These instructions assume a working miniconda or anaconda installation.

First, you'll need to set up a conda environment for the projects:

conda config --add channels conda-forge
conda create -n minecraft-assignments python=3.6 numpy scipy matplotlib ffmpeg openjdk future
# This is a line you'll be typing in every new terminal window
source activate minecraft-assignments

If you're on Linux or Mac OS, we are using the CrowdAI conda recipe for Malmo, so we'll need to install that separately:

conda install -c crowdai malmo

If you are on Windows, install Malmo and its dependencies using the instructions in the README here.

Next we'll throw in pulp, which has no Conda recipe, along with some Python helpers used by various development tools:

pip install pulp flake8 mypy jedi autopep8 yapf black pylint

In non-Malmo exercises, you'll need to run this in any Terminal window you want to use (or use emacs's pyenv-venv package, or…):

# One time in each terminal window you open
source activate minecraft-assignments 
# Anytime you want to run a program
python test_path_planning.py # For example

In the exercises which use Malmo, you'll need two terminal windows. In one, launch Minecraft:

source activate minecraft-assignments
cd "$MALMO_MINECRAFT_ROOT" && ./launchClient.sh

In the other, run your Python script (make sure the minecraft-assignments environment is active here too!):

source activate minecraft-assignments
python mission_script.py

On Windows, you can run your downloaded Malmo installation's launchClient.bat from a Powershell prompt and your Python script from Anaconda's cmd.