#!/bin/bash
# make_publish.bash 
# Extracts the latest HTM files from the web site and creates RSS feed XML file.
# Creates Send_Publish.bash file to ftp the contents.
# (c) Dennis Adams Associates Ltd.
# This file may be freely used and copied without prior permission, 
# providing that an acknowledgement of origin is attached.
# No warranty is expressed or implied regarding use !
#
# name of file to create:
MANIFEST=$HOME/Send_Publish.bash
# names of temporary files:
RSSFILE=rss/news.xml
RSSTMP1=$HOME/rss.tmp1
RSSEXTRACTS=$HOME/rssextracts.data
# root directory of local copy of web site:
WEBDIR="/web"
# URL of website for including in RSS feeds:
WEBSITE=www.yourdomain.com
# domain name and login id for FTP site for uploading:
FTPSITE=www.yourdomain.com
WEBUSER=yourusername
# email addresses of webmaster and RSS editor for showing on XML feed:
WEBMASTER=webmaster@yourdomain.com
RSSEDITOR=info@yourdomain.com
# no of htm files to include for RSS and for uploading to ftp site. 
RSSSIZE=5
TAILSIZE=10
if [ "$1" != "" ]
then 
	TAILSIZE="$1"
fi
echo "Starting..."
echo "No of files to include=$TAILSIZE"
echo "No of files in RSS feed=$RSSSIZE"
cd
cd "$WEBDIR"
echo "Default Directory is: "
pwd

# 
# 
# tidy up old rss file... 
DATESTAMP=`date "+%Y%m%d%H%M"`
if [ -f $RSSFILE ]
then
	echo "Found earlier file $RSSFILE"
# echo "Renaming $RSSFILE to $RSSFILE.$DATESTAMP"
# mv $RSSFILE $RSSFILE.$DATESTAMP
fi

# 
# 
#  create a new rss file and write the header to it...
echo "Creating $RSSTMP1..."
cat /dev/null > $RSSTMP1

cat >> $RSSTMP1 <<ENDOFTOP1
<?xml version="1.0" ?>
<!DOCTYPE rss PUBLIC "-//Netscape Communications//DTD RSS 0.91//EN" "http://my.netscape.com/publish/formats/rss-0.91.dtd">
<rss version="0.91">
<channel>
<title>Your Website Title</title>
<language>en-us</language>
<copyright>Copyright 2005. All rights reserved. This RSS file is offered to individuals and non-commercial organizations only. Commercial websites wishing to use RSS files from this site, please contact $RSSEDITOR.</copyright>
<managingEditor>$RSSEDITOR</managingEditor>
<webMaster>$WEBMASTER</webMaster>
<description>Your description of your site... .</description>
<link>http://$WEBSITE</link>
ENDOFTOP1

# 
# 
# work out the date properly and the rest of the header ...
echo "Publication date is..."
date
echo "<pubDate>`date "+%a, %e %b %Y %H:%M:%S GMT"`</pubDate>" >> $RSSTMP1

# 
# You can have as many <category> tags as you want...
cat >> $RSSTMP1 <<ENDOFTOP2
<category>Information Technology</category>
<category>Consulting</category>
<image>
	<title>   </title>
	<url>http://$WEBSITE/graphics/logo.jpg</url>
	<link>http://$WEBSITE</link>
	<width>144</width>
	<height>64</height>
	<description>Your description of your site..</description>
</image>
ENDOFTOP2

# 
# 
#  extract the names of the all the candidate htm files that are changed...
echo "Analysing Items..."
ls -tr *.htm | tail -$RSSSIZE | awk '{ print $1}' > $RSSEXTRACTS
#  Add index.html in all cases...
echo index.html >> $RSSEXTRACTS

# 
# 
#  cycle through all candidate htm files and extract item information...
while read HTMFILE 
do
	#echo found: $HTMFILE
	cat $HTMFILE | awk -v htmfile=$HTMFILE -v website=$WEBSITE '

	BEGIN { ttitle=""; para1=""
		p1=""; h1=""; h2=""; 
		gotp1=0; goth1=0; goth2=0 }
	
	/\<title\>/ { ttitle=$0 ; }
	/\<p\>/  { if (gotp1==0) { p1=$0; gotp1=1} }
	/\<h1\>/ { if (goth1==0) { h1=$0; goth1=1} }
	/\<h2\>/ { if (goth2==0) { h2=$0; goth2=1} }
	
	END { 	print "<item>"; 
		printf "\t%s\n", ttitle;
		printf "\t<description>" ;
		para1=p1;
		if ( para1=="" ) { para1=h1 }
		if ( para1=="" ) { para1=h2 }
		printf para1 ;
		printf "</description>\n" ;
		printf "\t<link>http://%s/%s</link>\n", website, htmfile ;
		print "</item>" }
	' 

done < $RSSEXTRACTS  >> $RSSTMP1

# 
# 
# Process the internal structure of the html files which have download links.
# Structure is assumed to be: <tr><td><strong> is the break to a new file title.
# The following line is the description.
# <a href=" contains the url for the linked file associated with it.
for HTMFILE in  papers.htm techpapers.htm utils.htm
do
	#echo found: $HTMFILE
	cat $HTMFILE | awk -v htmfile=$HTMFILE -v website=$WEBSITE '

	BEGIN { ttitle=""; desc=""; llink=""
		gottitle=0; gotdesc=0; gotlink=0; }
	/\<tr\>\<td\>\<strong\>/ { if (gottitle==0) { ttitle=$0; gottitle=1; next } }
	{ if (gottitle==1 && gotdesc==0 ) { desc=$0; gotdesc=1} }
	/\<a href=/ { if ( gottitle==1 && gotdesc==1 && gotlink==0 ) { llink=$0; gotlink=1 } }
	
	END { 	print "<item>"; 
		printf "\t<title>%s</title>\n", ttitle;
		printf "\t<description>" ;
		printf desc ;
		printf "</description>\n" ;
		printf "\t<link>%s</link>\n", llink ;
		print "</item>" }
	' | sed 's/<a href="//g; /">/s/">[^<>]*//g; ' >> $RSSTMP1
done

# 
# 
#  Add the channel footer...
#echo "Adding footer..."
echo "</channel>" >> $RSSTMP1
echo "</rss>" >> $RSSTMP1

# 
# 
# strip out all the extra HTML (non-XML) tags...
cat $RSSTMP1 | 	\
	sed '/<a.*>/s/<a[^<>]*>//g; 
		s/<p>//g; 
		s/<\/p>/ /g; 
		s/<h1>//g; 
		s/<\/h1>/ /g; 
		s/<h2>//g ; 
		s/<\/h2>/ /g; 
		s/<\/a>/ /g; 
		s/<tr>/ /g; 
		s/<\/tr>/ /g; 
		s/<td>/ /g; 
		s/<\/td>/ /g; 
		s/<br>/ /g; 
		s/<strong>//g 
		s/<\/strong>//g 
		s///g 
		s/&quot;//g ' > $RSSFILE
echo "Created $RSSFILE ..."

# 
# 
# create the ftp file...
echo "Creating $MANIFEST..."
echo "#!/bin/bash" > $MANIFEST
echo "echo \"Running ftp...\"" >> $MANIFEST
echo "ftp -v $FTPSITE <<ENDOFSTUFF" >>$MANIFEST
echo "$WEBUSER" >> $MANIFEST
echo "passive" >> $MANIFEST
echo "prompt" >> $MANIFEST
echo "ascii" >> $MANIFEST
echo "cd htdocs" >> $MANIFEST
echo "pwd" >> $MANIFEST
echo "lcd \"$WEBDIR\"" >> $MANIFEST
echo "lpwd" >> $MANIFEST
ls -tr *.htm | tail -$TAILSIZE | awk '{ print "put ",$1}' >> $MANIFEST
echo "put index.html" >> $MANIFEST
echo "cd rss" >> $MANIFEST
echo "pwd" >> $MANIFEST
echo "lcd rss" >> $MANIFEST
echo "lpwd" >> $MANIFEST
echo "put news.xml" >> $MANIFEST
echo "bye" >> $MANIFEST
echo "ENDOFSTUFF" >> $MANIFEST

chmod 775 $MANIFEST
echo "Created $MANIFEST ..."

# 
# 
# cleanup...
rm -f $RSSEXTRACTS
rm -f $RSSTMP1
exit