http://www.example.com/
http://www.example.com/catalog?item=12&desc=vacation_hawaii
http://www.example.com/catalog?item=73&desc=vacation_new_zealand
http://www.example.com/catalog?item=74&desc=vacation_newfoundland

sitemap1.xml http://www.example.com/ 2005-01-01 monthly 0.8 http://www.example.com/catalog?item=12&desc=vacation_hawaii weekly http://www.example.com/catalog?item=73&desc=vacation_new_zealand 2004-12-23 weekly http://www.example.com/catalog?item=74&desc=vacation_newfoundland 2004-12-23T18:00:15+00:00 0.3 http://www.example.com/catalog?item=83&desc=vacation_usa 2004-11-23...

        urlFilter.addInclude("http://test.com/.*");

        final String sessionId = "id1";
        urlFilter.init(sessionId);

        assertTrue(urlFilter.match("http://example.com/"));
        assertTrue(urlFilter.match("http://example.com/a"));
        assertTrue(urlFilter.match("http://test.com/"));
        assertTrue(urlFilter.match("http://test.com/a"));
    }

        // Add URLs to default set
        urlSet.add("http://example.com/robots.txt");
        urlSet.add("http://test.com/robots.txt");
        assertEquals(2, crawlerContext.getRobotsTxtUrlSet().size());

        // Set new set
        Set<String> newSet = new HashSet<>();
        newSet.add("http://new.com/robots.txt");
        crawlerContext.setRobotsTxtUrlSet(newSet);

<?xml version="1.0" encoding="UTF-8"?>
<address xmlns:hoge="http://www.example.com/hoge">
	<item sex="male" custid="A12345">
		<name>鈴木太郎</name>
		<access kind="email"></access>
		<access kind="url">http://www.taro.com/</access>
		<image file="taro.png" />
	</item>
	<item sex="male" custid="B23456">
		<name>佐藤二朗</name>
		<access kind="email">******@****.***</access>
		<image file="jiro.png" />
	</item>
	<item sex="female" custid="C34567">

        domainFilter.processUrl("http://example.com/");

        assertEquals(1, domainFilter.cachedIncludeSet.size());
        assertEquals(1, domainFilter.cachedExcludeSet.size());
        assertEquals("http://example.com/.*", domainFilter.cachedIncludeSet.toArray()[0]);
        assertEquals("http://example.com/.*", domainFilter.cachedExcludeSet.toArray()[0]);

        domainFilter.processUrl("https://test.com");

        String url;

        url = "http://hoge.com/http/fuga.html";
        assertEquals("http://hoge.com/http/fuga.html", urlConvertHelper.convert(url));

        url = "tp://hoge.com/http/fuga.html";
        assertEquals("tp://hoge.com/http/fuga.html", urlConvertHelper.convert(url));

        urlConvertHelper.add("^tp:", "http:");

        url = "http://hoge.com/http/fuga.html";

        urlFilter.addInclude("https://example.com/.*");
        urlFilter.addInclude("https://test.com/.*");
        urlFilter.addInclude(".*\\.pdf$");

        assertTrue(urlFilter.match("https://example.com/"));
        assertTrue(urlFilter.match("https://test.com/page"));
        assertTrue(urlFilter.match("https://any.com/document.pdf"));
        assertFalse(urlFilter.match("https://other.com/page.html"));
    }

    /**

        assertNotNull(dataService.getAccessResult("id2", "http://www.id2.com/"));

        dataService.store(accessResult1);
        assertNotNull(dataService.getAccessResult("id1", "http://www.id1.com/"));

        dataService.deleteAll();

        assertNull(dataService.getAccessResult("id1", "http://www.id1.com/"));

        assertEquals("http://www.example.com/", sitemaps[0].getLoc());
        assertEquals("monthly", ((SitemapUrl) sitemaps[0]).getChangefreq());
        assertEquals("0.8", ((SitemapUrl) sitemaps[0]).getPriority());

        assertNull(sitemaps[1].getLastmod());
        assertEquals("http://www.example.com/catalog?item=12&desc=vacation_hawaii", sitemaps[1].getLoc());