#lol_stats2

This is the data-harvesting, number-crunching, REST-serving backend. The frontend can be found here: https://github.com/chenwardT/lol_stats2-frontend

"There is an epidemic failure within the game to understand what is really happening"

--Peter Brand (Jonah Hill), "Moneyball"

lol_stats2 is a website that aims to provide unique insight into the competitive multiplayer game League of Legends, developed by Riot Games. Through querying the official League of Legends API it accesses match, player, and team data with the aim of finding patterns and making distinctions that may not be apparent to players/spectators of the game.

##Background The site is built on Django, a Python web framework.

API access is provided by RiotWatcher.

Asynchronous jobs, such as fetching and analyzing data, are performed by Celery, a distributed task queue. It is uses RabbitMQ as its broker and Redis as a result store.

Stored data is exposed via Django-Rest-Framework with and consumed by a JS frontend.

This project and its author are not affiliated with Riot Games.

##Setup

###Virtual Environment Install virtualenvwrapper if not already present:

[http://virtualenvwrapper.readthedocs.org/en/latest/install.html#basic-installation] (http://virtualenvwrapper.readthedocs.org/en/latest/install.html#basic-installation)

Find out where python3 lives:

which python3

Setup a development virtualenv:

mkvirtualenv --python=/path/to/python3 lol_stats2_dev

Install dependencies:

pip install -r requirements/development.txt

There is a requirements file for each environment; it is recommended that you create a virtualenv for each one (e.g. using production.txt will not install debugging packages).

virtualenvwrapper provides hooks for pre- and post- activation and deactivation of each virtualenv under $VIRTUAL_ENV/bin (by default this is in ~/.virtualenvs). It is recommended that you add the following to the development virtualenv "postactivate" script:

export DJANGO_SETTINGS_MODULE="lol_stats2.settings.development"

and "predeactivate" script:

unset DJANGO_SETTINGS_MODULE

You can follow suit for each virtualenv (production, staging, etc).

Executing manage.py runserver should now reflect the environment-specific settings when starting up.

You should configure Django's SECRET_KEY and RiotWatcher's RIOT_API_KEY environment variables via "postactivate" and "predeactivate" scripts.

Additionally, CELERY_ALWAYS_EAGER, should be set in your testing virtualenv, otherwise Celery tasks won't be run against the test database! The value of the variable does not matter, but it must exist.

###Database Caveat

Django does not create an index for the matches app's ParticipantTimeline model's participant field despite db_index=True being set.

To account for this, the following will create the proper index in Postgres:

 CREATE INDEX matches_participanttimeline_idx
  ON matches_participanttimeline
  USING btree
  (participant_id);

##Celery

Celery is configured to use a locally hosted AMQP broker with a Redis result backend. A single topic-type exchange, 'default', is configured to route messages to one of 3 queues. This is likely to change as development progresses. See settings/base.py for details.

Workers may be started independently due to an issue with celery multi:

celery -A lol_stats2 worker -l info -Q default -n default.%h
celery -A lol_stats2 worker -l info -Q match_ids -n match_ids.%h
celery -A lol_stats2 worker -l info -Q store -n store.%h

Alternatively, you may start (and restart!) workers via workers_restart.sh but beware of RuntimeError: Acquire on closed pool, as it uses celery multi.

The celery monitor is optional, and listens on port 5555 by default:

celery -A lol_stats2 flower

See http://celery.readthedocs.org/en/latest/userguide/monitoring.html#flower-real-time-celery-web-monitor.

Finally, the rabbitmq monitor (defaults; port: 15672, login: guest/guest) can be enabled via:

rabbitmq-plugins enable rabbitmq_management

This only needs to be run once and by default will start on boot.

See https://www.rabbitmq.com/management.html.

##Tests

Before running tests, ensure CELERY_ALWAYS_EAGER is set. See bottom of Setup section.

Selenium tests use PyVirtualDisplay to run in a headless environment. PyVirtualDisplay in turn requires Xvfb, Xephyr, or Xvnc to be installed.

To run tests:

./manage.py test