#!/usr/bin/perl
use LWP::Simple;
use XML::Writer;

use strict;

my $directory="/var/tmp/pacs2003" ;
my $url="http://www.aip.org/pacs/pacs03/all.txt";

# renew the download of the web page
my $renew=1; 
if($renew) {
    system("w3mir -drr -r $url 2>> /dev/null");
}


open(F,"cat all.txt |");

$_=join('',<F>);

s/\r//g;
s/\n\n/\n/g;
s/\n {13}//g;

my @lines=split(/\n/);
my %pacs;

foreach(@lines) {
    if(/  (\d{2}\.[^ ]*) +(.*)/) {
	$pacs{$1}=$2;
    }
    else {
	print "$_\n";
    }
}


my $out = new IO::File("> pacs2003.xsd");
my $x = new XML::Writer(OUTPUT => $out, DATA_MODE => 1,  DATA_INDENT => 1);
$x->xmlDecl();
$x->startTag('xs:schema', 'xmlns:xs' => "http://www.w3.org/2001/XMLSchema",
             'elementFormDefault'=>"unqualified",
             'attributeFormDefault'=>"unqualified",
             'xmlns'=>"http://amf.openlib.org",
             'targetNamespace'=>"http://amf.openlib.org");

$x->startTag('xs:simpleType', 'name'=>"pacs2003Element");
$x->startTag('xs:restriction', 'base'=>"xs:string");

foreach(sort keys %pacs) {
 $x->startTag('xs:enumeration', 'value'=>"$_");
 $x->startTag('xs:annotation');
 $x->startTag('xs:documentation');
 $x->characters($pacs{$_});
 $x->endTag;
 $x->endTag;
 $x->endTag;
}
$x->endTag;
$x->endTag;

# create the list type
$x->startTag('xs:simpleType', 'name'=>"pacs2003");
$x->emptyTag('xs:list', 'itemType'=>"pacs2003Element");
				       $x->endTag;
$x->endTag;
$out->close;

system("rm all.txt");

system("recode iso-8859-1..utf-8 pacs2003.xsd");

exit;