Extend google news

dfabulich · Jul 5, 2018 · 1ceac7f · 1ceac7f
1 parent d52043d
commit 1ceac7f
Show file tree

Hide file tree

Showing 5 changed files with 165 additions and 20 deletions.
diff --git a/src/main/java/com/redfin/sitemapgenerator/AbstractSitemapUrlRenderer.java b/src/main/java/com/redfin/sitemapgenerator/AbstractSitemapUrlRenderer.java
@@ -43,4 +43,10 @@ public void renderTag(StringBuilder sb, String namespace, String tagName, Object
 		sb.append(">\n");
 	}
 
+	public void renderSubTag(StringBuilder sb, String namespace, String tagName, Object value) {
+		if (value == null) return;
+		sb.append("  ");
+		renderTag(sb, namespace, tagName, value);
+	}
+
 }
diff --git a/src/main/java/com/redfin/sitemapgenerator/GoogleNewsPublication.java b/src/main/java/com/redfin/sitemapgenerator/GoogleNewsPublication.java
@@ -0,0 +1,28 @@
+package com.redfin.sitemapgenerator;
+
+public class GoogleNewsPublication {
+
+	private String name;
+	private String language;
+
+	public GoogleNewsPublication(String name, String language) {
+		this.name = name;
+		this.language = language;
+	}
+
+	public String getName() {
+		return name;
+	}
+
+	public void setName(String name) {
+		this.name = name;
+	}
+
+	public String getLanguage() {
+		return language;
+	}
+
+	public void setLanguage(String language) {
+		this.language = language;
+	}
+}
diff --git a/src/main/java/com/redfin/sitemapgenerator/GoogleNewsSitemapGenerator.java b/src/main/java/com/redfin/sitemapgenerator/GoogleNewsSitemapGenerator.java
@@ -100,7 +100,13 @@ public String getXmlNamespaces() {
 		public void render(GoogleNewsSitemapUrl url, StringBuilder sb, W3CDateFormat dateFormat) {
 			StringBuilder tagSb = new StringBuilder();
 			tagSb.append("    <news:news>\n");
+			tagSb.append("      <news:publication>\n");
+			renderSubTag(tagSb, "news", "name", url.getPublication().getName());
+			renderSubTag(tagSb, "news", "language", url.getPublication().getLanguage());
+			tagSb.append("      </news:publication>\n");
+			renderTag(tagSb, "news", "genres", url.getGenres());
 			renderTag(tagSb, "news", "publication_date", dateFormat.format(url.getPublicationDate()));
+			renderTag(tagSb, "news", "title", url.getTitle());
 			renderTag(tagSb, "news", "keywords", url.getKeywords());
 			tagSb.append("    </news:news>\n");
 			super.render(url, sb, dateFormat, tagSb.toString());

diff --git a/src/main/java/com/redfin/sitemapgenerator/GoogleNewsSitemapUrl.java b/src/main/java/com/redfin/sitemapgenerator/GoogleNewsSitemapUrl.java
@@ -15,22 +15,42 @@ public class GoogleNewsSitemapUrl extends WebSitemapUrl {
 
 	private final Date publicationDate;
 	private final String keywords;
+	private final String genres;
+	private final String title;
+	private final GoogleNewsPublication publication;
 
 	/** Options to configure Google News URLs */
 	public static class Options extends AbstractSitemapUrlOptions<GoogleNewsSitemapUrl, Options> {
 		private Date publicationDate;
 		private String keywords;
+		private String genres;
+		private String title;
+		private GoogleNewsPublication publication;
 
 		/** Specifies an URL and publication date (which is mandatory for Google News) */
-		public Options(String url, Date publicationDate) throws MalformedURLException {
-			this(new URL(url), publicationDate);
+		public Options(String url, Date publicationDate, String title, GoogleNewsPublication publication) throws MalformedURLException {
+			this(new URL(url), publicationDate, title, publication);
 		}
 
+		public Options(String url, Date publicationDate, String title, String name, String language) throws MalformedURLException {
+			this(new URL(url), publicationDate, title, new GoogleNewsPublication(name, language));
+		}
+
+		public Options(URL url, Date publicationDate, String title, String name, String language) {
+			this(url, publicationDate, title, new GoogleNewsPublication(name, language));
+		}
+
 		/** Specifies an URL and publication date (which is mandatory for Google News) */
-		public Options(URL url, Date publicationDate) {
+		public Options(URL url, Date publicationDate, String title, GoogleNewsPublication publication) {
 			super(url, GoogleNewsSitemapUrl.class);
 			if (publicationDate == null) throw new NullPointerException("publicationDate must not be null");
 			this.publicationDate = publicationDate;
+			if (title == null) throw new NullPointerException("title must not be null");
+			this.title = title;
+			if (publication == null) throw new NullPointerException("publication must not be null");
+			if (publication.getName() == null) throw new NullPointerException("publication name must not be null");
+			if (publication.getLanguage() == null) throw new NullPointerException("publication language must not be null");
+			this.publication = publication;
 		}
 
 		/** Specifies a list of comma-delimited keywords */
@@ -41,42 +61,73 @@ public Options keywords(String keywords) {
 
 		/** Specifies a list of comma-delimited keywords */
 		public Options keywords(Iterable<String> keywords) {
+			this.keywords = getListAsCommaSeparatedString(keywords);
+			return this;
+		}
+
+		public Options genres(String genres) {
+			this.genres = genres;
+			return this;
+		}
+
+		public Options genres(Iterable<String> genres) {
+			this.genres = getListAsCommaSeparatedString(genres);
+			return this;
+		}
+
+		private String getListAsCommaSeparatedString(Iterable<String> values) {
 			StringBuilder sb = new StringBuilder();
 			boolean first = true;
-			for (String keyword : keywords) {
+			for (String value : values) {
 				if (first) {
 					first = false;
 				} else {
 					sb.append(", ");
 				}
-				sb.append(keyword);
+				sb.append(value);
 			}
-			this.keywords = sb.toString();
-			return this;
+			return sb.toString();
 		}
 
 		/** Specifies a list of comma-delimited keywords */
 		public Options keywords(String... keywords) {
 			return keywords(Arrays.asList(keywords));
 		}
 
+		public Options genres(String... genres) {
+			return genres(Arrays.asList(genres));
+		}
+
 	}
 
-	/** Specifies an URL and publication date (which is mandatory for Google News) */
-	public GoogleNewsSitemapUrl(URL url, Date publicationDate) {
-		this(new Options(url, publicationDate));
+	/** Specifies an URL and publication date, title and publication (which are mandatory for Google News) */
+	public GoogleNewsSitemapUrl(URL url, Date publicationDate, String title, String name, String language) {
+		this(new Options(url, publicationDate, title, name, language));
 	}
 
-	/** Specifies an URL and publication date (which is mandatory for Google News) */
-	public GoogleNewsSitemapUrl(String url, Date publicationDate) throws MalformedURLException {
-		this(new Options(url, publicationDate));
+	/** Specifies an URL and publication date, title and publication (which are mandatory for Google News) */
+	public GoogleNewsSitemapUrl(URL url, Date publicationDate, String title, GoogleNewsPublication publication) {
+		this(new Options(url, publicationDate, title, publication));
+	}
+
+	/** Specifies an URL and publication date, title and publication (which are mandatory for Google News) */
+	public GoogleNewsSitemapUrl(String url, Date publicationDate, String title, String name, String language) throws MalformedURLException {
+		this(new Options(url, publicationDate, title, name, language));
+	}
+
+	/** Specifies an URL and publication date, title and publication (which are mandatory for Google News) */
+	public GoogleNewsSitemapUrl(String url, Date publicationDate, String title, GoogleNewsPublication publication) throws MalformedURLException {
+		this(new Options(url, publicationDate, title, publication));
 	}
 
 	/** Configures an URL with options */
 	public GoogleNewsSitemapUrl(Options options) {
 		super(options);
 		publicationDate = options.publicationDate;
 		keywords = options.keywords;
+		genres = options.genres;
+		title = options.title;
+		publication = options.publication;
 	}
 
 	/** Retrieves the publication date */
@@ -89,7 +140,26 @@ public String getKeywords() {
 		return keywords;
 	}
 
+	/**
+	 * Retrieves the Genres
+	 */
+	public String getGenres() {
+		return genres;
+	}
+
+	/**
+	 * Retrieves the title
+	 */
+	public String getTitle() {
+		return title;
+	}
 
+	/**
+	 * Retrieves the publication with name and language
+	 */
+	public GoogleNewsPublication getPublication() {
+		return publication;
+	}
 
 
 }
diff --git a/src/test/java/com/redfin/sitemapgenerator/GoogleNewsSitemapUrlTest.java b/src/test/java/com/redfin/sitemapgenerator/GoogleNewsSitemapUrlTest.java
@@ -6,9 +6,6 @@
 
 import junit.framework.TestCase;
 
-import com.redfin.sitemapgenerator.GoogleNewsSitemapGenerator;
-import com.redfin.sitemapgenerator.GoogleNewsSitemapUrl;
-import com.redfin.sitemapgenerator.W3CDateFormat;
 import com.redfin.sitemapgenerator.W3CDateFormat.Pattern;
 
 public class GoogleNewsSitemapUrlTest extends TestCase {
@@ -38,14 +35,19 @@ public void testSimpleUrl() throws Exception {
 		dateFormat.setTimeZone(W3CDateFormat.ZULU);
 		wsg = GoogleNewsSitemapGenerator.builder("http://www.example.com", dir)
 			.dateFormat(dateFormat).build();
-		GoogleNewsSitemapUrl url = new GoogleNewsSitemapUrl("http://www.example.com/index.html", new Date(0));
+		GoogleNewsSitemapUrl url = new GoogleNewsSitemapUrl("http://www.example.com/index.html", new Date(0), "Example Title", "The Example Times", "en");
 		wsg.addUrl(url);
 		String expected = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n" + 
 			"<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\" xmlns:news=\"http://www.google.com/schemas/sitemap-news/0.9\" >\n" + 
 			"  <url>\n" + 
 			"    <loc>http://www.example.com/index.html</loc>\n" + 
 			"    <news:news>\n" + 
-			"      <news:publication_date>1970-01-01T00:00:00Z</news:publication_date>\n" + 
+			"      <news:publication>\n" +
+			"        <news:name>The Example Times</news:name>\n" +
+			"        <news:language>en</news:language>\n" +
+			"      </news:publication>\n" +
+			"      <news:publication_date>1970-01-01T00:00:00Z</news:publication_date>\n" +
+			"      <news:title>Example Title</news:title>\n" +
 			"    </news:news>\n" + 
 			"  </url>\n" + 
 			"</urlset>";
@@ -58,7 +60,7 @@ public void testKeywords() throws Exception {
 		dateFormat.setTimeZone(W3CDateFormat.ZULU);
 		wsg = GoogleNewsSitemapGenerator.builder("http://www.example.com", dir)
 			.dateFormat(dateFormat).build();
-		GoogleNewsSitemapUrl url = new GoogleNewsSitemapUrl.Options("http://www.example.com/index.html", new Date(0))
+		GoogleNewsSitemapUrl url = new GoogleNewsSitemapUrl.Options("http://www.example.com/index.html", new Date(0), "Example Title", "The Example Times", "en")
 			.keywords("Klaatu", "Barrata", "Nicto")
 			.build();
 		wsg.addUrl(url);
@@ -67,14 +69,47 @@ public void testKeywords() throws Exception {
 			"  <url>\n" + 
 			"    <loc>http://www.example.com/index.html</loc>\n" + 
 			"    <news:news>\n" + 
-			"      <news:publication_date>1970-01-01T00:00:00Z</news:publication_date>\n" + 
+			"      <news:publication>\n" +
+			"        <news:name>The Example Times</news:name>\n" +
+			"        <news:language>en</news:language>\n" +
+			"      </news:publication>\n" +
+			"      <news:publication_date>1970-01-01T00:00:00Z</news:publication_date>\n" +
+			"      <news:title>Example Title</news:title>\n" +
 			"      <news:keywords>Klaatu, Barrata, Nicto</news:keywords>\n" +
 			"    </news:news>\n" + 
 			"  </url>\n" + 
 			"</urlset>";
 		String sitemap = writeSingleSiteMap(wsg);
 		assertEquals(expected, sitemap);
 	}
+
+	public void testGenres() throws Exception {
+		W3CDateFormat dateFormat = new W3CDateFormat(Pattern.SECOND);
+		dateFormat.setTimeZone(W3CDateFormat.ZULU);
+		wsg = GoogleNewsSitemapGenerator.builder("http://www.example.com", dir)
+			.dateFormat(dateFormat).build();
+		GoogleNewsSitemapUrl url = new GoogleNewsSitemapUrl.Options("http://www.example.com/index.html", new Date(0), "Example Title", "The Example Times", "en")
+			.genres("persbericht")
+			.build();
+		wsg.addUrl(url);
+		String expected = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n" +
+			"<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\" xmlns:news=\"http://www.google.com/schemas/sitemap-news/0.9\" >\n" +
+			"  <url>\n" +
+			"    <loc>http://www.example.com/index.html</loc>\n" +
+			"    <news:news>\n" +
+			"      <news:publication>\n" +
+			"        <news:name>The Example Times</news:name>\n" +
+			"        <news:language>en</news:language>\n" +
+			"      </news:publication>\n" +
+			"      <news:genres>persbericht</news:genres>\n" +
+			"      <news:publication_date>1970-01-01T00:00:00Z</news:publication_date>\n" +
+			"      <news:title>Example Title</news:title>\n" +
+			"    </news:news>\n" +
+			"  </url>\n" +
+			"</urlset>";
+		String sitemap = writeSingleSiteMap(wsg);
+		assertEquals(expected, sitemap);
+	}
 
 	private String writeSingleSiteMap(GoogleNewsSitemapGenerator wsg) {
 		List<File> files = wsg.write();