public function testGetSitemapsOrgXmlSitemapType()
{
$sitemap = new Sitemap();
$sitemap->setHttpResponse($this->getHttpFixture('SitemapsOrgXmlContent', 'application/xml'));
$sitemap->setUrl('http://webignition.net/sitemap.xml');
$this->assertEquals('sitemaps.org.xml', $sitemap->getType());
}
public function testIsIndexForSitemapsOrgXmlSitemapType()
{
$sitemap = new Sitemap();
$sitemap->setHttpResponse($this->getHttpFixture('SitemapsOrgXmlContent', 'application/xml'));
$sitemap->setUrl('http://webignition.net/sitemap.xml');
$this->assertFalse($sitemap->isIndex());
}
/**
*
* @return \webignition\WebResource\Sitemap\Sitemap
*/
protected function createSitemap()
{
$configuration = new SitemapConfiguration();
$configuration->setTypeToUrlExtractorClassMap(array('sitemaps.org.xml' => 'webignition\\WebResource\\Sitemap\\UrlExtractor\\SitemapsOrgXmlUrlExtractor', 'sitemaps.org.txt' => 'webignition\\WebResource\\Sitemap\\UrlExtractor\\SitemapsOrgTxtUrlExtractor', 'application/atom+xml' => 'webignition\\WebResource\\Sitemap\\UrlExtractor\\NewsFeedUrlExtractor', 'application/rss+xml' => 'webignition\\WebResource\\Sitemap\\UrlExtractor\\NewsFeedUrlExtractor', 'sitemaps.org.xml.index' => 'webignition\\WebResource\\Sitemap\\UrlExtractor\\SitemapsOrgXmlIndexUrlExtractor'));
$sitemap = new Sitemap();
$sitemap->setConfiguration($configuration);
return $sitemap;
}
public function testIsSitemapForNonSitemapContent()
{
$sitemap = new Sitemap();
$sitemap->setHttpResponse($this->getHttpFixture());
$sitemap->setUrl('http://webignition.net/sitemap.xml');
$this->assertFalse($sitemap->isSitemap());
$sitemap->setContent("one\ntwo\nthree");
$this->assertFalse($sitemap->isSitemap());
$sitemap->setContent("<!DOCTYPE html>\n<html>\n <head>\n <title>Hello HTML</title>\n </head>\n <body>\n <p>Hello World!</p>\n </body>\n</html>");
$this->assertFalse($sitemap->isSitemap());
}
/**
*
* @param \webignition\WebResource\Sitemap\Sitemap $sitemap
* @return boolean
*/
public function retrieve(Sitemap $sitemap)
{
if ($this->getConfiguration()->getShouldHalt()) {
return false;
}
$request = clone $this->getConfiguration()->getBaseRequest();
$request->setUrl($sitemap->getUrl());
$this->setRequestCookies($request);
$this->setRequestTimeout($request);
$events = $this->getPreAndPostTransferEvents();
$this->dispatcher->dispatch(Events::TRANSFER_PRE, $events['pre']);
$lastRequestException = null;
try {
$response = $request->send();
} catch (\Guzzle\Http\Exception\CurlException $curlException) {
$lastRequestException = $curlException;
} catch (\Guzzle\Http\Exception\RequestException $requestException) {
$lastRequestException = $requestException;
}
$this->dispatcher->dispatch(Events::TRANSFER_POST, $events['post']);
if ($lastRequestException instanceof \Exception || $response->getStatusCode() !== 200) {
return false;
}
$sitemap->setHttpResponse($response);
if ($sitemap->isIndex()) {
$childUrls = $sitemap->getUrls();
foreach ($childUrls as $childUrl) {
$childSitemap = new Sitemap();
$childSitemap->setConfiguration($sitemap->getConfiguration());
$childSitemap->setUrl($childUrl);
$sitemap->addChild($childSitemap);
if ($this->getConfiguration()->getRetrieveChildSitemaps()) {
$this->retrieve($childSitemap);
}
}
}
return true;
}
/**
*
* @param \webignition\WebResource\Sitemap\Sitemap $sitemap
* @return boolean
*/
public function addChild(Sitemap $sitemap)
{
if (!$this->isIndex()) {
return false;
}
$childUrl = new \webignition\NormalisedUrl\NormalisedUrl($sitemap->getUrl());
$childIndex = md5((string) $childUrl);
$this->children[$childIndex] = $sitemap;
return true;
}
