Ruby web crawler (See related posts)

NB. Again, this script was found in my old source code folder, it may not be fully working.

This Ruby script reads in a list of links from links.dat, it then picks out the ones it can easily spider and gets a list of URLs from each page listed in links.dat. Every new URL it finds will be added to newlinks.dat for later spidering by another bot running along side this one.

require 'socket' 
links = File.open("links.dat") 
while links.gets do 
        #domain = ($_ =~ /http:\/\/.*\.([0-9a-zA-Z\-]+\.com|net|org)/); 
        if %r{http://([^/]+)/([^/]+)}i =~ $_ 
                domain,path = $1, $2 
        end 
        if proto="http" 
                begin 
                        t = TCPSocket.new(domain, 'www') 
                rescue 
                        puts "error: #{$!}" 
                else 
                        t.print "GET /"+path+" HTTP/1.0\n\n" 
                        answer = t.gets(nil) 
                        t.close 
                end 
 
                if %r{<a\s+href="(\w+)://([^"]+)"[^>]*>([^<]*)</a>}i =~ answer 
                        proto, url, text = $1, $2, $3 
                end 
 
                print proto+"://"+url+"\n" 
                old = File.open("newlinks.dat") 
                new = File.open("links.dat.tmp", File::WRONLY|File::TRUNC|File::CREAT) 
                while old.gets do 
                        if $_ != proto+"://"+url 
                                new.print $_ 
                        end 
                end 
                new.print proto+"://"+url 
                old.close 
                new.close 
                File.rename("newlinks.dat", "links.dat.orig") 
                File.rename("links.dat.tmp", "newlinks.dat") 
        end 
end 
links.close