rss-bridge/bridges/NextgovBridge.php

91 lines
3.9 KiB
PHP
Raw Normal View History

2016-07-26 21:20:13 +02:00
<?php
class NextgovBridge extends BridgeAbstract {
public function loadMetadatas() {
$this->maintainer = 'ORelio';
$this->name = 'Nextgov Bridge';
$this->uri = 'https://www.nextgov.com/';
2016-07-26 21:20:13 +02:00
$this->description = 'USA Federal technology news, best practices, and web 2.0 tools.';
$this->parameters[] = array(
'category'=>array(
'name'=>'Category',
'type'=>'list',
'values'=>array(
'All'=>'all',
'Technology News'=>'technology-news',
'CIO Briefing'=>'cio-briefing',
'Emerging Tech'=>'emerging-tech',
'Cloud'=>'cloud-computing',
'Cybersecurity'=>'cybersecurity',
'Mobile'=>'mobile',
'Health'=>'health',
'Defense'=>'defense',
'Big Data'=>'big-data'
)
)
);
2016-07-26 21:20:13 +02:00
}
public function collectData(){
$param=$this->parameters[$this->queriedContext];
2016-07-26 21:20:13 +02:00
function ExtractFromDelimiters($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;
}
$category = $param['category']['value'];
2016-07-26 21:20:13 +02:00
if (empty($category))
$category = 'all';
if ($category !== preg_replace('/[^a-z-]+/', '', $category) || strlen($category > 32))
$this->returnClientError('Invalid "category" parameter.');
2016-07-26 21:20:13 +02:00
$url = $this->getURI().'rss/'.$category.'/';
$html = $this->getSimpleHTMLDOM($url) or $this->returnServerError('Could not request Nextgov: '.$url);
2016-07-26 21:20:13 +02:00
$limit = 0;
foreach ($html->find('item') as $element) {
if ($limit < 10) {
$article_url = ExtractFromDelimiters($element->innertext, '<link>', '</link>');
$article_author = ExtractFromDelimiters($element->innertext, 'dc/elements/1.1/">', '</dc:creator>');
$article_title = $element->find('title', 0)->plaintext;
$article_subtitle = $element->find('description', 0)->plaintext;
$article_timestamp = strtotime($element->find('pubDate', 0)->plaintext);
$article_thumbnail = ExtractFromDelimiters($element->innertext, '<media:content url="', '"');
$article = $this->getSimpleHTMLDOM($article_url) or $this->returnServerError('Could not request Nextgov: '.$article_url);
2016-07-26 21:20:13 +02:00
$contents = $article->find('div.wysiwyg', 0)->innertext;
$contents = StripWithDelimiters($contents, '<div class="ad-container">', '</div>');
$contents = StripWithDelimiters($contents, '<div', '</div>'); //ad outer div
$contents = StripWithDelimiters($contents, '<script', '</script>');
$contents = ($article_thumbnail == '' ? '' : '<p><img src="'.$article_thumbnail.'" /></p>')
.'<p><b>'.$article_subtitle.'</b></p>'
.trim($contents);
$item = array();
$item['uri'] = $article_url;
$item['title'] = $article_title;
$item['author'] = $article_author;
$item['timestamp'] = $article_timestamp;
$item['content'] = $contents;
2016-07-26 21:20:13 +02:00
$this->items[] = $item;
$limit++;
}
}
}
}