Markov

markov.rb

# Get phrases
phrases = File.readlines('phrases.txt')
               .map { |p| p.gsub(/["()*]/, '') }
               .map { |p| p.gsub(/ +/, ' ').strip }
               .reject { |p| p.split(' ').size < 2 }
               .uniq

# Get first words
starters = phrases.map { |p| p.split(' ').first }

# Create a chain { key => [word1, word2] }
chain = {}
phrases.each do |phrase|
  words = phrase.split(' ')
  words.each_with_index do |word, index|
    next_word = words[index + 1]
    chain[word] ||= []
    chain[word] << next_word
  end
end

10.times do
  phrase = []

  # first word from starters
  word = starters.sample
  while word
    phrase << word

    # next word from the chain
    word = chain[word].sample
  end

  phrase = phrase.join(' ')
  puts phrase unless phrases.include?(phrase)
end