caruby-core 1.4.1

data/lib/caruby/util/attribute_path.rb

@@ -0,0 +1,46 @@
+require 'caruby/util/validation'
+
+# An AttributePath encapsulates an array of attributes that can be evaluated on a source object.
+class AttributePath < Array
+  include Validation
+
+  # Creates an AttributePath from the path Array, String or Symbol. A path string is a period-delimited sequence
+  # of attributes, e.g. +person.name+.
+  def initialize(path)
+    raise ArgumentError.new("Path empty") if path.nil_or_empty?
+    # standardize the argument as a symbol array
+    attributes = case path
+    when Symbol then
+      [path]
+    when String then
+      path.split('.').map { |name| name.to_sym }
+    when Array then
+      path.map { |name| name.to_sym }
+    else
+      raise ArgumentError.new("Argument type unsupported - expected Symbol, String or Array; found #{path.class}")
+    end
+    # make the array
+    super(attributes)
+  end
+
+  # Returns the result of evaluating this evaluator's attribute path on the source object.
+  # If the evaluation results in a migratable object, then that object is migrated.
+  def evaluate(source)
+    # call the attribute path as far as possible
+    inject(source) do |current, attr|
+      return if current.nil?
+      evaluate_attribute(attr, current)
+    end
+  end
+
+  # Returns the result of evaluating attribute on the source object.
+  # If attr is +self+, then the source object is returned.
+  def evaluate_attribute(attr, source)
+    # call the attribute path as far as possible
+    attr == :self ? source : source.send(attr)
+  end
+
+  def to_s
+    join('.')
+  end
+end

data/lib/caruby/util/cache.rb

@@ -0,0 +1,53 @@
+require 'caruby/util/collection'
+
+module CaRuby
+  # Cache for objects held in memory and accessed by key.
+  class Cache
+    
+    # The classes which are not cleared when {#clear} is called without the +all+ flag.
+    attr_reader :sticky
+    
+    # Returns a new Cache whose value key is determined by calling the given
+    # extractor block on the cached value.
+    #
+    # If the value is not cached and there is a factory Proc, then the result of
+    # calling the factory on the missing value is cached with the value key.
+    #
+    # @param [Proc] optional factory Proc called with a missing value as argument
+    #   to create a cached object
+    def initialize(factory=nil, &extractor)
+      @factory = factory
+      # Make the class => { key => value } hash.
+      # The { key => value } hash takes a value as an argument and converts
+      # it to the key by calling the block given to this initializer.
+      @hash = LazyHash.new { KeyTransformerHash.new { |value| yield value } }
+      @sticky = Set.new
+    end
+
+    # Returns the object cached with the same class and key as the given value.
+    # If this Cache has a factory but does not have an entry for value, then the
+    # factory is called on the value to create a new entry. 
+    def [](value)
+      chash = @hash[value.class]
+      cached = chash[value] if chash
+      return cached unless cached.nil? and @factory
+      obj = @factory.call(value) || return
+      chash[value] = obj
+    end
+    
+    # Adds the given value to this cache.
+    def add(value)
+      @hash[value.class][value] = value
+    end
+
+    # Clears the cache key => object hashes. If all is true, then every class hash
+    # is cleared. Otherwise, only the non-sticky classes are cleared.
+    def clear(all=false)
+      if @sticky.empty? then
+        @hash.clear
+      else
+        @hash.each { |klass, chash| chash.clear unless @sticky.include?(klass) }
+      end
+    end
+  end
+end

data/lib/caruby/util/class.rb

@@ -0,0 +1,99 @@
+require 'enumerator'
+
+class Class
+  # Returns an Enumerable on this class and its ancestors.
+  def class_hierarchy
+    @hierarchy ||= Enumerable::Enumerator.new(self, :each_class_in_hierarchy)
+  end
+
+  # Returns this class's superclass, thereby enabling class ranges, e.g.
+  #   class A; end
+  #   class B < A; end
+  #   (B..Object).to_a #=> [B, A, Object]
+  alias :succ :superclass
+
+  private
+
+  # Creates an alias for each accessor method of the given attribute.
+  #
+  # @example
+  #   class Person
+  #    attr_reader :social_security_number
+  #    attr_accessor :postal_code
+  #    define_attribute_alias(:ssn, :social_security_number)
+  #    define_attribute_alias(:zip_code, :postal_code)
+  #   end
+  #   Person.method_defined?(:ssn) #=> true
+  #   Person.method_defined?(:ssn=) #=> false
+  #   Person.method_defined?(:zip_code) #=> true
+  #   Person.method_defined?(:zip_code=) #=> true
+  def define_attribute_alias(aliaz, attribute)
+    alias_method(aliaz, attribute) if method_defined?(attribute)
+    writer = "#{attribute}=".to_sym
+    alias_method("#{aliaz}=".to_sym, writer) if method_defined?(writer)
+  end
+
+  # Creates new accessor methods for each _method_ => _original_ hash entry.
+  # The new _method_ offsets the existing Number _original_ attribute value by the given
+  # offset (default -1).
+  #
+  # @example
+  #   class OneBased
+  #     attr_accessor :index
+  #     offset_attr_accessor :zero_based_index => :index
+  #   end
+  #@param [{Symbol => Symbol}] hash the offset => original method hash
+  #@param [Integer, nil] offset the offset amount (default is -1) 
+  def offset_attr_accessor(hash, offset=nil)
+    offset ||= -1
+    hash.each do |method, original|
+      define_method(method) { value = send(original); value + offset if value } if method_defined?(original)
+      original_writer = "#{original}=".to_sym
+      if method_defined?(original_writer) then
+        define_method("#{method}=".to_sym) do |value|
+          adjusted = value - offset if value
+          send(original_writer, adjusted)
+        end
+      end
+    end
+  end
+
+  def each_class_in_hierarchy
+    current = self
+    until current.nil?
+      yield current
+      current = current.superclass
+    end
+  end
+
+  # Redefines method using the given block. The block argument is a new alias for the old method.
+  # The block creates a proc which implements the new method body.
+  #
+  # @example
+  #   redefine_method(:ssn) { |old_method| lambda { send(old_method).delete('-').to_i } }
+  # @return [Symbol] an alias to the old method implementation
+  def redefine_method(method)
+    # make a new alias id method__base for the existing method.
+    # disambiguate with a counter suffix if necessary.
+    counter = 2
+    # make a valid alias base
+    old, eq = /^([^=]*)(=)?$/.match(method.to_s).captures
+    old.tr!('|', 'or')
+    old.tr!('&', 'and')
+    old.tr!('+', 'plus')
+    old.tr!('*', 'mult')
+    old.tr!('/', 'div')
+    old.gsub!(/[^\w]/, 'op')
+    base = "redefined__#{old}"
+    old_id = "#{base}#{eq}".to_sym
+    while method_defined?(old_id)
+      base = "#{base}#{counter}"
+      old_id = "#{base}#{eq}".to_sym
+      counter = counter + 1
+    end
+    alias_method(old_id, method)
+    body = yield old_id
+    define_method(method, body)
+    old_id
+  end
+end

data/lib/caruby/util/collection.rb

@@ -0,0 +1,1053 @@
+require 'set'
+require 'delegate'
+require 'enumerator'
+require 'generator'
+require 'caruby/util/class'
+require 'caruby/util/validation'
+require 'caruby/util/options'
+require 'caruby/util/pretty_print'
+
+class Object
+  # Returns whether this object is a collection capable of holding heterogenous items.
+  # An Object is a not a collection by default. Subclasses can override this method.
+  def collection?
+    false
+  end
+end
+
+module Enumerable
+  # Overrides {Object#collection?} to returns +true+, since an Enumerable is capable of
+  # holding heterogenous items by default. Subclasses can override this method.
+  def collection?
+    true
+  end
+end
+
+class String
+  # Overrides {Enumerable#collection?} to returns +false+, since a String is constrained
+  # to hold characters.
+  def collection?
+    false
+  end
+end
+
+module Enumerable
+  # Returns a new Hash generated from this Enumerable and an optional value generator block.
+  # This Enumerable contains the Hash keys. If the value generator block is given to this
+  # method then the block is called with each enumerated element as an argument to
+  # generate the associated hash value. If no block is given, then the values are nil.
+  #
+  # @example
+  #   [1, 2, 3].hashify { |item| item.modulo(2) } #=> { 1 => 1, 2 => 0, 3 => 1 }
+  #   [:a].hashify #=> { :a => nil }
+  # @return [Hash]
+  def hashify
+    hash = {}
+    each { |item| hash[item] = yield item if block_given? }
+    hash
+  end
+  
+  # Returns a new Hash generated from this Enumerable and a required value generator block.
+  # This Enumerable contains the Hash keys. The block is called with each enumerated
+  # element as an argument to generate the associated hash value.
+  # Only non-nil, non-empty values are included in the hash.
+  #
+  # @example
+  #   [1, 2, 3].to_compact_hash { |item| item.modulo(2) } #=> { 1 => 1, 2 => 0, 3 => 1 }
+  #   [1, 2, 3].to_compact_hash { |n| n.modulo(2) unless item > 2 } #=> {1 => 1, 2 => 0}
+  #   [1, 2, 3].to_compact_hash { |n| n > 2 } #=> {1 => false, 2 => false, 3 => true}
+  #   [1, 2, 3].to_compact_hash { |n| Array.new(n - 1, n) } #=> {2 => [2], 3 => [2, 3]}
+  # @return [Hash]
+  # @raise [ArgumentError] if the generator block is not given
+  # @see #hashify
+  def to_compact_hash
+    raise ArgumentError.new("Compact hash builder is missing the value generator block") unless block_given?
+    to_compact_hash_with_index { |item, index| yield item }
+  end
+
+  # Returns a new Hash generated from this Enumerable with a block whose arguments include the enumerated item
+  # and its index. nil or empty values are excluded.
+  def to_compact_hash_with_index
+    hash = {}
+    self.each_with_index do |item, index|
+      next if item.nil?
+      value = yield(item, index)
+      next if value.nil_or_empty?
+      hash[item] = value
+    end
+    hash
+  end
+
+  # Returns whether this Enumerable iterates over at least one item.
+  #
+  # This method is functionally equivalent to +to_a.empty+ but is more concise and efficient.
+  def empty?
+    not any? { true }
+  end
+
+  # Returns the first enumerated item in this Enumerable, or nil if this Enumerable is empty.
+  #
+  # This method is functionally equivalent to +to_a.first+ but is more concise and efficient.
+  def first
+    detect { true }
+  end
+
+  # Returns the count of items enumerated in this Enumerable.
+  #
+  # This method is functionally equivalent to +to_a.size+ but is more concise and efficient
+  # for an Enumerable which does not implement the {#size} method.
+  def size
+    inject(0) { |size, item| size + 1 }
+  end
+
+  alias :length :size
+
+  # Prints the content of this Enumerable as a series using {Array#to_series}.
+  def to_series(conjunction=nil)
+    to_a.to_series
+  end
+
+  # Returns the first non-nil, non-false enumerated value resulting from a call to the block given to this method,
+  # or nil if no value detected.
+  #
+  # @example
+  #   [1, 2].detect_value { |item| item / 2 if item % 2 == 0 } #=> 1
+  # @return [Object] the detected block result
+  # @see #detect_with_value
+  def detect_value
+    each do |*item|
+      value = yield(*item)
+      return value if value
+    end
+    nil
+  end
+
+  # Returns the first item and value for which an enumeration on the block given to this method returns
+  # a non-nil, non-false value.
+  #
+  # @example
+  #   [1, 2].detect_with_value { |item| item / 2 if item % 2 == 0 } #=> [2, 1]
+  # @return [(Object, Object)] the detected [item, value] pair
+  # @see #detect_value
+  def detect_with_value
+    value = nil
+    match = detect do |*item|
+      value = yield(*item)
+    end
+    [match, value]
+  end
+
+  # Returns a new Enumerable that iterates over the base Enumerable items for which filter evaluates to a non-nil,
+  #  non-false value, e.g.:
+  #   [1, 2, 3].filter { |n| n != 2 }.to_a #=> [1, 3]
+  #
+  # Unlike select, filter reflects changes to the base Enumerable, e.g.:
+  #   a = [1, 2, 3]
+  #   filter = a.filter { |n| n != 2 }
+  #   a << 4
+  #   filter.to_a #=> [1, 3, 4]
+  #
+  # In addition, filter has a small, fixed storage requirement, making it preferable to select for large collections.
+  # Note, however, that unlike select, filter does not return an Array.
+  # The default filter block returns the passed item.
+  #
+  # @return [Enumerable] the filtered result
+  # @example
+  #   [1, nil, 3].filter.to_a #=> [1, 3]
+  def filter(&filter) # :yields: item
+    Filter.new(self, &filter)
+  end
+
+  # @return an Enumerable which iterates over the non-nil items in this Enumerable
+  def compact
+    filter { |item| not item.nil? }
+  end
+
+  # Returns a new Flattener on this Enumerable, e.g.:
+  #   {:a => {:b => :c}, :d => [:e]}.enum_values.flatten.to_a #=> [:b, :c, :e]
+  #
+  # @return [Enumerable] the flattened result
+  def flatten
+    Flattener.new(self).to_a
+  end
+
+  # Returns an Enumerable which iterates over items in this Enumerable and the other Enumerable in sequence, e.g.:
+  #   [1, 2, 3] + [3, 4] #=> [1, 2, 3, 3, 4]
+  #
+  # Unlike Array#+, {#union} reflects changes to the underlying enumerators.
+  #
+  # @example
+  #   a = [1, 2]
+  #   b = [4, 5]
+  #   ab = a.union(b)
+  #   ab #=> [1, 2, 4, 5]
+  #   a << 3
+  #   ab #=> [1, 2, 3, 4, 5]
+  # @return [Enumerable] self followed by other
+  def union(other)
+    MultiEnumerator.new(self, other)
+  end
+
+  alias :+ :union
+
+  # @return an Enumerable which iterates over items in this Enumerable but not the other Enumerable
+  def difference(other)
+    filter { |item| not other.include?(item) }
+  end
+
+  alias :- :difference
+
+  # @return an Enumerable which iterates over items in this Enumerable which are also in the other Enumerable
+  def intersect(other)
+    filter { |item| other.include?(item) }
+  end
+
+  alias :& :intersect
+
+  # Returns a new Enumerable that iterates over the base Enumerable applying the transformer block to each item, e.g.:
+  #   [1, 2, 3].transform { |n| n * 2 }.to_a #=> [2, 4, 6]
+  #
+  # Unlike #collect, {#wrap} reflects changes to the base Enumerable, e.g.:
+  #   a = [2, 4, 6]
+``#   transformed = a.wrap { |n| n * 2 }
+  #   a << 4
+  #   transformed.to_a #=> [2, 4, 6, 8]
+  #
+  # In addition, transform has a small, fixed storage requirement, making it preferable to select for large collections.
+  # Note, however, that unlike collect, transform does not return an Array.
+  def wrap(&transformer) # :yields: item
+    Transformer.new(self, &transformer)
+  end
+
+  private
+
+  class Filter
+    include Enumerable
+
+    def initialize(enum=[], &filter)
+      @base = enum
+      @filter = filter
+    end
+
+    # Calls block on each item which passes this Filter's filter test.
+    def each(&block)
+      @base.each { |item| yield(item) if @filter ? @filter.call(item) : item }
+    end
+
+    # Optimized for a Set base.
+    def include?(item)
+      return false if Set === @base and not @base.include?(item)
+      super
+    end
+
+    # Adds value to the base Enumerable, if the base supports it.
+    def <<(value)
+      @base << value
+    end
+
+    # @return a new Array consisting of this Filter's filtered content merged with the other Enumerable
+    def merge(other)
+      to_a.merge(other)
+    end
+
+    # Merges the other Enumerable into the base Enumerable, if the base supports it.
+    def merge!(other)
+      @base.merge!(other)
+    end
+  end
+
+  class Transformer
+    include Enumerable
+
+    def initialize(enum=[], &transformer)
+      @base = enum
+      @xfm = transformer
+    end
+
+    # Sets the base Enumerable on which this Transformer operates and returns this transformer, e.g.:
+    #  transformer = Transformer.new { |n| n * 2 }
+    #  transformer.on([1, 2, 3]).to_a #=> [2, 4, 6]
+    def on(enum)
+      @base = enum
+      self
+    end
+
+    # Calls block on each item after this Transformer's transformer block is applied.
+    def each
+      @base.each { |item| yield(item.nil? ? nil : @xfm.call(item)) }
+    end
+  end
+
+  # A MultiEnumerator iterates over several Enumerators in sequence. Unlike Array#+, MultiEnumerator reflects changes to the
+  # underlying enumerators, e.g.:
+  #   a = [1, 2]
+  #   b = [4, 5]
+  #   ab = MultiEnumerator.new(a, b)
+  #   ab.to_a #=> [1, 2, 4, 5]
+  #   a << 3; b << 6; ab.to_a #=> [1, 2, 3, 4, 5, 6]
+  class MultiEnumerator
+    include Enumerable
+
+    # Creates a new MultiEnumerator on the given Enumerator enums.
+    def initialize(*enums)
+      super()
+      @enums = enums
+      @enums.compact!
+    end
+
+    # Iterates over each of this MultiEnumerator's Enumerators in sequence.
+    def each
+      @enums.each { |enum| enum.each { |item| yield item  } }
+    end
+  end
+end
+
+# The Collector utility module implements the {on} method to apply a block to
+# a collection transitive closure.
+module Collector
+  # Collects the result of applying the given block to the given obj.
+  # If obj is a collection, then collects the result of recursively calling this Collector on the enumerated members.
+  # If obj is nil, then returns nil.
+  # Otherwise, calls block on obj and returns the result.
+  #
+  # @example
+  #  Collector.on([1, 2, [3, 4]]) { |n| n * 2 } #=> [2, 4, [6, 8]]]
+  #  Collector.on(nil) { |n| n * 2 } #=> nil
+  #  Collector.on(1) { |n| n * 2 } #=> 2
+  def self.on(obj, &block)
+    obj.collection? ? obj.map { |item| on(item, &block) } : yield(obj) unless obj.nil?
+  end
+end
+
+class Object
+  # Visits this object's enumerable content as follows:
+  # * If this object is an Enumerable, then the block given to this method is called on each
+  # item in this Enumerable.
+  # * Otherwise, if this object is non-nil, then the the block is called on self.
+  # * Otherwise, this object is nil and this method is a no-op.
+  def enumerate(&block)
+    Enumerable === self ? each(&block) : yield(self) unless nil?
+  end
+
+  # Returns an enumerator on this Object. This default implementation returns an Enumerable::Enumerator
+  # on enumerate.
+  def to_enum
+    Enumerable::Enumerator.new(self, :enumerate)
+  end
+end
+
+module Enumerable
+  # @return self
+  def to_enum
+    self
+  end
+end
+
+# A Flattener applies a given block to flattened collection content.
+class Flattener
+  include Enumerable
+
+  # Visits the enumerated items in the given object's flattened content.
+  # block is called on the base itself if the base is neither nil nor a Enumerable.
+  # If the base object is nil or empty, then this method is a no-op and returns nil.
+  def self.on(obj, &block)
+    obj.collection? ? obj.each { |item| on(item, &block) } : yield(obj) unless obj.nil?
+  end
+
+  # Creates a new Flattener on the given object. obj can be an Enumerable,
+  # a single non-collection object or nil.
+  def initialize(obj)
+    @base = obj
+  end
+
+  # Calls the the given block on this Flattener's flattened content.
+  # If the base object is a collection, then the block is called on the flattened content.
+  # If the base object is nil, then this method is a no-op.
+  # If the base object is neither nil nor a collection, then the block given to this method
+  # is called on the base object itself.
+  #
+  # @example
+  #   Flattener.new(nil).each { |n| print n } #=>
+  #   Flattener.new(1).each { |n| print n } #=> 1
+  #   Flattener.new([1, [2, 3]]).each { |n| print n } #=> 123
+  def each(&block)
+    Flattener.on(@base, &block)
+  end
+end
+
+# ConditionalEnumerator applies a filter to another Enumerable, e.g.:
+#   ConditionalEnumerator.new([1, 2, 3]) { |i| i < 3 }.to_a #=> [1, 2]
+#
+class ConditionalEnumerator
+  include Enumerable
+
+  # Creates a ConditionalEnumerator which wraps the base Enumerator with a conditional filter.
+  def initialize(base, &filter)
+    @base = base
+    @filter = filter
+  end
+
+  # Applies the iterator block to each of this ConditionalEnumerator's base Enumerable items
+  # for which this ConditionalEnumerator's filter returns true.
+  def each
+    @base.each { |item| (yield item) if @filter.call(item) }
+  end
+end
+
+# Hashable is a Hash mixin that adds utility methods to a Hash.
+# Hashable can be included by any class or module which implements an _each_ method
+# with arguments _key_ and _value_.
+module Hashable
+  include Enumerable
+
+  # @see Hash#each
+  def each_pair(&block)
+    each(&block)
+  end
+
+  # @see Hash#[]
+  def [](key)
+    detect_value { |k, v| v if k == key }
+  end
+
+  # @see Hash#each_key
+  def each_key
+    each { |key, value| yield key }
+  end
+
+  # @return [Object,nil] the key for which the block given to this method returns a non-nil, non-false value,
+  #   or nil if none
+  def detect_key
+    each_key { |key| return key if yield key }
+    nil
+  end
+
+  # @see Hash#each_value
+  def each_value
+    each { |key, value| yield value }
+  end
+
+  # Returns a Hashable which composes each value in this Hashable with the key of
+  # the other Hashable, e.g.:
+  #   x = {:a => :c, :b => :d}
+  #   y = {:c => 1}
+  #   z = x.compose(y)
+  #   z[:a] #=> {:c => 1}
+  #   z[:b] #=> nil
+  #
+  # The accessor reflects changes to the underlying hashes, e.g. given the above example:
+  #   x[:b] = 2
+  #   z[:b] #=> {:c => 1}
+  #
+  # Update operations on the result are not supported.
+  #
+  # @param [Hashable] other the Hashable to compose with this Hashable
+  # @return [Hashable] the composed result
+  def compose(other)
+    transform { |value| {value => other[value]} if other.has_key?(value) }
+  end
+
+  # Returns a Hashable which joins each value in this Hashable with the key of
+  # the other Hashable, e.g.:
+  #   x = {:a => :c, :b => :d}
+  #   y = {:c => 1}
+  #   z = x.join(y)
+  #   z[:a] #=> 1
+  #   z[:b] #=> nil
+  #
+  # The accessor reflects changes to the underlying hashes, e.g. given the above example:
+  #   x[:b] = 2
+  #   z[:b] #=> 2
+  #
+  # Update operations on the result are not supported.
+  #
+  # @param [Hashable] other the Hashable to join with this Hashable
+  # @return [Hashable] the joined result
+  def join(other)
+    transform { |value| other[value] }
+  end
+
+  # Returns a Hashable which associates each key of both this Hashable and the other Hashable
+  # with the corresponding value in the first Hashable which has that key, e.g.:
+  #   x = {:a => 1, :b => 2}
+  #   y = {:b => 3, :c => 4}
+  #   z = x + y
+  #   z[:b] #=> 2
+  #
+  # The accessor reflects changes to the underlying hashes, e.g. given the above example:
+  #   x.delete(:b)
+  #   z[:b] #=> 3
+  #
+  # Update operations on the result are not supported.
+  #
+  # @param [Hashable] other the Hashable to form a union with this Hashable
+  # @return [Hashable] the union result
+  def union(other)
+    MultiHash.new(self, other)
+  end
+
+  alias :+ :union
+
+  # Returns a new Hashable that iterates over the base Hashable <key, value> pairs for which the block
+  # given to this method evaluates to a non-nil, non-false value, e.g.:
+  #   {:a => 1, :b => 2, :c => 3}.filter { |k, v| k != :b }.to_hash #=> {:a => 1, :c => 3}
+  #
+  # The default filter block tests the value, e.g.:
+  #   {:a => 1, :b => nil}.filter.to_hash #=> {:a => 1}
+  #
+  # @yield [key, value] the filter block
+  # @return [Hashable] the filtered result
+  def filter(&block)
+    Filter.new(self, &block)
+  end
+
+  # Optimization of {#filter} for a block that only uses the key.
+  #
+  # @example
+  #   {:a => 1, :b => 2, :c => 3}.filter_on_key { |k| k != :b }.to_hash #=> {:a => 1, :c => 3}
+  #
+  # @yield [key] the filter block
+  # @yieldparam key the hash key to filter
+  # @return [Hashable] the filtered result
+  def filter_on_key(&block)
+    KeyFilter.new(self, &block)
+  end
+  
+  # @return [Hashable] a {#filter} that only uses the value.
+  # @yield [value] the filter block
+  # @yieldparam value the hash value to filter
+  # @return [Hashable] the filtered result
+  def filter_on_value
+    filter { |key, value| yield value }
+  end
+  
+  # @return [Hash] a {#filter} of this Hashable which excludes the entries with a null value
+  def compact
+    filter_on_value { |value| not value.nil? }
+  end
+  
+  # Partitions this Hashable into two Hashables based on the given filter block
+  #
+  # @yield [key, value] the partition block
+  # @return [(Hashable, Hashable)] the partitioned result
+  def hash_partition(&block)
+    [filter(&block), filter { |k, v| not yield(k, v) }]
+  end
+  
+  # Returns the difference between this Hashable and the other Hashable in a Hash of the form:
+  #
+  # _key_ => [_mine_, _theirs_]
+  #
+  # where:
+  # * _key_ is the key of association which differs
+  # * _mine_ is the value for _key_ in this hash 
+  # * _theirs_ is the value for _key_ in the other hash 
+  #
+  # @param [Hashable] other the Hashable to subtract
+  # @yield [key, v1, v2] the optional block which determines whether values differ (default is equality)
+  # @yieldparam key the key for which values are compared
+  # @yieldparam v1 the value for key from this Hashable
+  # @yieldparam v2 the value for key from the other Hashable
+  # @return [{Object => (Object,Object)}] a hash of the differences
+  def diff(other)
+    (keys.to_set + other.keys).to_compact_hash do |key|
+       mine = self[key]
+       yours = other[key]
+       [mine, yours] unless block_given? ? yield(key, mine, yours) : mine == yours
+    end
+  end
+  
+  # @yield [key1, key2] the key sort block
+  # @return a Hashable whose #each and {#each_pair} enumerations are sorted by key
+  def sort(&sorter)
+    SortedHash.new(self, &sorter)
+  end
+
+  # Returns a hash which associates each key in this hash with the value mapped by the others.
+  #
+  # @example
+  #   {:a => 1, :b => 2}.assoc_values({:a => 3, :c => 4}) #=> {:a => [1, 3], :b => [2, nil], :c => [nil, 4]}
+  #   {:a => 1, :b => 2}.assoc_values({:a => 3}, {:a => 4, :b => 5}) #=> {:a => [1, 3, 4], :b => [2, nil, 5]}
+  #
+  # @param [<Hashable>] others the other Hashables to associate with this Hashable
+  # @return [Hash] the association hash
+  def assoc_values(*others)
+    all_keys = keys
+    others.each { |hash| all_keys.concat(hash.keys) }
+    all_keys.to_compact_hash do |key|
+      others.map { |other| other[key] }.unshift(self[key])
+    end
+  end
+
+  # Returns an Enumerable whose each block is called on each key which maps to a value which
+  # either equals the given target_value or satisfies the filter block.
+  #
+  # @param target_value the filter value
+  # @yield [value] the filter block
+  # @return [Enumerable] the filtered keys
+  def enum_keys_with_value(target_value=nil, &filter) # :yields: value
+    return enum_keys_with_value { |value| value == target_value } if target_value
+    filter_on_value(&filter).keys
+  end
+
+  # @return [Enumerable] Enumerable over this Hashable's keys
+  def enum_keys
+    Enumerable::Enumerator.new(self, :each_key)
+  end
+
+  # @return [Array] this Hashable's keys
+  def keys
+    enum_keys.to_a
+  end
+
+  # @param key search target
+  # @return whether this Hashable has the given key
+  def has_key?(key)
+    enum_keys.include?(key)
+  end
+
+  alias :include? :has_key?
+
+  # @return [Enumerable] an Enumerable over this Hashable's values
+  def enum_values
+    Enumerable::Enumerator.new(self, :each_value)
+  end
+
+  # @yield [key] the key selector
+  # @return the keys which satisfy the block given to this method
+  def select_keys(&block)
+    enum_keys.select(&block)
+  end
+  
+  # @yield [key] the key rejector
+  # @return the keys which do not satisfy the block given to this method
+  def reject_keys(&block)
+    enum_keys.reject(&block)
+  end
+  
+  # @yield [value] the value selector
+  # @return the values which satisfy the block given to this method
+  def select_values(&block)
+    enum_values.select(&block)
+  end
+  
+  # @yield [value] the value rejector
+  # @return the values which do not satisfy the block given to this method
+  def reject_values(&block)
+    enum_values.reject(&block)
+  end
+
+  # @return [Array] this Enumerable's values
+  def values
+    enum_values.to_a
+  end
+
+  # @param value search target
+  # @return whether this Hashable has the given value
+  def has_value?(value)
+    enum_values.include?(value)
+  end
+
+  # @return [Array] a flattened Array of this Hash
+  # @example
+  #   {:a => {:b => :c}, :d => :e, :f => [:g]} #=> [:a, :b, :c, :d, :e, :f, :g]
+  def flatten
+    Flattener.new(self).to_a
+  end
+
+  # @yield [key, value] hash splitter
+  # @return [Hash] two hashes split by whether calling the block on the
+  #   entry returns a non-nil, non-false value
+  # @example
+  #   {:a => 1, :b => 2}.partition { |key, value| value < 2 } #=> [{:a => 1}, {:b => 2}]
+  def partition(&block)
+    super(&block).map { |pairs| pairs.to_assoc_hash }
+  end
+
+  # Returns a new Hash that recursively copies this hash's values. Values of type hash are copied using copy_recursive.
+  # Other values are unchanged.
+  #
+  # This method is useful for preserving and restoring hash associations.
+  #
+  # @return [Hash] a deep copy of this Hashable 
+  def copy_recursive
+    copy = Hash.new
+    keys.each do |key|
+      value = self[key]
+      copy[key] = Hash === value ? value.copy_recursive : value
+    end
+    copy
+  end
+
+  # @return [Hash] a new Hash that transforms each value
+  # @example
+  #   {:a => 1, :b => 2}.transform { |n| n * 2 }.values #=> [2, 4]
+  def transform(&transformer)
+    ValueTransformerHash.new(self, &transformer)
+  end
+
+  def to_hash
+    inject({}) { |hash, pair| hash[pair.first] = pair.last; hash }
+  end
+
+  def to_set
+    to_a.to_set
+  end
+
+  def to_s
+    to_hash.to_s
+  end
+
+  def inspect
+    to_hash.inspect
+  end
+
+  def ==(other)
+    to_hash == other.to_hash rescue super
+  end
+
+  private
+
+  # @see #filter
+  class Filter
+    include Hashable
+
+    def initialize(base, &filter)
+      @base = base
+      @filter = filter
+    end
+
+    def each
+      @base.each { |k, v| yield(k, v) if @filter ? @filter.call(k, v) : v }
+    end
+  end
+
+  # @see #filter_on_key
+  class KeyFilter < Filter
+    include Hashable
+
+    def initialize(base)
+      super(base) { |k, v| yield(k) }
+    end
+
+    def [](key)
+      super if @filter.call(key, nil)
+    end
+  end
+  
+  # @see #sort
+  class SortedHash
+    include Hashable
+    
+    def initialize(base, &comparator)
+      @base = base
+      @comparator = comparator
+    end
+    
+    def each
+      @base.keys.sort { |k1, k2| @comparator ? @comparator.call(k1, k2) : k1 <=> k2 }.each { |k| yield(k, @base[k]) }
+    end
+  end
+
+  # Combines hashes. See Hash#+ for details.
+  class MultiHash
+    include Hashable
+
+    def initialize(*hashes)
+      @hashes = hashes
+    end
+
+    def [](key)
+      @hashes.each { |hash| return hash[key] if hash.has_key?(key) }
+      nil
+    end
+
+    def has_key?(key)
+      @hashes.any? { |hash| hash.has_key?(key) }
+    end
+
+    def has_value?(value)
+      @hashes.any? { |hash| hash.has_value?(value) }
+    end
+
+    def each
+      @hashes.each_with_index do |hash, index|
+        hash.each do |key, value|
+           yield(key, value) unless (0...index).any? { |i| @hashes[i].has_key?(key) }
+        end
+      end
+      self
+    end
+  end
+
+  # The ValueTransformerHash class pipes the value from a base Hashable into a transformer block.
+  class ValueTransformerHash
+    include Hashable
+
+    # Creates a ValueTransformerHash on the base hash and value transformer block.
+    def initialize(base, &transformer) # :yields: value
+      @base = base
+      @xfm = transformer
+    end
+
+    # Returns the value at key after this ValueTransformerHash's transformer block is applied, or nil
+    # if this hash does not contain key.
+    def [](key)
+      @xfm.call(@base[key]) if @base.has_key?(key)
+    end
+
+    def each
+      @base.each { |key, value| yield(key, @xfm.call(value)) }
+    end
+  end
+end
+
+# The KeyTransformerHash class pipes the key access argument into a transformer block before
+# accessing a base Hashable, e.g.:
+#   hash = KeyTransformerHash.new { |key| key % 2 }
+#   hash[1] = :a
+#   hash[3] # => :a
+class KeyTransformerHash
+  include Hashable
+
+  # Creates a KeyTransformerHash on the optional base hash and required key transformer block.
+  #
+  # Raises ArgumentError if there is no extractor block
+  def initialize(base={}, &transformer) # :yields: key
+    raise ArgumentError.new("Missing required Accessor block") unless block_given?
+    @base = base
+    @xfm = transformer
+  end
+
+  # Returns the value at key after this KeyTransformerHash's transformer block is applied to the key,
+  # or nil if the base hash does not contain an association for the transforemd key.
+  def [](key)
+    @base[@xfm.call(key)]
+  end
+
+  # Sets the value at key after this KeyTransformerHash's transformer block is applied, or nil
+  # if this hash does not contain an association for the transformed key.
+  def []=(key, value)
+    @base[@xfm.call(key)] = value
+  end
+
+  # Delegates to the base hash.
+  # Note that this breaks the standard Hash contract, since
+  #   all? { |k, v| self[k] }
+  # is not necessarily true because the key is transformed on access.
+  # @see Accessor for a KeyTransformerHash variant that restores this contract
+  def each(&block)
+    @base.each(&block)
+  end
+end
+
+class Hash
+  include Hashable
+
+  # The EMPTY_HASH constant is an immutable empty hash, used primarily as a default argument.
+  class << EMPTY_HASH = Hash.new
+    def []=(key, value)
+      raise NotImplementedError.new("Modification of the constant empty hash is not supported")
+    end
+  end
+end
+
+# Hashinator creates a Hashable from an Enumerable on [_key_, _value_] pairs.
+# The Hashinator reflects changes to the underlying Enumerable.
+#
+# @example
+#   base = [[:a, 1], [:b, 2]]
+#   hash = Hashinator.new(base)
+#   hash[:a] #=> 1
+#   base.first[1] = 3
+#   hash[:a] #=> 3
+class Hashinator
+  include Hashable
+
+  def initialize(enum)
+    @base = enum
+  end
+
+  def each
+    @base.each { |pair| yield(*pair) }
+  end
+end
+
+#
+# A Hash that creates a new entry on demand.
+#
+class LazyHash < Hash
+  #
+  # Creates a new Hash with the specified value factory proc.
+  # The factory proc has one argument, the key.
+  # If access by key fails, then a new association is created
+  # from the key to the result of calling the factory proc.
+  #
+  # Example:
+  #   hash = LazyHash.new { |key| key.to_s }
+  #   hash[1] = "1"
+  #   hash[1] #=> "1"
+  #   hash[2] #=> "2"
+  #
+  # If a block is not provided, then the default association value is nil, e.g.:
+  #   hash = LazyHash.new
+  #   hash.has_key?(1) #=> false
+  #   hash[1] #=> nil
+  #   hash.has_key?(1) #=> true
+  #
+  # A nil key always returns nil. There is no hash entry for nil, e.g.:
+  #   hash = LazyHash.new { |key| key }
+  #   hash[nil] #=> nil
+  #   hash.has_key?(nil) #=> false
+  #
+  # If the :compact option is set, then an entry is not created
+  # if the value initializer result is nil or empty, e.g.:
+  #   hash = LazyHash.new { |n| 10.div(n) unless n.zero? }
+  #   hash[0] #=> nil
+  #   hash.has_key?(0) #=> false
+  def initialize(options=nil)
+    reject_flag = Options.get(:compact, options)
+    # Make the hash with the factory block
+    super() do |hash, key|
+      if key then
+        value = yield key if block_given?
+        hash[key] = value unless reject_flag and value.nil_or_empty?
+      end
+    end
+  end
+end
+
+class Array
+  # The EMPTY_ARRAY constant is an immutable empty array, used primarily as a default argument.
+  class << EMPTY_ARRAY = Array.new
+    def <<(value)
+      raise NotImplementedError.new("Modification of the constant empty array is not supported")
+    end
+  end
+
+  # Relaxes the Ruby Array methods which take an Array argument to allow collection Enumerable arguments.
+  [:|, :+, :-, :&].each do |meth|
+    redefine_method(meth) do |old_meth|
+      lambda { |other| send(old_meth, other.collection? ? other.to_a : other) }
+    end
+  end
+
+  redefine_method(:flatten) do |old_meth|
+    # if an item is a non-Array collection, then convert it into an array before recursively flattening the list
+    lambda { map { |item| item.collection? ? item.to_a : item }.send(old_meth) }
+  end
+
+  # Returns an array containing all but the first item in this Array. This method is syntactic sugar for
+  # +self[1..-1]+ or +last(length-1)+
+  def rest
+    self[1..-1]
+  end
+
+  # Prints the content of this array as a series, e.g.:
+  #   [1, 2, 3].to_series #=> "1, 2 and 3"
+  #   [1, 2, 3].to_series('or') #=> "1, 2 or 3"
+  #
+  # If a block is given to this method, then the block is applied before the series is formed, e.g.:
+  #   [1, 2, 3].to_series { |n| n + 1 } #=> "2, 3 and 4"
+  def to_series(conjunction=nil)
+    conjunction ||= 'and'
+    return map { |item| yield item }.to_series(conjunction) if block_given?
+    padded_conjunction = " #{conjunction} "
+    # join all but the last item as a comma-separated list and append the conjunction and last item
+    length < 2 ? to_s : self[0...-1].join(', ') + padded_conjunction + last.to_s
+  end
+
+  # Returns a new Hash generated from this array of arrays by associating the first element of each
+  # member to the remaining elements. If there are only two elements in the member, then the first
+  # element is associated with the second element. If there is less than two elements in the member,
+  # the first element is associated with nil. An empty array is ignored.
+  #
+  # Example:
+  #   [[:a, 1], [:b, 2, 3], [:c], []].to_assoc_hash #=> { :a => 1, :b => [2,3], :c => nil }
+  def to_assoc_hash
+    hash = Hash.new
+    each do |item|
+      raise ArgumentError.new("Array member must be an array: #{item.pp_s(:single_line)}") unless Array === item
+      key = item.first
+      if item.size < 2 then
+        value = nil
+      elsif item.size == 2 then
+        value = item[1]
+      else
+        value = item[1..-1]
+      end
+      hash[key] = value unless key.nil?
+    end
+    hash
+  end
+
+  alias :base__flatten :flatten
+  private :base__flatten
+  # Recursively flattens this array, including any collection item that implements the +to_a+ method.
+  def flatten
+    # if any item is a Set or Java Collection, then convert those into arrays before recursively flattening the list
+    if any? { |item| Set === item or Java::JavaUtil::Collection === item } then
+      return map { |item| (Set === item or Java::JavaUtil::Collection === item) ? item.to_a : item }.flatten
+    end
+    base__flatten
+  end
+
+  # Adds the other Enumerable to this array.
+  #
+  # Raises ArgumentError if other does not respond to the +to_a+ method.
+  def add_all(other)
+    return concat(other) if Array === other
+    begin
+      return add_all(other.to_a)
+    rescue
+      raise ArgumentError.new("Can't convert #{other.class.name} to array") unless other.respond_to?(:to_a)
+    end
+  end
+
+  alias :merge! :add_all
+end
+
+# CaseInsensitiveHash accesses entries in a case-insensitive String comparison. The accessor method
+# key argument is converted to a String before look-up.
+#
+# @example
+#   hash = CaseInsensitiveHash.new
+#   hash[:UP] = "down"
+#   hash['up'] #=> "down"
+class CaseInsensitiveHash < Hash
+  def initialize
+    super
+  end
+
+  def [](key)
+    # if there is lower-case key association, then convert to lower-case and return.
+    # otherwise, delegate to super with the call argument unchanged. this ensures
+    # that a default block passed to the constructor will be called with the correct
+    # key argument.
+    has_key?(key) ? super(key.to_s.downcase) : super(key)
+  end
+
+  def []=(key, value)
+    super(key.to_s.downcase, value)
+  end
+
+  def has_key?(key)
+    super(key.to_s.downcase)
+  end
+
+  def delete(key)
+    super(key.to_s.downcase)
+  end
+
+  alias :store :[]=
+  alias :include? :has_key?
+  alias :key? :has_key?
+  alias :member? :has_key?
+end
+
+class Set
+  # The standard Set {#merge} is an anomaly among Ruby collections, since merge modifies the called Set in-place rather
+  # than return a new Set containing the merged contents. Preserve this unfortunate behavior, but partially address
+  # the anomaly by adding the merge! alias for in-place merge.
+  alias :merge! :merge
+end