# ###################################################################### Otakar Smrz, 2003/01/23 # # Encoding of Arabic: Tim Buckwalter's Notation ##################################### 2003/06/19 package Encode::Arabic::Buckwalter; our $VERSION = '14.1'; use 5.008; use strict; use warnings; use Scalar::Util 'blessed'; use Encode::Encoding; use base 'Encode::Encoding'; __PACKAGE__->Define('Buckwalter', 'Tim'); our $enmode; our $demode; our $optxml; our %modemap = ( 'default' => 0, 'undef' => 0, 'fullvocalize' => 0, 'full' => 0, 'nowasla' => 4, 'vocalize' => 3, 'nosukuun' => 3, 'novocalize' => 2, 'novowels' => 2, 'none' => 2, 'noshadda' => 1, 'noneplus' => 1, ); # use subs 'encoder', 'decoder'; # ignores later prototypes sub encoder ($); # respect prototypes sub decoder ($); # respect prototypes sub import { # perform import as if Encode were used one level before this module $optxml = defined $_[1] && $_[1] eq ':xml' ? 1 : 0; __PACKAGE__->enmode('full'); __PACKAGE__->demode('full'); splice @_, 1, 1; require Encode; push @Encode::ISA, 'Exporter' unless Encode->can('export_to_level'); Encode->export_to_level(1, @_); } sub encode ($$;$) { my (undef, $text, $check) = @_; $_[1] = '' if $check; # needed by in-place edit $text = encoder $text; $text = Encode::encode "utf8", $text if Encode::is_utf8($text); return $text; } sub decode ($$;$) { my (undef, $text, $check) = @_; $_[1] = '' if $check; # needed by in-place edit $text = Encode::decode "utf8", $text unless Encode::is_utf8($text); $text = decoder $text; return $text; } sub enmode ($$;$$) { my ($cls, $mode, $xml, $kshd) = @_; $cls = blessed $cls if ref $cls; $xml = $optxml unless defined $xml; $mode = 'undef' unless defined $mode; $mode = $modemap{$mode} if exists $modemap{$mode}; no strict 'refs'; my $return = ${ $cls . '::enmode' }; if (defined $mode) { ${ $cls . '::enmode' } = $mode; my @set = ( ( $kshd ? '' : q [\x{0640}] ) . q [\x{0623}\x{0624}\x{0625}] . q [\x{060C}\x{061B}\x{061F}] . q [\x{0621}\x{0622}\x{0626}-\x{063A}\x{0641}-\x{064A}] . q [\x{067E}\x{0686}\x{0698}\x{06A4}\x{06AF}] . q [\x{0660}-\x{0669}] . q [\x{0671}] . q [\x{0651}] . q [\x{064B}-\x{0650}\x{0670}] . q [\x{0652}] . ( $kshd ? q [\x{0640}] : '' ) , ( $kshd ? '' : q [_] ) . ( $xml ? q [OWI] : q [>&<] ) . q [,;?] . q ['|}AbptvjHxd*rzs$SDTZEgfqklmnhwYy] . q [PJRVG] . q [0-9] . ( $mode == 0 ? q [{] : q [A] ) . ( $mode == 1 ? '' : q [~] . ( $mode == 2 ? '' : q [FNKaui`] . ( $mode == 3 ? '' : q [o] ) ) ) ); undef &encoder; eval q / sub encoder ($) { $_[0] =~ tr[/ . $set[0] . q /] [/ . $set[1] . q /]d; return $_[0]; } /; } return $return; } sub demode ($$;$$) { my ($cls, $mode, $xml, $kshd) = @_; $cls = blessed $cls if ref $cls; $xml = $optxml unless defined $xml; $mode = 'undef' unless defined $mode; $mode = $modemap{$mode} if exists $modemap{$mode}; no strict 'refs'; my $return = ${ $cls . '::demode' }; if (defined $mode) { ${ $cls . '::demode' } = $mode; my @set = ( ( $kshd ? '' : q [_] ) . ( $xml ? q [OWI] : q [>&<] ) . q [,;?] . q ['|}AbptvjHxd*rzs$SDTZEgfqklmnhwYy] . q [PJRVG] . q [0-9] . q [{] . q [~] . q [FNKaui`] . q [o] . ( $kshd ? q [_] : '' ) , ( $kshd ? '' : q [\x{0640}] ) . q [\x{0623}\x{0624}\x{0625}] . q [\x{060C}\x{061B}\x{061F}] . q [\x{0621}\x{0622}\x{0626}-\x{063A}\x{0641}-\x{064A}] . q [\x{067E}\x{0686}\x{0698}\x{06A4}\x{06AF}] . q [\x{0660}-\x{0669}] . ( $mode == 0 ? q [\x{0671}] : q [\x{0627}] ) . ( $mode == 1 ? '' : q [\x{0651}] . ( $mode == 2 ? '' : q [\x{064B}-\x{0650}\x{0670}] . ( $mode == 3 ? '' : q [\x{0652}] ) ) ) ); undef &decoder; eval q / sub decoder ($) { $_[0] =~ tr[/ . $set[0] . q /] [/ . $set[1] . q /]d; return $_[0]; } /; } return $return; } 1; __END__ =head1 NAME Encode::Arabic::Buckwalter - Tim Buckwalter's transliteration of Arabic =head1 SYNOPSIS use Encode::Arabic::Buckwalter; # imports just like 'use Encode' would, plus more while ($line = <>) { # Tim Buckwalter's mapping into the Arabic script print encode 'utf8', decode 'buckwalter', $line; # 'Buckwalter' alias 'Tim' } # shell filter of data, e.g. in *n*x systems instead of viewing the Arabic script proper % perl -MEncode::Arabic::Buckwalter -pe '$_ = encode "buckwalter", decode "utf8", $_' # employing the modes of conversion for filtering and trimming Encode::Arabic::enmode 'buckwalter', 'nosukuun', '>&< xml'; Encode::Arabic::Buckwalter->demode(undef, undef, 'strip _'); $decode = "Aiqora>o h`*aA {l_n~a_S~a bi___{notibaAhK."; $encode = encode 'buckwalter', decode 'buckwalter', $decode; # $encode eq "AiqraO h`*aA Aln~aS~a biAntibaAhK." =head1 DESCRIPTION Tim Buckwalter's notation is a one-to-one transliteration of the Arabic script for Modern Standard Arabic, using lower ASCII characters to encode the graphemes of the original script. This system has been very popular in Natural Language Processing, however, there are limits to its applicability due to numerous non-alphabetic codes involved. =head2 IMPLEMENTATION The module takes care of the L<Encode::Encoding|Encode::Encoding> programming interface, while the effective code is Tim Buckwalter's C<tr>ick: $encode =~ tr[\x{060C}\x{061B}\x{061F}\x{0621}-\x{063A}\x{0640}-\x{0652} # !! no break in true perl !! \x{0670}\x{0671}\x{067E}\x{0686}\x{0698}\x{06A4}\x{06AF}\x{0660}-\x{0669}] [,;?'|>&<}AbptvjHxd*rzs$SDTZEg_fqklmnhwYyFNKaui~o`{PJRVG0-9]; $decode =~ tr[,;?'|>&<}AbptvjHxd*rzs$SDTZEg_fqklmnhwYyFNKaui~o`{PJRVG0-9] [\x{060C}\x{061B}\x{061F}\x{0621}-\x{063A}\x{0640}-\x{0652} # !! no break in true perl !! \x{0670}\x{0671}\x{067E}\x{0686}\x{0698}\x{06A4}\x{06AF}\x{0660}-\x{0669}]; =head2 EXPORTS & MODES If the first element in the list to C<use> is C<:xml>, the alternative mapping is introduced that suits the B<XML etiquette>. This option is there only to replace the C<< >&< >> reserved characters by C<OWI> while still having a one-to-one notation. There is no XML parsing involved, and the markup would get distorted if subject to C<decode>! $using_xml = eval q { use Encode::Arabic::Buckwalter ':xml'; decode 'buckwalter', 'OWI' }; $classical = eval q { use Encode::Arabic::Buckwalter; decode 'buckwalter', '>&<' }; # $classical eq $using_xml and $classical eq "\x{0623}\x{0624}\x{0625}" The module exports as if C<use Encode> also appeared in the package. The other C<import> options are just delegated to L<Encode|Encode> and imports performed properly. The B<conversion modes> of this module allow to override the setting of the C<:xml> option, in addition to filtering out diacritical marks and stripping off I<kashida>. The modes and aliases relate like this: our %Encode::Arabic::Buckwalter::modemap = ( 'default' => 0, 'undef' => 0, 'fullvocalize' => 0, 'full' => 0, 'nowasla' => 4, 'vocalize' => 3, 'nosukuun' => 3, 'novocalize' => 2, 'novowels' => 2, 'none' => 2, 'noshadda' => 1, 'noneplus' => 1, ); =over =item enmode (I<$obj,> $mode, $xml, $kshd) =item demode (I<$obj,> $mode, $xml, $kshd) These methods can be invoked directly or through the respective functions of L<Encode::Arabic|Encode::Arabic>. The meaning of the extra parameters follows from the L<examples of usage|/SYNOPSIS>. =back =head1 SEE ALSO L<Encode::Arabic|Encode::Arabic>, L<Encode|Encode>, L<Encode::Encoding|Encode::Encoding> Tim Buckwalter's Qamus L<http://www.qamus.org/> Buckwalter Arabic Morphological Analyzer L<http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2002L49> =head1 AUTHOR Otakar Smrz C<< <otakar-smrz users.sf.net> >>, L<http://otakar-smrz.users.sf.net/> =head1 COPYRIGHT AND LICENSE Copyright (C) 2003-2012 Otakar Smrz This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself. =cut