#!/usr/bin/perl

use strict;
use vars qw($opt_in_cs $opt_out_cs $opt_help $usage
	    $in_cs $out_cs
	    $i @chars
	    $char $char_id $out_char $omegadb_home
	    $ids $ids_argc %ids $idsdb
	    $idsdata_file $ids_start $font_start
	    %utf8mcs_map_from
	    %cmap_to
	    $inotp $perl56 $perl58
	    $useCDP $useHZK $useGT
	    @CDP @HZK @GT
	    );
use Getopt::Long;
use utf8;
use Chise_utils ':all';

my $strictly_forbidden_after = '「【『［（〈“‘‘（〔｛《{\[\(\x{3016}｛｢';
#       \x{3016} | # white 【

my $forbidden_after = "\x{0000}";

# ￥¥＄$〒♯＃#￠¢￡£＠@§
my $slightly_forbidden_after = '￥¥＄$〒♯＃#￠¢￡£＠@§';

# $strictly_forbidden_before
# All these characters are allowed to protrude
# in the right margin
my $strictly_forbidden_before=
    '!,.:;?、。！，．：；？。\)#}’”〉》」』】〕\x{3017}）］｝｝」\]';
#       \x{3017} | # white 】

my $forbidden_before
    = 'ー々ぁぃぅぇぉゃゅょっゎァィゥェォャュョッヮヵヶ';

my $slightly_forbidden_before
    = '\x{000a}\#\-‐−‰′″℃゛゜ゝゞヽヾ＂％－ﾞﾟ';

my $asian = '\x{1100}-\x{11FF}\x{2E80}-\x{D7AF}\x{F900}-\x{FAFF}\x{FE30}-\x{FE4F}\x{FF00}-';

my $space = '\x{0020}\x{0009}\x{000A}\x{000C}\x{000D}';

$useGT=1;
$useHZK=0;
$useCDP=0;

if($^V and $^V ge v5.8){
    $perl58=1;
}elsif($^V and $^V ge v5.6){
    $perl56=1;
}else{
    print STDERR "This versin is not supported.";
}
if($perl58){
    eval "use Encode";
    binmode(STDIN, ':encoding(utf8)');
    binmode(STDOUT, ':encoding(utf8)');
}

$omegadb_home="/usr/local/lib/chise/omega/db";

&GetOptions("in=s"=>\$opt_in_cs,
	    "i=s"=>\$opt_in_cs,
	    "out=s"=>\$opt_out_cs,
	    "o=s"=>\$opt_out_cs,
	    "help",\$opt_help,
	    "h",\$opt_help);

$usage=<<EOF;
Usage: $0 -i <input coding system> -o <cmap encoding>
    input coding system:
      Utf8mcs, Utf8cns, Utf8gb, Utf8jis, Utf8ks
    cmap encoding:
      UniCNS, UniGB, UniJIS, UniKS
EOF

if($opt_in_cs or $opt_out_cs){
    $in_cs=$opt_in_cs;
    $out_cs=$opt_out_cs;
}elsif(@ARGV==0){
    ($in_cs,$out_cs)=($0=~/(Utf8.+)To(\w+)/);
    $inotp=1;
}

# $in_cs:
#   utf-8-mcs,utf-8-cns,utf-8-gb,utf-8-jis,utf-8-ks,
# $out_cs:
#   UniCNS,UniGB,UniJIS,UniKS

if($opt_help
   or not defined($in_cs)
   or not defined($out_cs)){
    print $usage;
    exit 1;
}

$idsdata_file="idsdata.pl";
$ids_start=0x00; 
$font_start=0;

if(-e $idsdata_file){
    require $idsdata_file;
}

$ids_argc=0;
$ids="";

@GT=(#"=gt","=gt-k",
     "=gt-pj-1","=gt-pj-2","=gt-pj-3","=gt-pj-4","=gt-pj-5","=gt-pj-6","=gt-pj-7","=gt-pj-8","=gt-pj-9","=gt-pj-10","=gt-pj-11"
     #,"=gt-pj-k1","=gt-pj-k2"
     );
@HZK=("=hanziku-1","=hanziku-10","=hanziku-11","=hanziku-12","=hanziku-2","=hanziku-3","=hanziku-4","=hanziku-5","=hanziku-6","=hanziku-7","=hanziku-8","=hanziku-9");
@CDP=("=big5-cdp");

while(<>){
    # temporary fix for using in OTP for perl 5.6.
    s/(.)/pack("c",unpack("U",$1))/ge if($inotp
					 and $in_cs=~/utf8/i
					 and $perl56);
    # for perl 5.8.
    $_=decode('utf8', $_)  if ($inotp and $in_cs=~/utf8/i
			       and $perl58);
    s/(amp.+?;)/&tex_de_er($1)/ge;
#    s/(&.+?;)/&tex_de_er($1)/ge;
#    s/^(.*)$/&add_break($1)/e;
    @chars=split(//);
    for($i=0;$i<=$#chars;$i++){
#    while(m/(.)/g){
	$char=&get_char_in_utf8mcs($chars[$i],$in_cs);
	$char_id=unpack("U",$char);
	if($ids_argc>0){
	    # It's in IDS.
	    ($ids,$ids_argc)=&ids_rest($ids,$ids_argc,$char);
	    if($ids_argc==0){
		if(($char_id=&get_char_id_for_ids($ids))
		   and(($out_char=&get_output_char($char_id,$out_cs)))){
		    print $out_char,&add_break($i);
		}else{
		    print &replace_ids($ids),&add_break($i) if($perl56);
		    print encode('utf8', &replace_ids($ids)),&add_break($i) if($perl58);
		}
		$ids="";
	    }
	}elsif($char_id>=0x2ff0 and $char_id<=0x2fff){
	    ($ids,$ids_argc)=&ids_rest("",0,$char);
	    next;
	}else{
	    if($char_id<=0xff){
		print $char;
		next;
	    }
	    if(($out_char=&get_output_char($char_id,$out_cs))){
		print $out_char,&add_break($i);
	    }elsif($char_id >= 0x20000 && $char_id <=0x2a6df){
		unless(defined($ids{$char}) and $ids{$char}[1]>=0){
		    $ids{$char}[0]=$font_start;
		    $ids{$char}[1]=$ids_start;
		    $ids_start++;
		    if($ids_start>255){
			$ids_start=0;
			$font_start++;
		    }
		}
		print "{\\fontencoding{OT1}\\fontfamily{" .
		    sprintf("chise%03d",$ids{$char}[0]) .
		    "}\\selectfont\\char$ids{$char}[1]}",&add_break($i);
		next;
	    }else{
		print &replace_ids(&get_ids($char)),&add_break($i);
	    }
	}
    }
    if($ids_argc>0){
	print STDERR "IDS parse error: $ids\n";
#	print pack("U",0xfffd);
	print pack("U",0x3013) if($perl56);
	print encode('utf8',pack("U",0x3013)) if($perl58);
	$ids="";
	$ids_argc=0;
    }
}

open(IDSDATA,">$idsdata_file") or die;
print IDSDATA 'use utf8;',"\n";
foreach $ids (keys %ids){
    print IDSDATA '$ids{\'',$ids,'\'}=[',join ",",@{$ids{$ids}},"];\n" if($perl56);
    print IDSDATA '$ids{\'',encode('utf8',$ids),'\'}=[',join ",",@{$ids{$ids}},"];\n" if($perl58);
}
print IDSDATA '$font_start=',$font_start,";\n";
print IDSDATA '$ids_start=',$ids_start,";\n";
print IDSDATA "1;";

exit 0;

sub tex_de_er{
    my($er)=@_;
    my($out);
    $er=~s/^amp(.*);$/$1/;
#    $er=~s/^&(.*);$/$1/;
    $out=&de_er($er);
    if($out){
	return $out;
    }else{
	return "amp$er;";
    }
}

sub add_break{
    my($i)=@_;

    if($i<($#chars-1)){
	if(($chars[$i+1]=~m/[$strictly_forbidden_before]/o)
	   and($chars[$i+2]=~m/[$strictly_forbidden_before]/o)){
	    return "\\CJKunbreakablekernone ";
	}elsif(($chars[$i+1]=~m/[$strictly_forbidden_before]/o)
	       and($chars[$i+2]=~m/[^$strictly_forbidden_before]/o)){
	    return "\\CJKunbreakablekernone \\CJKprotrude ";
	}
    }
    if(($i<$#chars)
       and($chars[$i+1]=~m/[$strictly_forbidden_before]/o)){
	return "\\CJKunbreakablekernone ";
    }
    if($chars[$i]=~m/[$strictly_forbidden_after]/o){
	return "\\CJKunbreakablekernone ";
    }
    if(($i<$#chars)
       and($chars[$i+1]=~m/[$forbidden_before]/o)){
	return "\\CJKunbreakablekerntwo ";

    }
    if($chars[$i]=~m/[$forbidden_after]/o){
	return "\\CJKunbreakablekerntwo ";
    }
    if(($i<$#chars)
       and($chars[$i+1]=~m/[$slightly_forbidden_before]/o)){
	return "\\CJKunbreakablekernthree ";
    }
    if($chars[$i]=~m/[$slightly_forbidden_after]/o){
	return "\\CJKunbreakablekernthree ";
    }
    if($chars[$i]=~m/[$asian]/o){
	return "\\CJKbreakablekern ";
    }
    if(($i<$#chars)and($chars[$i+1]=~m/[$asian]/o)){
	return "\\CJKbreakablekern ";
    }
}

sub ids_rest{
    my($ids,$ids_argc,$char)=@_;
    my($argc);
    $argc=&ids_argc($char);
    if($argc){
	$ids_argc+= $ids_argc==0 ? $argc : $argc-1;
    }else{
	$ids_argc--;
    }
    $ids.=$char if($perl56);
    $ids.=encode('utf8',$char) if($perl58);
    return ($ids,$ids_argc);
}

sub replace_ids{
    my($ids)=@_;
    $ids=&normalize_ids($ids,"UniJIS");
#    return pack("U",0xfffd) if($ids!~/[$idc]/);
    return pack("U",0x3013) if(($ids!~/[$idc]/)
			       or($ids=~/[\x{10000}-]/));
                    #irregular for KAGE.
    unless(defined($ids{$ids}) and $ids{$ids}[1]>=0){
	$ids{$ids}[0]=$font_start;
	$ids{$ids}[1]=$ids_start;
	$ids_start++;
    }
    if($ids_start>255){
	$ids_start=0;
	$font_start++;
    }
    return "{\\fontencoding{OT1}\\fontfamily{".sprintf("chise%03d",$ids{$ids}[0])."}\\selectfont\\char$ids{$ids}[1]}";
}

sub normalize_ids{
    my($ids,$out_cs)=@_;
    $ids = decode('utf8', $ids) if $perl58;
    $out_cs=~s/Uni(.+)/"ucs-".lc($1)/e;
    my $output_ids="";
    my($char,$char_id,$output_char_id);
    while($ids=~m/(.)/g){
	$char=$1;
	$char_id=unpack("U",$char);
	if($char=~/[$idc]/){
	    $output_ids.=$char;
	}elsif($output_char_id=&get_char_attribute($char,$out_cs)){
	    $output_ids.=pack("U",$output_char_id);
	}elsif($output_char_id=&get_char_attribute($char,"=ucs")){
	    $output_ids.=pack("U",$output_char_id);
	}elsif($output_char_id=&get_char_attribute($char,"ucs")){
	    $output_ids.=pack("U",$output_char_id);
	}else{
	    return pack("U",0xfffe);
	}
    }
    return $output_ids;
}

sub get_output_char{
    my($char_id,$out_cs)=@_;
    my($out_char_id,$suffix);
    my($gt,$hzk,$cdp);
    if(not defined($cmap_to{$out_cs})){
	&get_cmap($out_cs);
    }
    if($out_char_id=$cmap_to{$out_cs}->{$char_id}){
	return pack("U",$out_char_id);
    }else{
	return $gt if($useGT and $gt=&get_macro_for_GT($char_id));
	return $hzk if($useHZK and $hzk=&get_macro_for_HZK($char_id));
	return $cdp if($useCDP and $cdp=&get_macro_for_CDP($char_id));
	return undef;
    }
}

sub get_cmap{
    my($out_cs)=@_;
    tie %{$cmap_to{$out_cs}}, "BerkeleyDB::Hash",
    -Filename => "$omegadb_home/$out_cs" or die $!;
}

sub get_ids{
    my($char)=@_;
    my $ids="";
    $ids=&get_char_attribute($char,"ids-aggregated")
	or &get_char_attribute($char,"ids");
#	or &get_char_attribute($char,"ideographic-structure");
    $ids=decode('utf8', $ids) if($perl58);
#    $ids=~s/[? ()]//g;
    return $ids;
}

sub get_char_id_for_ids{
    my($ids)=@_;
    my($char_id,$char);
    $ids=decode('utf8', $ids) if($perl58);
#    $ids="(?".(join " ?",(split(//,$ids))).")";
    &get_idsdb if(not defined($idsdb));
    $char=$idsdb->{$ids};
    $char=decode('utf8',$char) if($perl58);
    if($char){
	return unpack("U",$char);
    }else{
	return undef;
    }
}

sub get_idsdb{
    tie %{$idsdb}, "BerkeleyDB::Hash",
    -Filename => "$omegadb_home/idsdb" or die $!;
}

sub get_char_in_utf8mcs_bak{
    my($char,$in_cs)=@_;
    return $char if($in_cs eq "Utf8mcs");
    my($char_id,$output_char);
    $in_cs=~s/Utf8/ucs-/;
    $char_id=unpack("U",$char);
    if(($output_char)=&get_chars_matching("$in_cs",$char_id)){
	$output_char=decode('utf8', $output_char) if($perl58);
	return $output_char;
    }else{
	return $char;
    }
}

sub get_char_in_utf8mcs{
    # argument: <character>, <input coding system>
    # return:   character in UTF-8mcs.
    my($char,$in_cs)=@_;
    my($char_id,$output_char_id);
    return $char if($in_cs eq "Utf8mcs");
    $char_id=unpack("U",$char);
    &get_utf8mcs_map($in_cs) if(not defined($utf8mcs_map_from{$in_cs}));
    if($output_char_id=$utf8mcs_map_from{$in_cs}->{$char_id}){
	return pack("U",$output_char_id);
   }else{
	return $char;
   }
}

sub get_utf8mcs_map{
    my($in_cs)=@_;
    my($suffix);
    ($suffix=$in_cs)=~s/^Utf8//;
    tie %{$utf8mcs_map_from{$in_cs}}, "BerkeleyDB::Hash",
    -Filename => "$omegadb_home/ucs-$suffix" or die $!;
}

sub get_macro_for_GT{
    my($char_id)=@_;
    my($char,$gt,$GT);
    $char=pack("U",$char_id);
    foreach (@GT){
	if($gt=&get_char_attribute($char,$_)){
	    m/gt\-pj\-(\d+)/ and $GT=$1;
	    last;
	}
    }
    if($gt){
	return "{\\fontencoding{OT1}\\fontfamily{".sprintf("gt%02d",$GT)."}\\selectfont\\char".($gt|0x8080)."}";
#	return "\\GT{".sprintf("gt%02d",$GT)."}{\\char".($gt|0x8080)."}";
    }else{
	return undef;
    }
}

sub get_macro_for_HZK{
    my($char_id)=@_;
    my($char,$hzk,$HZK);
    $char=pack("U",$char_id);
    foreach (@HZK){
	if($hzk=&get_char_attribute($char,$_)){
	    m/hanziku\-(\d+)/ and $HZK=$1;
	    last;
	}
    }
    if($hzk){
	return "{\\fontencoding{OT1}\\fontfamily{".sprintf("hzk%02d",$HZK)."}\\selectfont\\char".($hzk|0x8080)."}";
    }else{
	return undef;
    }
}

sub get_macro_for_CDP{
    my($char_id)=@_;
    my($char,$cdp);
    $char=pack("U",$char_id);
    foreach (@CDP){
	if($cdp=&get_char_attribute($char,$_)){
	    last;
	}
    }
    if($cdp){
	return "{\\fontencoding{OT1}\\fontfamily{cdp}\\selectfont\\char".($cdp|0x8080)."}";
    }else{
	return undef;
    }
}