0.0
No release in over 3 years
Low commit activity in last 3 years
Pure-Ruby implemention of Red-Black tree, a self-balancing binary search tree with O(log n) search, insert and delete operations.
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
 Dependencies

Development

~> 2.0, >= 2.0.1
~> 13.0
~> 3.8
 Project Readme

Build Status

Tree::RedBlack

  • Description
  • Installation
  • Tree::RedBlack API
    • new(allow_duplicates = true) → red_black_tree
    • insert(value, ...) → red_black_tree
    • delete(value, ...) → red_black_tree
    • search(value, ifnone = nil) → red_black_node
    • bsearch { |node| block } → red_black_node
    • pre_order → node_enumerator
    • in_order → node_enumerator
    • dup → red_black_tree
  • Tree::RedBlackNode API
    • new(value = nil, color = :RED) → red_black_node
    • insert_red_black(value, allow_duplicates = true) → red_black_node
    • delete_red_black(value) → red_black_node
    • search(value, ifnone = nil) → red_black_node
    • bsearch(&block) → red_black_node
    • min() → red_black_node
    • max() → red_black_node
    • pred() → red_black_node
    • succ() → red_black_node
    • pre_order(&block) → red_black_node
    • in_order(&block) → red_black_node
    • dup() → red_black_node
  • Contributing
  • License

Description

The Tree::RedBlack library is a pure-Ruby implementation of a Red-Black tree -- i.e., a self-balancing binary tree with O(log n) search, insert and delete operations. It is appropriate for maintaining a sorted collection where insertion and deletion are desired at arbitrary positions.

This implementation differs slightly from the Wikipedia description referenced above. In particular, leaf nodes are nil, which affects the details of node deletion.

Included are two iRuby Jupyter notebooks for exploring Red-Black trees interactively:

Installation

With a recent version of the Ruby interpreter installed (e.g., ruby 2.5), run the following commands from a Unix shell:

git clone git@github.com:slewsys/tree-red_black.git
cd tree-red_black
sudo gem update --system
bundle
rake build
sudo gem install pkg/tree-red_black*.gem

The RSpec test suite can be run with:

bundle exec rspec spec

To build RDoc documentation, use:

rake rdoc

Tree::RedBlack API

Once a Red-Black tree has been instantiated (see new), any Comparable value can be stored, provided that every value in the tree is comparable with every other. Values are stored in nodes and referenced by the key attribute of the node. Nodes are added to a Red-Black tree by inserting values (see insert), and nodes are removed by deleting values (see delete).

A Red-Black tree's nodes can be enumerated in ascending order by value with the tree's each method. Additional enumeration methods are described in Enumerable.

new(allow_duplicates = true) → red_black_tree

Returns a new, empty Red-Black tree. If option allow_duplicates is false, then only unique values are inserted in a Red-Black tree.

The root attribute references the root node of the tree. The size attribute indicates the number of nodes in the tree. When size is 0, root is always nil.

Example:

require 'tree/red_black'

rbt = Tree::RedBlack.new
p rbt.root                #=> nil
p rbt.size                #=> 0
p rbt.allow_duplicates?   #=> true

insert(value, ...) → red_black_tree

Inserts a value or sequence of values in a Red-Black tree and increments the size attribute by the number of values inserted.

Since a Red-Black tree maintains a sorted, Enumerable collection, every value inserted must be comparable with every other value. Methods each, map, select, find, sort, etc., can be applied directly to the tree.

The individual nodes yielded by enumeration respond to method key to retrieve the value stored in that node. Method each, in particular, is aliased to in_order, so that nodes are sorted in ascending order by key value. Nodes can also be traversed by method pre_order, e.g., to generate paths in the tree.

Example:

require 'tree/red_black'

rbt = Tree::RedBlack.new
rbt.insert(*1..10)        # #<Tree::RedBlack:0x00...>
p rbt.size                #=> 10
rbt.map(&:key)            #=> [1, 2, ..., 10]
rbt.select { |node| node.key % 2 == 0 }.map(&:key)
                          #=> [2, 4, ..., 10]

delete(value, ...) → red_black_tree

Deletes a value or sequence of values from a Red-Black tree.

Example:

require 'tree/red_black'

rbt = Tree::RedBlack.new
rbt.insert(*1..10)        # #<Tree::RedBlack:0x00...>
p rbt.size                #=> 10
rbt.delete(*4..8)         # #<Tree::RedBlack:0x00...>
p rbt.size                #=> 5
rbt.map(&:key)            #=> [1, 2, 3, 9, 10]

search(value, ifnone = nil) → red_black_node

Returns a Red-Black tree node whose key matches value by binary search. If no match is found, calls non-nil ifnone, otherwise returns nil.

Example:

require 'tree/red_black'

shuffled_values = [*1..10].shuffle
rbt = shuffled_values.reduce(Tree::RedBlack.new) do |acc, v|
  acc.insert(v)
end
rbt.search(7)             #=> <Tree::RedBlackNode:0x00..., @key=7, ...>

bsearch { |node| block } → red_black_node

Returns a Red-Black tree node satisfying a criterion defined in block by binary search.

If block evaluates to true or false, returns the first node for which the block evaluates to true. In this case, the criterion is expected to return false for nodes preceding the matching node and true for subsequent nodes.

Example:

require 'tree/red_black'

shuffled_values = [*1..10].shuffle
rbt = shuffled_values.reduce(Tree::RedBlack.new) do |acc, v|
  acc.insert(v)
end
rbt.bsearch { |node| node.key >= 7 }
                          #=> <Tree::RedBlackNode:0x00..., @key=7, ...>

If block evaluates to <0, 0 or >0, returns first node for which block evaluates to 0. Otherwise returns nil. In this case, the criterion is expected to return <0 for nodes preceding the matching node, 0 for some subsequent nodes and >0 for nodes beyond that.

Example:

require 'tree/red_black'

shuffled_values = [*1..10].shuffle
rbt = shuffled_values.reduce(Tree::RedBlack.new) do |acc, v|
  acc.insert(v)
end
rbt.bsearch { |node| 7 <=> node.key }
                          #=> <Tree::RedBlackNode:0x00..., @key=7, ...>

If block is not given, returns an enumerator.

pre_order → node_enumerator

Returns an enumerator for nodes in a Red-Black tree by pre-order traversal.

Example:

require 'tree/red_black'

rbt = Tree::RedBlack.new
shuffled_values = [*1..10].shuffle  #=> [5, 9, 10, 8, 7, 6, 1, 2, 4, 3]
rbt.insert(*shuffled_values)        #=> #<Tree::RedBlack:0x00...>
rbt.pre_order.map(&:key)            #=> [7, 5, 2, 1, 4, 3, 6, 9, 8, 10]

in_order → node_enumerator

Returns an enumerator for nodes in a Red-Black tree by in-order traversal. The each method is aliased to in_order.

Example:

require 'tree/red_black'

rbt = Tree::RedBlack.new
shuffled_values = [*1..10].shuffle
rbt.insert(*shuffled_values)
rbt.in_order.map(&:key)             #=> [1, 2, ... 9, 10]

dup → red_black_tree

Returns a deep copy of a Red-Black tree, provided that the dup method for the key attribute of a tree node is also a deep copy.

Example:

require 'tree/red_black'

rbt = Tree::RedBlack.new
rbt.insert({a: 1, b: 2})
rbt_copy = rbt.dup
p rbt.root.key            #=> {:a=>1, :b=>2}
p rbt.root.key.delete(:a) #=> 1
p rbt.root.key            #=> {:b=>2}
p rbt_copy.root.key       #=> {:a=>1, :b=>2}

Tree::RedBlackNode API

A Red-Black tree is a collection of nodes arranged as a binary tree. In addition to the binary node attributes left and right, which reference the left and right sub-trees of a given node, and the attribute key which stores a node's data, a Red-Black tree node also has attributes color and parent.

The color attribute is used internally to balance the tree after a node is inserted or deleted. The parent attribute references a node's parent node, or nil in the case of the root node of a tree.

Not all implementations of Red-Black trees have nodes with parent attributes, but it's generally recognized as the most efficient way of balancing and re-coloring a Red-Black tree.

Since the data in a binary tree can be thought of as a sorted collection, it's convenient to be able to refer to a node's predecessor and successor (i.e., the node whose key is the predecessor or successor in the sorted ascending order. In general, this differs from parent or child node). This is provided as the node methods pred and succ. And for a given sub-tree, its convenient to be able to refer to its min and max nodes (i.e., the node whose key is a minimum or maximum in the sub-tree). This is provided as the node methods min and max.

While a Red-Black tree can be constructed from nodes alone, the Tree::RedBlack API provides a cleaner way of working with Red-Black trees. Start there if only using the Red-Black tree as a container.

new(value = nil, color = :RED) → red_black_node

Returns a new node with key parameter set to option value. The color option, if given, must be be either :RED or :BLACK.

Example:

require 'tree/red_black'

root = Tree::RedBlackNode.new(10)
p root.key              #=> 10
p root.color            #=> :RED
p root.parent           #=> nil
p root.left             #=> nil
p root.right            #=> nil

insert_red_black(value, allow_duplicates = true) → red_black_node

Inserts the given value in a tree whose root node is self. If the key attribute of the root node is nil, then value is assigned to key. Otherwise, value is used to instantiate Tree::RedBlackNode, and the node is inserted in the tree; the tree is then re-balanced as needed, and the root of the balanced tree returned.

Since a Red-Black tree maintains an ordered, Enumerable collection, every value inserted must be Comparable with every other value. Methods each, map, select, find, sort, etc., can be applied to a Red-Black tree's root node to iterate over all nodes in the tree.

Each node yielded by enumeration has a key attribute to retrieve the value stored in that node. Method each, in particular, is aliased to in_order, so that nodes are sorted in ascending order by key value. Nodes can also be traversed by method pre_order, e.g., to generate paths in the tree.

Example:

require 'tree/red_black'

root = Tree::RedBlackNode.new
p root.key                      #=> nil
root = root.insert_red_black(0)
p root.key                      #=> 0
root = root.insert_red_black(1)
p root.key                      #=> 0
p root.left                     #=> nil
p root.right.key                #=> 1
root = root.insert_red_black(2)
p root.key                      #=> 1
p root.left.key                 #=> 0
p root.right.key                #=> 2

delete_red_black(value) → red_black_node

Deletes the given value from a tree whose root node is self. If the tree has only one remaining node and its key attribute matches value, then the remaining node's key attribute is set to nil but the node itself is not removed. Otherwise, the first node found whose key matches value is removed from the tree, and the tree is re-balanced. The root of the balanced tree is returned.

Example:

require 'tree/red_black'

root = [*1..10].reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root = [*4..8].reduce(root) do |acc, v|
  acc.delete_red_black(v)
end
root.map(&:key)                 #=> [1, 2, 3, 9, 10]

search(value, ifnone = nil) → red_black_node

Returns a node whose key matches value by binary search. If no match is found, calls non-nil ifnone, otherwise returns nil.

Example:

require 'tree/red_black'

shuffled_values = [*1..10].shuffle
root = shuffled_values.reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root.search(7)        #=> <Tree::RedBlackNode:0x00..., @key=7, ...>

bsearch(&block) → red_black_node

Returns a node satisfying a criterion defined in block by binary search.

If block evaluates to true or false, returns the first node for which the block evaluates to true. In this case, the criterion is expected to return false for nodes preceding the matching node and true for subsequent nodes.

Example:

require 'tree/red_black'

shuffled_values = [*1..10].shuffle
rbt = shuffled_values.reduce(Tree::RedBlack.new) do |acc, v|
  acc.insert(v)
end
rbt.bsearch { |node| node.key >= 7 }
                      #=> <Tree::RedBlackNode:0x00..., @key=7, ...>

If block evaluates to <0, 0 or >0, returns first node for which block evaluates to 0. Otherwise returns nil. In this case, the criterion is expected to return <0 for nodes preceding the matching node, 0 for some subsequent nodes and >0 for nodes beyond that.

Example:

require 'tree/red_black'

shuffled_values = [*1..10].shuffle
rbt = shuffled_values.reduce(Tree::RedBlack.new) do |acc, v|
  acc.insert(v)
end
rbt.bsearch { |node| 7 <=> node.key }
                      #=> <Tree::RedBlackNode:0x00..., @key=7, ...>

If block is not given, returns an enumerator.

min() → red_black_node

Returns the node whose key is a minimum in the sub-tree with root self.

Example:

require 'tree/red_black'

root = [*0..10].reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root                  #=> <Tree::RedBlackNode:0x00..., @key=4, ...>
root.min              #=> <Tree::RedBlackNode:0x00..., @key=0, ...>
root.right            #=> <Tree::RedBlackNode:0x00..., @key=6, ...>
root.right.min        #=> <Tree::RedBlackNode:0x00..., @key=5, ...>

max() → red_black_node

Returns the node whose key is a maximum in the sub-tree with root self.

Example:

require 'tree/red_black'

root = [*0..10].reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root                  #=> <Tree::RedBlackNode:0x00..., @key=4, ...>
root.max              #=> <Tree::RedBlackNode:0x00..., @key=10, ...>
root.left             #=> <Tree::RedBlackNode:0x00..., @key=2, ...>
root.left.max         #=> <Tree::RedBlackNode:0x00..., @key=3, ...>

pred() → red_black_node

Returns the node preceding self, or nil if no predecessor exists. If duplicate keys are allowed, it's possible that pred.key == key.

Example:

require 'tree/red_black'

root = [*1..10].reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root.right.right.key            #=> 8
root.right.right.pred.key       #=> 7

succ() → red_black_node

Returns the node succeeding self, or nil if no successor exists. If duplicate keys are allowed, it's possible that succ.key == key.

Example:

require 'tree/red_black'

root = [*1..10].reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root.right.right.key            #=> 8
root.right.right.succ.key       #=> 9

pre_order(&block) → red_black_node

Returns an enumerator for nodes in the tree with root self by pre-order traversal.

Example:

require 'tree/red_black'

root = [*1..10].reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root.pre_order.map(&:key)       #=> [4, 2, 1, 3, 6, 5, 8, 7, 9, 10]

in_order(&block) → red_black_node

Returns an enumerator for nodes in the tree with root self by in-order traversal.

Example:

require 'tree/red_black'

root = [*1..10].reduce(Tree::RedBlackNode.new) do |acc, v|
  acc.insert_red_black(v)
end
root.in_order.map(&:key)        #=> [1, 2, ..., 10]

dup() → red_black_node

Returns a deep copy of the tree with root self, provided that the dup method for the key attribute of a node is also a deep copy.

Example:

require 'tree/red_black'

root = Tree::RedBlackNode.new({a: 1, b: 2})
root_copy = root.dup
p root.key                      #=> {:a=>1, :b=>2}
p root.key.delete(:a)           #=> 1
p root.key                      #=> {:b=>2}
p root_copy.key                 #=> {:a=>1, :b=>2}

Contributing

Bug reports and pull requests can be sent to GitHub tree-red_black.

License

This Rubygem is free software. It can be used and redistributed under the terms of the MIT License.