You may specify a topic found in some section URLs, else all topics are selected.';
public $parameters = array( array(
'topic'=>array('name'=>'Topic name')
));
public function collectData(){
$param=$this->parameters[$this->queriedContext];
function ExtractFromDelimiters($string, $start, $end) {
if (strpos($string, $start) !== false) {
$section_retrieved = substr($string, strpos($string, $start) + strlen($start));
$section_retrieved = substr($section_retrieved, 0, strpos($section_retrieved, $end));
return $section_retrieved;
} return false;
}
function StripWithDelimiters($string, $start, $end) {
while (strpos($string, $start) !== false) {
$section_to_remove = substr($string, strpos($string, $start));
$section_to_remove = substr($section_to_remove, 0, strpos($section_to_remove, $end) + strlen($end));
$string = str_replace($section_to_remove, '', $string);
} return $string;
}
function CleanArticle($article_html) {
$article_html = ''.substr($article_html, strpos($article_html, '
') + 3);
$article_html = StripWithDelimiters($article_html, '', '');
$article_html = StripWithDelimiters($article_html, '');
$article_html = StripWithDelimiters($article_html, '', '');
return $article_html;
}
$pageUrl = 'http://www.cnet.com/'.(empty($param['topic']['value']) ? '' : 'topics/'.$param['topic']['value'].'/');
$html = $this->getSimpleHTMLDOM($pageUrl) or $this->returnServerError('Could not request CNET: '.$pageUrl);
$limit = 0;
foreach($html->find('div.assetBody') as $element) {
if ($limit < 8) {
$article_title = trim($element->find('h2', 0)->plaintext);
$article_uri = 'http://www.cnet.com'.($element->find('a', 0)->href);
$article_timestamp = strtotime($element->find('time.assetTime', 0)->plaintext);
$article_author = trim($element->find('a[rel=author]', 0)->plaintext);
if (!empty($article_title) && !empty($article_uri) && strpos($article_uri, '/news/') !== false) {
$article_html = $this->getSimpleHTMLDOM($article_uri) or $this->returnServerError('Could not request CNET: '.$article_uri);
$article_content = trim(CleanArticle(ExtractFromDelimiters($article_html, '