Marky Markov and the Funky Sentences

Marky Markov is an experiment in Markov Chain generation implemented in Ruby. It can be used both from the command-line and as a library within your code.

NOTE: 0.3.0 now uses arrays with multiple entries per word instead of a hash key for each word with the value representing number of occurences. While a less elegant solution, it leads to much faster text generation.

Installation

gem install marky_markov

Imported Module Usage

Temporary Dictionary

A basic usage of the TemporaryDictionary, which parses strings and files into a temporary dictionary that will not be saved to disk.

require 'marky_markov'
markov = MarkyMarkov::TemporaryDictionary.new
markov.parse_string "These words will be added to the temporary dictionary."
markov.parse_file "filename.txt"
puts markov.generate_n_sentences 5
puts markov.generate_n_words 200
markov.clear! # Clear the temporary dictionary.

Persistent Dictionary

Dictionary creates or opens a persistent dictionary at a location defined by its initalizer, and will allow you to build and save a dictionary over multiple runs. to ensure existing files aren't overwritten, the system appends .mmd to the end of the dictionary name.

require 'marky_markov'
markov = MarkyMarkov::Dictionary.new('dictionary') # Saves/opens dictionary.mmd
markov.parse_file "ENV["HOME"]/Documents/largefileindocs.txt"
markov.parse_file "anotherfileyay.txt"
puts markov.generate_n_words 10
puts markov.generate_n_sentences 2
markov.save_dictionary! # Saves the modified dictionary/creates one if it didn't exist.

generate_20_words

If you keep looking at generate_n_words or generate_n_sentences and wonder why you can't put a number in there, well, you can!

markov.generate_7_sentences
markov.generate_20_words

The default dictionary depth is two words. {["I", "hope"] => ["this"], ["hope", "this"] => ["makes"], ["this", "makes"] => ["sense"]} but it can be set to a depth between 1 and 5 upon dictionary creation, though really any higher than 3 and it starts to simply print passages from the source text.

markov = MarkyMarkov::Dictionary.new('dictionary', 3)

creates a dictionary with a depth of three words. {["I", "hope", "this"] => ["makes"], ["hope", "this", "makes"] => ["sense"]

Delete a Dictionary

If you want to delete a dictionary you call it upon the Dictionary class itself while passing in the filename/location.

MarkyMarkov::Dictionary.delete_dictionary!('dictionary')

OR you can pass in a MarkyMarkov::Dictionary object directly.

MarkyMarkov::Dictionary.delete_dictionary!(markov)

Train from a string

If you want to add to your dictionary from a string instead of a file, easy!

markov.parse_string "I hope this makes sense."

Command-Line Usage

Build a Dictionary

marky_markov read textfile.txt

to build your word probability dictionary. You can run the command on different files to continue adding to your dictionary file.

Say Some Words

marky_markov speak -c 3

Will use the dictionary to create three sentences. If no number is passed it will default to five sentences..

Temporary Dictionaries

marky_markov speak -s other-file.txt -c 8

Generates a temporary dictionary based on the source file passed to it and uses that to speak. Here we're loading other-file.txt and restricting the generated text to 8 sentences.

STDIN, Pipe Away!

echo "Hello, how are you" | marky_markov

Marky-Markov is compatible with other STDIN/STDOUT command-line applications and can accept STDIN.

marky_markov listen "Bullfighting is difficult on the moon"

You can also supply a string as an argument to generate the text with, though the results are nonsense without a substantial text base to work from.

Usage: marky_markov COMMAND [OPTIONS]

Commands:
    speak: Generate Markov Chain sentence (5 sentences by default)
    listen [sentence]: Generate Markov Chain sentence from supplied string.
    read [file]: Add words to dictionary from supplied text file

Options
    -d, --dictionary LOCATION        Use custom dictionary location
    -c, --sentencecount NUMBER       Set number of sentences generated
    -s, --source FILE                Generate and use temporary dictionary from source text
        --reset                      WARNING: Deletes default dictionary.
    -h, --help                       Display this screen