affinity_propagation 0.1.0 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: 26f8c5ca0ecef4d7c7e89341a5039eadefef5871
-  data.tar.gz: cd466acdd9604ca6e547558438203b88283d21c9
+SHA256:
+  metadata.gz: 43c3b9be3a0ad3d3ac9a4c428194f66af06d9add87d9390fc70267d92631f04d
+  data.tar.gz: 91c8248f1dce46cd08a220114d5f74d01e2dad90e4ac4e5469c326f0cf60f0bb
 SHA512:
-  metadata.gz: 72c71cf377b691208ffeb714022216bb3d4b259ec09157feded788f49858d8e5260538ab73dec78e1a126b9914c9eb8c251a0c44818235ba35c366b81f24cd28
-  data.tar.gz: 4fdcd1c0742094f837c95ebcd4366045006deb97dd358fe55a6575244ad07eb9207a3b5331cf9f8035fb2ec9dca2ad0b46d68188eb23d79b0a87646e79542ab5
+  metadata.gz: 598aecb0dc0f1593dbe032527cb22e72903b75a1e5c9721a99145ce490a175e54ff6817db729328652aefbf6b93f1cd460fa6358a2cda167c5f0bdebfc607a89
+  data.tar.gz: fed4c922d48fd640889b98c1434cf39006b43628429406bd369a856e0b9c75df086b6b71d84f802753e7ed879fb451e0d8334ffde81914f65fb64a06f057cc9e

data/.gitignore

@@ -42,7 +42,7 @@ build-iPhoneSimulator/
 
 # for a library or gem, you might want to ignore these files since the code is
 # intended to run in multiple environments; otherwise, check them in:
-# Gemfile.lock
+Gemfile.lock
 # .ruby-version
 # .ruby-gemset
 
@@ -52,3 +52,4 @@ build-iPhoneSimulator/
 # IDEs
 .idea
 
+.byebug_history

data/affinity_propagation.gemspec

@@ -28,5 +28,6 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "bundler", "~> 1.15"
   spec.add_development_dependency "rake", "~> 10.0"
   spec.add_development_dependency "rspec", "~> 3.0"
-  spec.add_development_dependency "activesupport"
+
+  spec.add_runtime_dependency "concurrent-ruby", "~> 1.1.5"
 end

data/lib/affinity_propagation/calculator.rb

@@ -1,27 +1,8 @@
 require 'matrix'
+require 'concurrent'
+require 'thread'
 
 module AffinityPropagation
-  refine Array do
-    def median
-      relevant_elements = if size % 2 == 0
-        # Even number of items in this list => let's get the middle two and return their mean
-        slice(size / 2, 2)
-      else
-        slice(size / 2, 1)
-      end
-
-      relevant_elements.sum / relevant_elements.size
-    end
-  end
-
-  refine Matrix do
-    def to_matlab
-      output = row_vectors.map(&:to_a).map { |row| row.join(', ') }.join('; ')
-
-      "[#{output}]"
-    end
-  end
-
   class Calculator
     using AffinityPropagation
 
@@ -33,7 +14,7 @@ module AffinityPropagation
       @data = data
       @lambda = lambda
 
-      raise 'no block provide to calculate similarities within data!' unless block_given?
+      raise 'no block provided to calculate similarities within data!' unless block_given?
 
       @similarities = similarity_matrix(&block)
       reset
@@ -70,25 +51,47 @@ module AffinityPropagation
     end
 
     def run(iterations: 100, stable_iterations: 10)
-      iterate while @total_iterations < iterations && @stable_cluster_iterations < stable_iterations
+      while @total_iterations < iterations && @stable_cluster_iterations < stable_iterations
+        iterate
+
+        yield(@total_iterations, @stable_cluster_iterations) if block_given?
+      end
     end
 
     private
 
+    def median(array)
+      relevant_elements = if array.size % 2 == 0
+          # Even number of items in this list => let's get the middle two and return their mean
+          array.sort.slice(array.size / 2, 2)
+        else
+          array.sort.slice(array.size / 2, 1)
+        end
+
+      relevant_elements.sum / relevant_elements.size
+    end
+
     def similarity_matrix(&block)
       similarity_array = []
-      similarities = Matrix.build(@data.size, @data.size) do |row_idx, col_idx|
+      similarities_future = Matrix.build(@data.size, @data.size) do |row_idx, col_idx|
         exemplar = @data[row_idx]
         datum = @data[col_idx]
 
-        similarity = block.call(datum, exemplar)
-        similarity_array << similarity
-        similarity
+        similarity_future = Concurrent::Future.execute(executor: :fast) { block.call(datum, exemplar)}
+        similarity_array << similarity_future
+
+        similarity_future
       end
 
-      median_similarity = similarity_array.median
+      while similarity_array.any?(&:pending?)
+        sleep 0.1
+      end
+
+      similarity_array.map!(&:value)
+      similarities = similarities_future.map(&:value)
+
+      median_similarity = median(similarity_array)
 
-      # Matrices can't be modified once created so we have to use this hack
       (0...@data.size).each { |idx| similarities.send(:[]=, idx, idx, median_similarity) }
 
       similarities
@@ -99,49 +102,82 @@ module AffinityPropagation
     end
 
     def responsibility_matrix
-      Matrix.build(@similarities.row_count, @similarities.column_count) do |row_idx, col_idx|
+      responsibility_futures = []
+
+      responsibilities_future = Matrix.build(@similarities.row_count, @similarities.column_count) do |row_idx, col_idx|
         exemplar_idx = row_idx
         datum_idx = col_idx
 
+        current_similarity = @similarities[row_idx, col_idx]
+        current_responsibility = @responsibilities[exemplar_idx, datum_idx]
+
         availability_column = @availabilities.column(col_idx).to_a
-        availability_column.slice!(exemplar_idx)
         similarity_column = @similarities.column(col_idx).to_a
-        similarity_column.slice!(exemplar_idx)
 
-        availability_plus_similarity = []
-        availability_column.zip(similarity_column) { |data| availability_plus_similarity << data.sum }
+        responsibility_future = Concurrent::Future.execute(executor: :fast) do
+          availability_column.slice!(exemplar_idx)
+          similarity_column.slice!(exemplar_idx)
+
+          availability_plus_similarity = []
+          availability_column.zip(similarity_column) { |data| availability_plus_similarity << data.sum }
+
+          dampen(current_similarity - availability_plus_similarity.max, current_responsibility)
+        end
+
+        responsibility_futures << responsibility_future
+        responsibility_future
+      end
 
-        dampen(@similarities[row_idx, col_idx] - availability_plus_similarity.max, @responsibilities[exemplar_idx, datum_idx])
+      while responsibility_futures.any?(&:pending?)
+        sleep 0.1
       end
+
+      responsibilities_future.map(&:value)
     end
 
     def availability_matrix
-      Matrix.build(@responsibilities.row_count, @responsibilities.column_count) do |row_idx, col_idx|
+      availability_futures = []
+
+      availabilities_future = Matrix.build(@responsibilities.row_count, @responsibilities.column_count) do |row_idx, col_idx|
         exemplar_idx = row_idx
         datum_idx = col_idx
         responsibility_column = @responsibilities.row(exemplar_idx).to_a
 
-        if exemplar_idx == datum_idx
-          # self-availability
-          responsibility_column.slice!(exemplar_idx)
+        current_availability = @availabilities[exemplar_idx, datum_idx]
+        current_responsibility = @responsibilities[exemplar_idx, exemplar_idx]
 
-          dampen(responsibility_column.inject(0) { |sum, item| sum += [0, item].max },@availabilities[exemplar_idx, datum_idx])
-        else
-          self_responsibility = @responsibilities[exemplar_idx, exemplar_idx]
-          if datum_idx > exemplar_idx
-            # Slice out the datum index first since in this case it won't affect the exemplar index
-            responsibility_column.slice!(datum_idx)
+        availability_future = Concurrent::Future.execute(executor: :fast) do
+          if exemplar_idx == datum_idx
+            # self-availability
             responsibility_column.slice!(exemplar_idx)
+
+            dampen(responsibility_column.inject(0) { |sum, item| sum += [0, item].max }, current_availability)
           else
-            responsibility_column.slice!(exemplar_idx)
-            responsibility_column.slice!(datum_idx)
+            self_responsibility = current_responsibility
+            if datum_idx > exemplar_idx
+              # Slice out the datum index first since in this case it won't affect the exemplar index
+              responsibility_column.slice!(datum_idx)
+              responsibility_column.slice!(exemplar_idx)
+            else
+              responsibility_column.slice!(exemplar_idx)
+              responsibility_column.slice!(datum_idx)
+            end
+
+            responsibility_column_sum = responsibility_column.inject(0) { |sum, item| sum += [0, item].max }
+
+            dampen([0, self_responsibility + responsibility_column_sum].min, current_availability)
           end
+        end
 
-          responsibility_column_sum = responsibility_column.inject(0) { |sum, item| sum += [0, item].max }
+        availability_futures << availability_future
+        availability_future
+      end
 
-          dampen([0, self_responsibility + responsibility_column_sum].min, @availabilities[exemplar_idx, datum_idx])
-        end
+      while availability_futures.any?(&:pending?)
+        sleep 0.1
       end
+
+      availabilities_future.map(&:value)
     end
 
     def identify_raw_clusters

data/lib/affinity_propagation/version.rb

@@ -1,3 +1,3 @@
 module AffinityPropagation
-  VERSION = "0.1.0"
+  VERSION = "0.2.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: affinity_propagation
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors:
 - Shahbaz Javeed
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2017-09-10 00:00:00.000000000 Z
+date: 2019-08-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -53,19 +53,19 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '3.0'
 - !ruby/object:Gem::Dependency
-  name: activesupport
+  name: concurrent-ruby
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
-  type: :development
+        version: 1.1.5
+  type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 1.1.5
 description: "\n    Affinity Propagation is a clustering algorithm that does not require
   pre-specifying the number of clusters like\n    k-means and other similar algorithms
   do.  This is a ruby implementation of the original version defined by Frey and\n
@@ -83,7 +83,6 @@ files:
 - ".ruby-version"
 - ".travis.yml"
 - Gemfile
-- Gemfile.lock
 - LICENSE.txt
 - README.md
 - Rakefile
@@ -112,8 +111,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.6.11
+rubygems_version: 3.0.6
 signing_key: 
 specification_version: 4
 summary: An implementation of the affinity propagation clustering algorithm by Frey

data/Gemfile.lock

@@ -1,47 +0,0 @@
-PATH
-  remote: .
-  specs:
-    affinity_propagation (0.1.0)
-
-GEM
-  remote: https://rubygems.org/
-  specs:
-    activesupport (5.1.4)
-      concurrent-ruby (~> 1.0, >= 1.0.2)
-      i18n (~> 0.7)
-      minitest (~> 5.1)
-      tzinfo (~> 1.1)
-    concurrent-ruby (1.0.5)
-    diff-lcs (1.3)
-    i18n (0.8.6)
-    minitest (5.10.3)
-    rake (10.5.0)
-    rspec (3.6.0)
-      rspec-core (~> 3.6.0)
-      rspec-expectations (~> 3.6.0)
-      rspec-mocks (~> 3.6.0)
-    rspec-core (3.6.0)
-      rspec-support (~> 3.6.0)
-    rspec-expectations (3.6.0)
-      diff-lcs (>= 1.2.0, < 2.0)
-      rspec-support (~> 3.6.0)
-    rspec-mocks (3.6.0)
-      diff-lcs (>= 1.2.0, < 2.0)
-      rspec-support (~> 3.6.0)
-    rspec-support (3.6.0)
-    thread_safe (0.3.6)
-    tzinfo (1.2.3)
-      thread_safe (~> 0.1)
-
-PLATFORMS
-  ruby
-
-DEPENDENCIES
-  activesupport
-  affinity_propagation!
-  bundler (~> 1.15)
-  rake (~> 10.0)
-  rspec (~> 3.0)
-
-BUNDLED WITH
-   1.15.4