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

Example 6 with SAXTransformerFactory

use of javax.xml.transform.sax.SAXTransformerFactory in project intellij-community by JetBrains.

the class ExportTestResultsAction method getOutputText.

@Nullable
private String getOutputText(ExportTestResultsConfiguration config) throws IOException, TransformerException, SAXException {
    ExportTestResultsConfiguration.ExportFormat exportFormat = config.getExportFormat();
    SAXTransformerFactory transformerFactory = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
    TransformerHandler handler;
    if (exportFormat == ExportTestResultsConfiguration.ExportFormat.Xml) {
        handler = transformerFactory.newTransformerHandler();
        handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "yes");
        handler.getTransformer().setOutputProperty("{http://xml.apache.org/xslt}indent-amount", "4");
    } else {
        Source xslSource;
        if (config.getExportFormat() == ExportTestResultsConfiguration.ExportFormat.BundledTemplate) {
            URL bundledXsltUrl = getClass().getResource("intellij-export.xsl");
            xslSource = new StreamSource(URLUtil.openStream(bundledXsltUrl));
        } else {
            File xslFile = new File(config.getUserTemplatePath());
            if (!xslFile.isFile()) {
                showBalloon(myRunConfiguration.getProject(), MessageType.ERROR, ExecutionBundle.message("export.test.results.custom.template.not.found", xslFile.getPath()), null);
                return null;
            }
            xslSource = new StreamSource(xslFile);
        }
        handler = transformerFactory.newTransformerHandler(xslSource);
        handler.getTransformer().setParameter("TITLE", ExecutionBundle.message("export.test.results.filename", myRunConfiguration.getName(), myRunConfiguration.getType().getDisplayName()));
    }
    StringWriter w = new StringWriter();
    handler.setResult(new StreamResult(w));
    try {
        TestResultsXmlFormatter.execute(myModel.getRoot(), myRunConfiguration, myModel.getProperties(), handler);
    } catch (ProcessCanceledException e) {
        return null;
    }
    return w.toString();
}

Example 7 with SAXTransformerFactory

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

the class TikaResource method produceOutput.

private StreamingOutput produceOutput(final InputStream is, final MultivaluedMap<String, String> httpHeaders, final UriInfo info, final String format) {
    final Parser parser = createParser();
    final Metadata metadata = new Metadata();
    final ParseContext context = new ParseContext();
    fillMetadata(parser, metadata, context, httpHeaders);
    fillParseContext(context, httpHeaders, parser);
    logRequest(LOG, info, metadata);
    return new StreamingOutput() {

        public void write(OutputStream outputStream) throws IOException, WebApplicationException {
            Writer writer = new OutputStreamWriter(outputStream, UTF_8);
            ContentHandler content;
            try {
                SAXTransformerFactory factory = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
                TransformerHandler handler = factory.newTransformerHandler();
                handler.getTransformer().setOutputProperty(OutputKeys.METHOD, format);
                handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "yes");
                handler.getTransformer().setOutputProperty(OutputKeys.ENCODING, UTF_8.name());
                handler.setResult(new StreamResult(writer));
                content = new ExpandedTitleContentHandler(handler);
            } catch (TransformerConfigurationException e) {
                throw new WebApplicationException(e);
            }
            parse(parser, LOG, info.getPath(), is, content, metadata, context);
        }
    };
}

Example 8 with SAXTransformerFactory

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

the class TikaCLI method getTransformerHandler.

/**
     * Returns a transformer handler that serializes incoming SAX events
     * to XHTML or HTML (depending the given method) using the given output
     * encoding.
     *
     * @see <a href="https://issues.apache.org/jira/browse/TIKA-277">TIKA-277</a>
     * @param output output stream
     * @param method "xml" or "html"
     * @param encoding output encoding,
     *                 or <code>null</code> for the platform default
     * @return {@link System#out} transformer handler
     * @throws TransformerConfigurationException
     *         if the transformer can not be created
     */
private static TransformerHandler getTransformerHandler(OutputStream output, String method, String encoding, boolean prettyPrint) throws TransformerConfigurationException {
    SAXTransformerFactory factory = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
    TransformerHandler handler = factory.newTransformerHandler();
    handler.getTransformer().setOutputProperty(OutputKeys.METHOD, method);
    handler.getTransformer().setOutputProperty(OutputKeys.INDENT, prettyPrint ? "yes" : "no");
    if (encoding != null) {
        handler.getTransformer().setOutputProperty(OutputKeys.ENCODING, encoding);
    }
    handler.setResult(new StreamResult(output));
    return handler;
}

Example 9 with SAXTransformerFactory

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

the class OOXMLParserTest method testEmbeddedPDF.

// TIKA-989:
@Test
public void testEmbeddedPDF() throws Exception {
    Metadata metadata = new Metadata();
    StringWriter sw = new StringWriter();
    SAXTransformerFactory factory = (SAXTransformerFactory) SAXTransformerFactory.newInstance();
    TransformerHandler handler = factory.newTransformerHandler();
    handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "xml");
    handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "no");
    handler.setResult(new StreamResult(sw));
    try (InputStream input = OOXMLParserTest.class.getResourceAsStream("/test-documents/testWORD_embedded_pdf.docx")) {
        new OOXMLParser().parse(input, handler, metadata, new ParseContext());
    }
    String xml = sw.toString();
    int i = xml.indexOf("Here is the pdf file:");
    int j = xml.indexOf("<div class=\"embedded\" id=\"rId5\"/>");
    int k = xml.indexOf("Bye Bye");
    int l = xml.indexOf("<div class=\"embedded\" id=\"rId6\"/>");
    int m = xml.indexOf("Bye for real.");
    assertTrue(i != -1);
    assertTrue(j != -1);
    assertTrue(k != -1);
    assertTrue(l != -1);
    assertTrue(m != -1);
    assertTrue(i < j);
    assertTrue(j < k);
    assertTrue(k < l);
    assertTrue(l < m);
}

Example 10 with SAXTransformerFactory

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

the class OutlookParserTest method testOutlookHTMLVersion.

@Test
public void testOutlookHTMLVersion() 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/testMSG_chinese.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();
    assertContains("<dd>tests.chang@fengttt.com</dd>", content);
    assertContains("<p>Alfresco MSG format testing", content);
    assertContains("<li>1", content);
    assertContains("<li>2", content);
    // Make sure we don't have nested html docs
    assertEquals(2, content.split("<body>").length);
    assertEquals(2, content.split("<\\/body>").length);
    // Make sure that the Chinese actually came through
    assertContains("張毓倫", metadata.get(TikaCoreProperties.CREATOR));
    assertContains("陳惠珍", content);
    assertEquals("tests.chang@fengttt.com", metadata.get(Message.MESSAGE_TO_EMAIL));
    assertEquals("Tests Chang@FT (張毓倫)", metadata.get(Office.MAPI_FROM_REPRESENTING_NAME));
    assertEquals("/O=FT GROUP/OU=FT/CN=RECIPIENTS/CN=LYDIACHANG", metadata.get(Office.MAPI_FROM_REPRESENTING_EMAIL));
}