#!/bin/bash
#Copyright 2003 William Stearns <wstearns@pobox.com>
#Released under the GPL.
SaReportVer='0.1'
echo "Spamassassin Report version $SaReportVer"
echo "Permanent site for this tool: http://www.stearns.org/spamassassin/"

#Here's where you make confiuration changes:

#Spamassassin config files from which to read blacklist_from and
#whitelist_from entries.  Please use full paths.  The user running this
#script only needs read privileges to these.
SaConfigs="/etc/mail/spamassassin/local.cf /home/wstearns/.spamassassin/user_prefs /usr/src/sa-blacklist/sa-blacklist"

#How many sender email addresses there need to be in a given domain
#before we care about it.
MinSenders=2

#These are domains from which we might get ham or spam.  Commonly, 
#although not exclusively, free email providers.
CantColorDomains='@adelphia.net @aol.com @att.net @attbi.com @bigfoot.com @earthlink.net @email.com @eudoramail.com @excite.com @flashmail.com @hotmail.com @juno.com @lycos.com @mail.com @msn.com @pobox.com @rocketmail.com @verizon.com @verizon.net @yahoo.com'

#How many individuals and tokens to display in the spammiest and
#hammiest lists at the end
TopN=300

#You shouldn't need to work with anything below here.
#-----------------------------------------------------------------------

requireutil () {
	while [ -n "$1" ]; do
		if ! type -path "$1" >/dev/null 2>/dev/null ; then
			echo Missing utility "$1". Please install it. >&2
			return 1	#False, app is not available.
		fi
		shift
	done
	return 0			#True, app is there.
} #End of requireutil
requireutil cat check_bayes_db check_whitelist grep head sed sort tail uniq || exit 1


renice +15 -p $$ >/dev/null 2>&1
if [ -z "$LC_ALL" ]; then
	export LC_ALL=C
fi

check_whitelist ~/.spamassassin/auto-whitelist | sort -n >~/awl-raw
check_bayes_db | sort -n >~/bayes-raw

BlackDomains=`cat $SaConfigs | grep '^blacklist_from' | \
 sed -e 's/.* \*@/@/' | grep -v blacklist_from | grep -v '\*' | sort | uniq`
#echo BlackDomains are $BlackDomains
#echo

WhiteDomains=`cat $SaConfigs | grep '^whitelist_from' | \
 sed -e 's/.* \*@/@/' | grep -v whitelist_from | grep -v '\*' | sort | uniq`
#echo WhiteDomains are $WhiteDomains
#echo

echo ==============================
echo Domains with $MinSenders or more senders
echo ==============================

TopDomains=`cat ~/awl-raw | sed -e 's/.*@//' -e 's/|.*//'| sort | uniq -c | \
 sort -nr | while read COUNT DOMAIN ; do if [ $COUNT -ge $MinSenders ]; then \
 echo $DOMAIN ; fi ; done`

BlackListOut=''
WhiteListOut=''
CantColorOut=''
OtherOut=''
for Domain in $TopDomains ; do
	Handled=''
	for OneBlack in $BlackDomains ; do
		if [ "@$Domain" = "$OneBlack" ]; then
			Handled='yes'
			BlackListOut="$BlackListOut`echo ; echo ======== @$Domain \(Blacklisted\) ======== ; cat ~/awl-raw | grep @$Domain`"
		fi
	done

	for OneWhite in $WhiteDomains ; do
		if [ "@$Domain" = "$OneWhite" ]; then
			if [ -n "$Handled" ]; then
				echo "WARNING - $Domain is both blacklisted and whitelisted, please fix."
			fi
			Handled='yes'
			WhiteListOut="$WhiteListOut`echo ; echo ======== @$Domain \(Whitelisted\) ======== ; cat ~/awl-raw | grep @$Domain`"
		fi
	done

	for OneCantColor in $CantColorDomains ; do
		if [ "@$Domain" = "$OneCantColor" ]; then
			Handled='yes'
			CantColorListOut="$CantColorListOut`echo ; echo ======== @$Domain \(CantColor\) ======== ; cat ~/awl-raw | grep @$Domain`"
		fi
	done

	if [ -z "$Handled" ]; then
		OtherOut="$OtherOut`echo ; echo ======== @$Domain ======== ; cat ~/awl-raw | grep @$Domain ; echo`"
	fi

	#echo "======== @$Domain$BlackListed$WhiteListed ========"
	#cat ~/awl-raw | grep "@$Domain"
done

echo "================ Unclassified domains ================"
echo "$OtherOut"
echo
echo "================ Blacklisted domains ================"
echo "$BlackListOut"
echo
echo "================ Whitelisted domains ================"
echo "$WhiteListOut"
echo
echo "================ CantColor domains ================"
echo "$CantColorListOut"
echo

echo ============================
echo Top $TopN hammiest individuals
echo ============================
cat ~/awl-raw | head -$TopN

echo =============================
echo Top $TopN spammiest individuals
echo =============================
cat ~/awl-raw | tail -$TopN


echo ============================
echo Top $TopN hammiest tokens
echo ============================
cat ~/bayes-raw | grep -v 'non-token data:' | head -$TopN

echo =============================
echo Top $TopN spammiest tokens
echo =============================
cat ~/bayes-raw | grep -v 'non-token data:' | tail -$TopN