Examples with BodyContentHandler org.apache.tika.sax.BodyContentHandler used on opensource projects

Example 81 with BodyContentHandler

use of org.apache.tika.sax.BodyContentHandler in project tika by apache.

the class ArParserTest method testArParsing.

@Test
public void testArParsing() throws Exception {
    Parser parser = new AutoDetectParser();
    ContentHandler handler = new BodyContentHandler();
    Metadata metadata = new Metadata();
    try (InputStream stream = ArParserTest.class.getResourceAsStream("/test-documents/testARofText.ar")) {
        parser.parse(stream, handler, metadata, recursingContext);
    }
    assertEquals("application/x-archive", metadata.get(Metadata.CONTENT_TYPE));
    String content = handler.toString();
    assertContains("testTXT.txt", content);
    assertContains("Test d'indexation de Txt", content);
    assertContains("http://www.apache.org", content);
    try (InputStream stream = ArParserTest.class.getResourceAsStream("/test-documents/testARofSND.ar")) {
        parser.parse(stream, handler, metadata, recursingContext);
    }
    assertEquals("application/x-archive", metadata.get(Metadata.CONTENT_TYPE));
    content = handler.toString();
    assertContains("testAU.au", content);
}

Example 82 with BodyContentHandler

use of org.apache.tika.sax.BodyContentHandler in project tika by apache.

the class ChmParser method parsePage.

private void parsePage(byte[] byteObject, Parser htmlParser, ContentHandler xhtml, ParseContext context) throws TikaException {
    // throws IOException
    InputStream stream = null;
    Metadata metadata = new Metadata();
    // -1
    ContentHandler handler = new EmbeddedContentHandler(new BodyContentHandler(xhtml));
    try {
        stream = new ByteArrayInputStream(byteObject);
        htmlParser.parse(stream, handler, metadata, context);
    } catch (SAXException e) {
        throw new RuntimeException(e);
    } catch (IOException e) {
    // Pushback overflow from tagsoup
    }
}

Example 83 with BodyContentHandler

use of org.apache.tika.sax.BodyContentHandler in project tika by apache.

the class TIAParsingExample method testTeeContentHandler.

public static void testTeeContentHandler(String filename) throws Exception {
    InputStream stream = new ByteArrayInputStream(new byte[0]);
    Metadata metadata = new Metadata();
    ParseContext context = new ParseContext();
    Parser parser = new AutoDetectParser();
    LinkContentHandler linkCollector = new LinkContentHandler();
    try (OutputStream output = new FileOutputStream(new File(filename))) {
        ContentHandler handler = new TeeContentHandler(new BodyContentHandler(output), linkCollector);
        parser.parse(stream, handler, metadata, context);
    }
}

Example 84 with BodyContentHandler

use of org.apache.tika.sax.BodyContentHandler in project tika by apache.

the class StringsParserTest method testParse.

@Test
public void testParse() throws Exception {
    assumeTrue(canRun());
    String resource = "/test-documents/testOCTET_header.dbase3";
    String[] content = { "CLASSNO", "TITLE", "ITEMNO", "LISTNO", "LISTDATE" };
    String[] met_attributes = { "min-len", "encoding", "strings:file_output" };
    StringsConfig stringsConfig = new StringsConfig();
    FileConfig fileConfig = new FileConfig();
    Parser parser = new StringsParser();
    ContentHandler handler = new BodyContentHandler();
    Metadata metadata = new Metadata();
    ParseContext context = new ParseContext();
    context.set(StringsConfig.class, stringsConfig);
    context.set(FileConfig.class, fileConfig);
    try (InputStream stream = StringsParserTest.class.getResourceAsStream(resource)) {
        parser.parse(stream, handler, metadata, context);
    } catch (Exception e) {
        e.printStackTrace();
    }
    // Content
    for (String word : content) {
        assertTrue(handler.toString().contains(word));
    }
    // Metadata
    Arrays.equals(met_attributes, metadata.names());
}

Example 85 with BodyContentHandler

use of org.apache.tika.sax.BodyContentHandler in project tika by apache.

the class TXTParserTest method testEmptyText.

@Test
public void testEmptyText() throws Exception {
    ContentHandler handler = new BodyContentHandler();
    Metadata metadata = new Metadata();
    parser.parse(new ByteArrayInputStream(new byte[0]), handler, metadata, new ParseContext());
    assertEquals("text/plain; charset=UTF-8", metadata.get(Metadata.CONTENT_TYPE));
    assertEquals("\n", handler.toString());
}