Examples with TikaInputStream org.apache.tika.io.TikaInputStream used on opensource projects

Example 6 with TikaInputStream

use of org.apache.tika.io.TikaInputStream in project tika by apache.

the class NetworkParser method parse.

private void parse(TikaInputStream stream, ContentHandler handler, Metadata metadata, ParseContext context) throws IOException, SAXException, TikaException {
    if ("telnet".equals(uri.getScheme())) {
        try (Socket socket = new Socket(uri.getHost(), uri.getPort())) {
            new ParsingTask(stream, new FilterOutputStream(socket.getOutputStream()) {

                @Override
                public void close() throws IOException {
                    socket.shutdownOutput();
                }
            }).parse(socket.getInputStream(), handler, metadata, context);
        }
    } else {
        URL url = uri.toURL();
        URLConnection connection = url.openConnection();
        connection.setDoOutput(true);
        connection.connect();
        try (InputStream input = connection.getInputStream()) {
            new ParsingTask(stream, connection.getOutputStream()).parse(new CloseShieldInputStream(input), handler, metadata, context);
        }
    }
}

Example 7 with TikaInputStream

use of org.apache.tika.io.TikaInputStream in project tika by apache.

the class AutoDetectParser method parse.

public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context) throws IOException, SAXException, TikaException {
    TemporaryResources tmp = new TemporaryResources();
    try {
        TikaInputStream tis = TikaInputStream.get(stream, tmp);
        // Automatically detect the MIME type of the document
        MediaType type = detector.detect(tis, metadata);
        metadata.set(Metadata.CONTENT_TYPE, type.toString());
        // TIKA-216: Zip bomb prevention
        SecureContentHandler sch = handler != null ? new SecureContentHandler(handler, tis) : null;
        //the caller hasn't specified one.
        if (context.get(EmbeddedDocumentExtractor.class) == null) {
            Parser p = context.get(Parser.class);
            if (p == null) {
                context.set(Parser.class, this);
            }
            context.set(EmbeddedDocumentExtractor.class, new ParsingEmbeddedDocumentExtractor(context));
        }
        try {
            // Parse the document
            super.parse(tis, sch, metadata, context);
        } catch (SAXException e) {
            // Convert zip bomb exceptions to TikaExceptions
            sch.throwIfCauseOf(e);
            throw e;
        }
    } finally {
        tmp.dispose();
    }
}

Example 8 with TikaInputStream

use of org.apache.tika.io.TikaInputStream in project tika by apache.

the class DigestingParser method parse.

@Override
public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context) throws IOException, SAXException, TikaException {
    TemporaryResources tmp = new TemporaryResources();
    TikaInputStream tis = TikaInputStream.get(stream, tmp);
    try {
        if (digester != null) {
            digester.digest(tis, metadata, context);
        }
        super.parse(tis, handler, metadata, context);
    } finally {
        tmp.dispose();
    }
}

Example 9 with TikaInputStream

use of org.apache.tika.io.TikaInputStream in project tika by apache.

the class ZipContainerDetector method detect.

public MediaType detect(InputStream input, Metadata metadata) throws IOException {
    // Check if we have access to the document
    if (input == null) {
        return MediaType.OCTET_STREAM;
    }
    TemporaryResources tmp = new TemporaryResources();
    try {
        TikaInputStream tis = TikaInputStream.get(input, tmp);
        // enough for all known formats
        byte[] prefix = new byte[1024];
        int length = tis.peek(prefix);
        MediaType type = detectArchiveFormat(prefix, length);
        if (PackageParser.isZipArchive(type) && TikaInputStream.isTikaInputStream(input)) {
            return detectZipFormat(tis);
        } else if (!type.equals(MediaType.OCTET_STREAM)) {
            return type;
        } else {
            return detectCompressorFormat(prefix, length);
        }
    } finally {
        try {
            tmp.dispose();
        } catch (TikaException e) {
        // ignore
        }
    }
}

Example 10 with TikaInputStream

use of org.apache.tika.io.TikaInputStream in project tika by apache.

the class TikaDetectorConfigTest method testPSTDetectionWithoutZipDetector.

/**
     * TIKA-1708 - If the Zip detector is disabled, either explicitly,
     *  or via giving a list of detectors that it isn't part of, ensure
     *  that detection of PST files still works
     */
@Test
public void testPSTDetectionWithoutZipDetector() throws Exception {
    // Check the one with an exclude
    TikaConfig configWX = getConfig("TIKA-1708-detector-default.xml");
    assertNotNull(configWX.getParser());
    assertNotNull(configWX.getDetector());
    CompositeDetector detectorWX = (CompositeDetector) configWX.getDetector();
    // Check it has the POIFS one, but not the zip one
    assertDetectors(detectorWX, true, false);
    // Check the one with an explicit list
    TikaConfig configCL = getConfig("TIKA-1708-detector-composite.xml");
    assertNotNull(configCL.getParser());
    assertNotNull(configCL.getDetector());
    CompositeDetector detectorCL = (CompositeDetector) configCL.getDetector();
    assertEquals(2, detectorCL.getDetectors().size());
    // Check it also has the POIFS one, but not the zip one
    assertDetectors(detectorCL, true, false);
    // Check that both detectors have a mimetypes with entries
    assertTrue("Not enough mime types: " + configWX.getMediaTypeRegistry().getTypes().size(), configWX.getMediaTypeRegistry().getTypes().size() > 100);
    assertTrue("Not enough mime types: " + configCL.getMediaTypeRegistry().getTypes().size(), configCL.getMediaTypeRegistry().getTypes().size() > 100);
    // Now check they detect PST files correctly
    TikaInputStream stream = TikaInputStream.get(getResourceAsFile("/test-documents/testPST.pst"));
    assertEquals(OutlookPSTParser.MS_OUTLOOK_PST_MIMETYPE, detectorWX.detect(stream, new Metadata()));
    assertEquals(OutlookPSTParser.MS_OUTLOOK_PST_MIMETYPE, detectorCL.detect(stream, new Metadata()));
}