gnomes-at-night-gym 0.0.11

An environment for Gnomes at Night based on OpenAI gym.

Installation

You can install the package using pip:

pip install gnomes-at-night-gym

Usage

Here is an example of how to import and create the environment:

import gymnasium as gym
import gnomes_at_night_gym
env = gym.make("gnomes_at_night_gym:Board9-v0-multistep", render_mode="rgb_array", board_seed="A", round=1)

Then, we can use the environment by:

# Reset the environment
obs_n, info = env.reset(seed=1)
done = False

# Run the environment
while not done:
    action = env.action_space.sample()
    obs_n, reward, terminated, truncated, info = env.step(action)
    done = terminated or truncated
    env.render()
env.close()

Environment Details

Observation Space

Each player's observation space is a (149,) array containing:

• current_player (1 value)
• token_pos (2 values)
• flattened_walls (144 values) → {horizontal_walls (8x9) + vertical_walls (9x8)}
• treasure_pos (2 values)

Action Space

The action space allows the following actions:

• -1: Switch turn
• 0: Stay put
• 1: Move right
• 2: Move up
• 3: Move left
• 4: Move down

Coordinate System

The maze is zero-indexed, with coordinates (x, y) where:

• x represents the vertical axis (row)
• y represents the horizontal axis (column)
• Similar to traditional matrix notation

Below is a visual representation of the coordinate system for a 3x3 section of the maze:

(x, y)	0	1	2
0	(0,0)	(0,1)	(0,2)
1	(1,0)	(1,1)	(1,2)
2	(2,0)	(2,1)	(2,2)