################
# Spamassassin
#
# WARNING: Put as few recipes as possible above the spam recipes
#
# The condition line ensures that only messages smaller than 250 kB
# (250 * 1024 = 256000 bytes) are processed by SpamAssassin. Most spam
# isn't bigger than a few k and working with big messages can bring
# SpamAssassin to its knees.
#
# The lock file ensures that only 1 spamassassin invocation happens
# at 1 time, to keep the load down.

BEFORE=`date +'%s'`

# Changed on 24/10/17 - 2 emails in last 2 days over 256K limit
#:0fw: spamassassin.lock
#* < 256000
#| spamassassin
:0fw: spamassassin.lock
* < 1024000
| spamassassin


AFTER=`date +'%s'`

# Get the X-Spam-Status, X-Spam-Language, Content-Length & Message-Id
# Discard any leading and trailing blanks
SPAMSTAT=`formail -cx "X-Spam-Status: " \
          | sed -e 's/^[ ]*//g' -e 's/[ ]*$//g' -e 's/,[[:space:]]\{2,\}/,/g'`
SPAMSTAT=${SPAMSTAT:-"NULL"}

# Bayes score
BSCORE=`formail -cx "X-Spam-Report: " \
        | sed -e 's/.*\[score\://g' -e 's/\].*//g' -e 's/^[ ]*//g' -e 's/[ ]*$//g'`
BSCORE=${BSCORE:-"0"}

# Set VAR1 to VAR2 if VAR2 is set and non-null, and set VAR1 to default "value" otherwise
# VAR1 = ${VAR2:-"value"} 
SALANG=`formail -x "X-Spam-Languages: " \
        | sed -e 's/^[ ]*//g' -e 's/[ ]*$//g' || echo "NULL"`
SALANG=${SALANG:-"NULL"}

SACUST=`formail -cx "X-Spam-Custom: " \
        | sed -e 's/^[ ]*//g' -e 's/[ ]*$//g' -e 's/,[[:space:]]\{2,\}/,/g'`
SACUST=${SACUST:-"NULL"}

CONLEN=`formail -cx "Content-Length: " \
        | sed -e 's/^[ ]*//g' -e 's/[ ]*$//g'`
CONLEN=${CONLEN:-"0"}

MID=`formail -x "Message-Id: " \
     | sed -e 's/^[ ]*//g' -e 's/[ ]*$//g'`
MID=${MID:-"NULL"}

DATETIME=`date "+%Y.%m.%d-%H:%M:%S"`

# Strictly speaking should use a lock file here
FOO=`echo "$DATETIME\t$AFTER\t$BEFORE\t$CONLEN\t$MID\t$SALANG\t$SPAMSTAT\t$BSCORE" >> $SALOG`

# Replacing _TESTS_ with _TESTSSCORES_
CUSTLOG=$HOME/.spamassassin/log/spam-custom.log
BAR=`echo "$DATETIME\t$AFTER\t$BEFORE\t$CONLEN\t$MID\t$SALANG\t$SACUST" >> $CUSTLOG`


# Mails with a score of 15 or higher are almost certainly spam (with 0.05%
# false positives according to rules/STATISTICS.txt). Let's put them in a
# different mbox. (This one is optional.)
#
# 06/06/17 Permanently leaving this commented out to build more
#          representative training set
#:0:
#* ^X-Spam-Level: \*\*\*\*\*\*\*\*
#/dev/null

# All mail tagged as spam (eg. with a score higher than the set threshold)
# is moved to "probably-spam".
:0c
* ^X-Spam-Status: Yes
.spam/

# SA score lt 0 - ham
:0c
* ^X-Spam-Status: No, score=\/[0-9-]+
* ? test $MATCH -lt 0
.sure.ham/

# SA score 0 to 5 - unsure ham
:0c
* ^X-Spam-Status: No, score=\/[0-9-]+
* ? test $MATCH -gt 0
* ? test $MATCH -lt 6
.unsure.ham/


# Contradictory SA and Bayes classifications
#--------------------------------------------------------------
# Opposing SA (spam) and bayes (ham 0.0 or 0.1) classifications
:0c
* ^X-Spam-Status: Yes, score=\/[0-9-]+
* ? test $MATCH -gt 9
* BSCORE ?? 0\.[01]
.sa.spam.bayes.ham/

# Opposing SA (ham) and bayes (spam 0.9 or 1.0) classifications
:0c
* ^X-Spam-Status: No, score=\/[0-9-]+
* ? test $MATCH -lt 0
* BSCORE ?? (0\.9|1\.0)
.sa.ham.bayes.spam/
#--------------------------------------------------------------


# SA score 5 to 10 - unsure spam
:0
* ^X-Spam-Status: Yes, score=\/[0-9-]+
* ? test $MATCH -gt 4
* ? test $MATCH -lt 11
.unsure.spam/

# SA score gt 10 - spam
:0
* ^X-Spam-Status: Yes, score=\/[0-9-]+
* ? test $MATCH -gt 9
.sure.spam/


# End of spamassassin recipes
########################################################################

# vim: set ft=procmail :