import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
@Override
protected TripleCollection generateRdf(HttpRequestEntity entity) throws IOException {
	String rdfDataFormat = SupportedFormat.TURTLE;    	    	
	InputStream configIn = null;
	String queryString = entity.getRequest().getQueryString();
	log.info("Query string: " + queryString);
	//String configUri = getRequestParamValue(queryString, "config");	
	String configUri = entity.getRequest().getParameter("config");      	
	log.info("Config file URI: " + configUri);
	
	if(configUri != null) {    	  
	  configIn = getRemoteConfigFile(configUri);
	}
	
    final InputStream inputRdfData = entity.getData();
    TripleCollection duplicates = findSameEntities(inputRdfData, rdfDataFormat, configIn);
    return duplicates;
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
@Test
public void testSilkRdfTurtle() throws IOException {

    Response response
            = RestAssured.given().header("Accept", "text/turtle")
            .contentType("text/turtle;charset=UTF-8")
            .content(ttlData)
            .expect().statusCode(HttpStatus.SC_OK).content(new StringContains("http://www.w3.org/2002/07/owl#sameAs")).header("Content-Type", SupportedFormat.TURTLE).when()
            .post();

    
     Graph graph = Parser.getInstance().parse(response.getBody().asInputStream(), "text/turtle");
     Iterator<Triple> typeTriples = graph.filter(null, OWL.sameAs, null);
     Assert.assertTrue("No equivalent entities found", typeTriples.hasNext());
     
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
@Test
  public void testRemoteConfig() {
   // Set up a service in the mock server to respond to a get request that must be sent by the transformer 
      // to fetch the silk config file 
stubFor(get(urlEqualTo("/fusepoolp3/silk-config-file.xml"))
   	    .willReturn(aResponse()
   			.withStatus(HttpStatus.SC_OK)
   			.withHeader("Content-Type", "text/xml")
   			.withBody(silkconf)));

// The response object acts as a transformer's client. It sends a post request
// to the transformer with the url of the silk config file, the data to be interlinked
// and gets the result from the transformer.
  	Response response = 
   	RestAssured.given().header("Accept", "text/turtle")
   	.contentType("text/turtle")
   	.content(ttlData)
   	.expect().statusCode(HttpStatus.SC_OK).when()
   	.post("/?config=http://localhost:" + mockPort +"/fusepoolp3/silk-config-file.xml");
  	
  	Graph graph = Parser.getInstance().parse(response.getBody().asInputStream(), SupportedFormat.TURTLE);
      Iterator<Triple> typeTriples = graph.filter(null, OWL.sameAs, null);
      Assert.assertTrue("No equivalent entities found", typeTriples.hasNext());
  	
  }
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
/**
 * Load RDF data sent by HTTP POST. Use the Dataset custom header
 * to address the dataset in which to store the rdf data.
 * Use this service with the following curl command:
 *  curl -X POST -u admin: -H "Content-Type: application/rdf+xml" 
 *  	-H "Dataset: mydataset" -T <rdf_file> http://localhost:8080/dlcupload/rdf 
 */
@POST
@Path("rdf")
@Produces("text/plain")
public String uploadRdf(@Context final UriInfo uriInfo,  
		@HeaderParam("Content-Type") String mediaType,
		@HeaderParam("Dataset") String dataset,
        final InputStream stream) throws Exception {
	
    AccessController.checkPermission(new AllPermission());
    final MGraph graph = new SimpleMGraph();
    
    String message = "";
   
    if(mediaType.equals(SupportedFormat.RDF_XML)) {
    	parser.parse(graph, stream, SupportedFormat.RDF_XML);
    }
    else {
    	message = "Add header Content-Type: application/rdf+xml ";
    }
    
    return message + "Added " + graph.size() + " triples  to dataset " + dataset + "\n";
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
@BeforeClass
public static void initTest() throws Exception {
    //read the fise:Enhancements use by the test and validate that they do
    //confirm to the FISE enhancement structure
    InputStream in = Fise2FamEngineTest.class.getClassLoader().getResourceAsStream(TEST_ENHANCEMENTS);
    Assert.assertNotNull("Unable to load reaource '"+TEST_ENHANCEMENTS+"' via Classpath",in);
    origEnhancements = new IndexedMGraph();
    rdfParser.parse(origEnhancements, in, SupportedFormat.TURTLE, null);
    in.close();
    Assert.assertFalse(origEnhancements.isEmpty());
    //parse the ID of the ContentItem form the enhancements
    Iterator<Triple> it = origEnhancements.filter(null, Properties.ENHANCER_EXTRACTED_FROM, null);
    Assert.assertTrue(it.hasNext());
    Resource id = it.next().getObject();
    Assert.assertTrue(id instanceof UriRef);
    ciUri = (UriRef)id;
    //validate that the enhancements in the file are valid
    EnhancementStructureHelper.validateAllTextAnnotations(
        origEnhancements, CONTENT, null);
    
    //init the ContentItem for testing Keyword and Sentiment annotations
    //read the content for the KS test
    in = Fise2FamEngineTest.class.getClassLoader().getResourceAsStream(TEST_KEY_SENT_FILE);
    Assert.assertNotNull("Unable to load reaource '"+TEST_KEY_SENT_FILE+"' via Classpath",in);
    ksContent = IOUtils.toString(in, "UTF-8");
    in.close();
    //read RDF enhancements for the KS test
    in = Fise2FamEngineTest.class.getClassLoader().getResourceAsStream(TEST_KEY_SENT_ENHANCEMENTS);
    Assert.assertNotNull("Unable to load reaource '"+TEST_KEY_SENT_ENHANCEMENTS+"' via Classpath",in);
    ksOrigEnhancements = new IndexedMGraph();
    rdfParser.parse(ksOrigEnhancements, in, SupportedFormat.TURTLE, null);
    in.close();
    Assert.assertFalse(ksOrigEnhancements.isEmpty());
    //parse the ID of the ContentItem form the enhancements
    it = ksOrigEnhancements.filter(null, Properties.ENHANCER_EXTRACTED_FROM, null);
    Assert.assertTrue(it.hasNext());
    id = it.next().getObject();
    Assert.assertTrue(id instanceof UriRef);
    ksCiUri = (UriRef)id;
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
private static void logRdf(String title, TripleCollection graph) {
    if(log.isDebugEnabled()){
        ByteArrayOutputStream out = new ByteArrayOutputStream();
        rdfSerializer.serialize(out, graph, SupportedFormat.TURTLE);
        try {
            log.debug("{} {}",title == null ? "RDF:\n" : title, out.toString("UTF8"));
        } catch (UnsupportedEncodingException e) {/*ignore*/}
    }
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
/**
 * Logs the enhancements as TURTLE on DEBUG level
 * @param ci the contentItem
 */
private void logEnhancements(ContentItem ci){
    if(LOG.isDebugEnabled()){
        ByteArrayOutputStream bout = new ByteArrayOutputStream();
        serializer.serialize(bout, ci.getMetadata(), SupportedFormat.TURTLE);
        LOG.debug("Enhancements of {}",ci.getUri().getUnicodeString());
        LOG.debug(new String(bout.toByteArray(),Charset.forName("UTF8")));
    }
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
/**
 * The client RDF data is always used as the source data source, of type file, for the comparisons with a target data source. 
 * The target data source can be of type file or a SPARQL endpoint. If the target data source in the Silk config file
 * is set to be of type file then the same client data will be used and the task is a deduplication task (Silk works only with local files). 
 * The updated configuration file and the input RDF data and the output files are stored in the /tmp/ folder.
 * @param inputRdf
 * @return
 * @throws IOException
 */
protected TripleCollection findSameEntities(InputStream inputRdf, String rdfFormat, InputStream configIn) throws IOException {    	
	// Default silk config file
	File configFile = null;
	if(configIn != null){
		configFile = FileUtil.inputStreamToFile(configIn, "silk-config-", ".xml");
	}
	else {
		configFile = FileUtil.inputStreamToFile(getClass().getResourceAsStream("silk-config-file.xml"), "silk-config-", ".xml");
	}
	
    // file with original data serialized in N-TRIPLE format
    File ntFile = File.createTempFile("input-rdf", ".nt");
    // file containing the equivalences
    File outFile = File.createTempFile("output-", ".nt");
    
    // update the config file with the paths of the source datasource and output files and the format
    // if the type of target datasource is "file" update the path (deduplication) 
    SilkConfigFileParser silkParser = new SilkConfigFileParser(configFile.getAbsolutePath());
    silkParser.updateOutputFile(outFile.getAbsolutePath());
    silkParser.updateSourceDataSourceFile(ntFile.getAbsolutePath(), "N-TRIPLE");
    if (silkParser.getTargetDataSourcetype().equals("file")) {
        silkParser.updateTargetDataSourceFile(ntFile.getAbsolutePath(), "N-TRIPLE"); //deduplication
    }
    silkParser.saveChanges();
    
    // change the format into N-TRIPLE
    Parser parser = Parser.getInstance();
    TripleCollection origGraph =  parser.parse(inputRdf, rdfFormat);
    Serializer serializer = Serializer.getInstance();
    serializer.serialize(new FileOutputStream(ntFile), origGraph, SupportedFormat.N_TRIPLE);

    // interlink entities
    Silk.executeFile(configFile, null, 1, true);
    log.info("Interlinking task completed."); 
    TripleCollection equivalences = parseResult(outFile); 
    
    // add the equivalence set to the input rdf data to be sent back to the client
    TripleCollection resultGraph = new SimpleMGraph();
    resultGraph.addAll(origGraph);
    resultGraph.addAll(equivalences);
    
    // remove all temporary files
    configFile.delete();        
    ntFile.delete();
    outFile.delete();

    // returns the result to the client
    return resultGraph;
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
@Test
public void testGetPoint() throws IOException {
    TripleCollection graph = Parser.getInstance().parse(getClass().getResourceAsStream(TEST_DATASET), SupportedFormat.TURTLE);
    WGS84Point point = jenas.getPoint(graph);
    Assert.assertTrue(point != null);
}
 

import org.apache.clerezza.rdf.core.serializedform.SupportedFormat; //导入依赖的package包/类
/**
 * Reads the silk output (n-triples) and returns the owl:sameas statements
 * as a result
 *
 * @param file
 * @return
 * @throws IOException
 */
public TripleCollection parseResult(File file) throws IOException {
    Parser parser = Parser.getInstance();
    return parser.parse(new FileInputStream(file), SupportedFormat.N_TRIPLE);
}