diff --git a/CHANGELOG.txt b/CHANGELOG.txt
index 7fc6f6c8203be9d4e4ab210e497f978e61f6eedf..b90a628be7f3cd4dfdc4912f97f7296fdb3a720c 100644
--- a/CHANGELOG.txt
+++ b/CHANGELOG.txt
@@ -3,6 +3,7 @@
Feeds 6.x 1.0 xxxxxxxxxxxxxxxxxxxx
----------------------------------
+- #759904 lyricnz: Provide a Google Sitemap Parser.
- #774858 rjbrown99: Fix Node Processor updates node "created" time when
updating.
- #704236 jerdavis: Support mapping to CCK float field.
diff --git a/feeds.plugins.inc b/feeds.plugins.inc
index e1e7e4975608510b007e5f5e09d79e64e3f6ae59..6aa87d01a5d73c203c82ab4a6501438d6809be67 100644
--- a/feeds.plugins.inc
+++ b/feeds.plugins.inc
@@ -120,6 +120,16 @@ function _feeds_feeds_plugins() {
),
);
}
+ $info['FeedsSitemapParser'] = array(
+ 'name' => 'Sitemap parser',
+ 'description' => 'Parse Sitemap XML format feeds.',
+ 'handler' => array(
+ 'parent' => 'FeedsParser',
+ 'class' => 'FeedsSitemapParser',
+ 'file' => 'FeedsSitemapParser.inc',
+ 'path' => $path,
+ ),
+ );
$info['FeedsNodeProcessor'] = array(
'name' => 'Node processor',
'description' => 'Create nodes.',
diff --git a/plugins/FeedsSitemapParser.inc b/plugins/FeedsSitemapParser.inc
new file mode 100644
index 0000000000000000000000000000000000000000..b0180a57837c6783d91eaa0f1f567941333a59e9
--- /dev/null
+++ b/plugins/FeedsSitemapParser.inc
@@ -0,0 +1,52 @@
+<?php
+// $Id$
+
+/**
+ * A parser for the Sitemap specification http://www.sitemaps.org/protocol.php
+ */
+class FeedsSitemapParser extends FeedsParser {
+ /**
+ * Implementation of FeedsParser::parse().
+ */
+ public function parse(FeedsImportBatch $batch, FeedsSource $source) {
+ // Yes, using a DOM parser is a bit inefficient, but will do for now
+ $xml = new SimpleXMLElement($batch->getRaw());
+ foreach ($xml->url as $url) {
+ $item = array('url' => (string) $url->loc);
+ if ($url->lastmod) {
+ $item['lastmod'] = strtotime($url->lastmod);
+ }
+ if ($url->changefreq) {
+ $item['changefreq'] = $url->changefreq;
+ }
+ if ($url->priority) {
+ $item['priority'] = $url->priority;
+ }
+ $batch->addItem($item);
+ }
+ }
+
+ /**
+ * Implementation of FeedsParser::getMappingSources().
+ */
+ public function getMappingSources() {
+ return array(
+ 'url' => array(
+ 'name' => t('Item URL (link)'),
+ 'description' => t('URL of the feed item.'),
+ ),
+ 'lastmod' => array(
+ 'name' => t('Last modification date'),
+ 'description' => t('Last modified date as UNIX time GMT of the feed item.'),
+ ),
+ 'changefreq' => array(
+ 'name' => t('Change frequency'),
+ 'description' => t('How frequently the page is likely to change.'),
+ ),
+ 'priority' => array(
+ 'name' => t('Priority'),
+ 'description' => t('The priority of this URL relative to other URLs on the site.'),
+ ),
+ );
+ }
+}
diff --git a/tests/feeds.test b/tests/feeds.test
index 269bb58696b72c8721ad6f38ebafe62568743202..68a77a8e25afb6ed041b5151812483d301e074b7 100644
--- a/tests/feeds.test
+++ b/tests/feeds.test
@@ -841,3 +841,134 @@ class FeedsSyndicationParserTestCase extends FeedsWebTestCase {
);
}
}
+
+/**
+ * Test Sitemap parser.
+ */
+class FeedsSitemapParserTestCase extends FeedsWebTestCase {
+
+ /**
+ * Describe this test.
+ */
+ public function getInfo() {
+ return array(
+ 'name' => t('Sitemap parser'),
+ 'description' => t('Regression tests for Sitemap XML format parser.'),
+ 'group' => t('Feeds'),
+ );
+ }
+
+ /**
+ * Set up test.
+ */
+ public function setUp() {
+ parent::setUp('feeds', 'feeds_ui', 'ctools');
+
+ $this->drupalLogin(
+ $this->drupalCreateUser(
+ array(
+ 'administer feeds', 'administer nodes',
+ )
+ )
+ );
+ }
+
+ /**
+ * Run tests.
+ */
+ public function test() {
+ $this->createFeedConfiguration('Sitemap', 'sitemap');
+ $this->setPlugin('sitemap', 'FeedsSitemapParser');
+
+ $this->addMappings('sitemap',
+ array(
+ array(
+ 'source' => 'changefreq',
+ 'target' => 'title',
+ 'unique' => FALSE,
+ ),
+ array(
+ 'source' => 'priority',
+ 'target' => 'body',
+ 'unique' => FALSE,
+ ),
+ array(
+ 'source' => 'lastmod',
+ 'target' => 'created',
+ 'unique' => FALSE,
+ ),
+ array(
+ 'source' => 'url',
+ 'target' => 'url',
+ 'unique' => TRUE,
+ ),
+ array(
+ 'source' => 'url',
+ 'target' => 'guid',
+ 'unique' => TRUE,
+ ),
+ )
+ );
+
+
+ $path = $GLOBALS['base_url'] .'/'. drupal_get_path('module', 'feeds') .'/tests/feeds/';
+ $nid = $this->createFeedNode('sitemap', $path .'sitemap-example.xml', 'Testing Sitemap Parser');
+ $this->assertText('Created 5 Story nodes.');
+
+ // Assert DB status.
+ $count = db_result(db_query("SELECT COUNT(*) FROM {feeds_node_item}"));
+ $this->assertEqual($count, 5, 'Accurate number of items in database.');
+
+ // Check items against known content of feed.
+ $result = db_query('SELECT * FROM {feeds_node_item} WHERE feed_nid = %d ORDER BY nid', $nid);
+
+ // Check first item.
+ $item = db_fetch_object($result);
+ $node = node_load($item->nid);
+ $this->assertEqual($node->title, 'monthly', 'Feed item 1 changefreq is correct.');
+ $this->assertEqual($node->body, '0.8', 'Feed item 1 priority is correct.');
+ $this->assertEqual($node->created, strtotime('2005-01-01'), 'Feed item 1 lastmod is correct.');
+ $this->assertEqual($node->feeds_node_item->url, 'http://www.example.com/', 'Feed item 1 url is correct.');
+ $this->assertEqual($node->feeds_node_item->url, $node->feeds_node_item->guid, 'Feed item 1 guid is correct.');
+
+ // Check second item.
+ $item = db_fetch_object($result);
+ $node = node_load($item->nid);
+ $this->assertEqual($node->title, 'weekly', 'Feed item 2 changefreq is correct.');
+ $this->assertEqual($node->body, '', 'Feed item 2 priority is correct.');
+ // $node->created is... recently
+ $this->assertEqual($node->feeds_node_item->url, 'http://www.example.com/catalog?item=12&desc=vacation_hawaii', 'Feed item 2 url is correct.');
+ $this->assertEqual($node->feeds_node_item->url, $node->feeds_node_item->guid, 'Feed item 2 guid is correct.');
+
+ // Check third item.
+ $item = db_fetch_object($result);
+ $node = node_load($item->nid);
+ $this->assertEqual($node->title, 'weekly', 'Feed item 3 changefreq is correct.');
+ $this->assertEqual($node->body, '', 'Feed item 3 priority is correct.');
+ $this->assertEqual($node->created, strtotime('2004-12-23'), 'Feed item 3 lastmod is correct.');
+ $this->assertEqual($node->feeds_node_item->url, 'http://www.example.com/catalog?item=73&desc=vacation_new_zealand', 'Feed item 3 url is correct.');
+ $this->assertEqual($node->feeds_node_item->url, $node->feeds_node_item->guid, 'Feed item 3 guid is correct.');
+
+ // Check fourth item.
+ $item = db_fetch_object($result);
+ $node = node_load($item->nid);
+ $this->assertEqual($node->title, '', 'Feed item 4 changefreq is correct.');
+ $this->assertEqual($node->body, '0.3', 'Feed item 4 priority is correct.');
+ $this->assertEqual($node->created, strtotime('2004-12-23T18:00:15+00:00'), 'Feed item 4 lastmod is correct.');
+ $this->assertEqual($node->feeds_node_item->url, 'http://www.example.com/catalog?item=74&desc=vacation_newfoundland', 'Feed item 4 url is correct.');
+ $this->assertEqual($node->feeds_node_item->url, $node->feeds_node_item->guid, 'Feed item 1 guid is correct.');
+
+ // Check fifth item.
+ $item = db_fetch_object($result);
+ $node = node_load($item->nid);
+ $this->assertEqual($node->title, '', 'Feed item 5 changefreq is correct.');
+ $this->assertEqual($node->body, '', 'Feed item 5 priority is correct.');
+ $this->assertEqual($node->created, strtotime('2004-11-23'), 'Feed item 5 lastmod is correct.');
+ $this->assertEqual($node->feeds_node_item->url, 'http://www.example.com/catalog?item=83&desc=vacation_usa', 'Feed item 5 url is correct.');
+ $this->assertEqual($node->feeds_node_item->url, $node->feeds_node_item->guid, 'Feed item 5 guid is correct.');
+
+ // Check for more items.
+ $item = db_fetch_object($result);
+ $this->assertFalse($item, 'Correct number of feed items recorded.');
+ }
+}
diff --git a/tests/feeds/sitemap-example.xml b/tests/feeds/sitemap-example.xml
new file mode 100644
index 0000000000000000000000000000000000000000..7695060cc4e47c3a31a6e8f472ea150ec2959197
--- /dev/null
+++ b/tests/feeds/sitemap-example.xml
@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
+ <url>
+ <loc>http://www.example.com/</loc>
+ <lastmod>2005-01-01</lastmod>
+ <changefreq>monthly</changefreq>
+ <priority>0.8</priority>
+ </url>
+ <url>
+ <loc>http://www.example.com/catalog?item=12&desc=vacation_hawaii</loc>
+ <changefreq>weekly</changefreq>
+ </url>
+ <url>
+ <loc>http://www.example.com/catalog?item=73&desc=vacation_new_zealand</loc>
+ <lastmod>2004-12-23</lastmod>
+ <changefreq>weekly</changefreq>
+ </url>
+ <url>
+ <loc>http://www.example.com/catalog?item=74&desc=vacation_newfoundland</loc>
+ <lastmod>2004-12-23T18:00:15+00:00</lastmod>
+ <priority>0.3</priority>
+ </url>
+ <url>
+ <loc>http://www.example.com/catalog?item=83&desc=vacation_usa</loc>
+ <lastmod>2004-11-23</lastmod>
+ </url>
+</urlset>