Examples with XMLParser org.apache.tika.parser.xml.XMLParser used on opensource projects

Example 1 with XMLParser

use of org.apache.tika.parser.xml.XMLParser in project tika by apache.

the class TikaParserConfigTest method defaultParserBlacklist.

/**
     * TIKA-1558 It should be possible to exclude Parsers from being picked up by
     * DefaultParser.
     */
@Test
public void defaultParserBlacklist() throws Exception {
    TikaConfig config = new TikaConfig();
    assertNotNull(config.getParser());
    assertNotNull(config.getDetector());
    CompositeParser cp = (CompositeParser) config.getParser();
    List<Parser> parsers = cp.getAllComponentParsers();
    boolean hasXML = false;
    for (Parser p : parsers) {
        if (p instanceof XMLParser) {
            hasXML = true;
            break;
        }
    }
    assertTrue("Default config should include an XMLParser.", hasXML);
    // This custom TikaConfig should exclude XMLParser and all of its subclasses.
    config = getConfig("TIKA-1558-blacklistsub.xml");
    cp = (CompositeParser) config.getParser();
    parsers = cp.getAllComponentParsers();
    for (Parser p : parsers) {
        if (p instanceof XMLParser)
            fail("Custom config should not include an XMLParser (" + p.getClass() + ").");
    }
}

Example 2 with XMLParser

use of org.apache.tika.parser.xml.XMLParser in project tika by apache.

the class TIAParsingExample method useCompositeParser.

public static void useCompositeParser() throws Exception {
    InputStream stream = new ByteArrayInputStream(new byte[0]);
    ContentHandler handler = new DefaultHandler();
    ParseContext context = new ParseContext();
    Map<MediaType, Parser> parsersByType = new HashMap<MediaType, Parser>();
    parsersByType.put(MediaType.parse("text/html"), new HtmlParser());
    parsersByType.put(MediaType.parse("application/xml"), new XMLParser());
    CompositeParser parser = new CompositeParser();
    parser.setParsers(parsersByType);
    parser.setFallback(new TXTParser());
    Metadata metadata = new Metadata();
    metadata.set(Metadata.CONTENT_TYPE, "text/html");
    parser.parse(stream, handler, metadata, context);
}