Metadata returned for DOIs from CrossRef and DataCite

doi-metadata.php

<?php

require 'arc/ARC2.php';

$dois = array('10.1126/science.1157784', '10.5524/100005');

foreach ($dois as $doi) {
	$url = "http://dx.doi.org/$doi";
	print "$url:\n";
	
	$headers = array('Accept: text/turtle');
	$context = stream_context_create(array('http' => array('header' => $headers)));
	$data = file_get_contents($url, false, $context);
	
	$parser = ARC2::getTurtleParser();
	$parser->parse($url, $data);
	$index = $parser->getSimpleIndex();
	
	$metadata = $index[$url];
	if (!$metadata) $metadata = $index["http://dx.doi.org/\n$doi\n"]; // DataCite RDF is dodgy
	ksort($metadata);
	
	foreach ($metadata as $field => $value) {
		$value = str_replace("\n", '', $value[0]); // DataCite RDF is dodgy
		printf("\t<%s> %s\n", $field, $value);
	}	
}

/*
http://dx.doi.org/10.1126/science.1157784:
	<http://prismstandard.org/namespaces/basic/2.1/doi> 10.1126/science.1157784
	<http://prismstandard.org/namespaces/basic/2.1/endingPage> 832
	<http://prismstandard.org/namespaces/basic/2.1/startingPage> 828
	<http://prismstandard.org/namespaces/basic/2.1/volume> 325
	<http://purl.org/dc/terms/creator> http://id.crossref.org/contributor/a-h-renear-1z0zrfd0bp2b7
	<http://purl.org/dc/terms/date> 2009-08-13Z
	<http://purl.org/dc/terms/identifier> 10.1126/science.1157784
	<http://purl.org/dc/terms/isPartOf> http://id.crossref.org/issn/0036-8075
	<http://purl.org/dc/terms/title> Strategic Reading, Ontologies, and the Future of Scientific Publishing
	<http://purl.org/ontology/bibo/doi> 10.1126/science.1157784
	<http://purl.org/ontology/bibo/pageEnd> 832
	<http://purl.org/ontology/bibo/pageStart> 828
	<http://purl.org/ontology/bibo/volume> 325
	<http://www.w3.org/1999/02/22-rdf-syntax-ns#type> http://purl.org/ontology/bibo/Article
	<http://www.w3.org/2002/07/owl#sameAs> info:doi/10.1126/science.1157784
	
http://dx.doi.org/10.5524/100005:
	<http://purl.org/dc/terms/creator> Lambert, D
	<http://purl.org/dc/terms/date> 2011
	<http://purl.org/dc/terms/identifier> 10.5524/100005
	<http://purl.org/dc/terms/publisher> GigaScience
	<http://purl.org/dc/terms/title> Genomic data from the Emperor penguin (Aptenodytes forsteri)
	<http://www.w3.org/2002/07/owl#sameAs> doi:10.5524/100005
*/