module EmailCollector

# File lib/email_collector.rb, line 19
def self.collect(searchReq, domain = nil)
  @keywords.map { |keyword| collect_plain("#{searchReq} #{keyword}", domain) }.flatten.uniq.compact
end

# File lib/email_collector.rb, line 23
def self.collect_plain(searchReq, domain = nil)
  #@logger.debug("searching for #{searchReq}")
  #@logger.debug("domain = #{domain}")
      
      if (domain)
    res = google_search("#{searchReq} \"#{domain}\"")
    res_at = google_search("#{searchReq} \"at #{domain}\"")
    
    (res + res_at).map do |context|
      #@logger.debug("context = #{context}")
      context = filter_at_domain(context, domain)
      context.scan(/[a-z0-9._%+-]*[a-z0-9_%+]@#{Regexp.quote(domain)}/i)
    end
      else
    google_search(searchReq).map do |context|
      @logger.debug("context = #{context}")
      context.scan(/[a-z0-9._%+-]*[a-z0-9_%+]@(?:[a-z0-9.-]+\.)+[a-z0-9]{2,}/i)
    end
      end
end

# File lib/email_collector.rb, line 60
def self.filter_at(s)
  s.gsub(/\s+/, ' ').gsub(/[^a-z0-9_.%+-]+[ae]t[^a-z0-9.@-]+|([_+-]+)[ae]t\1/i, '@')
end

# File lib/email_collector.rb, line 65
def self.filter_at_domain(s, domain)
  s.gsub(/[^a-z0-9_%+-]+#{Regexp.quote(domain)}/, '@' + domain)
end

# File lib/email_collector.rb, line 75
def self.filter_b(s)
  s.gsub(/<\/?b>/, '')
end

# File lib/email_collector.rb, line 70
def self.filter_exclam(s)
  s.gsub(/[!:]/, '.')
end

# File lib/email_collector.rb, line 44
def self.google_search(searchReq)
  @logger.debug("searching for #{searchReq}")

  Google::Search::Web.new do |search|
    search.query = searchReq
    search.size = @size
  end.map do |item|
    #@logger.debug("URI = #{item.uri}")
    #@logger.debug(item.content)
    s = filter_b(item.content)
    s = filter_at(s)
    filter_exclam(s).split('...')
  end.flatten
end

# File lib/email_collector.rb, line 15
def self.keywords=(k)
  @keywords = k
end

# File lib/email_collector.rb, line 9
def self.size=(s)
  @size = s
end