Shinsuke Sugaya <******@****.***> 1452930635 +0900

                });
        emlExtractor = container.getComponent("emlExtractor");
    }

    public void test_getText() throws IOException {
        try (final InputStream in = ResourceUtil.getResourceAsStream("extractor/eml/sample1.eml")) {
            ExtractData data = emlExtractor.getText(in, null);
            final String content = data.getContent();
            logger.info(content);
            assertTrue(content.contains("プレイステーション"));

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...

        responseData.setUrl("http://example.com/");
        responseData.setMimeType("text/html");
        return responseData;
    }

    private ResponseData getTestData2() {
        final ResponseData responseData = new ResponseData();
        responseData.setHttpStatusCode(200);
        responseData.setUrl("https://example.com/");
        responseData.setMimeType("text/html");

<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>

        assertEquals("2005-01-01", sitemaps[0].getLastmod());
        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());

/**
 * Builder class for creating RequestData objects using a fluent interface.
 * This class provides a simple way to construct RequestData objects with method chaining.
 *
 * <p>Usage example:</p>
 * <pre>
 * RequestData request = RequestDataBuilder.newRequestData()
 *     .method("GET")
 *     .url("https://example.com")
 *     .weight(1.0f)
 *     .build();
 * </pre>
 *

        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]);

<?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>