random word generator (with multiple methods, based off of a previous markov generator)

markov.rb

require 'json'

TRUELOOKFORWARD = true
LOOKFORWARD = 2 #how many words forward it generates from a seed word

def prepareString(str)
	#return str.downcase.split(/(\W)+/).map{|word| word.gsub(/\s+/, "")}.reject{|word| word == ""} to remind you what it was
	return str.split("").push("").unshift("")
end

def addArrayToChain(strArray, chain)
	strArray.each_with_index do |currentStr, index|
		doneWithString = false
		LOOKFORWARD.times do |t|
			if strArray[index + t + 1].nil?
				doneWithString = true
			end
		end #break if no more strings can be found up ahead
		break if doneWithString
		if TRUELOOKFORWARD
			if index - (LOOKFORWARD - 1) <= 0
				currentStr = ""
			else
				currentStrArray = []
				LOOKFORWARD.times do |lookforwardIndex|
					currentStrArray.unshift(strArray[index - lookforwardIndex])
				end
				currentStr = currentStrArray.join("")
			end
		end
		nextStringsArray = [] #make an array of the next strings
		LOOKFORWARD.times do |t| nextStringsArray << strArray[index + t + 1] end
		nextStrings = nextStringsArray.join("")

		if !chain.key?(currentStr)
			chain[currentStr] = {}
		end
		if !chain[currentStr].key?(nextStrings)
			chain[currentStr][nextStrings] = 0
		end
		chain[currentStr][nextStrings] += 1
	end
	#File.open("markov2lookaheadbackup.rson", "w") { |file| file.write(JSON.generate(chain)) }
	return chain
end

def generateWords(chain)
	#topicWords = topic.split(/\s+/).reject{|word| word.include? BOTNAME}.sort_by(&:length).select{|word| word.length > 2} #select suitable-looking words for the bot to talk about
	topicWord = ""
	#topicWords.each do |word| #try to make the bot talk about them
	#	if !chain[word].nil?
	#		topicWord = word
	#		break
	#	end
	#end

	outputString = []
	outputString << topicWord #put the first seed word in
	loop do
		if !TRUELOOKFORWARD
			lastWord = outputString.last.to_s.split("").last
			lastWord = "" if lastWord.nil?
		else
			if outputString.length < LOOKFORWARD
				lastWord = ""
			else
				lastWord = outputString[-1]
			end
		end
		break if chain[lastWord].nil? #if there aren't words to complete the phrase

		chanceTable = chain[lastWord]

		cdf = 0 #calculating the weighted random, warning: gross
		chanceTable.each do |weight|
			cdf += weight[1]
		end
		randomInt = rand(cdf)
		runningSum = 0
		chanceTable.each_pair do |word, weight|
			runningSum += weight
			if runningSum > randomInt
				outputString << word
				break
			end
		end
		break if outputString.last == ""
	end
	return outputString.join("")
end

begin
	@permChain = JSON.parse(File.read("markov2truelookaheadbackup.rson"))
rescue
	puts "Backup file not found"
	@permChain = {}
	File.open("/usr/share/dict/words").each_line do |word|
	#`cat /usr/share/dict/words | head -n 50`.each_line do |word|
		@permChain = addArrayToChain(prepareString(word), @permChain)
	end
ensure
	10000.times do
		puts generateWords(@permChain)
	end
	File.open("markov2truelookaheadbackup.rson", "w") { |file| file.write(JSON.generate(@permChain)) }
end