Simple script for testing http://xkcd.com/903/ alltext trivia

wiki.rb

# encoding: UTF-8

require 'rubygems'
require 'nokogiri'
require 'open-uri'

start = 'C.A._Penarol'
target = 'Philosophy'
visited = []


def get_first_article(article)
  base = 'http://en.wikipedia.org/wiki/'
  doc = Nokogiri::HTML(open(base+article, 'User-Agent' => 'ruby'))
  p = doc.css('#bodyContent > p').first
  link = Nokogiri::HTML(formatParagraph(p.to_s)).css('a').find{ |l| l['href'].match(/^\//) }
  puts link['href']
  link ? link['href'].gsub(/^\/wiki\//, '') : nil
end

def formatParagraph(text)
  # remove parenthesis
  r = /\(.[^\(]*?\)/
  mark = '@#$'
  while text.match r
    text.gsub!(r) {|m| m.to_s.gsub(/<|>/, '').gsub('(', mark).gsub(')', mark.reverse)}
  end
  text.gsub(mark, '(').gsub(mark.reverse, ')')
end

current = start
while(current && current != target && !visited.include?(current))
  visited << current
  current = get_first_article(current)
end

if !current
  puts "We reached a dead end on #{visited.last}"
elsif current == target
  puts "We arrived from #{start} to #{target} in #{visited.length} steps"
else
  puts "We entered a loop starting on #{current}"
end

Note: I had to add 'User-Agent' => 'ruby' to line 15 (see: http://stackoverflow.com/questions/2305975/openurihttperror-403-forbidden )

Also, had to require 'rubygems'

Thanks for the script. I thought about writing the same thing but didn't think about it again until just now ;) Works great (Barack Obama -> Philosophy takes 20 steps, BTW)

Ok, I added those changes. We are probably using different ruby versions. I'm on 1.9.2, are you on 1.8.x?

Yeah, indeed, thanks. I'm on 1.8.7 because I've got an old, very large project still there. We're planning on upgrading ;)

Also, check out my fork, I've written a script which takes a number of source pages and creates a graph using GraphViz. Here's a sample graph:

http://semi-sentient.com/img/philosophy.pdf