<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE components PUBLIC "-//DBFLUTE//DTD LastaDi 1.0//EN"
    "http://dbflute.org/meta/lastadi10.dtd">
<components namespace="fessCrawler">
    <component name="crawler" class="org.codelibs.fess.crawler.Crawler" instance="prototype"/>
    <component name="httpClient" class="org.codelibs.fess.crawler.client.http.HcHttpClient" instance="singleton"/>

/**
 * The CrawlerContainer interface provides methods for managing components
 * within a crawler container. It includes methods to retrieve components,
 * check availability, and destroy the container. Additionally, it provides
 * a default method to initialize the container with specific protocol handlers.
 *
 */
public interface CrawlerContainer {

    /**
     * Retrieves a component by its name.
     *

/**
 * An abstract base class for implementing Extractor interfaces.
 * Provides common functionality such as access to CrawlerContainer components
 * and registration with the ExtractorFactory.
 *
 * <p>
 * This class handles the retrieval of essential crawler components like
 * {@link MimeTypeHelper} and {@link ExtractorFactory} from the
 * {@link CrawlerContainer}. It also provides a convenient method for

    }

    /**
     * Registers a prototype component with the specified name, class, and initializer.
     * A new instance will be created each time the component is requested.
     * @param <T> the type of the component
     * @param name the name of the component
     * @param cls the class of the component
     * @param component the initializer consumer for the component
     * @return this container instance for method chaining

            clientFactoryList.remove(0);
        }
    }

    /**
     * Registers a client component with a regular expression.
     * The component will be loaded into all registered CrawlerClientFactories.
     * @param regex The regular expression to match URLs.
     * @param componentName The name of the component to register.
     */
    public synchronized void register(final String regex, final String componentName) {

 * result and storing it. Unsuccessful responses are logged for debugging purposes.
 * </p>
 *
 * <p>
 * The class uses {@link CrawlerContainer} to access components like {@link AccessResult}
 * and {@link UrlQueue}. It also uses {@link CrawlingParameterUtil} to access services
 * like {@link UrlQueueService} and DataService, as well as the {@link CrawlerContext}.
 * </p>
 *
 * <p>

 *
 * <p>It implements the Runnable interface to be executed in a separate thread,
 * and the AutoCloseable interface to ensure resources are properly released after use.
 *
 * <p>The crawler uses several services and components, such as UrlQueueService, DataService,
 * UrlFilter, RuleManager, CrawlerContainer, IntervalController, and CrawlerClientFactory,
 * to perform its tasks.
 *
 * <p>The crawling process involves the following steps:
 * <ol>

 * <p>
 * The class uses a {@link CrawlerContainer} for managing crawler components and
 * supports dependency injection via the {@link Resource} annotation.
 * </p>
 */
public class ExtractorFactory {

    /** Logger instance for this class */
    private static final Logger logger = LogManager.getLogger(ExtractorFactory.class);

    /** Container for managing crawler components */
    @Resource
    protected CrawlerContainer crawlerContainer;

import org.codelibs.fess.suggest.util.SuggestUtil;
import org.opensearch.transport.client.Client;

/**
 * Builder class for creating instances of {@link Suggester}.
 * This class allows for the configuration of various components used by the {@link Suggester}.
 */
public class SuggesterBuilder {
    /**
     * Constructs a new {@link SuggesterBuilder}.
     */
    public SuggesterBuilder() {
        // nothing
    }

<!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">1</property>
		<property name="maxUncompressionSize">10000000</property>
	</component>
	<component name="msWordExtractor"