Skip to content
Simple AI for python games. In very early stages of development. Uses concepts of Goal Orientated Action Planing (GOAP).
Python
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
npc
pathfinding added some debug print statements, actions seem to work Apr 22, 2013
pygoap
16x16-overworld.png rewrite for version 3 Jan 19, 2012
NOTES
README
formosa.tmx
gpl-3.0.txt
test.py added comments in the test file Apr 22, 2013
tmxloader.py
tutorial.py

README

pygoap v.3
requires python 2.x

PyGoap is a small library for designing AI in python.  The basic library is based off of a well-known idea of using graph searches to to create realistic agents in real-time.  Behavior is determined in real-time and is extremely open ended and well suited for emergent behaviors.

Agents, or npcs, can be programmed very simply with an easy to understand api.  Check the 'npcs' folder and read through actions.py to get an idea of how it works.

This library is not complete, but is working as-is.  The demo gives you a simple diagram of the game world and the debugging information near the bottom lets you know what is going on 'in the heads' of the agents.  Note: the demo requires pygame.



I've build AI Agents around the concept of Actions and ActionBuilders.

Actions can be subclassed from a few different action types.
ActionBuilders will search a blackboard and return a list of actions that can be performed.



If you are unfamiliar with GOAP, I invite you to do some research online, it may help you to understand this module.



you can watch it work by running "test.py".


many thanks to opengameart.org for providing and hosting the tileset used in the project.  tiles for the map can be found at:
http://opengameart.org/content/worldmapoverworld-tileset
under the CC-BY 3.0 license.

have fun.
Something went wrong with that request. Please try again.