Examples of org.apache.tika.mime.MimeTypes

org.apache.tika.mime.MimeTypes
This class is a MimeType repository. It gathers a set of MimeTypes and enables to retrieves a content-type from its name, from a file name, or from a magic character sequence.
The MIME type detection methods that take an {@link InputStream} asan argument will never reads more than {@link #getMinLength()} bytesfrom the stream. Also the given stream is never {@link InputStream#close() closed}, {@link InputStream#mark(int) marked}, or {@link InputStream#reset() reset} by the methods. Thus a client canuse the {@link InputStream#markSupported() mark feature} of the stream(if available) to restore the stream back to the state it was before type detection if it wants to process the stream based on the detected type.


        return Normalizer.normalize(text, Normalizer.Form.NFC);
    }


    private String extractTextWithTika(byte[] textBytes, Metadata metadata) throws TikaException, SAXException, IOException {
        AutoDetectParser parser = new AutoDetectParser(new MimeTypes());
        ByteArrayOutputStream baos = new ByteArrayOutputStream();
        OutputStreamWriter writer = new OutputStreamWriter(baos, "UTF-8");
        ContentHandler handler = new BodyContentHandler(writer);
        ParseContext context = new ParseContext();
        context.set(PDFParserConfig.class, new LumifyParserConfig());

View Full Code Here

     * @return MIME type of the document
     * @throws IOException if the document stream could not be read
     */
    private MimeType getMimeType(InputStream stream, Metadata metadata)
            throws IOException {
        MimeTypes types = config.getMimeRepository();


        // Get type based on magic prefix
        stream.mark(types.getMinLength());
        try {
            byte[] prefix = getPrefix(stream, types.getMinLength());
            MimeType type = types.getMimeType(prefix);
            if (type != null) {
                return type;
            }
        } finally {
            stream.reset();
        }


        // Get type based on resourceName hint (if available)
        String resourceName = metadata.get(Metadata.RESOURCE_NAME_KEY);
        if (resourceName != null) {
            MimeType type = types.getMimeType(resourceName);
            if (type != null) {
                return type;
            }
        }


        // Get type based on metadata hint (if available)
        String typename = metadata.get(Metadata.CONTENT_TYPE);
        if (typename != null) {
            try {
                return types.forName(typename);
            } catch (MimeTypeException e) {
                // Malformed type name, ignore
            }
        }


        // Finally, use the default type if no matches found
        try {
            return types.forName(MimeTypes.DEFAULT);
        } catch (MimeTypeException e) {
            // Should never happen
            return null;
        }
    }

View Full Code Here

  private static final Logger LOG = LoggerFactory.getLogger(MimeUtil.class.getName());


  public MimeUtil(Configuration conf) {
    tika = new Tika();
    ObjectCache objectCache = ObjectCache.get(conf);
    MimeTypes mimeTypez = (MimeTypes) objectCache.getObject(MimeTypes.class
        .getName());
    if (mimeTypez == null) {
      try {
          String customMimeTypeFile = conf.get("mime.types.file");
          if (customMimeTypeFile!=null && customMimeTypeFile.equals("")==false){

View Full Code Here

             d.detect(tis, new Metadata())
       );
    }
    
    public void testTruncatedFiles() throws Exception {
        MimeTypes mimeTypes = MimeTypesFactory.create("tika-mimetypes.xml");
        ContainerAwareDetector detector = new ContainerAwareDetector(mimeTypes);
        
        // First up a truncated OOXML (zip) file
        InputStream input = getTestDoc("testEXCEL.xlsx");
        byte [] buffer = new byte[300];

View Full Code Here

    String productId = "123";
    int refIndex = 0;


    // Create a new ReferenceResource using a Reference instance.
    Reference reference = new Reference("original", "dataStore", 1000,
      new MimeTypes().forName("text/plain"));


    ReferenceResource resource = new ReferenceResource(productId, refIndex,
      reference, new File("/tmp"));

View Full Code Here

    metadataEntries.put("CAS.Test", "test value");
    Metadata metadata = new Metadata();
    metadata.addMetadata(metadataEntries);


    Reference reference = new Reference("original", "dataStore", 1000,
      new MimeTypes().forName("text/plain"));
    List<Reference> references = new ArrayList<Reference>();
    references.add(reference);


    ProductType productType = new ProductType("1", "GenericFile", "test type",
      "repository", "versioner");

View Full Code Here

    ProductType productType = new ProductType("1", "GenericFile", "test type",
      "repository", "versioner");


    // Create a ProductResource using Reference, Metadata and Product instances.
    Reference reference1 = new Reference("original1", "dataStore1", 500,
      new MimeTypes().forName("text/plain"));
    List<Reference> references1 = new ArrayList<Reference>();
    references1.add(reference1);


    Hashtable metadataEntries1 = new Hashtable<String, Object>();
    metadataEntries1.put("product1_meta", "test1");
    Metadata metadata1 = new Metadata();
    metadata1.addMetadata(metadataEntries1);


    Product product1 = new Product();
    product1.setProductId("123");
    product1.setProductName("test.txt");
    product1.setProductType(productType);


    ProductResource productResource1 = new ProductResource(product1, metadata1,
      references1, new File("/tmp"));




    // Create another ProductResource using Reference, Metadata and Product
    // instances.
    Reference reference2 = new Reference("original2", "dataStore2", 1000,
      new MimeTypes().forName("application/pdf"));
    List<Reference> references2 = new ArrayList<Reference>();
    references2.add(reference2);


    Hashtable metadataEntries2 = new Hashtable<String, Object>();
    metadataEntries2.put("product2_meta", "test2");

View Full Code Here

    metadataEntries1.put("CAS.ProductReceivedTime", "2013-09-12T16:25:50.662Z");
    Metadata metadata1 = new Metadata();
    metadata1.addMetadata(metadataEntries1);


    Reference reference1 = new Reference("original1", "dataStore1", 1000,
      new MimeTypes().forName("text/plain"));


    ProductType productType1 = new ProductType("1", "TestType", "test type 1",
      "repository1", "versioner1");


    Product product1 = new Product();
    product1.setProductId("123");
    product1.setProductName("test product");
    product1.setProductStructure("flat");
    product1.setProductType(productType1);


    FileTransferStatus status1 = new FileTransferStatus(reference1, 1000, 100,
      product1);




    // Create another FileTransferStatus instance using Metadata, Reference,
    // ProductType and Product instances.
    Hashtable metadataEntries2 = new Hashtable<String, Object>();
    metadataEntries2.put("CAS.ProductReceivedTime", "2011-04-11T11:59:59.662Z");
    Metadata metadata2 = new Metadata();
    metadata2.addMetadata(metadataEntries2);


    Reference reference2 = new Reference("original2", "dataStore2", 500,
      new MimeTypes().forName("application/pdf"));


    ProductType productType2 = new ProductType("2", "TestType2", "test type 2",
        "repository2", "versioner2");


    Product product2 = new Product();

View Full Code Here

    metadataEntries.put("CAS.ProductReceivedTime", "2013-09-12T16:25:50.662Z");
    Metadata metadata = new Metadata();
    metadata.addMetadata(metadataEntries);


    Reference reference = new Reference("original", "dataStore", 1000,
      new MimeTypes().forName("text/plain"));


    ProductType productType = new ProductType("1", "GenericFile", "test type",
      "repository", "versioner");


    Product product = new Product();

View Full Code Here

    */
   public static void main(String[] args) throws DataTransferException,
         IOException, URISyntaxException {
      String usage = "LocalFileTransfer --productName <name> --productRepo <repo> [--dir <dirRef>] [--files <origRef 1>...<origRef N>]\n";


      MimeTypes mimeTypeRepo;
      try {
         mimeTypeRepo = MimeTypesFactory
               .create(System
                     .getProperty("org.apache.oodt.cas.filemgr.mime.type.repository"));
      } catch (MimeTypeException e) {
         e.printStackTrace();
         throw new IOException(e.getMessage());
      }


      String productName = null;
      String productRepo = null;
      String transferType = null;
      Reference dirReference = null;


      List<Reference> fileReferences = null;


      for (int i = 0; i < args.length; i++) {
         if (args[i].equals("--dir")) {
            transferType = "dir";
            dirReference = new Reference();
            dirReference.setOrigReference(new File(new URI(args[++i])).toURI()
                  .toString());
            LOG.log(Level.FINER,
                  "LocalFileTransfer.main: Generated orig reference: "
                        + dirReference.getOrigReference());
         } else if (args[i].equals("--files")) {
            transferType = "files";
            fileReferences = new Vector<Reference>();
            for (int j = i + 1; j < args.length; j++) {
               LOG.log(Level.FINER, "LocalFileTransfer.main: Adding file ref: "
                     + args[j]);
               fileReferences.add(new Reference(args[j], null,
                     new File(args[j]).length(), mimeTypeRepo
                           .getMimeType(args[j])));
            }
         } else if (args[i].equals("--productName")) {
            productName = args[++i];
         } else if (args[i].equals("--productRepo")) {

View Full Code Here

0 1 2

TOP

Related Classes of org.apache.tika.mime.MimeTypes

com.cloudera.cdk.morphline.tika.DetectMimeTypeBuilder$DetectMimeType

dk.i2m.drupal.util.EntityUtils

io.lumify.tikaTextExtractor.TikaTextExtractorGraphPropertyWorker

org.apache.marmotta.platform.ldp.util.LdpUtils

org.apache.nutch.util.MimeUtil

org.apache.oodt.cas.filemgr.datatransfer.LocalDataTransferer

org.apache.oodt.cas.product.jaxrs.resources.DatasetResourceTest

org.apache.oodt.cas.product.jaxrs.resources.ProductResourceTest

org.apache.oodt.cas.product.jaxrs.resources.ReferenceResourceTest

org.apache.oodt.cas.product.jaxrs.resources.TransferResourceTest

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.