rss-bridge/bridges/GoogleSearchBridge.php

55 lines
1.8 KiB
PHP
Raw Normal View History

2013-08-11 15:30:41 +04:00
<?php
/**
* Returns the 100 most recent links in results in past year, sorting by date (most recent first).
* Example:
* http://www.google.com/search?q=sebsauvage&num=100&complete=0&tbs=qdr:y,sbd:1
* complete=0&num=100 : get 100 results
* qdr:y : in past year
* sbd:1 : sort by date (will only work if qdr: is specified)
*/
class GoogleSearchBridge extends BridgeAbstract{
const MAINTAINER = "sebsauvage";
const NAME = "Google search";
const URI = "https://www.google.com/";
const CACHE_TIMEOUT = 1800; // 30min
const DESCRIPTION = "Returns most recent results from Google search.";
const PARAMETERS = array( array(
'q'=>array(
'name'=>"keyword",
'required'=>true
)
));
public function collectData(){
2013-08-11 15:30:41 +04:00
$html = '';
$html = getSimpleHTMLDOM(self::URI
.'search?q=' . urlencode($this->getInput('q'))
.'&num=100&complete=0&tbs=qdr:y,sbd:1')
or returnServerError('No results for this query.');
2013-08-11 15:30:41 +04:00
$emIsRes = $html->find('div[id=ires]',0);
if( !is_null($emIsRes) ){
foreach($emIsRes->find('li[class=g]') as $element) {
$item = array();
// Extract direct URL from google href (eg. /url?q=...)
$t = $element->find('a[href]',0)->href;
$item['uri'] = ''.$t;
parse_str(parse_url($t, PHP_URL_QUERY),$parameters);
if (isset($parameters['q'])) { $item['uri'] = $parameters['q']; }
$item['title'] = $element->find('h3',0)->plaintext;
$item['content'] = $element->find('span[class=st]',0)->plaintext;
2013-08-11 15:30:41 +04:00
$this->items[] = $item;
}
}
}
public function getName(){
return $this->getInput('q') .' - Google search';
2013-08-11 15:30:41 +04:00
}
2014-05-21 21:15:52 +04:00
}