A script to collect (conserve? steal?) Constant Dullaart's "war" web pages

war_collector.sh

#!/bin/sh

# A script to collect (conserve? steal?) Constant Dullaart's "war" web pages
# See: https://twitter.com/constantdull/status/785797564167839744

# Usage: ./war_collector.sh

start_from="war.repair"

if [ ! -d src ] ; then
	mkdir src
fi

function collect_the_art() {

	echo "collecting $1"

	if [ ! -f "src/$1.html" ] ; then
		curl -s -o "src/$1.html" "http://$1/"
	fi

	img_num=`grep "<img" "src/$1.html" | sed 's/^.*<img src="//' | sed 's/.png" width="100%" height="100%"><\/img>//'`
	if [ ! -f src/$img_num.png ] ; then
		curl -s -o "src/$img_num.png" "http://$1/$img_num.png"
	fi

	next=`grep "URL=" src/$1.html | sed 's/^.*<meta http-equiv="refresh" content="0;URL=http:\/\///' | sed 's/\/" \/>//'`
	cat "src/$1.html" | sed "s/http:\/\/$next\//$next.html/" | sed "s/$img_num.png/src\/$img_num.png/" > "$1.html"

	if [ $next != $start_from ] ; then
		collect_the_art $next
	fi
}

collect_the_art $start_from

Worth pointing out here that sed is not the best tool for this job. A better option is pup.