DS - Data Structures for Ruby¶ ↑

<img src=“https://travis-ci.org/knife/ds.svg?branch=master” alt=“Build Status” />

DS provides some popular data structures not implemented in Ruby natively.

Data structures included in this gem:

• Pair

• Stacks

  • Stack

• Queues

  • SimpleQueue

  • PriorityQueue

• Lists

  • List

• Trees

  • Tree

  • BinaryTree

  • BinaryHeap

  • RedBlackTree

  • Trie

• Matrixes

  • Array2D

  • ExpandableArray

  • TriMatrix

• Sets

  • IndexedSet

gem install ds

require 'ds'
stack = DS::Stack.new

To not have to type “DS::” before each class, use:

include DS
stack = Stack.new

Pair is simple key-value data structure.

Creating new Pair

p = Pair.new(3, 9)

Accessors defined on Pair object:

p.key #=> 3
p.value #=> 9
p.value = 27

p.first #=> 3
p.second #=> 9
p.second = 27

Stack is very simple data structure which allows access only to the top element. More: Stack

Creating new Stack (implemented as Array).

stack = Stack.new

with initial values:

s = Stack.new(:first, :second)

Creating new Stack (implemented as List).

stack = Stack.create

The following methods are available on a Stack:

• push

• pop

• peek

• size

• empty?

Examples:

stack.empty? #=> true
stack.push :first
stack.push :second
stack.size #=> 2
stack.peek #=> :second
stack.empty? #=> false
stack.pop #=> :second
stack.size #=> 1

Queue is First-In-First-Out (FIFO) data structure. Which means that first element added to the queue will be the first one to be removed. More: Queue

Creating new SimpleQueue (implemented as Array).

q = SimpleQueue.new

with initial values:

q = SimpleQueue.new(1, 2, 3)

Creating new SimpleQueue (implemented as List)

q1 = SimpleQueue.create

The following methods are available on a Queue:

• enqueue or push

• dequeue or shift

• peek

• length or size

• empty?

Examples:

q.enqueue :first
q.push :second
q.peek #=> :first
q.length #=> 2
q.empty? #=> false
q.dequeue #=> :first
q.shift #=> :second
q.empty? #=> true

In opposite to simple Queue, in PriorityQueue each element is associated with a “priority”. More: Priority Queue

Creating new Priority Queue (implemented as BinaryHeap)

q = PriorityQueue.new

By default higher value means higher priority but you can define own priority order by passing block to constructor:

PriorityQueue.new { |a, b| a < b }

To add new element to priority queue use #unshift or #push method:

q.push(value, priority)

To remove element from priority queue use #shift or #pop method. The interface is very similar to SimpleQueue.

Examples:

q.push(:important, 3)
q.push(:very_important, 5)
q.push(:nevermind, 1)

q.shift #=> :very_important
q.peek #=> :important
q.length #=> 2
q.shift  #=> :important
q.peek  #=> :nevermind

Indexed Priority Queue is special form of PriorityQueue with constant access to any element. Additionaly you can easily change priority of any element stored on queue.

Creating new Indexed Priority Queue

q = IndexedPriorityQueue.new

or

IndexedPriorityQueue.new { |a, b| a.key < b.key }

IndexedPriorityQueue inherits from PriorityQueue so all methods from PriorityQueue are available.

Examples:

q.push(:important, 3)
q.push(:very_important, 5)
q.push(:nevermind, 1)

q.peek #=> :very_important

q.change(:nevermind, 10)
q.peek #=> :nevermind

Elements stored on priority queue are wrapped in Pair object, when you call get method this object is returned:

q.get(:very_important) #=> Pair.new(5, :very_important)
q.include?(:very_important) #=> true

List is an ordered collection of values. Each element of list has pointer to the next element (last element points to nil). This implementation uses doubly linked list underhood. More: List

Creating new List

l = List.new
l.append(2)

or

list = List.new(1, 2, 3, 4)

Examples:

Simple operations on lists

list.length #=> 4
list.append(5).to_a #=> [1, 2, 3, 4, 5]
list.unshift(0).to_a #=> [0, 1, 2, 3, 4, 5]
list.remove(list.head).to_a #=> [1, 2, 3, 4, 5]
list.shift #=> 1

Accessing first and last element

list.head.data #=> 2
list.tail.data #=> 5

list.first #=> 2
list.last #=> 5

Accessing by index

list[2].data #=> 2
list.at(2).data #=> 2
list[-1].data #=> 4
list[1..2].map(&:data) #=> [2, 3]
list[1,3].map(&:data) #=> [2, 3, 4]

Modifying elements on given index:

list[2].data = 8
list[2].data #=> 8
list[2] = [9, 10] #=> [1, 2, 9, 10, 4]
list[0,1] = 0 #=> [0, 2, 3, 4]
list[2..3] = ['x', 'x'] #=> [1, 2, 'x', 'x']

Checking if given element exists on list

list.get(el) #=> el or nil
list.get!(el) #=> raises Exception if not found

Reversing

list.reverse!.to_a #=> [5, 4, 3, 1, 0]

Enumerable methods are also available

list.map{ |e| e.data } #=> [1, 2, 3, 4]
list.inject(0){ |a, e|  a + e.data } #=> 10

Append one list to other list

list1.concat(list2)

Comparable module is included so you can:

Check if lists are equal

list1 == list2

Check if one list is greater than other (same rules as in Array class)

list1 > list2

Other operations

• clone

• clear

• insert_before

• insert_after

• zip?

• looped?

• cycle_size

• to_s

A tree is a data structure with nodes organised in hierarchy. More: Tree

Building Tree

t = Tree.new(2)
c1 = t << 5
c2 = t << 8
t << 9

c1 << 4
c1 << 10
c3 = c2 << 3

Methods:

t.leaf? #=> false
c3.leaf? #=> true

c1.sibblings.map &:data #=> [8, 9]
c1.parent.data #=> 2

t.height #=> 3
t.width #=> 3
t.leaf_count #=> 4

t.levels #=> {1=>1, 2=>3, 3=>3}

Other methods

• get_leaves

• isometric?

• mirror!

Enumerable Module is also included.

t.map { |node| node.data } #=> [2, 5, 8, 9, 4, 10, 3]

BinaryTree is sublass of Tree. In BinaryTree each node can have at most two children. More: BinaryTree

Building tree

bin_tree = BinaryTree.new
[2, 5, 8, 9, 11, 12, 14].each { |x| bin_tree.insert(x) } #builds complete binary Tree

Accessors defined on BinaryTree object:

bin_tree.left.data #=> 5
bin_tree.right.data #=> 8

Red-black tree is symbol table data structure. It’s very simmilar to hash, but internally uses tree (perfect balanced binary tree) and not depends on hash function. Red black trees aren’t as fast as hashes but supports ordered operations.

rb = RedBlackTree.new
rb.insert(:z, 3)
rb.insert(:p, 2)
rb.insert(:a, 1)
rb.get(:a) #=> 1

You can also create RBT by passing hash to constructor

rb = RedBlackTree.new(a: 1, p: 2, z: 3)

Hash like accessors are defined

rb[:z] = 3
rb[:z] #=> 3

You can convert RedBlackTree to Hash with to_h method:

rb.to_h #=> {a: 1, p: 2, z: 3}

Enumerable is included and traversing is ordered by key

rb.map(&:key) #=> [:a, :p, :z]

BinaryHeap is tree in which every node satisfies heap property. Binary Heap allows very fast access to maximum or minimum element of the tree (const access). More: Binary Heap

Creating

Maximum Binary Heap

max_heap = BinaryHeap.new(9, 8, 4, 5, 11, 6)

or

max_heap = BinaryHeap.max(9, 8, 4, 5, 11, 6)

Minimum Binary Heap

min_heap = BinaryHeap.min(9, 8, 4, 5, 11, 6)

or

BinaryHeap.new(9, 8, 4, 5, 11, 6){ |parent, child| parent < child }

You can set heap relation by passing block to BinaryHeap constructor.

Examples

max_heap.shift #returns max element (11)
max_heap.to_a #=> [9, 8, 6, 5, 4]
max_heap.insert 15
max_heap.shift #=> 15

min_heap.shift #returns min element (4)

Trie is an ordered tree data structure which allows very quick search: O(k), where k is word length. More: Trie

Creating

trie = Trie.new

Setting custom alphabet (memory usage depends on alphabet size)

trie.alphabet = %w(a b c d)

Examples

trie.insert('thing', true);
trie.find('thing')  # =>  true
trie.delete('thing')

or

trie['one'] = 'thing'
trie['one']  # =>  'thing'

Enumerable module is included so you can iterate through trie:

trie.map { |k, v| [k, v] } # => [['he', true], ['hello', true], ['help', true]]

Finding all words matching given prefix:

trie.with_prefix('th') # =>  ['the', 'thing']
trie.with_prefix('yeti') # =>  []

Alternatively you can pass block to this method:

trie.with_prefix('th'){ |word, val| res[word] = val } # =>  {'the' => true, 'thing' => true}

b = BinaryTree.new
[2, 5, 8, 9, 11, 12, 14].each{ |x| b.insert(x) }

walker = TreeWalker.new(b)

Iterating in postorder

walker.traverse(:postorder) #=> [9, 11, 5, 12, 14, 8, 2]

Iterating in inorder

walker.traverse(:inorder) #=> [9, 5, 11, 2, 12, 8, 14]

Iterating in preorder

walker.traverse(:preorder) #=> [2, 5, 9, 11, 8, 12, 14]

Iterating in BFS order

walker.each{ |x| x } #=> [2, 5, 8, 9, 11, 12, 14]

You can also pass block to traverse method

walker.traverse(:inorder){ |n| n.data**2 }

If you want to change value of tree nodes, use recalculate! method

walker.recalculate!(b, :preorder, 0) { |e, a| a + e.data }

Simple two dimensional array(matrix). Array2D extends automatically like simple Array.

Creating

discrete_matrix = Array2D.new(2, 0)

First argument is size of row(or column) and second is default value of matrix.

Examples

discrete_matrix.to_a  #=> [[0, 0], [0, 0]]
discrete_matrix[3, 3] #=> 0

Automaticaly fills empty slots with custom value:

arr = ExpandableArray.new(0, 0)
arr[4]  = 1 #=> [0, 0, 0, 0, 4]

Triangular matrix is a special kind of matrix where M = M. More: Triangular Matrix

Creating

tri_matrix = TriMatrix.new
tri_matrix[0, 1] = true
tri_matrix[0, 2] = true

Examples

tri_matrix[0, 1] == tri_matrix[1, 0] #=> true

IndexedSet is a set whose elements are indexed. In opposite to Array, duplicates are not allowed. Internally uses hash for fast access and array for ordering.

Creating new Indexed Set

set = IndexedSet.new

Examples

set.push(:first)  #=>  0
set.push(:second) #=> 1
set.index(:first) #=> 0
set.to_a #=> [:first, :second]