#!/bin/sh -x
DIRHEAD=/home/lconrad/homepage/
URLHEAD="./"
cd $HOME/Mail/wwwacct
#
# take the  messages and put them into alllines
#
grep -l wwwacct * | xargs cat >$HOME/homepage/stats/alllines
#
# get a count of how many uniq lines there are for each page
#
cd $HOME/homepage/stats
cat alllines | awk '{print $7}' | grep '~lconrad' | sed 's;//;/;'| grep -v '\.in\.' |  sort | uniq -c | sort -n -r >allpages
cat allpages | sed "s;/~lconrad/;$DIRHEAD;" | sed "s;/~lconrad$;$DIRHEAD/homepage.html;" >allfiles
#
# save the previous table
#
mv -f tablines.tex tablines.`date +%s`.tex
#
# For each file, get the creation date and modification date, and write a line
# LaTeX table line, 
#
for FILE in `cat allfiles | awk '{print $2}'` 
do
    BASEH=`basename $FILE .html`
    BASE=`basename $FILE`
    if [ "$BASEH" = "$BASE" ]
    then
# not html
	CREATION='---'
    else
	# there are lines in the html files like:
	#    <!-- Created: Fri Oct 30 20:44:23 EST 1998 -->
	#    <!-- Last modified: Fri Oct 30 20:47:28 EST 1998 -->
	# or ls -l gives a line like:
	# -rw-rw-r--   1 lconrad  lconrad      4076 Dec  3 10:21 ../serpent.html
	grep  "Created: " $FILE >created
	if [ "$?" = 0 ]
	then
		CREATION=`cat created | awk '{printf  ("%s %s, %s\n", $4, $5, $8)}'`
	else
		CREATION='---'
	fi
    fi
    if [ -f $FILE ]
    then
	MODIFIED=`ls -l $FILE | awk '{printf "%s %s", $6, $7
	 	if (length($8) == 4) {printf ", %s", $8}
					printf "\n"}'`
    else
	MODIFIED="---"
    fi
    ACCESSES=`grep "$FILE$" allfiles | awk '{print $1}'`
    PUBFILE=`echo $FILE | sed "s;$DIRHEAD;;"`
    URL=`echo $FILE | sed "s;$DIRHEAD;$URLHEAD;" | sed 's;_;\\_;'`
    if [ ! -d $FILE ]
    then
	if [ -s $FILE ]
	then
	    HREF="<a href=\"$URL\">$PUBFILE</a>"
	fi
    fi
    if [ -z $HREF ]
    then
	HREF=$PUBFILE
    fi
    echo "$HREF" '&' $ACCESSES '&' "$CREATION" '&' "$MODIFIED" '\\ \hline' >>tablines.tex

done
#
# convert the LaTeX table to a html table.  
# the sed is because tth wants to make my url's print correctly in html, and 
# doesn't care if they work.
# the awk is because tth puts titles into the body of the html, 
# and the w3 browser (and allegedly the dtd standard) 
# doesn't recognize them there.
#
cat table.tex | /home/lconrad/bin/tth | sed -e 's/&lt;/</g' -e 's/&nbsp;/~/g' -e 's/&#62;/>/g' >../usage_table.htm
chown  lconrad *