Fakedown -- A fake markdown post generator for Static Site Generators

This is a small, WSGI Flask-based web app that can be used as a text-based web service to generate fake markdown posts using Markov chains using two-word tuples as key.

The corpus has been created using text from various Project Gutenberg sites.

I made this app to quickly generate a huge amount of fake blog posts and test Static Site Genrators' performance.

It is geared towards Jekyll and Eleventy, but I guess it can be adapted to generate posts that can be read by other SSGs.

Usage

The services exposes the following GET endpoints

• /words -- generates some words
• /phrase -- like /words, but longer
• /paragraph -- some phrases combined in a single paragraph
• /paragraphs -- some markdown paragraphs
• /markdown -- a complete markdown post, sans front matter
• /markdown-post -- as above, with a complete fake fron matter
• /eleventy-post -- as above.

There's a live instance of an early version of this package running here:

https://fakedown.xoxarle.com/eleventy-post

You can use cUrl to generate and save a post this way:

$ curl https://localhost:5000/eleventy-post --output the-post.md

Offline usage

The module MarkovJekyll.py can also be executed as a standalone program. It will generate 10 markdown posts on 10 different files.

Using and creating a different corpus

If you want to generate post using different reference text, you can proceed this way.

1. Create a huge, UTF8-text file containing the text you want to "markov"-ize. Again, Project Gutenberg is your friend.

2. Change the following line in the file MarkovJekyll.py

markdownGenerator = MarkdownGenerator("word_cache.json")

to

markdownGenerator = MarkdownGenerator("my-text-reference.txt")

and run the program.

A new frequency map will be created and stored into word_cache.json and will serve as a reference for your fake markdown posts.

License

This app is licensed under LGPL 2.1