<?php
//  USIN.php  - Parsing tools for Universal Serial Item Names 
//  Version 0.91
//  Robert D. Cameron, March 21, 2001.
//  Copyright (c) 2001,  GNU Public License, Version 2, applies.
//  
//---------------------------------------------------------------
//
// 
//  parse_USIN($USIN_proto) {
//
//  Parse an arbitrary string $USIN_proto according to USIN syntax
//  and return a parse information map of the following form.
//   {principal_domain => 
//      The symbol naming the top-level domain of the USIN.
//    domain_extensions => 
//      An array of domain extensions, each of which is either
//      a parenthesized phrase, or a "." followed by a subdomain
//      symbol.
//    collection_label => 
//      The label (if any) of a publication or collection.  
//      The "/" separating the domain and the label is stripped off. 
//    item_extensions => 
//      An array of item extensions, each of which is either
//      a parenthesized phrase, or an operator (excluding "!") 
//      followed by a symbol.
//    attributes => 
//      An array of attributes at the end of the USIN, each
//      of which is a symbol followed by an optional parenthesized
//      phrase.
//      The "!" symbol introducing each attribute is stripped off.
//    junk => 
//      Junk following the longest legal USIN found from the input.
//
//  Example 1:
//    parse_USIN('ISSN/0361-526X:36(1/2)@175') =
//    {principal_domain => 'ISSN',
//     domain_extensions => [],
//     collection_label => '0361-526X',
//     item_extensions => [':36', '(1/2)', '@175'],
//     attributes => [],
//     junk => ''}
//    This example illustrates journal article citation,
//      with conventional item extension syntax for volume, issue and page. 
//
//  Example 2:
//    parse_USIN('RDNS(sfu.ca).CMPT/MSc:2000$SerbanTatu!author(1)') =
//    {principal_domain => 'RDNS',
//     domain_extensions => ['(sfu.ca)', '.CMPT'],
//     collection_label => 'MSc',
//     item_extensions => [':2000', '$SerbanTatu'],
//     attributes => ['author(1)'],
//     junk => ''}
//
//  Escape-encoding.  If a USIN is escape-encoded: it should be decoded
//    first, prior to application of parse_USIN.
//---------------------------------------------------------------
//
//  Regular Expressions for USIN Syntactic Components.
//
//
$symbolRE = '[A-Za-z0-9](?:[-_]?[A-Za-z0-9])*';
$phraseRE = '\([-A-Za-z0-9_/:!@$*~+,.]+\)';
//
// The BibP specification states that the letter, digit and extender
// classes are subject to future extension to include non-ASCII UTF-8
// characters.  The following two REs would be needed for such
// an extension, but this has not been tested.
//
// $symbolRE = '[A-Za-z0-9\x80-\xFF](?:[-_]?[A-Za-z0-9\x80-\xFF])*';
// $phraseRE = '\([-A-Za-z0-9_/:!@$*~+,.\x80-\xFF]+\)';
//
// The USIN hyphenation convention allows whitespace to be introduced
// in some contexts.
$hyphRE = '(?:-[\s]*)?';
//  
$domain_extensionRE = "$hyphRE(\\.$symbolRE|$phraseRE)";
//  Match operators of any length, except the single "!" for attributes.
$operatorRE= "(?:![/:!@$*~+,.]+|[/:@$*~+,.][/:!@$*~+,.]*)";
$item_extensionRE = "$hyphRE($operatorRE$symbolRE|$phraseRE)";
$attributeRE = "$hyphRE!($symbolRE(?:$phraseRE)?)";

function parse_USIN($USIN_proto) {
  global $symbolRE, $phraseRE, $domain_extensionRE, $operatorRE,
         $hyphRE, $item_extensionRE, $attributeRE;
//  NOTE: The regular expressions are delimited by "#" instead of
//  the typical "/".  This avoids the need to escape a literal
//  "/" which sometimes occurs in USIN syntax and hence in the REs
//  below. The "#" character is safe.
//
  $USIN_proto = rawurldecode($USIN_proto);
//  Trim whitespace.  
  $USIN_proto = trim($USIN_proto);
  if (!preg_match("#^$symbolRE#", $USIN_proto, $matches)) {
    return array('junk' => $USIN_proto);
  }
  $result_map['principal_domain'] = $matches[0];
  $USIN_proto = substr($USIN_proto, strlen($matches[0]));
  while (preg_match("#^$domain_extensionRE#", $USIN_proto, $matches)) {
    $USIN_proto = substr($USIN_proto, strlen($matches[0]));
    array_shift($matches);
    $result_map['domain_extensions'][] = join("", $matches);
  }
  if (preg_match("#^$hyphRE/($symbolRE)#", $USIN_proto, $matches)) {
    $result_map['collection_label'] = $matches[1];
    $USIN_proto = substr($USIN_proto, strlen($matches[0]));
    while (preg_match("#^$item_extensionRE#", $USIN_proto, $matches)) {
      $USIN_proto = substr($USIN_proto, strlen($matches[0]));
      array_shift($matches);
      $result_map['item_extensions'][] = join("", $matches);
    }
  }
  while (preg_match("#^$attributeRE#", $USIN_proto, $matches)) {
    $result_map['attributes'][] = $matches[1];
    $USIN_proto = substr($USIN_proto, strlen($matches[0]));
  }
  $result_map['junk'] = $USIN_proto;
  return $result_map;
}
?>