application/access             mdf              # Microsoft Access
#application/excel              xls              # Microsoft Excel
application/vnd.ms-excel       xls              # Microsoft Excel
application/font-tdpfr         pfr              # TrueDoc Portable Font Resource
application/futuresplash       spl              # Macromedia Flash
application/hep                hep              # Hummingbird Host Explorer Profiles

import org.codelibs.fess.crawler.entity.ExtractData;
import org.codelibs.fess.crawler.exception.CrawlerSystemException;
import org.codelibs.fess.crawler.exception.ExtractException;

/**
 * Gets a text from .xls file.
 *
 * @author shinsuke
 *
 */
public class MsExcelExtractor extends AbstractExtractor {

    /**
     * Creates a new MsExcelExtractor instance.
     */
    public MsExcelExtractor() {

            assertEquals("application/pdf", mimeMap.getMimeType("pdf"));
            assertEquals("application/msword", mimeMap.getMimeType("doc"));
            assertEquals("application/vnd.ms-excel", mimeMap.getMimeType("xls"));
            assertEquals("text/html", mimeMap.getMimeType("html"));
            assertEquals("text/html", mimeMap.getMimeType("htm"));
            assertEquals("image/jpeg", mimeMap.getMimeType("jpg"));

            assertEquals("filetype:pdf", queryMap.get("PDF"));
        }

        // Check for XLS - could be either key format
        if (queryMap.containsKey("labels.facet_filetype_xls")) {
            assertEquals("filetype:xls", queryMap.get("labels.facet_filetype_xls"));
        } else {
            assertEquals("filetype:xls", queryMap.get("XLS"));
        }

        // DOC type won't be duplicated since it already exists

        msExcelExtractor = container.getComponent("msExcelExtractor");
    }

    public void test_getText() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/msoffice/test.xls");
        final String content = msExcelExtractor.getText(in, null).getContent();
        CloseableUtil.closeQuietly(in);
        logger.info(content);
        assertTrue(content.contains("テスト"));
    }

application/access             mdf              # Microsoft Access
#application/excel              xls              # Microsoft Excel
application/vnd.ms-excel       xls              # Microsoft Excel
application/font-tdpfr         pfr              # TrueDoc Portable Font Resource
application/futuresplash       spl              # Macromedia Flash
application/hep                hep              # Hummingbird Host Explorer Profiles

    }

    public void test_getTika_msexcelx() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/msoffice/test.xlsx");
        final ExtractData extractData = tikaExtractor.getText(in, null);
        final String content = extractData.getContent();
        CloseableUtil.closeQuietly(in);
        logger.info(content);

    }

    @Test
    @DisplayName("Test getFileName returns correct value")
    void testGetFileName() throws SMBProtocolDecodingException {
        String expectedFileName = "important_document.xlsx";
        byte[] buffer = createValidNotificationBuffer(expectedFileName, FileNotifyInformation.FILE_ACTION_ADDED);

        notifyInfo.decode(buffer, 0, buffer.length);

    // public ResponseData call() throws Exception {
    // String[] urls =
    // new String[] {
    // "http://.../",
    // "http://.../test.pdf",
    // "http://.../test.doc",
    // "http://.../test.xls",
    // "http://.../test.ppt",
    // "http://.../test.txt", };
    // for (String url : urls) {
    // ResponseData responseData = httpClient.doGet(url);
    // // assertEquals(200, responseData.getHttpStatusCode());

				"application/vnd.xmi+xml",
				"application/vnd.xmpie.cpkg",
				"application/vnd.xmpie.dpkg",
				"application/vnd.xmpie.plan",
				"application/vnd.xmpie.ppkg",
				"application/vnd.xmpie.xlim",
				"application/vnd.yamaha.hv-dic",
				"application/vnd.yamaha.hv-script",
				"application/vnd.yamaha.hv-voice",
				"application/vnd.yamaha.openscoreformat",
				"application/vnd.yamaha.openscoreformat.osfpvg+xml",