twins 0.0.2 → 0.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 20e9e6421c19c6abccbb7f121b1ba3a36ae56a7e
-  data.tar.gz: 660fdd54b8e6acb0d2a947c329e0293d147919af
+  metadata.gz: 330a31cfe1d6f44addbd04f07ef8ce599c1e7ab5
+  data.tar.gz: c34f97c80fc1c360fbbb4c9a6ba4e0b3b34ef180
 SHA512:
-  metadata.gz: e2dababf31cb5d844ae4ff9588758e27538fad29fd086354a6f43e66640647f0c8339c838a17c34515339bd6191286ea6618e4fd042a82bee70140980c905710
-  data.tar.gz: fd4442cd82dbc5d41df4d605d724c03c3070e8e363dd4b0b97b54454e298b059289de248e1c6284b6109efdb1812f580f2977af6bb48a00dbb1da2c08b05aefc
+  metadata.gz: 7a3eec8cb57b212ca23e94b40a26fd077f0552ce0918691dd51994e45b0bf0bcf5d154c0a28c38a8815959db19e8388cd53625942fa488d6aa08960d69303422
+  data.tar.gz: f6c9119803bdb1661d57433088b3f6454d14da1720428cf94752573efd61ec8da020e0b2d1d7efd362bd127394af3f3de076a891c244c7e4b88becc6bf35c1cd

data/lib/twins.rb

@@ -3,34 +3,32 @@ require 'twins/utilities'
 
 module Twins
 
-  # @param collection [Enumerable] A collection of Hash objects
+  # Consolidates keys with mode or lowest distance
+  #
+  # @param collection [Enumerable] A collection of Hash or Hash-like objects
   # @param options [Hash]
-  # @return [Hash, Nil]
+  # @return [HashWithIndifferentAccess, Nil]
   def consolidate(collection, options = {})
     return nil unless collection.any?
+    ensure_collection_uniformity!(collection)
 
-    if collection.all? { |e| e.is_a?(Hash) }
-      # noop
-    elsif collection.all? { |e| e.is_a?(collection.first.class) }
-      collection = collection.map do |element|
-        Hash[element.instance_variables.map { |name| [name.to_s.sub(/\A@/, ''), element.instance_variable_get(name)] }]
-      end
+    if collection.first.is_a?(Hash)
+      indiff_collection = collection
     else
-      raise ArgumentError, "The collection's elements must all be of the same Class"
+      indiff_collection = collection.map { |element| element.to_hash }
     end
 
     options = options.with_indifferent_access
     consolidated = Hash.new
 
-    collection.each do |hash|
-      hash.each_pair do |key, value|
-
+    indiff_collection.each do |element|
+      element.each_pair do |key, value|
         # Recursively consolidate nested hashes
         if value.is_a?(Hash) && !consolidated[key]
-          consolidated[key] = consolidate(collection.map { |element| element[key] })
+          consolidated[key] = consolidate(indiff_collection.map { |el| el[key] })
         else
           # Filter elements without a given key to avoid unintentionally nil values
-          values = collection.select { |element| element.has_key?(key) }.map { |element| element[key] }
+          values = indiff_collection.select { |el| el.has_key?(key) }.map { |el| el[key] }
 
           if options[:priority].try(:[], key)
             # Compute each element's distance from the given priority
@@ -49,4 +47,91 @@ module Twins
     consolidated.with_indifferent_access
   end
   module_function :consolidate
+
+  # Find element with the highest count of modes or the lowest overall distances
+  #
+  # @param collection [Enumerable] A collection of Hash or Hash-like objects
+  # @param options [Hash]
+  # @return [Object, Nil]
+  def pick(collection, options = {})
+    return nil unless collection.any?
+    ensure_collection_uniformity!(collection)
+
+    options = options.with_indifferent_access
+
+    if options[:priority]
+      pick_by_priority(collection, options[:priority])
+    else
+      pick_by_mode(collection)
+    end
+  end
+  module_function :pick
+
+  # Find the element with the highest count of modes
+  #
+  # @param collection [Enumerable]
+  # @return [Object, Nil]
+  def pick_by_mode(collection)
+    return nil unless collection.any?
+
+    if collection.first.is_a?(Hash)
+      indiff_collection = collection
+    else
+      indiff_collection = collection.map { |element| element.to_hash.with_indifferent_access }
+    end
+
+    collection.max_by do |element|
+      if collection.first.is_a?(Hash)
+        indiff_element = element
+      else
+        indiff_element = element.to_hash.with_indifferent_access
+      end
+
+      # Build a map of modes for each existing key
+      modes = indiff_element.map do |key, value|
+        # Filter elements without a given key to avoid unintentionally nil values
+        values = indiff_collection.select { |el| el.has_key?(key) }.map { |el| el[key] }
+        [key, Twins::Utilities.mode(values)]
+      end
+      modes = Hash[modes]
+
+      # Count the number of modes present in element
+      modes.select { |key, mode| indiff_element[key] == mode }.count
+    end
+  end
+  module_function :pick_by_mode
+
+  # Find the element with the lowest overall distances
+  #
+  # @param collection [Enumerable]
+  # @param options [Hash]
+  # @return [Object, Nil]
+  def pick_by_priority(collection, priorities)
+    return nil unless collection.any?
+    raise ArgumentError unless priorities.is_a?(Hash)
+
+    collection.min_by do |element|
+      if collection.first.is_a?(Hash)
+        indiff_element = element
+      else
+        indiff_element = element.to_hash.with_indifferent_access
+      end
+
+      priorities.map do |key, value|
+        Twins::Utilities.distance(value, indiff_element[key])
+      end.sum
+    end
+  end
+  module_function :pick_by_priority
+
+  # @private
+  def ensure_collection_uniformity!(collection)
+    if collection.none? { |e| e.is_a?(Hash) || e.is_a?(collection.first.class) }
+      raise ArgumentError, "The collection's elements must all be of the same Class"
+    elsif collection.none? { |e| e.respond_to?(:to_hash) }
+      raise ArgumentError, "The collection's elements must respond to '#to_hash'"
+    end
+  end
+  module_function :ensure_collection_uniformity!
+  private_class_method :ensure_collection_uniformity!
 end

data/lib/twins/version.rb

@@ -1,3 +1,3 @@
 module Twins
-  VERSION = "0.0.2"
+  VERSION = "0.0.3"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: twins
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
 platform: ruby
 authors:
 - Philippe Dionne
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-03-25 00:00:00.000000000 Z
+date: 2014-03-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport