Guest User

Untitled

a guest
Feb 12th, 2014
167
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
Ruby 1.61 KB | None | 0 0
  1. #!/usr/bin/env ruby
  2. require 'nokogiri'
  3. require 'open-uri'
  4. require 'sequel'
  5. require 'sqlite3'
  6. require 'pry'
  7. require 'uri'
  8. require 'cgi'
  9.  
  10.  
  11.  
  12. clubland = "http://www.clublandlv.com/forum.php"
  13. doc = Nokogiri::HTML(open(clubland))
  14.  
  15. puts doc.css('title').text
  16. puts
  17.  
  18.  
  19. doc.css("#cat10").each do |grab|
  20.   genres = grab.css(".forumtitle").map(&:text)
  21.   genre_links = grab.css(".forumtitle"){[:href]}
  22.   subgenre = grab.css(".subforum a").map(&:text)
  23.   subgenre_links = grab.css(".subforum a"){[:href]}
  24.    
  25.  
  26.   puts "This is an array of the genres:"
  27.   puts genres
  28.   puts
  29.   puts "These are the genre links:"
  30.   puts genre_links
  31.   puts
  32. #  puts "This is an array of the subgenres:"
  33. #  puts subgenre
  34. #  puts
  35. #  puts "These are the subgenre links:"
  36. #  puts subgenrelink
  37.  
  38.  
  39.    
  40. # Table has 3 columns. ID(auto), genre, forum_id
  41.         DB = Sequel.connect('sqlite://clublandlv.sqlite')
  42. # Create a dataset
  43.         main_genre = DB[:genres]
  44. # Insert the results
  45.  
  46. #  genres.each do |genre|
  47. #   main_genre.insert(:genre => genre)
  48. #    unless db_has_genre?
  49. #    genre;
  50.  
  51.  
  52.   genre_links.map { |g_link|
  53.         link_id = g_link.children.first["href"]
  54.         link_cut = CGI.parse(URI.parse(link_id).query)['f'].first.to_i #returns Numeric,Integer,Fixnum
  55.         genre_links << (link_cut) #`push': node must be a Nokogiri::XML::Node or Nokogiri::XML::Namespace (ArgumentError)
  56.         binding.pry
  57.  
  58.  
  59.   }
  60.         #   main_genre.insert(:forum_id => link_cut)
  61.         g_link;
  62.  
  63.              
  64.      #  main_genre.insert(:forum_id => genrelink)
  65.      #  unless db_has_genre?
  66.      #   genrelink;
  67.  
  68.       end
Advertisement
Add Comment
Please, Sign In to add comment