Class: Immutable::Set

Inherits:
Object
  • Object
show all
Includes:
Enumerable
Defined in:
lib/immutable/set.rb

Overview

Immutable::Set is a collection of unordered values with no duplicates. Testing whether an object is present in the Set can be done in constant time. Set is also Enumerable, so you can iterate over the members of the set with #each, transform them with #map, filter them with #select, and so on. Some of the Enumerable methods are overridden to return immutable-ruby collections.

Like the Set class in Ruby's standard library, which we will call RubySet, Immutable::Set defines equivalency of objects using #hash and #eql?. No two objects with the same #hash code, and which are also #eql?, can coexist in the same Set. If one is already in the Set, attempts to add another one will have no effect.

Sets have no natural ordering and cannot be compared using #<=>. However, they define #<, #>, #<=, and #>= as shorthand for #proper_subset?, #proper_superset?, #subset?, and #superset? respectively.

The basic set-theoretic operations #union, #intersection, #difference, and #exclusion work with any Enumerable object.

A Set can be created in either of the following ways:

Immutable::Set.new([1, 2, 3]) # any Enumerable can be used to initialize
Immutable::Set['A', 'B', 'C', 'D']

The latter 2 forms of initialization can be used with your own, custom subclasses of Immutable::Set.

Unlike RubySet, all methods which you might expect to "modify" an Immutable::Set actually return a new set and leave the existing one unchanged.

Examples:

set1 = Immutable::Set[1, 2] # => Immutable::Set[1, 2]
set2 = Immutable::Set[1, 2] # => Immutable::Set[1, 2]
set1 == set2              # => true
set3 = set1.add("foo")    # => Immutable::Set[1, 2, "foo"]
set3 - set2               # => Immutable::Set["foo"]
set3.subset?(set1)        # => false
set1.subset?(set3)        # => true

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Enumerable

#<=>, #compact, #each_index, #grep, #grep_v, #group_by, #inspect, #join, #partition, #product, #reject, #sum

Methods included from Enumerable

#to_list

Constructor Details

#initialize(items = []) ⇒ Set

Returns a new instance of Set.


77
78
79
80
81
# File 'lib/immutable/set.rb', line 77

def initialize(items=[])
  @trie = Trie.new(0)
  items.each { |item| @trie.put!(item, nil) }
  freeze
end

Class Method Details

.[](*items) ⇒ Set

Create a new Set populated with the given items.

Returns:


55
56
57
# File 'lib/immutable/set.rb', line 55

def [](*items)
  items.empty? ? empty : new(items)
end

.emptySet

Return an empty Set. If used on a subclass, returns an empty instance of that class.

Returns:


63
64
65
# File 'lib/immutable/set.rb', line 63

def empty
  @empty ||= new
end

Instance Method Details

#add(item) ⇒ Set Also known as: <<

Return a new Set with item added. If item is already in the set, return self.

Examples:

Immutable::Set[1, 2, 3].add(4) # => Immutable::Set[1, 2, 4, 3]
Immutable::Set[1, 2, 3].add(2) # => Immutable::Set[1, 2, 3]

Parameters:

  • item (Object)

    The object to add

Returns:


105
106
107
# File 'lib/immutable/set.rb', line 105

def add(item)
  include?(item) ? self : self.class.alloc(@trie.put(item, nil))
end

#add?(item) ⇒ Set, false

If item is not a member of this Set, return a new Set with item added. Otherwise, return false.

Examples:

Immutable::Set[1, 2, 3].add?(4) # => Immutable::Set[1, 2, 4, 3]
Immutable::Set[1, 2, 3].add?(2) # => false

Parameters:

  • item (Object)

    The object to add

Returns:


119
120
121
# File 'lib/immutable/set.rb', line 119

def add?(item)
  !include?(item) && add(item)
end

#clearSet

Return an empty Set instance, of the same class as this one. Useful if you have multiple subclasses of Set and want to treat them polymorphically.

Returns:


504
505
506
# File 'lib/immutable/set.rb', line 504

def clear
  self.class.empty
end

#delete(item) ⇒ Set

Return a new Set with item removed. If item is not a member of the set, return self.

Examples:

Immutable::Set[1, 2, 3].delete(1)  # => Immutable::Set[2, 3]
Immutable::Set[1, 2, 3].delete(99) # => Immutable::Set[1, 2, 3]

Parameters:

  • item (Object)

    The object to remove

Returns:


132
133
134
135
# File 'lib/immutable/set.rb', line 132

def delete(item)
  trie = @trie.delete(item)
  new_trie(trie)
end

#delete?(item) ⇒ Set, false

If item is a member of this Set, return a new Set with item removed. Otherwise, return false.

Examples:

Immutable::Set[1, 2, 3].delete?(1)  # => Immutable::Set[2, 3]
Immutable::Set[1, 2, 3].delete?(99) # => false

Parameters:

  • item (Object)

    The object to remove

Returns:


146
147
148
# File 'lib/immutable/set.rb', line 146

def delete?(item)
  include?(item) && delete(item)
end

#difference(other) ⇒ Set Also known as: subtract, -

Return a new Set with all the items in other removed. other can be any Enumerable object.

Examples:

Immutable::Set[1, 2] - Immutable::Set[2, 3] # => Immutable::Set[1]

Parameters:

  • other (Enumerable)

    The collection to subtract from this set

Returns:


346
347
348
349
350
351
352
353
# File 'lib/immutable/set.rb', line 346

def difference(other)
  trie = if (@trie.size <= other.size) && (other.is_a?(Immutable::Set) || (defined?(::Set) && other.is_a?(::Set)))
    @trie.select { |key, _| !other.include?(key) }
  else
    @trie.bulk_delete(other)
  end
  new_trie(trie)
end

#disjoint?(other) ⇒ Boolean

Return true if this Set and other do not share any items.

Examples:

Immutable::Set[1, 2].disjoint?(Immutable::Set[8, 9]) # => true

Parameters:

Returns:

  • (Boolean)

448
449
450
451
452
453
454
455
456
457
458
459
# File 'lib/immutable/set.rb', line 448

def disjoint?(other)
  if other.size <= size
    other.each { |item| return false if include?(item) }
  else
    # See comment on #subset?
    if other.size >= 150 && @trie.size >= 190 && !(other.is_a?(Immutable::Set) || other.is_a?(::Set))
      other = ::Set.new(other)
    end
    each { |item| return false if other.include?(item) }
  end
  true
end

#dupSet Also known as: clone

Return self. Since this is an immutable object duplicates are equivalent.

Returns:


533
534
535
# File 'lib/immutable/set.rb', line 533

def dup
  self
end

#each {|item| ... } ⇒ self, Enumerator

Call the block once for each item in this Set. No specific iteration order is guaranteed, but the order will be stable for any particular Set. If no block is given, an Enumerator is returned instead.

Examples:

Immutable::Set["Dog", "Elephant", "Lion"].each { |e| puts e }
Elephant
Dog
Lion
# => Immutable::Set["Dog", "Elephant", "Lion"]

Yields:

  • (item)

    Once for each item.

Returns:

  • (self, Enumerator)

163
164
165
166
167
# File 'lib/immutable/set.rb', line 163

def each
  return to_enum if not block_given?
  @trie.each { |key, _| yield(key) }
  self
end

#empty?Boolean

Return true if this Set contains no items.

Returns:

  • (Boolean)

85
86
87
# File 'lib/immutable/set.rb', line 85

def empty?
  @trie.empty?
end

#eql?(other) ⇒ Boolean Also known as: ==

Return true if other has the same type and contents as this Set.

Parameters:

  • other (Object)

    The object to compare with

Returns:

  • (Boolean)

512
513
514
515
516
517
518
519
520
521
# File 'lib/immutable/set.rb', line 512

def eql?(other)
  return true if other.equal?(self)
  return false if not instance_of?(other.class)
  other_trie = other.instance_variable_get(:@trie)
  return false if @trie.size != other_trie.size
  @trie.each do |key, _|
    return false if !other_trie.key?(key)
  end
  true
end

#exclusion(other) ⇒ Set Also known as: ^

Return a new Set which contains all the items which are members of this Set or of other, but not both. other can be any Enumerable object.

Examples:

Immutable::Set[1, 2] ^ Immutable::Set[2, 3] # => Immutable::Set[1, 3]

Parameters:

  • other (Enumerable)

    The collection to take the exclusive disjunction of

Returns:


365
366
367
# File 'lib/immutable/set.rb', line 365

def exclusion(other)
  ((self | other) - (self & other))
end

#firstObject

Return a member of this Set. The member chosen will be the first one which would be yielded by #each. If the set is empty, return nil.

Examples:

Immutable::Set["A", "B", "C"].first # => "C"

Returns:

  • (Object)

242
243
244
# File 'lib/immutable/set.rb', line 242

def first
  (entry = @trie.at(0)) && entry[0]
end

#flattenSet

Recursively insert the contents of any nested Sets into this Set, and remove them.

Examples:

Immutable::Set[Immutable::Set[1, 2], Immutable::Set[3, 4]].flatten
# => Immutable::Set[1, 2, 3, 4]

Returns:


480
481
482
483
484
485
# File 'lib/immutable/set.rb', line 480

def flatten
  reduce(self.class.empty) do |set, item|
    next set.union(item.flatten) if item.is_a?(Set)
    set.add(item)
  end
end

#hashInteger

See Object#hash.

Returns:

  • (Integer)

526
527
528
# File 'lib/immutable/set.rb', line 526

def hash
  reduce(0) { |hash, item| (hash << 5) - hash + item.hash }
end

#include?(object) ⇒ Boolean Also known as: member?

Return true if the given item is present in this Set. More precisely, return true if an object with the same #hash code, and which is also #eql? to the given object is present.

Examples:

Immutable::Set["A", "B", "C"].include?("B") # => true
Immutable::Set["A", "B", "C"].include?("Z") # => false

Parameters:

  • object (Object)

    The object to check for

Returns:

  • (Boolean)

230
231
232
# File 'lib/immutable/set.rb', line 230

def include?(object)
  @trie.key?(object)
end

#intersect?(other) ⇒ Boolean

Return true if this Set and other have at least one item in common.

Examples:

Immutable::Set[1, 2].intersect?(Immutable::Set[2, 3]) # => true

Parameters:

Returns:

  • (Boolean)

468
469
470
# File 'lib/immutable/set.rb', line 468

def intersect?(other)
  !disjoint?(other)
end

#intersection(other) ⇒ Set Also known as: &

Return a new Set which contains all the items which are members of both this Set and other. other can be any Enumerable object.

Examples:

Immutable::Set[1, 2] & Immutable::Set[2, 3] # => Immutable::Set[2]

Parameters:

  • other (Enumerable)

    The collection to intersect with

Returns:


323
324
325
326
327
328
329
330
331
332
333
334
335
# File 'lib/immutable/set.rb', line 323

def intersection(other)
  if other.size < @trie.size
    if other.is_a?(Immutable::Set)
      trie = other.instance_variable_get(:@trie).select { |key, _| include?(key) }
    else
      trie = Trie.new(0)
      other.each { |obj| trie.put!(obj, nil) if include?(obj) }
    end
  else
    trie = @trie.select { |key, _| other.include?(key) }
  end
  new_trie(trie)
end

#map {|item| ... } ⇒ Set Also known as: collect

Call the block once for each item in this Set. All the values returned from the block will be gathered into a new Set. If no block is given, an Enumerator is returned instead.

Examples:

Immutable::Set["Cat", "Elephant", "Dog", "Lion"].map { |e| e.size }
# => Immutable::Set[8, 4, 3]

Yields:

  • (item)

    Once for each item.

Returns:


213
214
215
216
217
# File 'lib/immutable/set.rb', line 213

def map
  return enum_for(:map) if not block_given?
  return self if empty?
  self.class.new(super)
end

#proper_subset?(other) ⇒ Boolean Also known as: <

Returns true if other contains all the items in this Set, plus at least one item which is not in this set.

Examples:

Immutable::Set[2, 3].proper_subset?(Immutable::Set[1, 2, 3])    # => true
Immutable::Set[1, 2, 3].proper_subset?(Immutable::Set[1, 2, 3]) # => false

Parameters:

Returns:

  • (Boolean)

417
418
419
420
421
422
423
424
# File 'lib/immutable/set.rb', line 417

def proper_subset?(other)
  return false if other.size <= size
  # See comments above
  if other.size >= 150 && @trie.size >= 190 && !(other.is_a?(Immutable::Set) || other.is_a?(::Set))
    other = ::Set.new(other)
  end
  all? { |item| other.include?(item) }
end

#proper_superset?(other) ⇒ Boolean Also known as: >

Returns true if this Set contains all the items in other, plus at least one item which is not in other.

Examples:

Immutable::Set[1, 2, 3].proper_superset?(Immutable::Set[2, 3])    # => true
Immutable::Set[1, 2, 3].proper_superset?(Immutable::Set[1, 2, 3]) # => false

Parameters:

Returns:

  • (Boolean)

436
437
438
# File 'lib/immutable/set.rb', line 436

def proper_superset?(other)
  other.proper_subset?(self)
end

#reverse_each {|item| ... } ⇒ self

Call the block once for each item in this Set. Iteration order will be the opposite of #each. If no block is given, an Enumerator is returned instead.

Examples:

Immutable::Set["Dog", "Elephant", "Lion"].reverse_each { |e| puts e }
Lion
Dog
Elephant
# => Immutable::Set["Dog", "Elephant", "Lion"]

Yields:

  • (item)

    Once for each item.

Returns:

  • (self)

182
183
184
185
186
# File 'lib/immutable/set.rb', line 182

def reverse_each
  return enum_for(:reverse_each) if not block_given?
  @trie.reverse_each { |key, _| yield(key) }
  self
end

#sampleObject

Return a randomly chosen item from this Set. If the set is empty, return nil.

Examples:

Immutable::Set[1, 2, 3, 4, 5].sample # => 3

Returns:

  • (Object)

496
497
498
# File 'lib/immutable/set.rb', line 496

def sample
  empty? ? nil : @trie.at(rand(size))[0]
end

#select {|item| ... } ⇒ Set Also known as: find_all, keep_if

Return a new Set with all the items for which the block returns true.

Examples:

Immutable::Set["Elephant", "Dog", "Lion"].select { |e| e.size >= 4 }
# => Immutable::Set["Elephant", "Lion"]

Yields:

  • (item)

    Once for each item.

Returns:


195
196
197
198
199
# File 'lib/immutable/set.rb', line 195

def select
  return enum_for(:select) unless block_given?
  trie = @trie.select { |key, _| yield(key) }
  new_trie(trie)
end

#sizeInteger Also known as: length

Return the number of items in this Set.

Returns:

  • (Integer)

91
92
93
# File 'lib/immutable/set.rb', line 91

def size
  @trie.size
end

#sort {|a, b| ... } ⇒ SortedSet

Return a Immutable::SortedSet which contains the same items as this Set, ordered by the given comparator block.

Examples:

Immutable::Set["Elephant", "Dog", "Lion"].sort
# => Immutable::SortedSet["Dog", "Elephant", "Lion"]
Immutable::Set["Elephant", "Dog", "Lion"].sort { |a,b| a.size <=> b.size }
# => Immutable::SortedSet["Dog", "Lion", "Elephant"]

Yields:

  • (a, b)

    Any number of times with different pairs of elements.

Yield Returns:

  • (Integer)

    Negative if the first element should be sorted lower, positive if the latter element, or 0 if equal.

Returns:


260
261
262
# File 'lib/immutable/set.rb', line 260

def sort(&comparator)
  SortedSet.new(to_a, &comparator)
end

#sort_by {|item| ... } ⇒ SortedSet

Return a Immutable::SortedSet which contains the same items as this Set, ordered by mapping each item through the provided block to obtain sort keys, and then sorting the keys.

Examples:

Immutable::Set["Elephant", "Dog", "Lion"].sort_by { |e| e.size }
# => Immutable::SortedSet["Dog", "Lion", "Elephant"]

Yields:

  • (item)

    Once for each item to create the set, and then potentially again depending on what operations are performed on the returned Immutable::SortedSet. As such, it is recommended that the block be a pure function.

Yield Returns:

  • (Object)

    sort key for the item

Returns:


278
279
280
# File 'lib/immutable/set.rb', line 278

def sort_by(&mapper)
  SortedSet.new(to_a, &mapper)
end

#subset?(other) ⇒ Boolean Also known as: <=

Return true if all items in this Set are also in other.

Examples:

Immutable::Set[2, 3].subset?(Immutable::Set[1, 2, 3]) # => true

Parameters:

Returns:

  • (Boolean)

377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
# File 'lib/immutable/set.rb', line 377

def subset?(other)
  return false if other.size < size

  # This method has the potential to be very slow if 'other' is a large Array, so to avoid that,
  #   we convert those Arrays to Sets before checking presence of items
  # Time to convert Array -> Set is linear in array.size
  # Time to check for presence of all items in an Array is proportional to set.size * array.size
  # Note that both sides of that equation have array.size -- hence those terms cancel out,
  #   and the break-even point is solely dependent on the size of this collection
  # After doing some benchmarking to estimate the constants, it appears break-even is at ~190 items
  # We also check other.size, to avoid the more expensive #is_a? checks in cases where it doesn't matter
  #
  if other.size >= 150 && @trie.size >= 190 && !(other.is_a?(Immutable::Set) || other.is_a?(::Set))
    other = ::Set.new(other)
  end
  all? { |item| other.include?(item) }
end

#superset?(other) ⇒ Boolean Also known as: >=

Return true if all items in other are also in this Set.

Examples:

Immutable::Set[1, 2, 3].superset?(Immutable::Set[2, 3]) # => true

Parameters:

Returns:

  • (Boolean)

403
404
405
# File 'lib/immutable/set.rb', line 403

def superset?(other)
  other.subset?(self)
end

#to_setself

Return self.

Returns:

  • (self)

544
545
546
# File 'lib/immutable/set.rb', line 544

def to_set
  self
end

#union(other) ⇒ Set Also known as: |, +, merge

Return a new Set which contains all the members of both this Set and other. other can be any Enumerable object.

Examples:

Immutable::Set[1, 2] | Immutable::Set[2, 3] # => Immutable::Set[1, 2, 3]

Parameters:

  • other (Enumerable)

    The collection to merge with

Returns:


290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
# File 'lib/immutable/set.rb', line 290

def union(other)
  if other.is_a?(Immutable::Set)
    if other.size > size
      small_set_pairs = @trie
      large_set_trie = other.instance_variable_get(:@trie)
    else
      small_set_pairs = other.instance_variable_get(:@trie)
      large_set_trie = @trie
    end
  else
    if other.respond_to?(:lazy)
      small_set_pairs = other.lazy.map { |e| [e, nil] }
    else
      small_set_pairs = other.map { |e| [e, nil] }
    end
    large_set_trie = @trie
  end

  trie = large_set_trie.bulk_put(small_set_pairs)
  new_trie(trie)
end