import org.codelibs.fess.crawler.util.XPathAPI;
import org.codelibs.nekohtml.parsers.DOMParser;
import org.w3c.dom.Document;
import org.w3c.dom.Node;
import org.xml.sax.InputSource;

/**
 * Extracts text content from HTML documents.
 */
public class HtmlExtractor extends AbstractXmlExtractor {
    /** Logger for this class. */

import org.codelibs.fess.opensearch.config.exentity.LabelType;
import org.codelibs.fess.opensearch.config.exentity.WebConfig;
import org.dbflute.optional.OptionalEntity;
import org.xml.sax.Attributes;
import org.xml.sax.InputSource;
import org.xml.sax.SAXException;
import org.xml.sax.helpers.DefaultHandler;

/**
 * Parser for Google Search Appliance (GSA) configuration files.

    }
          </code>
        </codeSegment>
      </codeSegments>
    </class>
    <class sourceTracker="source">
      <name>InputSource</name>
      <version>1.2.0+</version>
      <fields>
        <field>
          <name>location</name>
          <version>1.2.0+</version>
          <type>String</type>
          <description>

            bis.mark(UTF8_BOM_SIZE);
            final int size = bis.read(bomBytes);
            if (size < 3 || !isUtf8BomBytes(bomBytes)) {
                bis.reset();
            }
            final InputSource is = new InputSource(bis);
            if (responseData.getCharSet() != null) {
                is.setEncoding(responseData.getCharSet());
            }
            parser.parse(is);
        } catch (final Exception e) {

import org.codelibs.fess.crawler.util.CharUtil;
import org.codelibs.fess.crawler.util.XPathAPI;
import org.codelibs.nekohtml.parsers.DOMParser;
import org.w3c.dom.Document;
import org.w3c.dom.Node;
import org.xml.sax.InputSource;

import jakarta.annotation.Resource;

/**
 * The {@code HtmlTransformer} class is responsible for transforming HTML responses
 * during the crawling process. It extracts data, identifies child URLs, and handles

import org.lastaflute.di.core.exception.ComponentNotFoundException;
import org.lastaflute.di.core.factory.SingletonLaContainerFactory;
import org.w3c.dom.Document;
import org.w3c.dom.Node;
import org.xml.sax.InputSource;

public class FessXpathTransformerTest extends UnitFessTestCase {
    private static final Logger logger = LogManager.getLogger(FessXpathTransformerTest.class);

    @Override
    public void setUp() throws Exception {

    }
          </code>
        </codeSegment>
      </codeSegments>
    </class>
    <class sourceTracker="source">
      <name>InputSource</name>
      <version>2.0.0+</version>
      <fields>
        <field>
          <name>location</name>
          <version>2.0.0+</version>
          <type>String</type>
          <description>

    }
            ]]>
          </code>
        </codeSegment>
      </codeSegments>
    </class>
    <class sourceTracker="source">
      <name>InputSource</name>
      <version>4.0.0/4.0.99</version>
      <fields>
        <field>
          <name>modelId</name>
          <version>4.0.0+</version>
          <type>String</type>
          <description>