Package org.apache.tika.sax

Examples of org.apache.tika.sax.XHTMLContentHandler

49
50
51
52
53
54
55
56
57
58
59
60
     */
    public void parse(
            final InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        XHTMLContentHandler xhtml =
            new XHTMLContentHandler(handler, metadata);

        Process process = Runtime.getRuntime().exec(command);
        try {
            sendInput(process, stream);
            ignoreError(process);
View Full Code Here
62
63
64
65
66
67
68
69
70
71
     * @see org.apache.tika.parser.microsoft.ooxml.OOXMLExtractor#getXHTML(org.xml.sax.ContentHandler,
     *      org.apache.tika.metadata.Metadata)
     */
    public void getXHTML(ContentHandler handler, Metadata metadata)
            throws SAXException, XmlException, IOException {
        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();
        buildXHTML(xhtml);
        xhtml.endDocument();
    }
View Full Code Here
39
40
41
42
43
44
45
46
47
    public void parse(
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws SAXException {
        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();
        xhtml.endDocument();
    }
View Full Code Here
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
            int bom = reader.read();
            if (bom != '\ufeff') { // zero-width no-break space
                reader.reset();
            }

            XHTMLContentHandler xhtml =
                new XHTMLContentHandler(handler, metadata);
            xhtml.startDocument();

            xhtml.startElement("p");
            char[] buffer = new char[4096];
            int n = reader.read(buffer);
            while (n != -1) {
                xhtml.characters(buffer, 0, n);
                n = reader.read(buffer);
            }
            xhtml.endElement("p");

            xhtml.endDocument();
        } catch (UnsupportedEncodingException e) {
            throw new TikaException(
                    "Unsupported text encoding: " + encoding, e);
        }
    }
View Full Code Here
65
66
67
68
69
70
71
72
73
74
75
    private final XHTMLContentHandler handler;

    private PDF2XHTML(ContentHandler handler, Metadata metadata)
            throws IOException {
        this.handler = new XHTMLContentHandler(handler, metadata);

        // TIKA-292: Ignore unneeded PDF operators
        // TODO: Remove this once PDFBox is no longer so verbose
        OperatorProcessor ignore = new OperatorProcessor() {
            @Override @SuppressWarnings("unchecked")
View Full Code Here
43
44
45
46
47
48
49
50
51
52
53
54
    public void parse(
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        final XHTMLContentHandler xhtml =
            new XHTMLContentHandler(handler,metadata);

        try {
            SAXParserFactory factory = SAXParserFactory.newInstance();
            factory.setValidating(false);
            factory.setNamespaceAware(true);
View Full Code Here
64
65
66
67
68
69
70
        }
    }

    public HtmlHandler(
            HtmlParser parser, ContentHandler handler, Metadata metadata) {
        this(parser, new XHTMLContentHandler(handler, metadata), metadata);
    }
View Full Code Here
66
67
68
69
70
71
72
73
74
            throw new TikaException("Can't read JPEG metadata", e);
        } catch (MetadataException e) {
            throw new TikaException("Can't read JPEG metadata", e);
        }

        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();
        xhtml.endDocument();
    }
View Full Code Here
50
51
52
53
54
55
56
57
58
                metadata.set("width", Integer.toString(reader.getWidth(0)));
                reader.dispose();
            }
        }

        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();
        xhtml.endDocument();
    }
View Full Code Here
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
            InputStream stream, ContentHandler handler,
            Metadata metadata, ParseContext context)
            throws IOException, SAXException, TikaException {
        metadata.set(Metadata.CONTENT_TYPE, "application/x-gzip");

        XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
        xhtml.startDocument();

        // At the end we want to close the gzip stream to release any associated
        // resources, but the underlying document stream should not be closed
        InputStream gzip =
            new GZIPInputStream(new CloseShieldInputStream(stream));
        try {
            Metadata entrydata = new Metadata();
            String name = metadata.get(Metadata.RESOURCE_NAME_KEY);
            if (name != null && name.length() > 0) {
                entrydata.set(
                        Metadata.RESOURCE_NAME_KEY,
                        GzipUtils.getUncompressedFilename(name));
            }
            // Use the delegate parser to parse the compressed document
            super.parse(
                    new CloseShieldInputStream(gzip),
                    new EmbeddedContentHandler(
                            new BodyContentHandler(xhtml)),
                    entrydata, context);
        } finally {
            gzip.close();
        }

        xhtml.endDocument();
    }
View Full Code Here
TOP

Related Classes of org.apache.tika.sax.XHTMLContentHandler

  • de.innosystec.unrar.tika.RARParser
  • org.apache.jackrabbit.core.query.lucene.BlockingParser
  • org.apache.jackrabbit.core.query.pdf.PDF2XHTML
  • org.apache.jackrabbit.oak.http.HtmlRepresentation
  • org.apache.tika.fork.ForkTestParser
  • org.apache.tika.parser.asm.XHTMLClassVisitor
  • org.apache.tika.parser.audio.AudioParser
  • org.apache.tika.parser.audio.MidiParser
  • org.apache.tika.parser.chm.ChmParser
  • org.apache.tika.parser.dwg.DWGParser
    • Copyright © 2018 www.massapicom. All rights reserved.
    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.