Planner Interface for PDDLGym

This is a lightweight Python interface for using off-the-shelf classical planners like FastForward and FastDownward with PDDLGym.

This library is under development by Tom Silver and Rohan Chitnis. Extensions to this library have been made by Mohamed Khodeir and Christopher Agia to support a broader range of satisficing and optimal symbolic planners encapsulated in a pip-installable package. Correspondance: tslvr@mit.edu and ronuchit@mit.edu, m.khodeir@mail.utoronto.ca and cagia@stanford.edu.

Setup

System Requirements

This repository has been mostly tested on MacOS Mojave and Catalina with Python 3.6. We would like to make it accessible on more systems; please let us know if you try another and run into any issues.

Installation

We recommend creating a virtual environment, e.g. with venv or anaconda3 before proceeding with the following installation steps.

# if on macOS
brew install coreutils 
# clone and install package
git clone https://github.com/agiachris/pddlgym_planners.git --recurse-submodules
cd ./pddlgym_planners && pip install .
# if your virtual env does not have pddlgym
pip install -r requirements.txt

Instructions

Basic Usage

Important Note: When you invoke a planner for the first time, the respective external package will be installed automatically. This will take up to a few minutes. This step will be skipped the next time you run the same planner.

import pddlgym
from pddlgym_planners.ff import FF  # FastForward
from pddlgym_planners.fd import FD  # FastDownward

# Planning with FastForward
ff_planner = FF()
env = pddlgym.make("PDDLEnvBlocks-v0")
state, _ = env.reset()
print("Plan:", ff_planner(env.domain, state))
print("Statistics:", ff_planner.get_statistics())

# Planning with FastDownward (--alias seq-opt-lmcut)
fd_planner = FD()
env = pddlgym.make("PDDLEnvBlocks-v0")
state, _ = env.reset()
print("Plan:", fd_planner(env.domain, state))
print("Statistics:", fd_planner.get_statistics())

# Planning with FastDownward (--alias lama-first)
lama_first_planner = FD(alias_flag="--alias lama-first")
env = pddlgym.make("PDDLEnvBlocks-v0")
state, _ = env.reset()
print("Plan:", lama_first_planner(env.domain, state))
print("Statistics:", lama_first_planner.get_statistics())

Extended Usage

Upon importing this package in your python script, you'll have easy access to both satisficing and optimal planners through the pddlgym_planners.PlannerHandler object; a dictionary hashing a planner name to its corresponding PDDLPlanner object. You may also directly import a planner with get_planner should you know its name and alias. All necessary dependencies are auto-installed for planners being used the first time.

# pyexample.py script

import pddlgym
import pddlgym_planners

# Instantiate planner handler (dict)
planners = pddlgym_planners.PlannerHandler()
print(planners.keys())      # [optional] check for short-form planner names (keys)

# Planning with Cerberus-agl (satisficing)
cerberus_planner = planners["Cerberus-seq-agl"]
env = pddlgym.make("PDDLEnvBlocks-v0")
state, _ = env.reset()
print("Plan:", cerberus_planner(env.domain, state))
print("Statistics:", cerberus_planner.get_statistics())

# Planning with Delfi (optimal)
delfi_planner = planners["Delfi"]
env = pddlgym.make("PDDLEnvBlocks-v0")
state, _ = env.reset()
print("Plan:", delfi_planner(env.domain, state))
print("Statistics:", delfi_planner.get_statistics())

# Can also directly access planners with get_planner
planner_data = {
    "name": "FD", 
    "kwargs": {"alias_flag": "--alias seq-opt-lmcut"}
}
fd_planner = pddlgym_planners.get_planner(planner_data["name"], **planner_data["kwargs"])
env = pddlgym.make("PDDLEnvBlocks-v0")
state, _ = env.reset()
print("Plan:", fd_planner(env.domain, state))
print("Statistics:", fd_planner.get_statistics())

Please refer to pddlgym_planners/__init__.py for the names of possible planners to choose from. Additional samples are provided in test.py.