#!/bin/bash
#
# pak-web (was txt2gzip) -- 2008-08-21, last edit 2008-09-13
#
# part of the pak-web-site project, this is a wrapper script for the 
#  pak-web-site gawk script to coordinate the compression of web and 
#  text files as well as provde a restore option
#
# Copyright (C) 2008 Grant Coady <grant@bugsplatter.id.au> GPLv2
#
# Home site: http://bugsplatter.id.au/pakweb/
#
# call:
#  to shrink the web site and compress .txt files:
#	pak-web
#
#  to shrink a single page:
#	pak-web usedir <directory_below_doc_root> 
#
#  to restore the web site and remove gzipped .txt files:
#	pak-web restore
#

# adjust to suit your site
web_doc_root="/home/web/bugsplatter"

PATH="/usr/local/bin:$PATH"		# for root cron job (junkshow)

# gzrate -- how to display shrink vs source filesize
#  1:	saved:		(filesize - gzipsize) / filesize
#  0:	by size:	gzipsize / filesize
gzrate=1

rename_src2html() # -
{
	find . -name "*.html.src" | while read name
	do
		printf "  mv -f %-40s %s\n" $name ${name%.*}
		mv -f $name ${name%.*}
	done
	echo ""
}
restore_html_files() # "-" or pak-web caller's $PWD
{
	echo "
restore *.html.src files"
	if [ -z "$1" ]
	then
		rename_src2html
	else
		[ -d $1 ] || exit 1
		cd $1 && rename_src2html && cd - > /dev/null
	fi
}

# remove all gzipped files, restore site for random review and editing
restore_web_site() # -
{
	echo "
pak-web remove gzipped files"
	find . -name "*.html.gz" -delete
	find . -name "*.txt.gz"	| xargs rm -f
	rm -f bugs.css.gz
	rm -f msie.css.gz
	restore_html_files $1
}

gzip_text_files() # "" or pak-web caller's $PWD
{
	echo "gzip text files"

	if [ -n "$1" ]
	then
		[ -d $1 ] || exit 1
		cd $1 || exit 1
	fi

	[ -r bugs.css ] && gzip -c bugs.css > bugs.css.gz
	[ -r msie.css ] && gzip -c msie.css > msie.css.gz

	find . -name "*.txt" | \
		grep -v networkmonitor | \
		grep -v public | \
		grep -v robots.txt | \
		while read name
		do
			gzip -c $name > $name.gz
		done
}

# drive the pak-web-site script, needs pak-web-scan to make a list of ssi 
#  include files and the find command to provide a list of target files.
# may be called from below the $doc_root directory to refresh single web 
#  page, examples: junkshow for hourly page updates, netdraw for the after 
#  midnight summary image updates

shrink_web_site() # "" or pak-web caller's $PWD
{
	if [ -n "$1" ]
	then
		[ -d $1 ] || exit 1
		cd $1 || exit 1
		local ssi_list="$(pak-web-scan)"
		cd - > /dev/null
		local thisdir=$(basename $1)

		pak-web-site -v gzrate=$gzrate -v ssi_list="$ssi_list" \
			$(find . -name "*.html" -type f | \
				grep $thisdir | sort | \
				egrep -v 'archive|public')
	else
		pak-web-site -v gzrate=$gzrate -v ssi_list="$(pak-web-scan)" \
			$(find . -name "*.html" -type f | \
				sort | egrep -v 'archive|public')
	fi
}

# script driver
callerpwd=$PWD
cd $web_doc_root || exit 1
case $1 in
	-d|dir)
	# rebuild current directory and below only
	restore_html_files	$callerpwd
	shrink_web_site		$callerpwd
	gzip_text_files		$callerpwd
	;;
	-r|restore)
	# undo all merged web pages and remove gzipped text files
	restore_web_site
	;;
	-u|usedir)
	[ -d $2 ] || exit
	restore_html_files	$2
	shrink_web_site		$2
	gzip_text_files		$2
	;;
	*)
	# default is to shrink entire web site and gzip text files
	shrink_web_site
	gzip_text_files
	;;
esac
# end