/ 
AVLTree-0.01
River stage zero No dependents
/ AVLTree

NAME

AVLTree - Perl extension for efficient creation and manipulation of AVL balanced binary trees.

VERSION

Version 0.01

DESCRIPTION

This module provides a simple and fast implementation of AVL balanced trees. It uses the Perl XS extension mechanism by providing a tiny wrapper around an efficient C library which does the core of the work. Preliminary benchmarking shows this module one order of magnitude faster than a pure perl implementation.

The nodes of an AVL tree object can hold any kind of item, as long as each one of these has an element which can be used to define a partial order on the set of possible items. This is specified by providing, upon tree construction, a reference to a function for comparing any two of the possible items.

The underlying C library is a reinterpretation of the C library originally developed by Julienne Walker. This library has been adapted for dealing directly with Perl (SV) variables.

The module at the moment is in beta stage but it is usable. It provides methods for creating and querying an AVL tree, get its size and insert and remove elements from it. No methods exist to traverse the tree at this stage, but I promise this functionality is going to be implemented very soon.

SYNOPSIS

 use AVLTree;

 # Define a function to compare two numbers i1 and i2,
 # return -1 if i1 < i2, 1 if i2 > i1 and 0 otherwise 

 sub cmp_f = sub {
   my ($i1, $i2) = @_;

   return $i1<$i2?-1:($i1>$i2)?1:0;
 }

 # Instantiate a tree which holds numbers
 my $tree = AVLTree->new(\&cmp_f);
 
 # Add some numbers to the tree
 map { $tree->insert($_) } qw/10 20 30 40 50 25/;

 # Now invoke some useful methods
 # Size of the tree
 printf "Size of the tree: %d\n", $tree->size();
 
 # Query the tree
 my $query = 30;
 print "Query: %d, Found: %d\n", $query, $tree->find($query)?1:0;

 # Remove an item
 my $item = 1
 if($tree->remove($item)) {
   print "Item $item has been removed\n";
 } else {
   print "Item $item was not in the tree so it's not been removed\n";
 }
 
 printf "Size of tree is now: %d\n", $tree->size();

 ...

 # Suppose you want the tree to hold generic data items, e.g. hashrefs
 # which hold some data. We can deal with these by definying a custom
 # comparison function based on one of the attributes of these data items, 
 # e.g. 'id':

 sub compare {
   my ($i1, $i2) = @_;
   my ($id1, $id2) = ($i1->{id}, $i2->{id});

   croak "Cannot compare items based on id"
     unless defined $id1 and defined $id2;
 
   return $id1<$id2?-1:($id1>$id2)?1:0;
 }

 # Now can do the same as with numbers
 my $tree = AVLTree->new(\&compare);

 my $insert_ok = $tree->insert({ id => 10, data => 'ten' });
 croak "Could not insert item" unless $insert_ok;

 $insert_ok = $tree->insert({ id => 20, data => 'twenty' });
 
 ...

 my $id = 10;
 my $result = $tree->find({ id => $id });
 if($result) {
   printf "Item with id %d found\nData: %s\n", $id, $result->{data};
 } else {
   print "Item with id $id not found\n";
 }

 ...

METHODS

new

Arg [1]     : (required) A reference to a subroutine

Example     : my $tree->new(\&compare);
              carp "Unable to instantiate tree" unless defined $tree;

Description : Creates a new AVL tree object.
              The objects hold by the tree are implicitly defined
              by the provided callback.

Returntype  : AVLTreePtr or undef if unable to instantiate
Exceptions  : None
Caller      : General
Status      : Unstable, interface might change to accomodate suitable defaults, 
              e.g. numbers

find

Arg [1]     : Item to search, can be defined just in terms of the attribute
              with which the items in the tree are compared. 

Example     : $tree->find({ id => 10 }); # objects in the tree can hold data as well
              if($result) {
                printf "Item with id %d found\nData: %s\n", $id, $result->{data};
              } else { print "Item with id $id not found\n"; }

Description : Query if an item exists in the tree.

Returntype  : The item, if found, as stored in the tree or undef
              if the item was not found or the query was not provided
              or it was undefined.
Exceptions  : None
Caller      : General
Status      : Unstable

insert

Arg [1]     : An item to insert in the tree.

Example     : my $ok = $tree->insert({ id => 10, data => 'ten' });
              croak "Unable to insert 10" unless $ok;

Description : Insert an item in the tree, use the provided, upon tree construction,
              comparison function to determine the position of the item in the tree

Returntype  : Bool, true if the item was successfully installed, false otherwise 
Exceptions  : None
Caller      : General
Status      : Unstable

remove

Arg[1]      : An item to remove from the tree

Example     : my $ok = $tree->remove({ id => 10 });
              croak "Unable to remove 10" unless $ok;

Description : Remove an item from the tree.

Returntype  : Bool, true if the item was successfully installed, false otherwise 
Exceptions  : None
Caller      : General
Status      : Unstable

size

 Arg[...]    : None
 
 Example     : print "Size of the tree is: %d\n", $tree->size();

 Description : Returns the size of the tree (number of nodes)

 Returntype  : Int, the size of the tree

 Exceptions  : None
 Caller      : General
 Status      : Unstable

DEPENDENCIES

AVLTree requires Carp and Test::More and Test::LeakTrace to run the tests during installation. If you want to run the benchmarks in the scripts directory, you need to install the Benchmark and List::Util modules.

CHANGES

0.01 2017-12-18 * First release to CPAN, partial port of jsw_avltree AVL tree C library, traversal methods not implemented.

EXPORT

None

SEE ALSO

There are of course other modules which provide this functionality, see e.g. Tree::AVL, Btrees.

You can appreciate the power of this module by running some benchmarking against the above. If you've installed from source, go to the installation directory and:

cd scripts
perl benchmarking.pl

Preliminary experiments suggest speed gains of one order of magnitude.

To the best of my knowledge, there are no modules using Perl XS attempting to implement AVL trees in the most efficient possible way. The closest thing is Tree::Fat, a Perl extension to implement Fat-Node trees.

AUTHOR

Alessandro Vullo, <avullo at cpan.org>

BUGS

Please report any bugs or feature requests to bug-avltree at rt.cpan.org, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=AVLTree. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.

CONTRIBUTING

You can obtain the most recent development version of this module via the GitHub repository at https://github.com/avullo/AVLTree. Please feel free to submit bug reports, patches etc.

SUPPORT

You can find documentation for this module with the perldoc command.

perldoc AVLTree

You can also look for information at:

ACKNOWLEDGEMENTS

I am very grateful to Julienne Walker for generously providing the source code of his production quality C library for handling AVL balanced trees.

Julienne's library can be found at:

http://www.eternallyconfuzzled.com/Libraries.aspx

LICENSE AND COPYRIGHT

Copyright 2017 Alessandro Vullo.

This program is free software; you can redistribute it and/or modify it under the terms of either: the GNU General Public License as published by the Free Software Foundation; or the Artistic License.

See http://dev.perl.org/licenses/ for more information.