jtebert / arena-gen

Programmatically generated environments for gridsim projects

Arena Generator

Programmatically generated environments for gridsim projects

This generates rectangular grayscale images based on Perlin noise, using the Python noise package.

This is designed to generate environments (images) for use with Gridsim.

Installation and setup

Create a Python3 virtual environment:

python3 -m venv venv

Activate the virtual environment (you can deactivate it with deactivate):

source venv/bin/activate

Install the dependencies:

pip install -r requirements.txt

Running

With the virtual environment active, run:

python gen.py CONFIG_FILE.yml

where CONFIG_FILE.yml is a YAML configuration file containing the configuration parameters specified below. An example is provided as demo_config.yml.

Parameters

• output_dir (str, optional): Directory where to save the generated images. If not specified, images will be placed in the current directory.

• img_filename_prefix (str, optional): Every image filename will start with this, followed by a 3-digit number (and .png suffix). If not specified, images will just be named as a number.

• allow_overwrite (bool, optional): Whether images with the same name will be overwritten. If true and an image exists, the script will quit instead of overwriting. Defaults to false

• img_dim (List[int, int]): Dimensions (width and height, in pixels) of the output images

• num_imgs (int): Number of images to generate and save

• img_ind_offset (int, optional): Offset of the numbering of saved images. This also determines the output image (see note below). Defaults to 0.

• frequency (float): Base scale of the output noise (lower means smaller features)

• octaves (int): How many different frequencies (scales) of noise to add.

• persistence (float): Amplitude of each additional octave (relative to the preceding octave). This should be > 0 and < 1.

• lacunarity (float): Frequency (scale) of each layer (relative to the previous octave). This should be > 1.

• make_valleys (bool, optional): Whether to generate valleys (inverted ridges), instead of completely random noise/hills. Defaults to False.

Notes

Because of the nature of Perlin and Simplex noise, they are deterministic. That means that if you provide the same parameters, you'll get the same image out. Generating different images is done by specifying an "offset" for the coordinates. That's done here with the img_ind_offset parameters. That means that images generated with the same parameters and the same index, will be the same, but you can generate new images by changing this offset.

Useful reading:

• Making maps with noise functions