blallo/rss-bridge
1
0
Fork 1
Code Issues Pull requests Releases Wiki Activity

Add http-context to simulate Mozilla user-agent

Browse source 
This fixes "Warning: file_get_contents(http://www.ledauphine.com/rss):
failed to open stream: HTTP request failed! HTTP/1.1 403 Forbidden"
This commit is contained in:
logmanoriginal 2016-08-02 15:57:01 +02:00
parent 4420906a7a
commit c8c3e9ef65
1 changed files with 21 additions and 11 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

32
bridges/DauphineLibereBridge.php
View file

@ -81,8 +81,8 @@ class DauphineLibereBridge extends BridgeAbstract {
]'; ]';
} }
function ExtractContent($url) { function ExtractContent($url, $context) {
$html2 = $this->file_get_html($url); $html2 = $this->file_get_html($url, false, $context);
$text = $html2->find('div.column', 0)->innertext; $text = $html2->find('div.column', 0)->innertext;
$text = preg_replace('@<script[^>]*?>.*?</script>@si', '', $text); $text = preg_replace('@<script[^>]*?>.*?</script>@si', '', $text);
return $text; return $text;
@ -90,24 +90,34 @@ class DauphineLibereBridge extends BridgeAbstract {
public function collectData(array $param){ public function collectData(array $param){
// Simulate Mozilla user-agent to fix error 403 (Forbidden)
$opts = array('http' =>
array(
'method' => 'GET',
'header' => 'User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.102 Safari/537.36'
)
);
$context = stream_context_create($opts);
if (isset($param['u'])) { /* user timeline mode */ if (isset($param['u'])) { /* user timeline mode */
$this->request = $param['u']; $this->request = $param['u'];
$html = $this->file_get_html('http://www.ledauphine.com/'.$this->request.'/rss') or $this->returnError('Could not request DauphineLibere.', 404); $html = $this->file_get_html('http://www.ledauphine.com/'.$this->request.'/rss', false, $context) or $this->returnError('Could not request DauphineLibere.', 404);
} }
else { else {
$html = $this->file_get_html('http://www.ledauphine.com/rss') or $this->returnError('Could not request DauphineLibere.', 404); $html = $this->file_get_html('http://www.ledauphine.com/rss', false, $context) or $this->returnError('Could not request DauphineLibere.', 404);
} }
$limit = 0; $limit = 0;
foreach($html->find('item') as $element) { foreach($html->find('item') as $element) {
if($limit < 10) { if($limit < 10) {
$item = new \Item(); $item = new \Item();
$item->title = $element->find('title', 0)->innertext; $item->title = $element->find('title', 0)->innertext;
$item->uri = $element->find('guid', 0)->plaintext; $item->uri = $element->find('guid', 0)->plaintext;
$item->timestamp = strtotime($element->find('pubDate', 0)->plaintext); $item->timestamp = strtotime($element->find('pubDate', 0)->plaintext);
$item->content = $this->ExtractContent($item->uri); $item->content = $this->ExtractContent($item->uri, $context);
$this->items[] = $item; $this->items[] = $item;
$limit++; $limit++;
} }
} }
} }