:
##########################################################################
# Shellscript:	newstop - top 10 news posters
# Version    :	2.3
# Author     :	Heiner Steven <heiner.steven@odn.de>
# Category   :	News
# Date       :	1995-09-24
# SCCS-Id.   :	@(#) newstop	2.3 04/02/18
##########################################################################
# Description
#	Writes a "top 10 posters" newsgroup statistics to
#	standard output. Directly accesses the news spool
#	directory.
#
# Notes
#    o	Uses "edate" to print a date
#
# See also
#	newssub
##########################################################################

PN=`basename "$0"`			# Program name
VER='2.3'

: ${GREP:=egrep}
: ${NEWSSPOOL:=/var/spool/news}
: ${NEWSSTAT:=$HOME/.newsstat}

# Search a "new" awk (nawk) or GNU awk
for awk in mawk gawk nawk
do
    for path in . `echo "$PATH" | tr : ' '`
    do
        [ -x "$path/$awk" ] || continue
        AWK=$path/$awk
        break 2
    done
done
: ${AWK:=awk}

Period=7				# Number of days for one period
TopCnt=10				# Number of posters to print

Usage () {
    echo >&2 "$PN - create newsgroup statistics $VER (hs '95)
usage: $PN [-n number] [-d days] newsgroup
   -n:	number of authors to print (default is $TopCnt)
   -d:	length of a period (default is $Period days)

News articles are searched in the spool directory
NEWSSPOOL=$NEWSSPOOL.

The statistic are printed to standard output.

Example:
    $PN comp.unix.shell"
    exit 1
}

Msg () {
    for MsgLine
    do echo "$PN: $MsgLine" >&2
    done
}

Fatal () { Msg "$@"; exit 1; }

set -- `getopt n:d:h "$@"`
while [ $# -gt 0 ]
do
    case "$1" in
	-n)	TopCnt="$2"; shift;;
	-d)	Period="$2"; shift;;
	--)	shift; break;;
	-h)	Usage;;
	-*)	Usage;;
	*)	break;;			# First news group name
    esac
    shift
done

[ $# -ne 1 ] && Usage
Group="$1"

TmpDir="${TMPDIR:=/tmp}/nt$$"		# Directory with all tmp. files
Authors="$TmpDir/authors"		# List of Authors
Files="$TmpDir/files"			# List of Files
List="$TmpDir/list"			# List for statistics
Sizes="$TmpDir/sizes"			# List of file sizes
Tmp="$TmpDir/tmp"

NewsStat="$NEWSSTAT/$Group"		# Last statistics
Top="$NewsStat/top$TopCnt.authors.new"	# Last Top 10
OldAuthors="$NewsStat/authors"
NewAuthors="$TmpDir/newauthors"

[ -d "$NewsStat" ] || mkdir -p "$NewsStat" || exit $?

trap 'rm -rf "$TmpDir" >/dev/null 2>&1' 0
trap "exit 2" 1 2 3 13 15

[ -d "$TmpDir" ] || mkdir -p "$TmpDir" ||
	Fatal "cannot create temporary directory: $TmpDir"

Dir="$NEWSSPOOL/`echo \"$Group\" | sed 's:\.:/:g'`"
[ -d "$Dir" -a -r "$Dir" -a -x "$Dir" ] ||
    Fatal "cannot read news group $Group" "(Dir=$Dir)"

cd "$Dir" || exit $?

# Create file list
find * -type f -mtime -$Period -print |
	sort > "$Files" || exit $?

[ -s "$Files" ] || exit 0

# Create list with file sizes (SYSV/BSD ls)
# 	file	size
xargs < $Files ls -ld |
	$AWK 'NF == 8 || NF == 9 { print $NF "	" $(NF-4) }' |
	sort > "$Sizes" || exit $?

# Create author list (there could be more than one "From:" line
# in each file - only take the first line)
#	file	author
xargs < $Files $GREP '^From: ' /dev/null |
    sed 's/:[ 	]*From: /	/' |
    $AWK -F'	' '++A[$1] == 1' > "$Authors" || exit $?

# Rewrite Author list in the format
#	file	address	realname
$AWK -F'	' '
    {
	if ( $2 ~ /\(.*\)[ 	]*$/ ) {
	    # Address format "heiner@hsysnbg.nbg.sub.org (Heiner Steven)"
	    Name=$2; Addr=$2
	    gsub (/.*\(/, "", Name)
	    gsub (/\)$/, "", Name)
	    gsub (/[ 	]*\(.*\)[ 	]*/, "", Addr)
	    print $1 "	" Addr "	" Name
	} else if ( $2 ~ /<.*>[ 	]*$/ ) {
	    # Address format "Heiner Steven <heiner@hsysnbg.nbg.sub.org>"
	    Name=$2; Addr=$2
	    gsub (/[ 	]*<.*>.*/, "", Name)
	    sub (/^"/, "", Name);
	    sub (/"$/, "", Name);
	    gsub (/.*</, "", Addr)
	    gsub (/>.*/, "", Addr)
	    print $1 "	" Addr "	" Name
	} else {
	    # Address format "heiner@hsysnbg.nbg.sub.org"
	    print $1 "	" $2 "	" $2
	}
    }
' "$Authors" | sort > "$Tmp" && mv "$Tmp" "$Authors"

join -t'	' "$Sizes" "$Authors" > "$Tmp" || exit $?

# Format of list:
# FILE	SIZE	Addr	Realname
$AWK -F'	' '
    {
	Posts [$3]++
	Bytes [$3] += $2
	Name  [$3] = $4
    }
    END {
	for ( Author in Posts ) {
	    printf "%d\t%f\t%s\t%s\n",
		Posts [Author], Bytes [Author], Author, Name [Author]
	}
    }
' "$Tmp" |
    sort -t'	' +2 > "$List"		# Sorted by name

#
# Special handling for new posters:
# Get number of posters, articles and bytes.
#

# Create sorted author list
cut -d'	' -f3 < "$List" | sort -u > "$Authors"

# List only new authors
if [ `ls $OldAuthors.[1-3] 2>/dev/null | wc -l` -gt 0 ]
then
    sort -u "$OldAuthors".[1-3] | comm -23 "$Authors" - > "$NewAuthors"
else
    > "$NewAuthors"
fi

# Extract lines of new authors from list
join -t'	' -j1 1 -j2 3 -o "2.1 2.2 2.3 2.4" "$NewAuthors" "$List" > "$Tmp"

# Put totals into shell variables
set -- `awk '{ n++; a+=$1; b+=$2 }; END { print n, a, b }' "$Tmp"`

if [ $# -eq 3 ]
then
    NewPosters="$1"
    NewArticles="$2"
    NewBytes="$3"
else
    NewPosters=0 NewArticles=0 NewBytes=0
fi

# Age author files
i=1
while [ $i -le 2 ]
do
    old=`expr $i + 1`
    [ -f "$OldAuthors.$old" ] && mv "$OldAuthors.$old" "$OldAuthors.$i"
    i=$old
done
cp "$Authors" "$OldAuthors.3"

# Sort table by posts and bytes
sort -t'	' -nbr +0 -1 +1 -2 +2 "$List" > "$Tmp" &&
	mv "$Tmp" "$List"

echo "Top $TopCnt posters for the $Period day period ending `edate`.
"

# Format of Tmp:
#	Posts	Bytes	Address	Name
$AWK -F '	' '
    BEGIN {
	TopCnt = '"$TopCnt"'
	print "posts kbytes  name                         address"
    }
    {
	TPosters++
	TArticles += $1
	TBytes    += $2
	if ( n < TopCnt ) {
	    n++
	    printf "%3d %8.1f  %-28.28s %-.35s\n", $1, $2 / 1024, $4, $3
	    Articles += $1
	    Bytes    += $2
	} 
    }
    END {
	print "---    -----"
	printf "%3d %8.1f  Total for top %d\n\n",
		Articles, Bytes / 1024, TopCnt
	print "Totals for the newsgroup:"
	printf "  %d posters\n", TPosters
	printf "  %d articles\n", TArticles
	printf "  %.1f kbytes\n\n", TBytes / 1024
	printf "The top %d accounted for:\n", TopCnt
	if ( TPosters ) 
	    printf "  %5.1f%% of the posters\n", n * 100 / TPosters
	if ( TArticles )
	    printf "  %5.1f%% of the articles\n", Articles * 100 / TArticles
	if ( TBytes )
	    printf "  %5.1f%% of the bytes\n\n", Bytes * 100 / TBytes
	print "Averages:"
	if ( TPosters )
	    printf "  %.1f articles / poster\n", TArticles / TPosters
	if ( TArticles )
	    printf "  %.1f kbytes / article\n", (TBytes / 1024) / TArticles
	if ( TPosters )
	    printf "  %.1f kbytes / poster\n", (TBytes / 1024) / TPosters

	if ( '"$NewPosters"' ) {
	    printf "\n%d people posted for the first time this period.\n",
		    '"$NewPosters"'
	    printf "They went on to post %d articles altogether\n\n",
		    '"$NewArticles"'
	    print "The new posters accounted for:"
	    if ( TPosters )
		printf "  %5.1f%% of the posters\n",
			'"$NewPosters"' * 100 / TPosters
	    if ( TArticles )
		printf "  %5.1f%% of the articles\n",
			'"$NewArticles"' * 100 / TArticles
	    if ( TBytes )
		printf "  %5.1f%% of the bytes\n\n",
			'"$NewBytes"' * 100 / TBytes
	}
	print "-- "
	printf "Your UNIX shell scripting resource: Heiner%cs SHELLdorado\n", 39
	print "http://www.shelldorado.com/"
    }
' "$List" | tee "$Top"