*/
package org.codelibs.fess.crawler;

/**
 * Enum representing the status of a crawler.
 * It can be INITIALIZING, RUNNING, or DONE.
 */
public enum CrawlerStatus {
    /**
     * The crawler is initializing.
     */
    INITIALIZING,

    /**
     * The crawler is currently running.
     */
    RUNNING,

    /**
     * The crawler has completed its task.
     */
    DONE;

 */
package org.codelibs.fess.crawler.extractor.impl;

import java.io.IOException;
import java.io.InputStream;
import java.util.Map;

import org.apache.poi.hpbf.extractor.PublisherTextExtractor;
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 . file.

<!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="encodingHelper" class="org.codelibs.fess.crawler.helper.EncodingHelper">
		<postConstruct name="addEncodingMapping">
			<arg>"unicode"</arg>
			<arg>"UTF-16LE"</arg>
		</postConstruct>
	</component>

#smart.deploy.mode.location = maihama_env.properties: lasta_di.smart.deploy.mode

# package for smart deploy target e.g. org.docksidestage.app
smart.package1 = org.codelibs.fess.crawler

index.document.suggest.index=fess
# Index name for crawler documents.
index.document.crawler.index=fess_crawler
# Number of primary shards for crawler queue index.
index.document.crawler.queue.number_of_shards=10
# Number of primary shards for crawler data index.
index.document.crawler.data.number_of_shards=10
# Number of primary shards for crawler filter index.
index.document.crawler.filter.number_of_shards=10

 *   <li>org.codelibs.fess.crawler.Constants</li>
 *   <li>org.codelibs.fess.crawler.exception.CrawlerSystemException</li>
 *   <li>org.codelibs.fess.crawler.entity.AccessResultData</li>
 *   <li>org.codelibs.core.lang.StringUtil</li>
 * </ul>
 *
 * @see org.codelibs.fess.crawler.Constants
 * @see org.codelibs.fess.crawler.exception.CrawlerSystemException
 * @see org.codelibs.fess.crawler.entity.AccessResultData

    /** The key of the configuration. e.g. UTF-8 */
    String CRAWLER_DOCUMENT_SITE_ENCODING = "crawler.document.site.encoding";

    /** The key of the configuration. e.g. unknown */
    String CRAWLER_DOCUMENT_UNKNOWN_HOSTNAME = "crawler.document.unknown.hostname";

    /** The key of the configuration. e.g. false */

import org.codelibs.fess.Constants;
import org.codelibs.fess.crawler.client.FesenClient;
import org.codelibs.fess.mylasta.direction.FessConfig;
import org.codelibs.fess.util.ComponentUtil;
import org.opensearch.common.settings.Settings;
import org.opensearch.common.settings.Settings.Builder;
import org.opensearch.transport.client.Client;

/**
 * OpenSearch client implementation specifically for crawler operations.

    assertThat(response.trailers()).isEmpty()
  }

  @Test
  fun worksIfTrailersSet() {
    val response =
      newResponse("".toResponseBody()) {
        trailers(
          object : TrailersSource {
            override fun get() = Headers.headersOf("a", "b")
          },
        )
      }

    assertThat(response.trailers()["a"]).isEqualTo("b")
  }

 */
package org.codelibs.fess.crawler.rule.impl;

import java.io.File;

import org.codelibs.core.io.CloseableUtil;
import org.codelibs.core.io.ResourceUtil;
import org.codelibs.fess.crawler.container.StandardCrawlerContainer;
import org.codelibs.fess.crawler.entity.ResponseData;
import org.codelibs.fess.crawler.helper.SitemapsHelper;
import org.codelibs.fess.crawler.rule.Rule;