Skip to content

Instantly share code, notes, and snippets.

Embed
What would you like to do?
#!/usr/bin/env ruby
require 'open-uri'
require 'nokogiri'
base_url = 'https://en.wikipedia.org'
site = ARGV.empty? ? 'Film_adaptation' : ARGV.shift
url = base_url + "/wiki/#{site}"
Signal.trap('INT') { exit }
# fetch the first link from the first paragraph of the site
CSS_SELECTOR = '#mw-content-text p a'
link_num = 0
# maximum links to skip
MAX_SKIPS = 10
# array of visited sites
sites = [site]
# skip pronounciation links
class PronunciationLink < StandardError; end
# skip links to files like images or audio
class FileLink < StandardError; end
class NoPhilosophy < StandardError; end
puts site
#require 'byebug'; byebug
loop do
html = Nokogiri::HTML(URI.open(url))
raise NoPhilosophy if link_num > MAX_SKIPS
element = html.css(CSS_SELECTOR)[link_num]
# raise exception of there is no link in the first paragraph
raise NoPhilosophy if element.nil?
site, site_href = [element.text, element['href']]
raise PronunciationLink if site.match(%r(^[\[/].*[\]/]$)) || site_href.downcase.match(/help:pronun/)
raise FileLink if site_href.match(/.*\.(ogg|png|jpg)/)
sites << site
puts sites.last
url = base_url + site_href
(puts "Reached philosophy!"; exit) if sites.last.downcase == 'philosophy'
# exit if there is a repeated site
raise NoPhilosophy if sites.uniq!
link_num = 0
rescue NoMethodError, TypeError, PronunciationLink, FileLink, OpenURI::HTTPError
link_num += 1
puts "Skipped #{link_num} link#{link_num > 1 ? 's' : ''} in #{sites.last}"
rescue NoPhilosophy
puts "Did not reached philosophy :("
exit
end
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.