1. If you're running Windows, install Cygwin. As a part of the installation, select git and python as packages to install. Boot Cygwin and proceed to step 2. If you're not on Windows, open a terminal.

2. Install pip if you do not already have it.

3. Follow these instructions:

sudo pip install virtualenv
git clone git@github.com:willcrichton/awap-2015.git
cd awap-2015
virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

You should look at the following files:

• src/game/player.py - You will implement your algorithm here.
• src/game/base_player.py - Utilities for the Player class.
• src/game/state.py - All the state maintained for the game.
• src/game/order.py - Representation of orders in the game.
• src/game/settings.py - Constants used in the game.
• src/game/graphs.py - Functions for generating graphs.

You don't need to look at any other files.

The graphs used in the code are NetworkX graphs. Some useful links:

• Short tutorial
• Full documentation

To run your algorithm quickly with no visuals, do

./run.sh shell

To step through your algorithm and see it work, do

./run.sh web

Then visit http://localhost:5000 in your browser.

To submit your algorithm for scoring, go to Autolab: https://autolab.cs.cmu.edu/courses/15097-f15/assessments/awapcompetition

Then click "Submit File" and upload src/game/player.py. Do not upload anything else (e.g. a zip/tar file, a different Python file, etc.). Your Autolab submission history will now have a new entry. If your score is -1, then your code did not compile or run correctly. Click on your score to see the error logs. Note that you can only import from src/game/ or the Python 2.7 standard library. You cannot use any custom packages besides networkx.

Submissions are closed at 7:00pm. Please submit your final bot early (like 6pm) to ensure that it works on our servers--the queue times will grow as we get close to 7:00pm.