enjuus / margopher Goto Github PK

Markov chain random text generator

TL;DR You give it some text, it gives you a random (and sane) chunk out of this text.

Basically margopher is based on Markov chain statistical model, except that it starts by choosing a random prefix from the states dictionary to ensure more randomness.

It starts by parsing the input text and putting it into states dictionary then starts generating the output sentence.

The generator will keep generating words till it encounters a terminal word (a word that ends in any of .,:;?!)

States is a map contains prefix as keys and suffix as values.

Prefix is an array of two consecutive words from the original text.

Suffix is a slice of all the words that occur after a given prefix.

go get github.com/aonemd/margopher

1. Import the package

import "github.com/aonemd/margopher"

2. Create new margopher object

m := margopher.New()

3. Read input text using one of three parsing methods:

• ReadText(text string)

text := "Cats are nice. Cats love pizza, and Cats hates dogs."
fmt.Println(m.ReadText(text))

• ReadFile(filePath string)

filePath := "../file.txt"
fmt.Println(m.ReadFile(filePath))

• ReadURL(url string)

url := "https://github.com/aonemd/margopher"
fmt.Println(m.ReadURL(url))

4. You can see the input parsed into a dictionary of States using margopher.StateDictionary():

• It returns a dicitonary of this signature map[[2]string][]string.
• The words are unordered because Go maps do not keep order.

fmt.Println(m.StateDictionary())

I wrote this simple project particularly to learn Go so any feedback is more than welcome. If you have any, please open an issue or send a pull request.

See LICENSE.