Replace tripadvisor sitemap example with stormcrawler.apache.org content

apache · Nov 22, 2024 · fa509a5 · fa509a5
1 parent ee50de8
commit fa509a5
Show file tree

Hide file tree

Showing 4 changed files with 24 additions and 24 deletions.
diff --git a/core/src/test/java/org/apache/stormcrawler/bolt/SiteMapParserBoltTest.java b/core/src/test/java/org/apache/stormcrawler/bolt/SiteMapParserBoltTest.java
@@ -68,8 +68,8 @@ void testSitemapIndexParsing() throws IOException {
         // and its mime-type
         metadata.setValue(HttpHeaders.CONTENT_TYPE, "application/xml");
         parse(
-                "http://www.tripadvisor.com/sitemap-index.xml",
-                "tripadvisor.sitemap.index.xml",
+                "http://stormcrawler.apache.org/sitemap-index.xml",
+                "stormcrawler.sitemap.index.xml",
                 metadata);
         for (List<Object> fields : output.getEmitted(Constants.StatusStreamName)) {
             Metadata parsedMetadata = (Metadata) fields.get(1);
@@ -85,8 +85,8 @@ void testGzipSitemapParsing() throws IOException {
         Metadata metadata = new Metadata();
         // specify that it is a sitemap file
         metadata.setValue(SiteMapParserBolt.isSitemapKey, "true");
-        parse("https://www.tripadvisor.com/sitemap.xml.gz", "tripadvisor.sitemap.xml.gz", metadata);
-        Assertions.assertEquals(50001, output.getEmitted(Constants.StatusStreamName).size());
+        parse("https://stormcrawler.apache.org/sitemap.xml.gz", "stormcrawler.sitemap.xml.gz", metadata);
+        Assertions.assertEquals(7, output.getEmitted(Constants.StatusStreamName).size());
     }
 
     @Test

diff --git a/...t/resources/tripadvisor.sitemap.index.xml → .../resources/stormcrawler.sitemap.index.xml b/...t/resources/tripadvisor.sitemap.index.xml → .../resources/stormcrawler.sitemap.index.xml
@@ -18,24 +18,24 @@ specific language governing permissions and limitations
 under the License.
 -->
 <sitemapindex
-  xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
-  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
-  xsi:schemaLocation="http://www.sitemaps.org/schemas/sitemap/0.9
+        xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
+        xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+        xsi:schemaLocation="http://www.sitemaps.org/schemas/sitemap/0.9
                       http://www.sitemaps.org/schemas/sitemap/0.9/siteindex.xsd">
-<sitemap>
-  <loc>https://www.tripadvisor.com/sitemap/2/en_US/sitemap-1806509-en_US-hotel_review-1686849999.xml.gz</loc>
-  <lastmod>2023-06-15T17:26:39Z</lastmod>
-</sitemap>
-<sitemap>
-  <loc>https://www.tripadvisor.com/sitemap/2/en_US/sitemap-1806530-en_US-hotel_review-1686850054.xml.gz</loc>
-  <lastmod>2023-06-15T17:27:34Z</lastmod>
-</sitemap>
-<sitemap>
-  <loc>https://www.tripadvisor.com/sitemap/2/en_US/sitemap-1806537-en_US-hotel_review-1686850072.xml.gz</loc>
-  <lastmod>2023-06-15T17:27:52Z</lastmod>
-</sitemap>
-<sitemap>
-  <loc>https://www.tripadvisor.com/sitemap/2/en_US/sitemap-1841024-en_US-hotel_review-1694976638.xml.gz</loc>
-  <lastmod>2023-09-17T18:50:38Z</lastmod>
-</sitemap>
-</sitemapindex>
+  <sitemap>
+    <loc>https://stormcrawler.apache.org/sitemap-001.xml.gz</loc>
+    <lastmod>2024-10-19T11:21:53Z</lastmod>
+  </sitemap>
+  <sitemap>
+    <loc>https://stormcrawler.apache.org/sitemap-002.xml.gz</loc>
+    <lastmod>2024-10-19T11:21:53Z</lastmod>
+  </sitemap>
+  <sitemap>
+    <loc>https://stormcrawler.apache.org/sitemap-003.xml.gz</loc>
+    <lastmod>2024-10-19T11:21:53Z</lastmod>
+  </sitemap>
+  <sitemap>
+    <loc>https://stormcrawler.apache.org/sitemap-004.xml.gz</loc>
+    <lastmod>2024-10-19T11:21:53Z</lastmod>
+  </sitemap>
+</sitemapindex>
diff --git a/core/src/test/resources/stormcrawler.sitemap.xml.gz b/core/src/test/resources/stormcrawler.sitemap.xml.gz
diff --git a/core/src/test/resources/tripadvisor.sitemap.xml.gz b/core/src/test/resources/tripadvisor.sitemap.xml.gz