AndrewPaulChester / sage-code

This repository contains the code for the manuscript "SAGE: Generating Symbolic Goals for Myopic Models in Deep Reinforcement Learning", currently under review.
MIT License
3 stars 0 forks source link

PLANNER_PATH not provided #1

Open tianpeiyang opened 1 year ago

tianpeiyang commented 1 year ago

Hi there,

I tried to run the source code, but it misses two directories/files in the /gym-agent/planner.py, such as "downward/fast-downward.py" and "ENHSP/enhsp" Could you fix this issue? Thanks

AndrewPaulChester commented 1 year ago

Hi,

Thanks for your interest! Due to platform compatibility reasons, the two planners cannot be bundled with this code release. They need to be installed according to their instructions: https://www.fast-downward.org/ and https://gitlab.com/enricos83/ENHSP-Public . When they are installed, create an environment variable PLANNER_PATH that is the root directory in which you have installed the planners, and you should be good to go.

Full instructions can be found in the README of this repository.