マルコフ連鎖でツイートするやつ

Gemfile

# -*- coding: utf-8 -*-

source 'https://rubygems.org'

gem 'twitter'
gem 'natto'

markov_bot.rb

# -*- coding: utf-8 -*-

require 'natto'
require 'twitter'
require 'uri'

MARKER_BEGIN = '__BEGIN__'
MARKER_END   = '__END__'

Twitter.configure do |config|
  config.consumer_key       = '<!!! consumer key !!!>'
  config.consumer_secret    = '<!!! consumer secret !!!>'
  config.oauth_token        = '<!!! oauth token !!!>'
  config.oauth_token_secret = '<!!! oauth token secret !!!>'
end

tagger = Natto::MeCab.new('-O wakati')
client = Twitter.client
statuses = client.user_timeline '<!!! scree_name !!!>', :count => 200

original_texts = []
nodes = []
statuses.each do |status|
  text = status.text
  # url が含まれてたらスキップ
  next if text.match(URI.regexp)
  # screen_name を除外
  text = text.gsub(/@[\w_]+/, '')
  # RT / QR を除外
  text = text.gsub(/(RT|QT):? @[\w_]+.*(\n|$)/, '')
  # ハッシュタグを除外
  text = text.gsub(/#[\wＡ-Ｚａ-ｚ０-９ぁ-ヶ亜-黑]+/, '')
  # 前後の空白削除
  text = text.strip
  original_texts << text
  words = tagger.parse(text).split(' ')
  nodes.concat([MARKER_BEGIN, *words, MARKER_END].each_cons(3).to_a)
end

# 原文そのままでないやつが出来るまで繰り返す
text = nil
loop do
  # マルコフ連鎖
  node = nodes.select { |n| n[0] == MARKER_BEGIN }.sample
  text = node.join
  loop do
    node = nodes.select { |n| n[0] == node[1] && n[1] == node[2] }.sample
    break unless node
    text += node[2]
    break if node[2] == MARKER_END
  end
  text = text.gsub(MARKER_BEGIN, '').gsub(MARKER_END, '')
  break unless original_texts.include?(text)
end

client.update text
puts text