Examples of org.apache.tika.io.CloseShieldInputStream

org.apache.tika.io.CloseShieldInputStream
Proxy stream that prevents the underlying input stream from being closed.
This class is typically used in cases where an input stream needs to be passed to a component that wants to explicitly close the stream even if more input would still be available to other components. @since Apache Tika 0.4, copied from Commons IO 1.4

            stream = new BufferedInputStream(stream);
        }


        // Protect the stream from being closed by CyberNeko
        // TODO: Is this still needed, given our use of TagSoup?
        stream = new CloseShieldInputStream(stream);


        // Prepare the input source using the encoding hint if available
        InputSource source = new InputSource(stream); 
        source.setEncoding(getEncoding(stream, metadata));

View Full Code Here

               }


               metadata.add(Metadata.CONTENT_TYPE, type.getType().toString());
               xhtml.startDocument();
               context.getSAXParser().parse(
                       new CloseShieldInputStream(entryStream),
                       new OfflineContentHandler(contentHandler)
               );
               xhtml.endDocument();
            }

View Full Code Here

                // the exception is fine here, deployments without this feature
                // are inherently vulnerable to XML denial-of-service attacks.
            }
            SAXParser parser = factory.newSAXParser();
            parser.parse(
                    new CloseShieldInputStream(stream),
                    new OfflineContentHandler(
                            new NSNormalizerContentHandler(dh)));
        } catch (ParserConfigurationException e) {
            throw new TikaException("XML parser configuration error", e);
        }

View Full Code Here

    public void parse(
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        PDDocument pdfDocument =
            PDDocument.load(new CloseShieldInputStream(stream), true);
        try {
            if (pdfDocument.isEncrypted()) {
                try {
                    String password = metadata.get(PASSWORD);
                    if (password == null) {

View Full Code Here

                        (OPCPackage) tis.getOpenContainer());
            } else if (tis != null && tis.hasFile()) {
                poiExtractor = (POIXMLTextExtractor)
                        ExtractorFactory.createExtractor(tis.getFile());
            } else {
                InputStream shield = new CloseShieldInputStream(stream);
                poiExtractor = (POIXMLTextExtractor)
                        ExtractorFactory.createExtractor(shield);
            }


            POIXMLDocument document = poiExtractor.getDocument();

View Full Code Here

                Iterator<ImageReader> iterator =
                    ImageIO.getImageReadersByMIMEType(type);
                if (iterator.hasNext()) {
                    ImageReader reader = iterator.next();
                    reader.setInput(ImageIO.createImageInputStream(
                            new CloseShieldInputStream(stream)));
                    
                    metadata.set(Metadata.IMAGE_WIDTH, Integer.toString(reader.getWidth(0)));
                    metadata.set(Metadata.IMAGE_LENGTH, Integer.toString(reader.getHeight(0)));
                    metadata.set("height", Integer.toString(reader.getHeight(0)));
                    metadata.set("width", Integer.toString(reader.getWidth(0)));

View Full Code Here

            pipeMode = false;
            if (serverMode) {
                new TikaServer(Integer.parseInt(arg)).start();
            } else if (arg.equals("-")) {
                InputStream stream =
                    TikaInputStream.get(new CloseShieldInputStream(System.in));
                try {
                    type.process(stream, System.out, new Metadata());
                } finally {
                    stream.close();
                }

View Full Code Here

            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        // Automatically detect the character encoding
        AutoDetectReader reader = new AutoDetectReader(
                new CloseShieldInputStream(stream), metadata, LOADER);
        try {
            Charset charset = reader.getCharset();
            MediaType type = new MediaType(MediaType.TEXT_PLAIN, charset);
            metadata.set(Metadata.CONTENT_TYPE, type.toString());
            // deprecated, see TIKA-431

View Full Code Here

            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        // Automatically detect the character encoding
        AutoDetectReader reader = new AutoDetectReader(
                new CloseShieldInputStream(stream), metadata, LOADER);
        try {
            Charset charset = reader.getCharset();
            String previous = metadata.get(Metadata.CONTENT_TYPE);
            if (previous == null || previous.startsWith("text/html")) {
                MediaType type = new MediaType(MediaType.TEXT_HTML, charset);

View Full Code Here

            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        // At the end we want to close the archive stream to release
        // any associated resources, but the underlying document stream
        // should not be closed
        stream = new CloseShieldInputStream(stream);


        // Ensure that the stream supports the mark feature
        stream = new BufferedInputStream(stream);


        ArchiveInputStream ais;

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of org.apache.tika.io.CloseShieldInputStream

com.cloudera.cdk.morphline.tika.decompress.DecompressBuilder$Decompress

com.cloudera.cdk.morphline.tika.decompress.EmbeddedExtractor

com.cloudera.cdk.morphline.tika.decompress.UnpackBuilder$Unpack

org.apache.tika.cli.TikaCLI

org.apache.tika.detect.POIFSContainerDetector

org.apache.tika.detect.XmlRootExtractor

org.apache.tika.extractor.EmbeddedDocumentExtractor

org.apache.tika.extractor.ParsingEmbeddedDocumentExtractor

org.apache.tika.parser.code.SourceCodeParser

org.apache.tika.parser.crypto.Pkcs7Parser

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.