predictor 2.0.0.rc1 → 2.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 29f61606a156b3a6132dc9212f3d027492285d2e
-  data.tar.gz: acf97ff88f34ca518536e18b7753fe27e69956ec
+  metadata.gz: 22ffddcaceb5b1a189aa7f0e237fad7469cb537e
+  data.tar.gz: 21c1b3aa6902f605c01e1de0cd91fa37c4e046e1
 SHA512:
-  metadata.gz: cd25db9f133ee47f44703e8e6cab7d0daed59b9a035abfac75d733d6891d0c312981f76875b0278a9b66285c3f374650b01fe4466f45b51c0d42685dafae4783
-  data.tar.gz: 4b01b5f70cdb3d4de6267e72502b50e8c8c7210f3d13855477b9d1dd9a4556b57787bca2827ddc7ac3060c915f032a2b32a50b5feb13709b9ecd5a86167c092c
+  metadata.gz: 1e6231526449192f10c34b73bee3500e52b13c7f4e9a3f6699b40c13e361dd7c098c28e199dd2eba2c8490532bf5c07f79af39027b865e31f7100fd5219ba855
+  data.tar.gz: d07e1994e3221003cc17fda719f2964ab35478a0f396735018fab0a12e278831a710eca7512ea538328a6473c0296a3a79295a48322f5b215a8b535be7e962b4

data/Changelog.md

@@ -1,7 +1,7 @@
 =======
 Predictor Changelog
 =========
-2.0.0 (2014-03-07)
+2.0.0 (2014-04-17)
 ---------------------
 **Rewrite of 1.0.0 and contains several breaking changes!**
 

data/Gemfile

@@ -1,10 +1,3 @@
 source 'https://rubygems.org'
 
-gem "redis"
-
-group :development do
-  gem "rake"
-  gem "rspec"
-  gem "yard"
-  gem "pry"
-end
+gemspec

data/README.md

@@ -11,7 +11,7 @@ Originally forked and based on [Recommendify](https://github.com/paulasmuth/reco
 * Provide item similarities such as "Users that read this book also read ..."
 * Provide personalized predictions based on a user's past history, such as "You read these 10 books, so you might also like to read ..."
 
-At the moment, Predictor uses the [Jaccard index](http://en.wikipedia.org/wiki/Jaccard_index) to determine similarities between items. There are other ways to do this, which we intend to implement eventually, but if you want to beat us to the punch, pull requests are quite welcome :)
+At the moment, Predictor uses the [Jaccard index](http://en.wikipedia.org/wiki/Jaccard_index) or the [Sorenson-Dice coefficient](http://en.wikipedia.org/wiki/S%C3%B8rensen%E2%80%93Dice_coefficient) (default is Jaccard) to determine similarities between items. There are other ways to do this, which we intend to implement eventually, but if you want to beat us to the punch, pull requests are quite welcome :)
 
 Notice
 ---------------------
@@ -19,11 +19,8 @@ This is the readme for Predictor 2.0, which contains a few breaking changes from
 
 Installation
 ---------------------
+In your Gemfile:
 ```ruby
-gem install predictor
-````
-or in your Gemfile:
-````
 gem 'predictor'
 ```
 Getting Started
@@ -59,7 +56,7 @@ class CourseRecommender
   limit_similarities_to 500   # Optional, but if specified, Predictor only caches the top x similarities for an item at any given time. Can greatly help with efficient use of Redis memory
   input_matrix :users, weight: 3.0
   input_matrix :tags, weight: 2.0
-  input_matrix :topics, weight: 1.0
+  input_matrix :topics, weight: 1.0, measure: :sorensen_coefficient # Use Sorenson over Jaccard
 end
 ```
 
@@ -191,6 +188,7 @@ predictor.topics.delete_item!("course-1")
 # to delete_from_matrix! if you want to update similarities to account for the deleted item (in v1, this was a bug and didn't occur)
 predictor.delete_from_matrix!(:topics, "course-1")
 ```
+* Regenerate your recommendations, as redis keys have changed for Predictor 2. You can use the recommender.clean! to clear out old similarities, then run your rake task (or whatever you've setup) to create new similarities.
 
 Problems? Issues? Want to help out?
 ---------------------
@@ -215,5 +213,4 @@ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
 FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
 COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
 IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
-CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

data/Rakefile

@@ -1,12 +1 @@
-require "rubygems"
-require "rspec"
-require 'rspec/core/rake_task'
-require "yard"
-
-desc "Run all examples"
-task RSpec::Core::RakeTask.new('spec')
-
-task :default => "spec"
-
-desc "Generate documentation"
-task YARD::Rake::YardocTask.new
+require "bundler/gem_tasks"

data/lib/predictor.rb

@@ -1,4 +1,5 @@
-require "predictor/version"
+require "redis"
 require "predictor/predictor"
+require "predictor/distance"
 require "predictor/input_matrix"
 require "predictor/base"

data/lib/predictor/base.rb

@@ -79,7 +79,7 @@ module Predictor::Base
       keys.concat(sets.map { |set| matrix.redis_key(:items, set) })
     end
 
-    keys.empty? ? [] : (Predictor.redis.sunion(keys) - [item])
+    keys.empty? ? [] : (Predictor.redis.sunion(keys) - [item.to_s])
   end
 
   def predictions_for(set=nil, item_set: nil, matrix_label: nil, with_scores: false, offset: 0, limit: -1, exclusion_set: [])
@@ -185,7 +185,7 @@ module Predictor::Base
   def cache_similarity(item1, item2)
     score = 0
     input_matrices.each do |key, matrix|
-      score += (matrix.calculate_jaccard(item1, item2) * matrix.weight)
+      score += (matrix.score(item1, item2) * matrix.weight)
     end
     if score > 0
       add_similarity_if_necessary(item1, item2, score)
@@ -213,4 +213,4 @@ module Predictor::Base
     end
     Predictor.redis.zadd(key, score, similarity) if store
   end
-end
+end

data/lib/predictor/distance.rb

@@ -0,0 +1,31 @@
+module Predictor
+  module Distance
+    extend self
+
+    def jaccard_index(key_1, key_2, redis = Predictor.redis)
+      x, y = nil
+
+      redis.multi do |multi|
+        x = multi.sinterstore 'temp', [key_1, key_2]
+        y = multi.sunionstore 'temp', [key_1, key_2]
+        multi.del 'temp'
+      end
+
+      y.value > 0 ? (x.value.to_f/y.value.to_f) : 0.0
+    end
+
+    def sorensen_coefficient(key_1, key_2, redis = Predictor.redis)
+      x, y, z = nil
+
+      redis.multi do |multi|
+        x = multi.sinterstore 'temp', [key_1, key_2]
+        y = multi.scard key_1
+        z = multi.scard key_2
+        multi.del 'temp'
+      end
+
+      denom = (y.value + z.value)
+      denom > 0 ? (2 * (x.value) / denom.to_f) : 0.0
+    end
+  end
+end

data/lib/predictor/input_matrix.rb

@@ -1,85 +1,82 @@
-class Predictor::InputMatrix
-  def initialize(opts)
-    @opts = opts
-  end
-
-  def parent_redis_key(*append)
-    ([@opts.fetch(:redis_prefix)] + append).flatten.compact.join(":")
-  end
+module Predictor
+  class InputMatrix
+    def initialize(opts)
+      @opts = opts
+    end
 
-  def redis_key(*append)
-    ([@opts.fetch(:redis_prefix), @opts.fetch(:key)] + append).flatten.compact.join(":")
-  end
+    def parent_redis_key(*append)
+      ([@opts.fetch(:redis_prefix)] + append).flatten.compact.join(":")
+    end
 
-  def weight
-    (@opts[:weight] || 1).to_f
-  end
+    def redis_key(*append)
+      ([@opts.fetch(:redis_prefix), @opts.fetch(:key)] + append).flatten.compact.join(":")
+    end
 
-  def add_to_set(set, *items)
-    items = items.flatten if items.count == 1 && items[0].is_a?(Array)
-    Predictor.redis.multi do
-      items.each { |item| add_single_nomulti(set, item) }
+    def weight
+      (@opts[:weight] || 1).to_f
     end
-  end
 
-  def add_set(set, items)
-    add_to_set(set, *items)
-  end
+    def add_to_set(set, *items)
+      items = items.flatten if items.count == 1 && items[0].is_a?(Array)
+      Predictor.redis.multi do
+        items.each { |item| add_single_nomulti(set, item) }
+      end
+    end
 
-  def add_single(set, item)
-    add_to_set(set, item)
-  end
+    def add_set(set, items)
+      add_to_set(set, *items)
+    end
 
-  def items_for(set)
-    Predictor.redis.smembers redis_key(:items, set)
-  end
+    def add_single(set, item)
+      add_to_set(set, item)
+    end
 
-  def sets_for(item)
-    Predictor.redis.sunion redis_key(:sets, item)
-  end
+    def items_for(set)
+      Predictor.redis.smembers redis_key(:items, set)
+    end
 
-  def related_items(item)
-    sets = Predictor.redis.smembers(redis_key(:sets, item))
-    keys = sets.map { |set| redis_key(:items, set) }
-    keys.length > 0 ? Predictor.redis.sunion(keys) - [item] : []
-  end
+    def sets_for(item)
+      Predictor.redis.sunion redis_key(:sets, item)
+    end
 
-  # delete item from the matrix
-  def delete_item(item)
-    Predictor.redis.watch(redis_key(:sets, item)) do
+    def related_items(item)
       sets = Predictor.redis.smembers(redis_key(:sets, item))
-      Predictor.redis.multi do |multi|
-        sets.each do |set|
-          multi.srem(redis_key(:items, set), item)
-        end
+      keys = sets.map { |set| redis_key(:items, set) }
+      keys.length > 0 ? Predictor.redis.sunion(keys) - [item.to_s] : []
+    end
+
+    # delete item from the matrix
+    def delete_item(item)
+      Predictor.redis.watch(redis_key(:sets, item)) do
+        sets = Predictor.redis.smembers(redis_key(:sets, item))
+        Predictor.redis.multi do |multi|
+          sets.each do |set|
+            multi.srem(redis_key(:items, set), item)
+          end
 
-        multi.del redis_key(:sets, item)
+          multi.del redis_key(:sets, item)
+        end
       end
     end
-  end
 
-  def calculate_jaccard(item1, item2)
-    x = nil
-    y = nil
-    Predictor.redis.multi do |multi|
-      x = multi.sinterstore 'temp', [redis_key(:sets, item1), redis_key(:sets, item2)]
-      y = multi.sunionstore 'temp', [redis_key(:sets, item1), redis_key(:sets, item2)]
-      multi.del 'temp'
+    def score(item1, item2)
+      measure_name = @opts.fetch(:measure, :jaccard_index)
+      Distance.send(measure_name, redis_key(:sets, item1), redis_key(:sets, item2), Predictor.redis)
     end
 
-    if y.value > 0
-      return (x.value.to_f/y.value.to_f)
-    else
-      return 0.0
+    def calculate_jaccard(item1, item2)
+      warn 'InputMatrix#calculate_jaccard is now deprecated. Use InputMatrix#score instead'
+      Distance.jaccard_index(redis_key(:sets, item1), redis_key(:sets, item2), Predictor.redis)
     end
-  end
 
-  private
+    private
+
+    def add_single_nomulti(set, item)
+      Predictor.redis.sadd(parent_redis_key(:all_items), item)
+      Predictor.redis.sadd(redis_key(:items, set), item)
+      # add the set to the item's set--inverting the sets
+      Predictor.redis.sadd(redis_key(:sets, item), set)
+    end
 
-  def add_single_nomulti(set, item)
-    Predictor.redis.sadd(parent_redis_key(:all_items), item)
-    Predictor.redis.sadd(redis_key(:items, set), item)
-    # add the set to the item's set--inverting the sets
-    Predictor.redis.sadd(redis_key(:sets, item), set)
   end
-end
+end

data/lib/predictor/version.rb

@@ -1,3 +1,3 @@
 module Predictor
-  VERSION = "2.0.0.rc1"
+  VERSION = "2.0.0"
 end

data/predictor.gemspec

@@ -13,7 +13,10 @@ Gem::Specification.new do |s|
 
   s.add_dependency "redis", ">= 3.0.0"
 
-  s.add_development_dependency "rspec", "~> 2.8.0"
+  s.add_development_dependency "rspec", "~> 2.14.0"
+  s.add_development_dependency "rake"
+  s.add_development_dependency "pry"
+  s.add_development_dependency "yard"
 
   s.files         = `git ls-files`.split("\n") - [".gitignore", ".rspec", ".travis.yml"]
   s.test_files    = `git ls-files -- spec/*`.split("\n")

data/spec/base_spec.rb

@@ -1,4 +1,4 @@
-require ::File.expand_path('../spec_helper', __FILE__)
+require 'spec_helper'
 
 describe Predictor::Base do
   class BaseRecommender

data/spec/input_matrix_spec.rb

@@ -1,9 +1,13 @@
-require ::File.expand_path('../spec_helper', __FILE__)
+require 'spec_helper'
 
 describe Predictor::InputMatrix do
+  let(:options) { @default_options.merge(@options) }
+
+  before(:each) { @options = {} }
 
   before(:all) do
-    @matrix = Predictor::InputMatrix.new(:redis_prefix => "predictor-test", :key => "mymatrix")
+    @default_options = { redis_prefix: "predictor-test", key: "mymatrix" }
+    @matrix = Predictor::InputMatrix.new(@default_options)
   end
 
   before(:each) do
@@ -92,15 +96,54 @@ describe Predictor::InputMatrix do
     end
   end
 
-  it "should calculate the correct jaccard index" do
-    @matrix.add_to_set "item1", "foo", "bar", "fnord", "blubb"
-    @matrix.add_to_set "item2", "bar", "fnord", "shmoo", "snafu"
-    @matrix.add_to_set "item3", "bar", "nada", "snafu"
+  describe "#score" do
+    let(:matrix) { Predictor::InputMatrix.new(options) }
+
+    context "default" do
+      it "scores as jaccard index by default" do
+        matrix.add_to_set "item1", "foo", "bar", "fnord", "blubb"
+        matrix.add_to_set "item2", "bar", "fnord", "shmoo", "snafu"
+        matrix.add_to_set "item3", "bar", "nada", "snafu"
+
+        matrix.score("bar", "snafu").should == 2.0/3.0
+      end
+
+      it "scores as jaccard index when given option" do
+        matrix = Predictor::InputMatrix.new(options.merge(measure: :jaccard_index))
+        matrix.add_to_set "item1", "foo", "bar", "fnord", "blubb"
+        matrix.add_to_set "item2", "bar", "fnord", "shmoo", "snafu"
+        matrix.add_to_set "item3", "bar", "nada", "snafu"
 
-    @matrix.calculate_jaccard("bar", "snafu").should == 2.0/3.0
+        matrix.score("bar", "snafu").should == 2.0/3.0
+      end
+
+      it "should handle missing sets" do
+        matrix.add_to_set "item1", "foo", "bar", "fnord", "blubb"
+
+        matrix.score("is", "missing").should == 0.0
+      end
+    end
+
+    context "sorensen_coefficient" do
+      before { @options[:measure] = :sorensen_coefficient }
+
+      it "should calculate the correct sorensen index" do
+        matrix.add_to_set "item1", "foo", "bar", "fnord", "blubb"
+        matrix.add_to_set "item2", "fnord", "shmoo", "snafu"
+        matrix.add_to_set "item3", "bar", "nada", "snafu"
+
+        matrix.score("bar", "snafu").should == 2.0/4.0
+      end
+
+      it "should handle missing sets" do
+        matrix.add_to_set "item1", "foo", "bar", "fnord", "blubb"
+
+        matrix.score("is", "missing").should == 0.0
+      end
+    end
   end
 
-private
+  private
 
   def add_two_item_test_data!(matrix)
     matrix.add_to_set("user42", "fnord", "blubb")
@@ -118,4 +161,4 @@ private
     matrix.add_to_set("user50", "fnord", "shmoo")
   end
 
-end
+end

data/spec/predictor_spec.rb

@@ -1,4 +1,4 @@
-require ::File.expand_path('../spec_helper', __FILE__)
+require 'spec_helper'
 
 describe Predictor do
 

data/spec/spec_helper.rb

@@ -1,9 +1,6 @@
-require "rspec"
-require "redis"
+require "predictor"
 require "pry"
 
-require ::File.expand_path('../../lib/predictor', __FILE__)
-
 def flush_redis!
   Predictor.redis = Redis.new
   Predictor.redis.keys("predictor-test*").each do |k|
@@ -19,22 +16,20 @@ module Predictor::Base
 
 end
 
-
 class TestRecommender
   include Predictor::Base
 
   input_matrix :jaccard_one
-
 end
 
 class Predictor::TestInputMatrix
 
   def initialize(opts)
-  	@opts = opts
+    @opts = opts
   end
 
   def method_missing(method, *args)
-  	@opts[method]
+    @opts[method]
   end
 
-end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: predictor
 version: !ruby/object:Gem::Version
-  version: 2.0.0.rc1
+  version: 2.0.0
 platform: ruby
 authors:
 - Pathgather
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-03-08 00:00:00.000000000 Z
+date: 2014-04-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: redis
@@ -30,14 +30,56 @@ dependencies:
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 2.8.0
+        version: 2.14.0
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 2.8.0
+        version: 2.14.0
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: pry
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: yard
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Fast and efficient recommendations and predictions using Redis
 email:
 - tech@pathgather.com
@@ -53,6 +95,7 @@ files:
 - docs/READMEv1.md
 - lib/predictor.rb
 - lib/predictor/base.rb
+- lib/predictor/distance.rb
 - lib/predictor/input_matrix.rb
 - lib/predictor/predictor.rb
 - lib/predictor/version.rb
@@ -76,9 +119,9 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>'
+  - - '>='
     - !ruby/object:Gem::Version
-      version: 1.3.1
+      version: '0'
 requirements: []
 rubyforge_project: 
 rubygems_version: 2.1.11