db_clustering 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 7da4fe20fbc049e8449e67437319f746dc368099
-  data.tar.gz: ac84dd76c9e9cbe21cabb52e2425a90796717112
+  metadata.gz: ddf1c0c19cc0a107f1727d260d1101a2b4bf49f6
+  data.tar.gz: f6ab2b977b9759aaa69d215240c12df3f1fc9426
 SHA512:
-  metadata.gz: 2fb3221634bbc0f02558d45f376f383c34e94a7fccd936d87fc7388e4bed3e868f26da552b1a7f11f57feb7fc4bb72949ac62f5be26a3d1bc5d77cad28afb716
-  data.tar.gz: 1b81cd46f2d49efaea7081df28beaf361c2a4d1106d5f184c83fb85cd208a4e8df61037f141aa7590d28d7690347ee27cdc4b0bc0fb53fb73b390227841fb3b6
+  metadata.gz: 706097cbe232cef96549c0e078f72b60b224d194ad5249ff5a95c6dcb44a6213fb0a5df8742f2e8e65d58aee2385ddd7412a2ef2188e1f1715dfb9c5167d2a92
+  data.tar.gz: 6eef6d390e20f4e9d762060dca6d4f87194ef0cc5f3c4b59180d06558af9f00994a7d622930a82273cef258b6be500412cd760156a6036127ba8f67a0a30b749

data/.rspec

@@ -1 +1,2 @@
+--format Fuubar
 --color

data/.travis.yml

@@ -0,0 +1,5 @@
+language: ruby
+rvm:
+  - 2.1.6
+  - 2.2.2
+  - ruby-head

data/Gemfile

@@ -1,14 +1,18 @@
-source "http://rubygems.org"
-# Add dependencies required to use your gem here.
-# Example:
-#   gem "activesupport", ">= 2.3.5"
+source 'http://rubygems.org'
 
-# Add dependencies to develop your gem here.
-# Include everything needed to run rake, tests, features, etc.
+# Dependencies required to use the gem
+gem 'activerecord', '~> 4.2', '>= 4.2.1'
+
+# Dependencies to develop the gem
 group :development do
-  gem "rspec", "~> 2.8.0"
-  gem "rdoc", "~> 3.12"
-  gem "bundler", "~> 1.0"
-  gem "jeweler", "~> 2.0.1"
-  gem "simplecov", ">= 0"
+  gem 'rspec', '~> 3.0', '>= 3.2.0'
+  gem 'rdoc', '~> 4.0', '>= 4.2.0'
+  gem 'bundler', '~> 1.0', '>= 1.10.2'
+  gem 'jeweler', '~> 2.0', '>= 2.0.1'
+  gem 'simplecov', '>= 0'
+  gem 'guard-rspec', '~> 4.0', '>= 4.5.1'
+  gem 'byebug', '~> 5.0'
+  gem 'awesome_print', '~> 1.6', '>= 1.6.1', require: 'awesome_print'
+  gem 'fuubar', '~> 2.0', '>= 2.0.0.rc1'
+  gem 'simple-random', '~> 1.0', '>= 1.0.0'
 end

data/Gemfile.lock

@@ -0,0 +1,149 @@
+GEM
+  remote: http://rubygems.org/
+  specs:
+    activemodel (4.2.3)
+      activesupport (= 4.2.3)
+      builder (~> 3.1)
+    activerecord (4.2.3)
+      activemodel (= 4.2.3)
+      activesupport (= 4.2.3)
+      arel (~> 6.0)
+    activesupport (4.2.3)
+      i18n (~> 0.7)
+      json (~> 1.7, >= 1.7.7)
+      minitest (~> 5.1)
+      thread_safe (~> 0.3, >= 0.3.4)
+      tzinfo (~> 1.1)
+    addressable (2.3.8)
+    arel (6.0.1)
+    awesome_print (1.6.1)
+    builder (3.2.2)
+    byebug (5.0.0)
+      columnize (= 0.9.0)
+    coderay (1.1.0)
+    columnize (0.9.0)
+    descendants_tracker (0.0.4)
+      thread_safe (~> 0.3, >= 0.3.1)
+    diff-lcs (1.2.5)
+    docile (1.1.5)
+    faraday (0.9.1)
+      multipart-post (>= 1.2, < 3)
+    ffi (1.9.10)
+    formatador (0.2.5)
+    fuubar (2.0.0)
+      rspec (~> 3.0)
+      ruby-progressbar (~> 1.4)
+    git (1.2.9.1)
+    github_api (0.12.3)
+      addressable (~> 2.3)
+      descendants_tracker (~> 0.0.4)
+      faraday (~> 0.8, < 0.10)
+      hashie (>= 3.3)
+      multi_json (>= 1.7.5, < 2.0)
+      nokogiri (~> 1.6.3)
+      oauth2
+    guard (2.12.8)
+      formatador (>= 0.2.4)
+      listen (>= 2.7, <= 4.0)
+      lumberjack (~> 1.0)
+      nenv (~> 0.1)
+      notiffany (~> 0.0)
+      pry (>= 0.9.12)
+      shellany (~> 0.0)
+      thor (>= 0.18.1)
+    guard-compat (1.2.1)
+    guard-rspec (4.6.1)
+      guard (~> 2.1)
+      guard-compat (~> 1.1)
+      rspec (>= 2.99.0, < 4.0)
+    hashie (3.4.2)
+    highline (1.7.2)
+    i18n (0.7.0)
+    jeweler (2.0.1)
+      builder
+      bundler (>= 1.0)
+      git (>= 1.2.5)
+      github_api
+      highline (>= 1.6.15)
+      nokogiri (>= 1.5.10)
+      rake
+      rdoc
+    json (1.8.3)
+    jwt (1.5.1)
+    listen (3.0.2)
+      rb-fsevent (>= 0.9.3)
+      rb-inotify (>= 0.9)
+    lumberjack (1.0.9)
+    method_source (0.8.2)
+    mini_portile (0.6.2)
+    minitest (5.7.0)
+    multi_json (1.11.2)
+    multi_xml (0.5.5)
+    multipart-post (2.0.0)
+    nenv (0.2.0)
+    nokogiri (1.6.6.2)
+      mini_portile (~> 0.6.0)
+    notiffany (0.0.6)
+      nenv (~> 0.1)
+      shellany (~> 0.0)
+    oauth2 (1.0.0)
+      faraday (>= 0.8, < 0.10)
+      jwt (~> 1.0)
+      multi_json (~> 1.3)
+      multi_xml (~> 0.5)
+      rack (~> 1.2)
+    pry (0.10.1)
+      coderay (~> 1.1.0)
+      method_source (~> 0.8.1)
+      slop (~> 3.4)
+    rack (1.6.4)
+    rake (10.4.2)
+    rb-fsevent (0.9.5)
+    rb-inotify (0.9.5)
+      ffi (>= 0.5.0)
+    rdoc (4.2.0)
+    rspec (3.3.0)
+      rspec-core (~> 3.3.0)
+      rspec-expectations (~> 3.3.0)
+      rspec-mocks (~> 3.3.0)
+    rspec-core (3.3.1)
+      rspec-support (~> 3.3.0)
+    rspec-expectations (3.3.0)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.3.0)
+    rspec-mocks (3.3.1)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.3.0)
+    rspec-support (3.3.0)
+    ruby-progressbar (1.7.5)
+    shellany (0.0.1)
+    simple-random (1.0.0)
+    simplecov (0.10.0)
+      docile (~> 1.1.0)
+      json (~> 1.8)
+      simplecov-html (~> 0.10.0)
+    simplecov-html (0.10.0)
+    slop (3.6.0)
+    thor (0.19.1)
+    thread_safe (0.3.5)
+    tzinfo (1.2.2)
+      thread_safe (~> 0.1)
+
+PLATFORMS
+  ruby
+
+DEPENDENCIES
+  activerecord (~> 4.2, >= 4.2.1)
+  awesome_print (~> 1.6, >= 1.6.1)
+  bundler (~> 1.0, >= 1.10.2)
+  byebug (~> 5.0)
+  fuubar (~> 2.0, >= 2.0.0.rc1)
+  guard-rspec (~> 4.0, >= 4.5.1)
+  jeweler (~> 2.0, >= 2.0.1)
+  rdoc (~> 4.0, >= 4.2.0)
+  rspec (~> 3.0, >= 3.2.0)
+  simple-random (~> 1.0, >= 1.0.0)
+  simplecov
+
+BUNDLED WITH
+   1.10.5

data/LICENSE.txt

@@ -1,4 +1,4 @@
-Copyright (c) 2015 Cihat Gündüz
+Copyright (c) 2015 Cihat Gündüz
 
 Permission is hereby granted, free of charge, to any person obtaining
 a copy of this software and associated documentation files (the

data/README.md

@@ -0,0 +1,90 @@
+# DBClustering [![Build Status](https://travis-ci.org/Dschee/db_clustering.svg?branch=develop)](https://travis-ci.org/Dschee/db_clustering)
+
+Please note that this gem is still in its very early stages and should not considered stable.
+Also it currently only supports the in-memory datasource adapter. In future versions an ActiveRecord adapter is planned but this is not yet implemented. Stay tuned.
+
+## Requirements
+
+Ruby 2.1+ is required, earlier Rubies may work but are not officially supported.
+
+## Getting Started
+
+This gem was developed to work best in Ruby on Rails projects.
+
+1. Add this gem to your Gemfile
+
+        gem 'db_clustering'
+
+2. Rund `bundle install` in your terminal
+
+3. Implement the `clustering_vector` method in your model class and return either:
+   - an **array** with numeric values for similarity comparison
+   - a **hash** with numeric values for similarity comparison between keys existing in both hashes
+
+   See `TestModel` class within the `spec/support` directory for a very simple example.
+
+4. Decide for a *distance metric* and initialize it, e.g.:
+
+   ``` ruby
+   average_difference = DbClustering::DistanceMetrics::AverageDifference.new
+
+   # Instead you can also use one of the following:
+   cosine_similarity = DbClustering::DistanceMetrics::CosineSimilarity.new
+   euclidean_distance = DbClustering::DistanceMetrics::EuclideanDistance.new
+   pearson_correlation = DbClustering::DistanceMetrics::PearsonCorrelation.new
+   ```
+
+5. Decide for a datasource adapter (currently only in-memory datasource available), e.g.:
+
+   ``` ruby
+    in_memory_datasource = DbClustering::DatasourceAdapters::InMemory.new(array: your_array)
+   ```
+
+   Please note that `your_array` should be an array filled with objects of the class type that implements the `clustering_vector` method from step 3.
+
+   An **ActiveRecord datasource** type is planned but not yet implemented. Please stay tuned.
+
+6. Decide for an **algorithm** and initialize it:
+
+   ``` ruby
+   dbscan = DbClustering::Algorithms::Dbscan.new(datasource: in_memory_datasource, distance_metric: average_difference)
+   ```
+   Please note that currently **only one algorithm is available**. More algorithms aren't currently planned but may be added if needed. Contributions are welcome, of course.
+
+7. Decide for the **algorithm parameters** and start the process of clustering your data:
+
+   ``` ruby
+   dbscan.cluster(max_distance: 10, min_neighbors: 5)
+   ```
+   The `max_distance` is the epsilon parameter and the `min_neighbors` the minPts parameter from the usual DBSCAN algorithm documentation (e.g. Wikipedia). You might want to try different values here first before you decide for the right values for your purpose.
+
+   Plase also take note that the `max_distance` value is **highly dependent on the type of metric** you decided to go for. For the `AverageDifference` and `EuclideanDistance` metrics it can be an **open-ended positive value**. For the `CosineSimilarity` and `PearsonCorrelation` types it needs to be a value between 0 and 2 where a value of `0` means "100% positive correlation/similarity", a value of `1` means "no correlation/similarity at all" and a value of `2` means "100% negative correlation/similarity". You can use any decimal value in between (e.g. 0.25) as a partly positive/negative correlation.
+
+8. Wait for the calculations to finish and use the results the way you want:
+
+   ``` ruby
+   clusters = dbscan.clusters # the resulting Clusters, each cluster contains Points
+   first_cluster = clusters.first
+   point = first_cluster.points.first
+   # a point knows its cluster, and its position in there
+   point.cluster # will return the same object as `first_cluster`
+   point.is_edge_point? # boolean specifying if it's an edge point of its cluster
+   point.is_core_point? # boolean specifying if it's a core point of its cluster
+   point.is_noise_point? # boolean specifiying if it's a noise point without a cluster
+
+   # a point also contains the source object specifying the `clustering_vector` method
+   your_model = point.datasource_point
+   ```
+
+   For more please don't hesitate to have a look into the underlying models under the `lib/models` directory as well as the corresponding specs.
+
+ That's it, it **looks more complicated than it actually** is, just try it out! You can find complete usage examples within the `spec/algorithms/density_based/dbscan_spec.rb` file.
+
+## Contributing
+
+Contributions are welcome. Please fork this project, make your changes and file a pull request. Please also make sure to write tests to ensure your changes persist over time.
+
+
+## License
+
+This gem is released under the [MIT License](http://www.opensource.org/licenses/MIT).

data/Rakefile

@@ -14,13 +14,13 @@ require 'rake'
 require 'jeweler'
 Jeweler::Tasks.new do |gem|
   # gem is a Gem::Specification... see http://guides.rubygems.org/specification-reference/ for more options
-  gem.name = "clustering"
-  gem.homepage = "http://github.com/Dschee/clustering"
+  gem.name = "db_clustering"
+  gem.homepage = "http://github.com/Dschee/db_clustering"
   gem.license = "MIT"
-  gem.summary = %Q{TODO: one-line summary of your gem}
-  gem.description = %Q{TODO: longer description of your gem}
+  gem.summary = %Q{Big-Data clustering algorithms in Ruby (on Rails).}
+  gem.description = %Q{Big-Data clustering algorithms in Ruby (on Rails). Currently only supports DBSCAN algorithm and ActiveRecord database.}
   gem.email = "CihatGuenduez@posteo.de"
-  gem.authors = ["Cihat Gündüz"]
+  gem.authors = ["Cihat Gündüz"]
   # dependencies defined in Gemfile
 end
 Jeweler::RubygemsDotOrgTasks.new
@@ -44,7 +44,7 @@ Rake::RDocTask.new do |rdoc|
   version = File.exist?('VERSION') ? File.read('VERSION') : ""
 
   rdoc.rdoc_dir = 'rdoc'
-  rdoc.title = "clustering #{version}"
+  rdoc.title = "db_clustering #{version}"
   rdoc.rdoc_files.include('README*')
   rdoc.rdoc_files.include('lib/**/*.rb')
 end

data/VERSION

@@ -0,0 +1 @@
+0.1.1

data/lib/algorithms/density_based/dbscan.rb

@@ -0,0 +1,48 @@
+# For more information see https://en.wikipedia.org/wiki/DBSCAN
+
+module DbClustering
+  module Algorithms
+    class Dbscan
+
+      attr_accessor :datasource, :clusters
+
+      def initialize(datasource:, distance_metric:)
+        @datasource = datasource
+        @distance_metric = distance_metric
+        @clusters = []
+      end
+
+      def cluster(max_distance:, min_neighbors:)
+        @clusters = []
+        cluster = nil
+
+        @datasource.iterate_all_points do |point|
+          neighbors = @datasource.neighbors(point: point, distance_metric: @distance_metric, max_distance: max_distance)
+
+          if neighbors.count < min_neighbors
+            point.is_noise = true
+          else
+            if point.cluster.nil?
+              cluster = DbClustering::Models::Cluster.new
+              @clusters << cluster
+            else
+              cluster = point.cluster
+            end
+            expand_cluster(cluster: cluster, neighbors: neighbors)
+          end
+
+        end
+      end
+
+      def expand_cluster(cluster:, neighbors:)
+        # Important: If neighbors do not include point itself, then point must be added to cluster, too.
+        neighbors.each do |neighbor|
+          if !neighbor.visited?
+            cluster.add(neighbor)
+          end
+        end
+      end
+
+    end
+  end
+end

data/lib/datasource_adapters/active_record.rb

@@ -0,0 +1,32 @@
+module DbClustering
+  module DatasourceAdapters
+    class ActiveRecord
+
+      def initialize(relation:)
+        @relation = relation
+      end
+
+      def iterate_all_points
+        @relation.find_each do |datasource_point|
+          point = DbClustering::Models::Point.new(datasource_point)
+          yield(point)
+        end
+      end
+
+      def neighbors(point:, distance_metric:, max_distance:)
+        neighbors = []
+
+        @relation.find_each do |neighbor_candidate|
+          candidate_point = DbClustering::Models::Point.new(neighbor_candidate)
+
+          if distance_metric.distance(point.vector, candidate_point.vector) <= max_distance
+            neighbors << candidate_point
+          end
+        end
+
+        neighbors
+      end
+
+    end
+  end
+end

data/lib/datasource_adapters/in_memory.rb

@@ -0,0 +1,29 @@
+module DbClustering
+  module DatasourceAdapters
+    class InMemory
+
+      def initialize(array:)
+        @array = array.map{ |datasource_point| DbClustering::Models::Point.new(datasource_point) }
+      end
+
+      def iterate_all_points
+        @array.each do |point|
+          yield(point)
+        end
+      end
+
+      def neighbors(point:, distance_metric:, max_distance:)
+        neighbors = []
+
+        @array.each do |neighbor_candidate|
+          if distance_metric.distance(point.vector, neighbor_candidate.vector) <= max_distance
+            neighbors << neighbor_candidate
+          end
+        end
+
+        neighbors
+      end
+
+    end
+  end
+end