For simple action games! Hopefully! It's a big work in progress.
✨ I am not maintaining this in the usual sense-- it's a hobby project that you can clone / take a look at! If you want to add in a feature, clone it and go wild, but please don't request it. I won't add it in. I make no promises about it running on your machine or my timeline to fix bugs.✨
This repository is a Python implementation of Analyzing Action Games: A Hybrid Systems Approach. Currently the only game I have a model for is Flappy Bird, so I hope you're into a mid-aughts mobile game hit. I'm not 100% sure it's correct yet, some lower bounds still look pretty suspicious to me.
To Do Some Stuff With Flappy, Use:
# download and install requirements
# (matplotlib for graphs, scipy for ODE solving, tqdm isn't needed but helps my sanity with dev work)
$ pip install -r requirements.txt
# for a single run on a generated level
$ python3 cli.py --sample_rate 0.02 single --samples 0 0 0 1 1 1
# for a reachability analysis on a generated level
$ python3 cli.py --sample_rate 0.02 reachability --num_samples 20Some cases can run really slowly-- it can take up to 15 minutes to get 1 second of game play sometimes :c. The reachability graphs can get real chonky because the currently plot both the good edge boundaries and all the failed runs.
Blog Post Pending on What Is Going On Here and Why You Might Care