#!/usr/bin/perl

my @osis_books = qw(Gen Exod Lev Num Deut Josh Judg Ruth 1Sam 2Sam 1Kgs 2Kgs
		    1Chr 2Chr Ezra Neh Esth Job Ps Prov Eccl Song Isa Jer
		    Lam Ezek Dan Hos Joel Amos Obad Jonah Mic Nah Hab Zeph
		    Hag Zech Mal
		    Matt Mark Luke John Acts Rom 1Cor 2Cor Gal Eph Phil Col
		    1Thess 2Thess 1Tim 2Tim Titus Phlm Heb Jas 1Pet 2Pet
		    1John 2John 3John Jude Rev);

my $last_ref = undef;
my $last_book = undef;
my $last_chapter = undef;
my $last_verse = undef;

my $_workID = 'MorphGNT';
my $_workName = 'MorphGNT';
my $_lang = 'gr';
my $_year = '2006';


#
# Print the OSIS header.
#

print "<?xml version=\"1.0\" encoding=\"UTF-8\" ?>\n";
print "<osis xmlns=\"http://www.bibletechnologies.net/2003/OSIS/namespace\" xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xsi:schemaLocation=\"http://www.bibletechnologies.net/osisCore.2.1.xsd\">\n";
print "<osisText osisIDWork=\"$_workID\" osisRefWork=\"defaultReferenceScheme\" xml:lang=\"$_lang\">\n";
print "<header>\n";
print "<work osisWork=\"$_workID\">\n";
print "<title>$_workName</title>\n";
print "<identifier type=\"OSIS\">Bible.$_lang.$_workID.$_year</identifier>\n";
print "<refSystem>Bible</refSystem>\n";
print "</work>\n";
print "<work osisWork=\"defaultReferenceScheme\">\n";
print "<refSystem>Bible</refSystem>\n";
print "</work>\n";
print "</header>\n";


#
# Print the text.
# XXX morph
# XXX part of speech
#

while (<STDIN>) {

    chomp;
    my ($ref, $pos, $parse, $form, $lemma) = split(/\s+/, $_, 5);
    my ($book, $chapter, $verse) = $ref =~ /(\d\d)(\d\d)(\d\d)/;
    $book += 39; # 01 = Matthew

    my $osisID = sprintf("%s.%d.%d", $osis_books[$book - 1], $chapter, $verse);

    if ($ref != $last_ref) {

	print $close_verse, "\n" if $close_verse and $last_verse != $verse;
	print $close_chapter, "\n" if $close_chapter and $last_chapter != $chapter;
	print "</div>\n" if defined $last_book and $last_book != $book;

	my $osis_id = $osis_books[$book - 1];

	if ($last_book != $book) {
	    print "<div type=\"book\" osisID=\"$osis_id\">\n";
	    $last_book = $book;
	}

	$osis_id .= '.' . $chapter;
	if ($last_chapter != $chapter) {

	    print "<chapter sID=\"$osis_id\" osisID=\"$osis_id\" n=\"$chapter\"/><div type=\"section\" scope=\"$osis_id\">\n";

	    $close_chapter = "</div>\n<chapter eID=\"$osis_id\"/>";
	    $last_chapter = $chapter;
	}

	$osis_id .= '.' . $verse;
	if ($last_verse != $verse) {

	    print "<verse sID=\"$osis_id\" osisID=\"$osis_id\"/>\n";

	    $close_verse = "<verse eID=\"$osis_id\"/>";
	    $last_verse = $verse;
	}

    }
    # XXX Convert morphological codes to Robinson or Packard?
    printf('<w lemma="%s" morph="%s">%s</w> ', $lemma, $parse, $form) if $form;
    $last_ref = $ref;
}

print $close_verse, "\n" if $close_verse;
print $close_chapter, "\n" if $close_chapter;
print "</div>\n" if defined $last_book;

print "</osisText>\n";
print "</osis>\n";