class WikiReader

# File lib/wikipedia.rb, line 29
def self.article(title, language = 'en')
  raise Errno::ENOENT if ! db.key?(language) # invalid language
  db[language] = fetch_vital(language) if ! db[language] # not initialized

  art = fetch_article(title, language) # read from the internet
  raise Errno::ENOENT if ! art
  db[language] << title if art && ! db[language].index(title) # adds in our list

  @@sizes[language] ||= {}
  @@sizes[language][title] = art.bytesize

  art
end

# File lib/wikipedia.rb, line 18
def self.articles(language = 'en')
  return nil if ! db.key?(language) # invalid language
  db[language] = fetch_vital(language) if ! db[language] # not initialized

  db[language]
end

# File lib/wikipedia.rb, line 48
def self.db
  return @@db if @@db
  @@db = Hash[fetch_languages.collect { |item| [item, nil] }]
end

# File lib/wikipedia.rb, line 25
def self.languages
  db.keys
end

# File lib/wikipedia.rb, line 43
def self.size(title, language = 'en')
  @@sizes[language] ||= {}
  @@sizes[language][title] || 1
end

# File lib/wikipedia.rb, line 55
def self.fetch_article(title, language = 'en')
  begin
  url = "https://#{language}.wikipedia.org/wiki/#{URI::encode(title)}?action=raw"
  @@http.get_content(url)
  rescue => e
    puts e
    nil
  end
end

# File lib/wikipedia.rb, line 65
def self.fetch_languages
  # hardcoded list (used as fallback)
  default = [
    'en', 'ceb', 'sv', 'de', 'nl', 'fr', 'ru', 'it', 'es', 'war', 'pl', 'vi',
    'ja', 'pt', 'zh', 'uk', 'ca', 'fa', 'ar', 'no', 'sh', 'fi', 'hu', 'id',
    'cs', 'ko', 'ro', 'sr', 'tr', 'ms', 'eu', 'eo', 'bg', 'da', 'min', 'kk',
    'hy', 'sk', 'zh-min-nan', 'he', 'lt', 'hr', 'ce', 'et', 'sl', 'be', 'gl',
    'nn', 'el', 'uz', 'la', 'simple', 'vo', 'ur', 'hi', 'az', 'th', 'ka'
  ]

  # fetch from the internet
  lst = nil
  begin
  url = 'https://meta.wikimedia.org/wiki/List_of_Wikipedias'
  doc = Nokogiri::HTML(@@http.get_content(url))
  lst  = doc.at_css('[id="1_000_000.2B_articles"]').xpath('./ancestor::h3')[0].next.next.children.xpath('td[4]').map {|e| e.text}
  lst += doc.at_css('[id="100_000.2B_articles"]').xpath('./ancestor::h3')[0].next.next.children.xpath('td[4]').map {|e| e.text}
  rescue
  end

  # uses default list if theres no results
  lst = (lst.class == Array) && lst.any? ? lst : default
  lst[0, 15]
end

# File lib/wikipedia.rb, line 90
def self.fetch_vital(language = 'en')
  lst = []

  begin
  url = 'https://en.wikipedia.org/wiki/Wikipedia:Vital_articles'
  doc = Nokogiri::HTML(@@http.get_content(url))

  # not english, click on "Languages" section (left column)
  if language != 'en'
    link = doc.at_css("a.interlanguage-link-target[lang='#{language}']")
    return [] if ! link
    url = link.attribute('href')
    doc = Nokogiri::HTML(@@http.get_content(url))
    return [] if ! doc
  end

  # get all links starting with "/wiki/" and with no class (it will get some extra non-vital articles, but it's ok)
  lst = doc.xpath('//li/a[not (@class)][starts-with(@href, "/wiki/")]').map do |e|
    href = e.attribute('href').text[6..-1] # remove initial '/wiki/'
    href = URI::decode(href) # convert non-ascii chars
    pos = href.index('#')
    href = href[0, href.index('#')] if pos # remove anchor
    href
  end

  lst.uniq!

  rescue
  end # rescue

  lst
end