#!/bin/bash
 
 # $Id: nb-build-index.bash,v 1.2 2005/01/02 06:20:57 schwehr Exp $
 
 # nanoblogger extension by Kurt Schwehr
 
 # This script goes through the .txt files in the current directory.
 # It takes the list of phrases and for each phrase generates a list of
 # links.  In the future, these numbers will hopefully match the blog
 # entries, but for now, they are just sequential for each entry.
 # It also does not currently handle wrapping or tags between words.
 
 # To get an initial list of words, you might do something like this,
 # but know that it will not catch phrases like "San Diego"
 #
 # cat *.txt | tr '#"=<>,.;:/\\' ' ' | tr ' \t' '\n\n'  | egrep '[a-z]' | sort -u > words
 
 # Sees my "AUTHOR: Kurt" in each.  Nedd to strip out the first 6 lines
 
 phrases=$1
 
 declare -i num=0
 
 cat $phrases | while read phrase; do
     echo -n "$num "
     echo $phrase > tmp-$num.phrase
     # FIX: turn files into one line each
     grep -l "$phrase" *.txt > tmp-$num.matches
     num=$[num+1]
 done
 
 echo 
 echo "Assempting index from matches"
 
 
 out=phrase-index.html
 rm -f $out
 for file in tmp-*.phrase; do
     num=1
     base=${file%%.phrase}
     phrase=`cat $file`
     #echo phrase is $phrase
     echo "<b>$phrase:</b>" >> $out
     cat $base.matches | while read filematch; do
 	month=`echo $filematch | cut -c1-7`
 	echo "<a href=\"$month.html#e$filematch\">$num</a>" >> $out
 	num=$[num+1]
     done
     echo "<br>" >> $out
 done