NAME

Statistics::SparseVector - Perl5 extension for manipulating sparse bitvectors

SYNOPSIS

 use Statistics::SparseVector;

 # methods that create new bitvectors
 $vec = Statistics::SparseVector->new($n);
 $vec2 = $vec1->Clone();
 $vec = Statistics::SparseVector->new_Enum($s, $n);
 $vec = Statistics::SparseVector->new_Bin($s, $n);

 # miscellaneous
 $vec2->Substitute_Vector($vec1, $of22, $len2, $off1, $len1);
 $vec->Size();
 $vec->to_Enum();
 $vec->to_Bin();
 $vec->Fill();
 $vec->Empty();
 $vec->increment();
 $n = $vec->Norm();
 @list = $vec->indices();

 # manipulation on the bit level
 $vec->Bit_Off($i);
 $vec->Bit_On($i);
 $vec->bit_flip($i);
 $vec->bit_test($i);

 # overloaded operators
 # increment
 $vec++;
 # stringify
 "$vec"

DESCRIPTION

This module implements sparse bitvectors. Several methods for manipulating bitvectors are implemented. On purpose the naming of the methods is identical to the Bit::Vector package by Stephen Beyer; if you find your vectors to be sparse (have little bits that are on) you can easily switch to a less memory consuming representation.

Creation of bitvectors

new

 $vec = Statistics::BitVector->new($n);

A bitvector of length $n is created. All bits are zero.

Clone

 $clone = $vec->Clone();

A copy of $vec is returned.

new_Enum

 $vec = Statistics::BitVector->new_Enum($enumstring, $n);

A new bitvector of length $n is created from the comma-separated list of in $enumstring.

new_Bin

 $vec = Statistics::BitVector->new_Bin($bitstring, $n);

A new bitvector of length $n is created from bitstring $bitstring.

Vector-wide manipulation of bits

Substitute_Vector

 $vec2->Substitute_Vector($vec1, $off2, $len2, $off1, $len1);

$len2 contiguous bits in target vector $vec2 starting from $off2 are replaced by $len1 contiguous bits from source vector $vec1 starting at bit $off1. If $off2 equals the length of $vec2 the bits from $vec1 are appended. If $len1 is zero the $len2 bits from $vec2 are deleted.

Fill

 $vec->Fill();

All bits of $vec are set to one.

Empty

 $vec->Empty();

All bits of $vec are set to zero.

increment

 $vec->increment(); $vec++;

The integer value of the bitvector is increased by one.

Bit_Off

 $vec->Bit_Off($i);

Bit $i is set to zero.

Bit_On

 $vec->Bit_On($i);

Bit $i is set to one.

bit_flip

 $vec->bit_flip($i);

Bit $i is flipped.

bit_test

 $vec->bit_test($i);

Returns 1 if bit $i is one, 0 otherwise.

Miscellany

Size

 $n = $vec->Size();

Returns the size of the vector.

to_Enum

 $enumstring = $vec->to_Enum();

Returns a comma-separated list of bits that are set.

indices

Returns an array of indices of bits that are set.

to_Bin

 $bitstring = $vec->to_Bin();

Returns a bitstring; bits should be read from right to left.

Norm

Returns the number of set bits.

Overloaded operators

++

 $vec++;

Same as method increment.

Double quotes

 $string = "$vec";

Data::Dumper wants to stringify vectors. Probably because Statistics::SparseVector is an overloaded package it expects double quotes to be overloaded as well.

Remarks about the implementation

• Internally sparse vectors are represented by hashes.

• Only a few methods from Bit::Vector are implemented. Maybe new ones will follow in the future.

• Method Substitute_Vector is not thorougly debugged.

VERSION

Version 0.1.

AUTHOR

SEE ALSO

perl(1), Bit::Vector(3), Statistics::MaxEntropy(3), Statistics::ME.wrapper.pl(3), Statistics::Candidates(3).

COPYRIGHT

Statistics::MaxEntropy comes with ABSOLUTELY NO WARRANTY and may be copied only under the terms of the GNU Library General Public License (version 2, or later), which may be found in the distribution.