import org.xml.sax.SAXException;

import jakarta.annotation.PostConstruct;

/**
 * @author shinsuke
 *
 */
public class TikaExtractor extends PasswordBasedExtractor {

    private static final Logger logger = LoggerFactory.getLogger(TikaExtractor.class);

    public static final String TIKA_TESSERACT_CONFIG = "tika.tesseract.config";

    public static final String TIKA_PDF_CONFIG = "tika.pdf.config";

<!DOCTYPE components PUBLIC "-//DBFLUTE//DTD LastaDi 1.0//EN"
	"http://dbflute.org/meta/lastadi10.dtd">
<components namespace="fessCrawler">
	<include path="crawler/container.xml" />
	<component name="tikaExtractor"
		class="org.codelibs.fess.crawler.extractor.impl.TikaExtractor">
		<property name="maxCompressionRatio">2</property>
		<property name="maxUncompressionSize">10000000</property>
	</component>

                    TikaExtractor tikaExtractor = container.getComponent("tikaExtractor");
                    factory.addExtractor("text/plain", tikaExtractor);
                    factory.addExtractor("text/html", tikaExtractor);
                })//
        ;

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

    public void test_getTika_text() {

                .singleton("tikaExtractor", TikaExtractor.class).singleton("zipExtractor", ZipExtractor.class)
                .<ExtractorFactory> singleton("extractorFactory", ExtractorFactory.class, factory -> {
                    TikaExtractor tikaExtractor = container.getComponent("tikaExtractor");
                    ZipExtractor zipExtractor = container.getComponent("zipExtractor");

        }
        try {
            final TikaExtractor tikaExtractor = ComponentUtil.getComponent("tikaExtractor");
            if (tikaExtractor != null) {
                tikaExtractor.setMaxAlphanumTermSize(getMaxAlphanumTermSize());
                tikaExtractor.setMaxSymbolTermSize(getMaxSymbolTermSize());
                tikaExtractor.setReplaceDuplication(isDuplicateTermRemoved());
                tikaExtractor.setSpaceChars(getSpaceChars());

        container.singleton("mimeTypeHelper", MimeTypeHelperImpl.class).singleton("tikaExtractor", TikaExtractor.class)
                .singleton("lhaExtractor", LhaExtractor.class)
                .<ExtractorFactory> singleton("extractorFactory", ExtractorFactory.class, factory -> {
                    TikaExtractor tikaExtractor = container.getComponent("tikaExtractor");
                    LhaExtractor lhaExtractor = container.getComponent("lhaExtractor");

                .singleton("textTransformer", TextTransformer.class).singleton("tikaExtractor", TikaExtractor.class);
        textTransformer = container.getComponent("textTransformer");
        textTransformer.setName("textTransformer");
        ExtractorFactory extractorFactory = container.getComponent("extractorFactory");
        TikaExtractor tikaExtractor = container.getComponent("tikaExtractor");
        extractorFactory.addExtractor("text/plain", tikaExtractor);

                .singleton("mimeTypeHelper", MimeTypeHelperImpl.class)//
                .singleton("tikaExtractor", TikaExtractor.class)//
                .<ExtractorFactory> singleton("extractorFactory", ExtractorFactory.class, factory -> {
                    TikaExtractor tikaExtractor = container.getComponent("tikaExtractor");
                    PdfExtractor pdfExtractor = container.getComponent("pdfExtractor");

                }).singleton("tikaExtractor", TikaExtractor.class)
                .<ExtractorFactory> singleton("extractorFactory", ExtractorFactory.class, factory -> {
                    TikaExtractor tikaExtractor = container.getComponent("tikaExtractor");
                    factory.addExtractor("text/plain", tikaExtractor);
                    factory.addExtractor("text/html", tikaExtractor);
                })//

        extractorFactory = container.getComponent("extractorFactory");
        TikaExtractor tikaExtractor = container.getComponent("tikaExtractor");
        LhaExtractor lhaExtractor = container.getComponent("lhaExtractor");
        PasswordBasedExtractor pdfExtractor = container.getComponent("pdfExtractor");
        extractorFactory.addExtractor("application/msword", tikaExtractor);
        extractorFactory.addExtractor("application/vnd.ms-excel", tikaExtractor);