Experiments into markov chains, n-grams, and text generation. This repository is the result of the markov chain blog post mini-series on my blog.
- Part 1 - n-grams
- Part 2 - unweighted markov chains
- Markov Chains Part 3: Weighted Chains
- Markov Chains Part 4: Test Data
Building this project should be easy - I can provide prebuilt binaries upon request though if asked (see my website for contact information).
Open the solution in Visual Studio / MonoDevelop and hit build.
Linux and everyone else
As above with MonoDevelop, or simply use the command
xbuild whilst inside the root of the repository.
Detailed usage help can be seen by simply running the tool without any arguments:
Linux users might need to explicitly run it with
wordlists/ directory contains a few interesting wordlists I used when writing and testing this program.
||List of cool sciencey-type words||Compiled from scratch by Starbeamrainbowlabs - that's me! This list falls under the Mozilla Public License 2.0, as described below in the license section.|
||List of items in Cross-Code||A list of all items in the game CrossCode. Scraped from here by a clever bit of bash in
||List of items in Final Fantasy 15||Another list of in-game items - this time from Final Fantasy XV. Scraped from this wiki page and related pages. The even cleverer bit of bash that does this is also in
||List of items in No Man's Sky||Yep, you guessed it. Don't own this. Bash available in
||List of blocks and items in Starbound||From the Official Wiki - I don't own it, I just wrote the scraper :P|
The following pages & websites looks like they show promise, but I haven't imported them yet.
- The code was written by Starbeamrainbowlabs - that's me too!
- I found the Markov Chain Text Generation post by Chris Wellons rather useful when writing this.
- Wordlists - See the table above
This repository (except where stated above - the wordlists except the science words) is licensed under the Mozilla Public License 2.0 (MPL-2.0). A copy of the license text can be found in the LICENSE file in this repository, and tldr-legal have a summary if you can't speak legalese :-)