ActsAsMongoRateable (with weights!)

Inspired by the old Rails+AR standby "acts_as_rateable," this rating plugin works with MongoDB+MongoMapper and has weighted ratings, as well as bayesian and straight averages, and some friendly class-level helpers.

Intends to be super-performant by taking advantage of the benefits of document-driven db denormalization.

Requirements

MongoDB
MongoMapper gem
Expects you to have a User model that includes MongoMapper::Document

Installation

Install the plugin:

./script/plugin install git://github.com/mepatterson/acts_as_mongo_rateable.git

Add the following 2 lines to the Model class that you want to make rateable:

include ActsAsMongoRateable
RATING_RANGE = (1..5)

Obviously, change the rating range if you want to rate on a 10-star system or a 14-star or whatever.

Usage

class User
  include MongoMapper::Document
end

class Widget
  include ActsAsMongoRateable
  RATING_RANGE = (1..5)
  include MongoMapper::Document
end

widget = Widget.first

To rate it:

widget.rate(score, user, weight)

score must be an Integer within your RATING_RANGE
user is the User who is rating this widget
weight is optional; defaults to 1)

Now try all these fun methods:

widget.average_rating

widget.bayesian_rating

widget.rating_stats

And some useful class methods:

Widget.highest_rated(how_many)

Widget.most_rated(how_many)

Widget.most_rated_by_authorities(how_many)

Widget.highest_bayesian_rated(how_many)

('how_many' is a limit and is optional. i.e. Do you want a highest_rated list of 5, 10, 15 widgets?
Defaults to just 1 if you don't pass any argument.)

Future

Tests (I have tests in the project I cut this from, but they need to be extracted out and I'm lazy)
More helper methods
Performance improvements as I come across the need
Investigate using map/reduce to improve the efficiency of the bayesian calc (?)

Thanks To...