# Installation instructions.
#
# The Pommerman homepage is located at
# https://www.pommerman.com/
#
# These instructions are based on the docs at
# https://github.com/suphoff/pommerman/tree/master/docs
#

# To setup a virtual environment for the necessary python packages
# install anaconda from
https://www.anaconda.com/download/

# If anaconda is too big you can also install miniconda from
# https://conda.io/miniconda.html or use virtual environments in another way

# Create virtual environment for pommerman
conda create -n pommerman python=3.6

# Activate pommerman virtual environment
source activate pommerman

# Git clone pommerman playground using a specific git version so that everyone has the same version
git clone https://github.com/MultiAgentLearning/playground
cd playground
git reset 164ff38f7844696fc4a5e97487e57c6f358cb94a

# Install playground
pip install -U .

# Try out the example script (should display 4 agents running around and exploding things)
python examples/simple_ffa_run.py

# You can try out different ready made agents by modifying the "simple_ffa_run.py" script
- SimpleAgent includes some smart heuristics
- RandomAgent chooses random actions
- With PlayerAgent you can use the keyboard to control the agent
# To disable visualization change "render_mode='rgb_array'"