#!/usr/bin/perl -w
# -*- coding: iso-8859-1 -*-
# 	$Id: ispellaff2myspell,v 1.29 2005/07/04 12:21:55 agmartin Exp $
# 
#   (C) 2002-2005 Agustin Martin Domingo <agustin.martin@hispalinux.es> 
# 
#    This program is free software; you can redistribute it and/or modify
#    it under the terms of the GNU General Public License as published by
#    the Free Software Foundation; either version 2 of the License, or
#    (at your option) any later version.
#
#    This program is distributed in the hope that it will be useful,
#    but WITHOUT ANY WARRANTY; without even the implied warranty of
#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#    GNU General Public License for more details.
#
#    You should have received a copy of the GNU General Public License
#    along with this program; if not, write to the Free Software
#    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.


sub usage {
    print "ispellaff2myspell: A program to convert ispell affix tables to myspell format
(C) 2002-2005 Agustin Martin Domingo <agustin.martin\@hispalinux.es>         License: GPL

Usage:
	ispellaff2myspell [options] <affixfile>

      Options:
	--affixfile=s      Affix file
	--bylocale         Use current locale setup for upper/lowercase 
                           conversion
	--charset=s        Use specified charset for upper/lowercase 
                           conversion (defaults to latin1)
 	--debug            Print debugging info
 	--extraflags       Allow some non alphabetic flags
	--lowercase=s      Lowercase string
        --myheader=s       Header file
	--printcomments    Print commented lines in output 
        --replacements=s   Replacements file 
        --split=i          Split flags with more that i entries
	--uppercase=s      Uppercase string
	--wordlist=s       Still unused

  Currently allowed valued for charset are: latin1, latin2, latin3

This script does not create the dict file. Something like

( echo `cat mydict.words+ | wc -l`; cat mydict.words+ ) > mydict.dict

should do the work, with mydict.words+ being the ispell munched wordlist

";
    exit;
}

sub debugprint {
    if ( $debug ){
	print STDERR "@_";
    }
}

sub shipoutflag{
    my $flag_entries=scalar @flag_array;
	
    if ( $flag_entries != 0 ){
	if ( $split ){
	    while ( @flag_array ){
		my @flag_subarray=splice(@flag_array,0,$split);
		my $subflag_entries=scalar @flag_subarray;
		if ( scalar @flag_array ){
		    print "$myaffix $flagname $flagcombine $subflag_entries S\n";
		} else {
		    print "$myaffix $flagname $flagcombine $subflag_entries\n";
		}
		print join("\n",@flag_subarray);
		print "\n\n";
	    }
	} else {
	    print "$myaffix $flagname $flagcombine $flag_entries\n";
	    print join("\n",@flag_array);
	    print "\n\n";
	}
    }
    @flag_array=();
    $flagname='';
    $flagcombine='';
}

sub mylc{
    my $inputstring=shift;
    my $outputstring;

    if ( $bylocale ){
	{ 
	    use locale;
	    $outputstring =  lc $inputstring;
	}
    } else {
	if ( $charset eq "latin0" ){
	    $lowercase='a-zàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþ½¨¸';
	    $uppercase='A-ZÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞ¼¦´';
	} elsif ( $charset eq "latin1" ){
	    $lowercase='a-zàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþ';
	    $uppercase='A-ZÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞ';
	} elsif ( $charset eq "latin2" ){
	    $lowercase='a-z±³µ¶¹º»¼¾¿àáâãäåæçèéêëìíîïðñòóôõöøùúûüýþ';
	    $uppercase='A-Z¡£¥¦©ª«¬®¯ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞ';
	} elsif ( $charset eq "latin3" ){
	    $lowercase='a-z±¶¹º»¼¿àáâäåæçèéêëìíîïñòóôõö÷øùúûüýþ';
	    $uppercase='A-Z¡¦©ª«¬¯ÀÁÂÄÅÆÇÈÉÊËÌÍÎÏÑÒÓÔÕÖ×ØÙÚÛÜÝÞ';
#	} elsif ( $charset eq "other_charset" ){
#	    die "latin2 still unimplemented";
	} else {
	    if ( not $lowercase and not $uppercase ){
		die "Unsupported charset [$charset]

use explicitely --lowercase=string and --uppercase=string
options. Remember that both string must match exactly, but 
case changed.
";
	    }
	}
	$outputstring=$inputstring;
	eval "\$outputstring=~tr/$uppercase/$lowercase/";
    }
    return $outputstring;
}

sub validate_flag (){
    my $flag = shift;
    if ($flag=~m/[a-zA-Z]+/){
	return $flag;
    } elsif ( $hasextraflags ){
	foreach ( keys %theextraflags ){
	    if ($flag =~ m/^$_/){
		$flag =~ s/^$_//;
		return $flag;
	    } 
	}
    } 
    return '';
}

sub process_replacements{
    my $file = shift;
    my @replaces = ();
    
    open (REPLACE,"< $file") || 
	die "Error: Could not open replacements file: $file\n";
    while (<REPLACE>){
	next unless m/^REP[\s\t]*\D.*/;
	next if m/^REP\s+[0-9]+/;
	s/\015\012//;
	s/\015//;
	chomp;
	push @replaces, $_;
    }
    close REPLACE;
    my $number = scalar @replaces;
    print "REP $number\n";
    foreach ( @replaces ){
	print $_ . "\n";
    }
}

# -----------------------------------------------------------
# Now the progran start, after the functions are defined
# -----------------------------------------------------------

use Getopt::Long;

# Initializing option values
$affixfile     = '';
$bylocale      = '';
$charset       = '';
$debug         = '';
$lowercase     = '';
$myheader      = '';
$printcomments = '';
$replacements  = ''; 
$split         = '';
$uppercase     = '';
$wordlist      = '';
$hasextraflags = '';
@flag_array    = ();
%theextraflags = ();
# Initializing root values
$rootremove    = "0";
$rootname      = '';
$addtoroot     = '';
$comment       = '';
# Initializing flag values
$flagname      = '';
$flagcombine   = '';
$inflags       = '';

GetOptions ('affixfile=s'   => \$affixfile,
	    'bylocale'      => \$bylocale,
	    'charset=s'     => \$charset,
	    'debug'         => \$debug,
	    'extraflags:s'  => sub {
		$hasextraflags = 1;
		shift;
		$theflag = shift;
		$theextraflags{$theflag}++ if $theflag},
	    'lowercase=s'   => \$lowercase,
	    'myheader=s'    => \$myheader,
	    'printcomments' => \$printcomments,
	    'replacements=s'=> \$replacements,
	    'split=i'       => \$split,
	    'uppercase=s'   => \$uppercase,
	    'wordlist=s'    => \$wordlist) or usage;

if ( not $affixfile ){
    $affixfile=shift or usage;
}

if ( $charset and ( $lowercase or $uppercase )){
    die "Error: charset and lowercase/uppercase options
are incompatible. Use either charset or lowercase/uppercase options to 
specify the patterns
"
} elsif ( not $lowercase and not $uppercase and not $charset ){
    $charset="latin1";
}

if ( scalar(keys %theextraflags) == 0 && $hasextraflags ){
    $theextraflags{"\\\\"}++;
}

debugprint "$affixfile $charset";

open (AFFIXFILE,"< $affixfile") || 
    die "Error: Could not open affix file: $affixfile";

if ( $myheader ){
    my $myspell_header=`cat $myheader`;
    print $myspell_header . "\n";
}

while (<AFFIXFILE>){
    chomp;
    if (/^\s*\#.*/){
	debugprint "Ignoring line $.\n";
	print "$_\n" if $printcomments;
    } elsif (/^\s*$/){
	debugprint "Ignoring line $.\n";
    } elsif (/^\s*prefixes/){
	debugprint "Prefixes starting in line $.\n";
	$affix="PFX";
    } elsif (/^\s*suffixes/){
	debugprint "Suffixes starting in line $.\n";
	$affix="SFX";
    } elsif (/^[\s\t]*flag.*/){
	next if not $affix;         # In case we are still in the preamble
	shipoutflag if $inflags;
	$inflags="yes";
	s/^[\s\t]*flag[\s\t]*//;
	s/[\s\t]*:.*$//;
	debugprint "Found flag $_ in line $.\n";
	
	if (/\*/){
	    s/[\*\s]//g;
	    $flagcombine="Y";
	    debugprint "Flag renamed to $_ with combine=$flagcombine\n";
	} else {
	    $flagcombine="N";
	}
	
	if ( $flagname = &validate_flag($_) ){
	    $myaffix  = $affix;
	} else {
	    $myaffix  = "\# $affix";
	    $flagname = $_;
	    print STDERR "Ignoring invalid flag $flagname in line $.\n";
	}
    } elsif ( $affix and $inflags ) {
	($rootname,@comments)   =  split('#',$_);
	$comment                =  '# ' . join('#',@comments);
	
	$rootname               =~ s/\s*//g;
	$rootname               =  mylc $rootname;
	($rootname,$addtoroot)  =  split('>',$rootname);
	
	if ( $addtoroot =~ s/^\-//g ){
	    ($rootremove,$addtoroot)  = split(',',$addtoroot);
	    $addtoroot                = "0" unless $addtoroot;
	    $addtoroot                = "0" if ( $addtoroot eq "-");
	} else {
	    $rootremove = "0";
	}
	$addtoroot =~ s/\\\-/\-/g; # prefix ANTI\- to anti-

	if ( $rootname eq '.' && $rootremove ne "0" ){
	    $rootname = $rootremove;
	}
	
	debugprint "$rootname, $addtoroot, $rootremove\n";
	if ( $printcomments ){
	    $affix_line=sprintf("%s %s   %-5s %-11s %-24s %s",
				$myaffix, $flagname, $rootremove, 
				$addtoroot, $rootname, $comment);
	} else {
	    $affix_line=sprintf("%s %s   %-5s %-11s %s",
				$myaffix, $flagname, $rootremove, 
				$addtoroot, $rootname);
	}
	$rootremove = "0";
	$rootname   = '';
	$addtoroot  = '';
	$comment    = '';
	@comments   = ();
	push @flag_array,$affix_line;
	debugprint "$affix_line\n";
    } else {
	#
    }
}
shipoutflag;

close AFFIXFILE;

if ( $replacements ){
    &process_replacements($replacements);
}

__END__

=head1 NAME

B<ispellaff2myspell> - A program to convert ispell affix tables to myspell format.

=head1 SYNOPSIS

 ispellaff2myspell [options] <affixfile> --myheader your_header

   Options:

    --affixfile=s      Affix file
    --bylocale         Use current locale setup for upper/lowercase 
                       conversion
    --charset=s        Use specified charset for upper/lowercase 
                       conversion (defaults to latin1)
    --debug            Print debugging info
    --extraflags=s     Allow some non alphabetic flags
    --lowercase=s      Lowercase string
    --myheader=s       Header file 
    --printcomments    Print commented lines in output 
    --replacements=s   Replacements file 
    --split=i          Split flags with more that i entries
    --uppercase=s      Uppercase string

=head1 DESCRIPTION

B<ispellaff2myspell> is a script that will convert ispell affix tables 
to myspell format in a more or less successful way. 

This script does not create the dict file. Something like

( echo `cat mydict.words+ | wc -l`; cat mydict.words+ ) > mydict.dict

should do the work, with mydict.words+ being the munched wordlist

=head1 OPTIONS

=over 8

=item B<--affixfile=s>  

Affix file. You can put it directly in the command line.

=item B<--bylocale> 

Use current locale setup for upper/lowercase conversion. Make sure 
that the selected locale match the dictionary one, or you might get 
into trouble.

=item B<--charset=s>        

Use specified charset for upper/lowercase conversion (defaults to latin1). 
Currently allowed values for charset are: latin0, latin1, latin2, latin3.

=item B<--debug>            

Print some debugging info.

=item B<--extraflags:s>       

Allows some non alphabetic flags. 

When invoked with no value the supported flags are currently those 
corresponding to chars represented with the escape char B<\> as 
first char. B<\> will be stripped.

When given with the flag prefix will allow that flag and strip the 
given prefix. Be careful when giving the prefix to properly escape chars, 
e.g. you will need B<-e "\\\\"> or B<-e '\\'> for flags like B<\[> to be stripped to 
B<[>. Otherwise you might even get errors. Use B<-e "^"> to allow all 
flags and pass them unmodified.

You will need a call to -e for each flag type, e.g., 
B<-e "\\\\" -e "~\\\\"> (or B<-e '\\' -e '~\\'>). 

When a prefix is explicitely set, the default value (anything starting by B<\>) 
is disabled and you need to enable it explicitely as in previous example.

=item B<--lowercase=s>      

Lowercase string. Manually set the string of lowercase chars. This 
requires B<--uppercase> having exactly that string but uppercase.
 
=item B<--myheader=s>       

Header file. The myspell aff header. You need to write it 
manually. This can contain everything you want to be before the affix table

=item B<--printcomments>    

Print commented lines in output.

=item B<--replacements=file>      

Add a pre-defined replacements table taken from 'file' to the .aff file.
Will skip lines not beginning with REP, and set the replacements number
appropriately.

=item B<--split=i>          

Split flags with more that i entries. This can be of interest for flags 
having a lot of entries. Will split the flag in chunks containing B<i> 
entries.

=item B<--uppercase=s>      

Uppercase string. Manually set the sring of uppercase chars. This 
requires B<--lowercase> having exactly that string but lowercase.

=back

If your encoding is currently unsupported you can send me a file with 
the two strings of lower and uppercase chars. Note that they must match 
exactly but case changed. It will look something like

  $lowercase='a-zàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþ';
  $uppercase='A-ZÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞ';

=head1 SEE ALSO

The OpenOffice.org Lingucomponent Project home page

L<http://lingucomponent.openoffice.org/index.html>

and the document

L<http://lingucomponent.openoffice.org/affix.readme>

that provides information about the basics of the myspell affix file format.

You can also take a look at 

 /usr/share/doc/libmyspell-dev/affix.readme.gz
 /usr/share/doc/libmyspell-dev/README.compoundwords
 /usr/share/doc/libmyspell-dev/README.replacetable

in your Debian system.

=head1 AUTHORS

Agustin Martin <agustin.martin@hispalinux.es>

=cut