import org.codelibs.fess.crawler.Constants;
import org.codelibs.fess.crawler.CrawlerContext;
import org.codelibs.fess.crawler.client.AbstractCrawlerClient;
import org.codelibs.fess.crawler.client.AccessTimeoutTarget;
import org.codelibs.fess.crawler.client.http.conn.IdnDnsResolver;
import org.codelibs.fess.crawler.client.http.form.FormScheme;
import org.codelibs.fess.crawler.entity.ResponseData;

 * governing permissions and limitations under the License.
 */
package org.codelibs.fess.crawler.helper;

import java.util.HashMap;
import java.util.Map;

import org.codelibs.core.lang.StringUtil;
import org.codelibs.fess.crawler.exception.CrawlerSystemException;

/**
 * @author shinsuke
 *
 */
public class ContentLengthHelper {

 * governing permissions and limitations under the License.
 */
package org.codelibs.fess.crawler.extractor.impl;

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

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 .doc file.
 *

import org.codelibs.fess.app.service.FailureUrlService;
import org.codelibs.fess.crawler.CrawlerContext;
import org.codelibs.fess.crawler.entity.UrlQueue;
import org.codelibs.fess.crawler.exception.CrawlingAccessException;
import org.codelibs.fess.crawler.exception.MultipleCrawlingAccessException;
import org.codelibs.fess.crawler.helper.impl.LogHelperImpl;
import org.codelibs.fess.crawler.log.LogType;
import org.codelibs.fess.es.config.exentity.CrawlingConfig;

import javax.xml.xpath.XPathNodes;

import org.codelibs.core.lang.StringUtil;
import org.codelibs.core.stream.StreamUtil;
import org.codelibs.fess.crawler.entity.ExtractData;
import org.codelibs.fess.crawler.exception.CrawlerSystemException;
import org.codelibs.fess.crawler.util.XPathAPI;
import org.codelibs.nekohtml.parsers.DOMParser;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.w3c.dom.Document;

 * governing permissions and limitations under the License.
 */
package org.codelibs.fess.crawler.filter.impl;

import org.codelibs.fess.crawler.container.StandardCrawlerContainer;
import org.codelibs.fess.crawler.helper.MemoryDataHelper;
import org.codelibs.fess.crawler.service.impl.UrlFilterServiceImpl;
import org.dbflute.utflute.core.PlainTestCase;

/**
 * @author shinsuke
 *
 */

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

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

import org.apache.poi.hssf.usermodel.HSSFWorkbook;
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.

                    factory.setFeature(org.codelibs.fess.crawler.Constants.FEATURE_SECURE_PROCESSING, true);
                    factory.setFeature(org.codelibs.fess.crawler.Constants.FEATURE_EXTERNAL_GENERAL_ENTITIES, false);
                    factory.setFeature(org.codelibs.fess.crawler.Constants.FEATURE_EXTERNAL_PARAMETER_ENTITIES, false);

<project name="deps" basedir=".">
	<property name="target.dir" value="${basedir}/target/deps" />
	<property name="webinf.dir" value="${basedir}/src/main/webapp/WEB-INF" />
	<property name="crawler.dir" value="${basedir}/src/main/webapp/WEB-INF/env/crawler" />
	<property name="suggest.dir" value="${basedir}/src/main/webapp/WEB-INF/env/suggest" />
	<property name="thumbnail.dir" value="${basedir}/src/main/webapp/WEB-INF/env/thumbnail" />

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

import java.util.List;
import java.util.Map;
import java.util.function.Consumer;

import org.apache.lucene.search.TotalHits;
import org.codelibs.core.beans.util.BeanUtil;
import org.codelibs.fess.crawler.entity.OpenSearchAccessResult;
import org.codelibs.fess.crawler.entity.OpenSearchAccessResultData;
import org.codelibs.fess.crawler.exception.OpenSearchAccessException;