import java.io.Serializable;

/**
 * The Sitemap interface represents a sitemap entity with location and last modification date.
 * It extends the Serializable interface to allow sitemap objects to be serialized.
 */
public interface Sitemap extends Serializable {

    /**
     * Retrieves the location (URL) of the sitemap.
     *
     * @return the location of the sitemap as a String.
     */
    String getLoc();

User-agent: Crawler
Disallow: /aaa

User-agent: Crawler/1.0
Disallow: /bbb

User-agent: Crawler/2.0
Disallow: /ccc

User-agent: Hoge Crawler
Disallow: /ddd

sitemap: http://www.example.com/sitmap.xml

        Map<String, Object> dataMap = new HashMap<>();
        dataMap.put("a", "A0");
        Map<String, Object> responseDataMap = new HashMap<>();
        responseDataMap.put("b", "B1");

        indexUpdateCallback.mergeResponseData(dataMap, responseDataMap);

        assertEquals(2, dataMap.size());
        assertEquals("A0", dataMap.get("a"));
        assertEquals("B1", dataMap.get("b"));
    }

    }

    private static class DocHandler extends DefaultHandler {
        private final Map<String, Object> dataMap = new HashMap<>();

        private String fieldName;

        private final StringBuilder buffer = new StringBuilder(1000);

        @Override
        public void startDocument() {
            dataMap.clear();
        }

        @Override

        sitemapsRule.addRule("url", Pattern.compile(".*sitemap.*"));
        assertTrue(sitemapsRule.match(responseData));
        InputStream is = responseData.getResponseBody();
        assertTrue(is instanceof InputStream);
        CloseableUtil.closeQuietly(responseData);
    }

    private void assertMatchFalse(ResponseData responseData) {
        sitemapsRule.addRule("url", Pattern.compile(".*sitemap.*"));

     *
     * @param paramMap the data store parameters containing configuration and metadata
     * @param dataMap the document data to be indexed as key-value pairs
     */
    void store(DataStoreParams paramMap, Map<String, Object> dataMap);

    /**
     * Returns the total number of documents processed by this callback.
     *
     * @return the document count
     */

        sitemapList.add(sitemap);
    }

    /**
     * Removes a sitemap from this set.
     * @param sitemap the sitemap to remove
     */
    public void removeSitemap(final Sitemap sitemap) {
        sitemapList.remove(sitemap);
    }

    /**
     * Gets all sitemaps in this set as an array.
     * @return an array of sitemaps
     */
    public Sitemap[] getSitemaps() {

            final Set<RequestData> requestDataSet = new LinkedHashSet<>();
            for (final Sitemap sitemap : sitemapSet.getSitemaps()) {
                if (sitemap != null) {
                    requestDataSet.add(RequestDataBuilder.newRequestData().get().url(sitemap.getLoc()).build()); // TODO priority
                }
            }

<?xml version="1.0" encoding="UTF-8"?>
<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
   <sitemap>
      <loc>http://www.example.com/sitemap1.xml.gz</loc>
      <lastmod>2004-10-01T18:23:17+00:00</lastmod>
   </sitemap>
   <sitemap>
      <loc>http://www.example.com/sitemap2.xml.gz</loc>
      <lastmod>2005-01-01</lastmod>
   </sitemap>

package org.codelibs.fess.crawler.entity;

import org.codelibs.core.lang.StringUtil;

/**
 * Represents a Sitemap file entry, conforming to the Sitemap XML format.
 * This class holds information about a single Sitemap, including its location and last modification timestamp.
 * It implements the {@link Sitemap} interface.
 *
 * <p>
 * A Sitemap file provides search engines with a list of URLs available for crawling.