Examples with ContentHandler org.xml.sax.ContentHandler used on opensource projects

Example 76 with ContentHandler

use of org.xml.sax.ContentHandler in project tika by apache.

the class OOXMLParserTest method testPowerPointMetadataEarly.

/**
     * Test that the metadata is already extracted when the body is processed.
     * See TIKA-1109
     */
@Test
public void testPowerPointMetadataEarly() throws Exception {
    String[] extensions = new String[] { "pptx", "pptm", "ppsm", "ppsx", "potm" };
    final String[] mimeTypes = new String[] { "application/vnd.openxmlformats-officedocument.presentationml.presentation", "application/vnd.ms-powerpoint.presentation.macroenabled.12", "application/vnd.ms-powerpoint.slideshow.macroenabled.12", "application/vnd.openxmlformats-officedocument.presentationml.slideshow", "application/vnd.ms-powerpoint.template.macroenabled.12" };
    for (int i = 0; i < extensions.length; i++) {
        String extension = extensions[i];
        final String filename = "testPPT." + extension;
        Parser parser = new AutoDetectParser();
        final Metadata metadata = new Metadata();
        // Allow the value to be access from the inner class
        final int currentI = i;
        ContentHandler handler = new BodyContentHandler() {

            public void startDocument() {
                assertEquals("Mime-type checking for " + filename, mimeTypes[currentI], metadata.get(Metadata.CONTENT_TYPE));
                assertEquals("Attachment Test", metadata.get(TikaCoreProperties.TITLE));
                assertEquals("Rajiv", metadata.get(TikaCoreProperties.CREATOR));
                assertEquals("Rajiv", metadata.get(Metadata.AUTHOR));
            }
        };
        ParseContext context = new ParseContext();
        try (InputStream input = getTestDocument(filename)) {
            parser.parse(input, handler, metadata, context);
        }
    }
}

Example 77 with ContentHandler

use of org.xml.sax.ContentHandler in project tika by apache.

the class OOXMLParserTest method testUnsupportedPowerPoint.

/**
     * For the PowerPoint formats we don't currently support, ensure that
     * we don't break either
     */
@Test
public void testUnsupportedPowerPoint() throws Exception {
    String[] extensions = new String[] { "xps", "thmx" };
    String[] mimeTypes = new String[] { "application/vnd.ms-xpsdocument", // Is this right?
    "application/vnd.openxmlformats-officedocument" };
    for (int i = 0; i < extensions.length; i++) {
        String extension = extensions[i];
        String filename = "testPPT." + extension;
        Parser parser = new AutoDetectParser();
        Metadata metadata = new Metadata();
        metadata.set(Metadata.RESOURCE_NAME_KEY, filename);
        ContentHandler handler = new BodyContentHandler();
        ParseContext context = new ParseContext();
        try (InputStream input = getTestDocument(filename)) {
            parser.parse(input, handler, metadata, context);
            // Should get the metadata
            assertEquals("Mime-type checking for " + filename, mimeTypes[i], metadata.get(Metadata.CONTENT_TYPE));
        // But that's about it
        }
    }
}

Example 78 with ContentHandler

use of org.xml.sax.ContentHandler in project tika by apache.

the class OldExcelParserTest method testMetadata.

// Disabled, until we can get the POI code to tell us the version
@Test
@Ignore
public void testMetadata() throws Exception {
    TikaInputStream stream = getTestFile(file);
    Metadata metadata = new Metadata();
    ContentHandler handler = new BodyContentHandler();
    OldExcelParser parser = new OldExcelParser();
    parser.parse(stream, handler, metadata, new ParseContext());
    // We can get the content type
    assertEquals("application/vnd.ms-excel.sheet.4", metadata.get(Metadata.CONTENT_TYPE));
    // But no other metadata
    assertEquals(null, metadata.get(TikaCoreProperties.TITLE));
    assertEquals(null, metadata.get(Metadata.SUBJECT));
}

Example 79 with ContentHandler

use of org.xml.sax.ContentHandler in project tika by apache.

the class OutlookParserTest method testOutlookParsing.

@Test
public void testOutlookParsing() throws Exception {
    // Should auto-detect!
    Parser parser = new AutoDetectParser();
    ContentHandler handler = new BodyContentHandler();
    Metadata metadata = new Metadata();
    try (InputStream stream = OutlookParserTest.class.getResourceAsStream("/test-documents/test-outlook.msg")) {
        parser.parse(stream, handler, metadata, new ParseContext());
    }
    assertEquals("application/vnd.ms-outlook", metadata.get(Metadata.CONTENT_TYPE));
    assertEquals("Microsoft Outlook Express 6", metadata.get(TikaCoreProperties.TITLE));
    assertEquals("Nouvel utilisateur de Outlook Express", metadata.get(Metadata.MESSAGE_RECIPIENT_ADDRESS));
    assertEquals("L'Équipe Microsoft Outlook Express", metadata.get(TikaCoreProperties.CREATOR));
    assertEquals("L'Équipe Microsoft Outlook Express", metadata.get(Metadata.AUTHOR));
    //ensure that "raw" header is correctly decoded
    assertEquals("L'Équipe Microsoft Outlook Express <msoe@microsoft.com>", metadata.get(Metadata.MESSAGE_RAW_HEADER_PREFIX + "From"));
    assertEquals("Nouvel utilisateur de Outlook Express", metadata.get(Message.MESSAGE_TO_EMAIL));
    assertEquals("", metadata.get(Message.MESSAGE_TO_NAME));
    assertEquals("Nouvel utilisateur de Outlook Express", metadata.get(Message.MESSAGE_TO_DISPLAY_NAME));
    // Stored as Thu, 5 Apr 2007 09:26:06 -0700
    assertEquals("2007-04-05T16:26:06Z", metadata.get(TikaCoreProperties.CREATED));
    String content = handler.toString();
    assertContains("Microsoft Outlook Express 6", content);
    assertContains("L'Équipe Microsoft Outlook Express", content);
    assertContains("Nouvel utilisateur de Outlook Express", content);
    assertContains("Messagerie et groupes de discussion", content);
}

Example 80 with ContentHandler

use of org.xml.sax.ContentHandler in project tika by apache.

the class OutlookParserTest method testOutlookNew.

/**
     * Test case for TIKA-395, to ensure parser works for new Outlook formats.
     *
     * @see <a href="https://issues.apache.org/jira/browse/TIKA-395">TIKA-395</a>
     */
@Test
public void testOutlookNew() throws Exception {
    Parser parser = new AutoDetectParser();
    ContentHandler handler = new BodyContentHandler();
    Metadata metadata = new Metadata();
    try (InputStream stream = OutlookParserTest.class.getResourceAsStream("/test-documents/test-outlook2003.msg")) {
        parser.parse(stream, handler, metadata, new ParseContext());
    }
    assertEquals("application/vnd.ms-outlook", metadata.get(Metadata.CONTENT_TYPE));
    assertEquals("Welcome to Microsoft Office Outlook 2003", metadata.get(TikaCoreProperties.TITLE));
    String content = handler.toString();
    assertContains("Outlook 2003", content);
    assertContains("Streamlined Mail Experience", content);
    assertContains("Navigation Pane", content);
    //make sure these are parallel
    assertEquals("", metadata.get(Message.MESSAGE_TO_EMAIL));
    assertEquals("New Outlook User", metadata.get(Message.MESSAGE_TO_NAME));
    assertEquals("New Outlook User", metadata.get(Message.MESSAGE_TO_DISPLAY_NAME));
}