Shinsuke Sugaya <******@****.***> 1444529815 +0900

Shinsuke Sugaya <******@****.***> 1444529815 +0900

        assertContentType("application/msword", "extractor/msoffice/test.doc", "hoge.doc");
        assertContentType("application/vnd.ms-excel", "test/text1.txt", "hoge.xls");
        assertContentType("application/vnd.ms-excel", "extractor/msoffice/test.xls", "hoge.xls");
        assertContentType("application/vnd.ms-powerpoint", "test/text1.txt", "hoge.ppt");

     */

    public void test_getText_msexcel() {
        InputStream in = ResourceUtil.getResourceAsStream("extractor/msoffice/test.xls");
        Map<String, String> params = new HashMap<String, String>();
        params.put("resourceName", "test.xls");
        ExtractData extractData = jodExtractor.getText(in, params);
        String content = extractData.getContent();
        CloseableUtil.closeQuietly(in);

        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"));
    }

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() {

    // 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());

        extensionMap.put("html", "txt");
        // Spreadsheet Formats
        extensionMap.put("ods", "tsv");
        extensionMap.put("ots", "tsv");
        extensionMap.put("sxc", "tsv");
        extensionMap.put("xls", "tsv");
        extensionMap.put("xlsx", "tsv");
        extensionMap.put("csv", "tsv");
        extensionMap.put("tsv", "tsv");
        // Presentation Formats
        extensionMap.put("odp", "pdf");

        assertEquals("かいしゃ", extractData.getValues("dc:publisher")[0]);
    }

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

				"application/vnd.sealed.eml",
				"application/vnd.sealed.mht",
				"application/vnd.sealed.net",
				"application/vnd.sealed.ppt",
				"application/vnd.sealed.tiff",
				"application/vnd.sealed.xls",
				"application/vnd.sealedmedia.softseal.html",
				"application/vnd.sealedmedia.softseal.pdf",
				"application/vnd.seemail",
				"application/vnd.sema",
				"application/vnd.semd",
				"application/vnd.semf",