Web scraping sample 青空文庫

download.rb

require 'rubygems'
require 'bundler'
Bundler.require
require 'open-uri'

aozora_uri = 'http://www.aozora.gr.jp/cards/000148/files/789_14547.html'
file = 'sample.txt'
unless File.exist?(file)
    html = open(aozora_uri, 'r:Shift_JIS').read.encode('utf-8', universal_newline: true)
    text = Oga.parse_html(html).css('.main_text').text.gsub(/（.*?）/, '')
    File.write(file, text)
    puts "#{file} created"
end