witter 0.0.4

Witter - generate random text in the style of the document you provide.

Witter

What's Witter?

Witter accepts text input, splits it into sections, or "chains", and then works out which character is most likely to follow that "chain" whenever it appears in the source text.

Once it has analyzed the source text, it generates a number of random sample pieces of text based on it.

Because of the way it uses the source text, the text that it produces will be "in the style" of the original text. It may, or may not, make sense.

It's worth noting that this isn't Machine Learning (ML), or any kind of Artificial Intelligence (AI). It's statistics, with some Monte Carlo methods thrown in.

How Do I Use Witter?

Run witter from within a virtual environment using the syntax:

witter --help

or from the command line using:

python -m witter --help

Both of these examples will display the command line options.

A simple example to get you started is:

• download a large text file - perhaps The Complete Works of Shakespeare from Project Gutenberg
• pass the contents of the file to witter

In Windows or Linux, in a Python virtual environment, the following command will produce samples:

witter filepath.txt

or, if you want to use piping in Linux:

cat filepath.txt | witter

So, if you want to generate text from The Complete Works of William Shakespeare, and why wouldn't you?

wget https://www.gutenberg.org/cache/epub/100/pg100.txt
cat pg100.txt | witter

or if A Christmas Carol by Charles Dickens is more your thing:

wget https://www.gutenberg.org/ebooks/24022.txt.utf-8
cat 24022.txt.utf-8 | witter

You've probably noticed the utf-8 extension on one of the text files. This is a way that Project Gutenberg uses to denote UTF-8-encoded text files. witter can handle UTF-8 with no problems.

Download Statistics