Examples with SAXTransformerFactory javax.xml.transform.sax.SAXTransformerFactory used on opensource projects

Example 21 with SAXTransformerFactory

use of javax.xml.transform.sax.SAXTransformerFactory in project jackrabbit by apache.

the class ExportDocViewTest method doTestExportDocView.

/**
     * Tests session.exportDocView with the different argument possibilities.
     * The flag withHandler decides if the method requiring a ContentHandler as
     * argument is called. The class org.apache.xml.serialize.XMLSerializer is
     * taken as ContentHandler in this case. In both cases ( export with a
     * ContentHandler and export with Stream) the test node is exported to the
     * file defined in the setUp. This exported file is parsed using
     * javax.xml.transform package and the receiving document is compared with
     * the test node and its properties and child nodes in the repository.
     *
     * @param withHandler boolean, decides to call method requiring a
     *                    ContentHandler as argument
     * @param skipBinary
     * @param noRecurse
     */
public void doTestExportDocView(boolean withHandler, boolean skipBinary, boolean noRecurse) throws RepositoryException, IOException, SAXException, TransformerException {
    this.skipBinary = skipBinary;
    this.noRecurse = noRecurse;
    this.withHandler = withHandler;
    BufferedOutputStream os = new BufferedOutputStream(new FileOutputStream(file));
    try {
        if (withHandler) {
            SAXTransformerFactory stf = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
            TransformerHandler th = stf.newTransformerHandler();
            th.setResult(new StreamResult(os));
            session.exportDocumentView(testPath, th, skipBinary, noRecurse);
        } else {
            session.exportDocumentView(testPath, os, skipBinary, noRecurse);
        }
    } finally {
        os.close();
    }
    // build the DOM tree
    InputStream in = new BufferedInputStream(new FileInputStream(file));
    doc = readDocument(in);
    compareTree();
}

Example 22 with SAXTransformerFactory

use of javax.xml.transform.sax.SAXTransformerFactory in project jackrabbit-oak by apache.

the class HtmlRepresentation method startResponse.

private XHTMLContentHandler startResponse(HttpServletResponse response, String title) throws IOException {
    try {
        response.setContentType("text/html");
        response.setCharacterEncoding("UTF-8");
        SAXTransformerFactory factory = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
        TransformerHandler handler = factory.newTransformerHandler();
        Transformer transformer = handler.getTransformer();
        transformer.setOutputProperty(OutputKeys.METHOD, "html");
        transformer.setOutputProperty(OutputKeys.ENCODING, "UTF-8");
        handler.setResult(new StreamResult(response.getOutputStream()));
        Metadata metadata = new Metadata();
        metadata.set(Metadata.TITLE, title);
        return new XHTMLContentHandler(handler, metadata);
    } catch (TransformerConfigurationException e) {
        throw new IOException(e);
    }
}

Example 23 with SAXTransformerFactory

use of javax.xml.transform.sax.SAXTransformerFactory in project lucene-solr by apache.

the class TikaEntityProcessor method getHtmlHandler.

private static ContentHandler getHtmlHandler(Writer writer) throws TransformerConfigurationException {
    SAXTransformerFactory factory = (SAXTransformerFactory) TransformerFactory.newInstance();
    TransformerHandler handler = factory.newTransformerHandler();
    handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "html");
    handler.setResult(new StreamResult(writer));
    return new ContentHandlerDecorator(handler) {

        @Override
        public void startElement(String uri, String localName, String name, Attributes atts) throws SAXException {
            if (XHTMLContentHandler.XHTML.equals(uri)) {
                uri = null;
            }
            if (!"head".equals(localName)) {
                super.startElement(uri, localName, name, atts);
            }
        }

        @Override
        public void endElement(String uri, String localName, String name) throws SAXException {
            if (XHTMLContentHandler.XHTML.equals(uri)) {
                uri = null;
            }
            if (!"head".equals(localName)) {
                super.endElement(uri, localName, name);
            }
        }

        @Override
        public void startPrefixMapping(String prefix, String uri) {
        /*no op*/
        }

        @Override
        public void endPrefixMapping(String prefix) {
        /*no op*/
        }
    };
}

Example 24 with SAXTransformerFactory

use of javax.xml.transform.sax.SAXTransformerFactory in project tika by apache.

the class HtmlParserTest method makeHtmlTransformer.

/**
     * Create ContentHandler that transforms SAX events into textual HTML output,
     * and writes it out to <writer> - typically this is a StringWriter.
     *
     * @param writer Where to write resulting HTML text.
     * @return ContentHandler suitable for passing to parse() methods.
     * @throws Exception
     */
private ContentHandler makeHtmlTransformer(Writer writer) throws Exception {
    SAXTransformerFactory factory = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
    TransformerHandler handler = factory.newTransformerHandler();
    handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "html");
    handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "no");
    handler.getTransformer().setOutputProperty(OutputKeys.ENCODING, "utf-8");
    handler.setResult(new StreamResult(writer));
    return handler;
}

Example 25 with SAXTransformerFactory

use of javax.xml.transform.sax.SAXTransformerFactory in project tika by apache.

the class OutlookParserTest method testOutlookHTMLfromRTF.

@Test
public void testOutlookHTMLfromRTF() throws Exception {
    Parser parser = new AutoDetectParser();
    Metadata metadata = new Metadata();
    // Check the HTML version
    StringWriter sw = new StringWriter();
    SAXTransformerFactory factory = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
    TransformerHandler handler = factory.newTransformerHandler();
    handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "xml");
    handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "yes");
    handler.setResult(new StreamResult(sw));
    try (InputStream stream = OutlookParserTest.class.getResourceAsStream("/test-documents/test-outlook2003.msg")) {
        parser.parse(stream, handler, metadata, new ParseContext());
    }
    // As the HTML version should have been processed, ensure
    //  we got some of the links
    String content = sw.toString().replaceAll("<p>\\s+", "<p>");
    assertContains("<dd>New Outlook User</dd>", content);
    assertContains("designed <i>to help you", content);
    assertContains("<p><a href=\"http://r.office.microsoft.com/r/rlidOutlookWelcomeMail10?clid=1033\">Cached Exchange Mode</a>", content);
    // Link - check text around it, and the link itself
    assertContains("sign up for a free subscription", content);
    assertContains("Office Newsletter", content);
    assertContains("newsletter will be sent to you", content);
    assertContains("http://r.office.microsoft.com/r/rlidNewsletterSignUp?clid=1033", content);
    // Make sure we don't have nested html docs
    assertEquals(2, content.split("<body>").length);
    assertEquals(2, content.split("<\\/body>").length);
}