A simple class for RDF MIME detection using sesame Rio Parser.

gistfile1.java

package formats;

import java.io.IOException;
import java.net.HttpURLConnection;
import java.net.URL;
import java.util.HashMap;
import java.util.Map;

import org.openrdf.rio.RDFFormat;
import org.openrdf.rio.RDFParser;
import org.openrdf.rio.Rio;

/**
 * A simple class for RDF MIME detection using sesame Rio Parser.
 * @author seralf
 *
 */
public class RdfMimeDetect {
	
	
	private static class NoMIMEException extends Exception {

		public NoMIMEException() {
			super();
		}

		public NoMIMEException(String message, Throwable cause) {
			super(message, cause);
		}

		public NoMIMEException(String message) {
			super(message);
		}

		public NoMIMEException(Throwable cause) {
			super(cause);
		}
		
	}
	
	private static class MimeAlias {

		private String wrong;
		private String right;

		public MimeAlias(String wrongName, String rightName) {
			this.wrong = wrongName;
			this.right = rightName;
		}

		@Override
		public String toString() {
			return "(" + wrong + ", " + right + ")";
		}
		
	}

	private final static Map<String, String> aliases = new HashMap<String, String>();

	public static void setMimeAlias(final MimeAlias... mimeAliases){;
		for (MimeAlias mimeAlias : mimeAliases) {
			aliases.put(mimeAlias.wrong, mimeAlias.right);
		}
	}
	
	public static RDFFormat getMime(final URL url, boolean useExtension) throws IOException, NoMIMEException {
		
		final RDFFormat format;
		
		// we could obtaind the correct RDFFormat from file extension or directly from MIME name
		if(useExtension){
			format = RDFFormat.forFileName(url.toString());	
		} else {
			String mimeType = ((HttpURLConnection) url.openConnection()).getContentType();
			// in case we have MIME in the form: text/turtle; utf-8 or similar...
			String mimeTypeSimple = mimeType.split(";")[0].trim();
			// check if the mime name is in the exception or not
			format = aliases.containsKey(mimeTypeSimple) ? RDFFormat.forMIMEType(aliases.get(mimeTypeSimple)) : RDFFormat.forMIMEType(mimeTypeSimple);
		}
		
		if (format == null) throw new NoMIMEException("NO MIME found");

		return format;
	}

	public static void main(String[] args) throws Exception {

		// map of some alternative mime names...
		final MimeAlias[] aliases = {
			new MimeAlias("text/n3", "text/rdf+n3")
		};
		
		// list of example URLs
		final URL[] urls = {
			new URL("http://dbpedia.org/data/Banco_del_Mutuo_Soccorso.rdf"),
			new URL("http://dbpedia.org/data/Banco_del_Mutuo_Soccorso.json"),
			new URL("http://dbpedia.org/data/Banco_del_Mutuo_Soccorso.n3"),
			new URL("http://dbpedia.org/data/Banco_del_Mutuo_Soccorso.ntriples")
		};

		// set the mime name exception list 
		RdfMimeDetect.setMimeAlias(aliases);

		for (URL url : urls) {
			try {
				RDFFormat mime = getMime(url, false);
				System.out.printf("URL: %s, MIME: %s\n", url, mime);
				
				RDFParser rdfParser = Rio.createParser(mime);
				System.out.printf("TODO: SOMETHING WITH PARSER FOR: %s\n",  rdfParser.getRDFFormat());
				
			} catch (Exception e) {
				System.err.printf("Can't detect MIME for url: %s (%s)\n", url, e.getMessage());
			}
		}

	}

}

/*
 * 
 * 
 * 
 * 
 * RDF/XML application/rdf+xml .rdf .rdfs .owl UTF-8 RDF/XML application/xml
 * .xml UTF-8
 * 
 * N-TRIPLES text/plain .nt 7-bit US-ASCII
 * 
 * TURTLE text/turtle .ttl UTF-8 TURTLE application/x-turtle .ttl UTF-8
 * 
 * N3 text/rdf+n3 .n3 UTF-8
 * 
 * TRIX application/trix .trix UTF-8
 * 
 * TRIG application/x-trig .trig UTF-8
 * 
 * BINARY application/x-binary-rdf .brf binary
 * 
 * NQUADS text/x-nquads .nq 7-bit US-ASCII
 */