You may specify a
topic found in some section URLs, else all topics are selected.';
const PARAMETERS = array( array(
'topic' => array(
'name' => 'Topic name'
)
));
public function collectData(){
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;
}
$pageUrl = self::URI . (empty($this->getInput('topic')) ? '' : 'topics/' . $this->getInput('topic') . '/');
$html = getSimpleHTMLDOM($pageUrl) or 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);
$article_uri = self::URI . ($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) {
$article_html = getSimpleHTMLDOM($article_uri)
or returnServerError('Could not request CNET: ' . $article_uri);
$article_content = trim(
cleanArticle(
extractFromDelimiters(
$article_html,
'