Open jkelly467 opened 10 years ago
https://github.com/jkelly467/nanogenmo/blob/master/samples/horror.md
Ok, got something at least ready to show. So far, it grabs a bunch of tweets containing horror themed words, and uses those as chapter headings.
For the chapters themselves, I grab a random section of the timeline of the user who supplied the chapter heading and use those as topic sentences. The rest of the paragraph is just Markov chains based on the topic sentence.
The source material for the Markov chain is old horror novels pulled from Project Gutenberg.
For the future, I'd like to:
A random idea: since it is a horror story, corrupt some text with something like http://eeemo.net/.
For real.