Galaxy OJ is an online judge system ready to be deployed on Linux systems.

• python3
• postgresql
• redis

First install python dependencies (using a virtualenv is recommended):

pip install -r requirements.txt

Then, you probably should set the following environment variables for your web server or change the values in server/config.py:

# the configuration set you want to use. Always use 'PRODUCTION' in production environment
CONFIG_NAME
# database uri for sqlalchemy (The app is only tested with PostgreSQL).
DATABASE_URI
# secret key for the app. You probably want a random one (maybe using 'openssl rand').
SECRET_KEY
# path to the data folder used by the server.
DATA_FOLDER
# the url to redis server
REDISTOGO_URL

Replace $DATA_FOLDER to the path you specified in your configuration:

$ mkdir -p $DATA_FOLDER/{submissions,testcases,uploaded_images}

After these, you should create the database:

# here take postgresql as an example. Note that the database name should be the
#   same as specified in DATABASE_URI
$ createdb galaxyoj

And then create the tables:

# use migration scripts
$ python manager.py db upgrade
# or import server.db and call 'db.create_all()'

Currently the OJ has three types of users: normal, admin, root. You will need a root user to manage user privileges.

Launch a python shell in app context:

$ python manager.py shell

Run the following in that shell:

u = User()
u.login_name = 'root'
u.password = 'whatever'
u.email = '[email protected]'
u.privilege_level = 10
db.session.add(u)
db.session.commit()

For a development environment, you will want to just run a redis server, a judge worker, and the web server, all on your local machine.

Example (assuming default connection):

$ redis-server

$ python -m judge

$ python manager.py runserver --threaded

The procedures for running a server for production use is similar to that for development use, except that you should not use the web server of flask, but make use of real web servers such as uwsgi.

Since the environments are different and we won't and can't cover all cases, this section is deliberately left blank.

In the future, we may employ docker to provide easy production set up.

Dual license:

• For non-commercial use, the software is licensed under GPLv3.0.
• For commercial use, please contact the authors for licensing.

• Jan 3rd, 2016: Version 1.0 is released.