boyska/rss-bridge
1
0
Fork 0
forked from blallo/rss-bridge
Code Releases Activity

[pull] Fix Youtube, Twitter

Browse source
This commit is contained in:
Mitsukarenai 2014-05-14 14:34:06 +02:00
parent e81a19a04b
commit 1e40d6f492
2 changed files with 16 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
bridges/TwitterBridge.php
View file

@ -26,15 +26,15 @@ class TwitterBridge extends BridgeAbstract{
$this->returnError('You must specify a keyword (?q=...) or a Twitter username (?u=...).', 400); $this->returnError('You must specify a keyword (?q=...) or a Twitter username (?u=...).', 400);
} }
foreach($html->find('div.tweet') as $tweet) { foreach($html->find('div.js-stream-tweet') as $tweet) {
$item = new \Item(); $item = new \Item();
$item->username = trim(substr($tweet->find('span.username', 0)->plaintext, 1)); // extract username and sanitize $item->username = $tweet->getAttribute('data-screen-name'); // extract username and sanitize
$item->fullname = $tweet->getAttribute('data-name'); // extract fullname (pseudonym) $item->fullname = $tweet->getAttribute('data-name'); // extract fullname (pseudonym)
$item->avatar = $tweet->find('img', 0)->src; // get avatar link $item->avatar = $tweet->find('img', 0)->src; // get avatar link
$item->id = $tweet->getAttribute('data-tweet-id'); // get TweetID $item->id = $tweet->getAttribute('data-tweet-id'); // get TweetID
$item->uri = 'https://twitter.com'.$tweet->find('a.details', 0)->getAttribute('href'); // get tweet link $item->uri = 'https://twitter.com'.$tweet->find('a.js-permalink', 0)->getAttribute('href'); // get tweet link
$item->timestamp = $tweet->find('span._timestamp', 0)->getAttribute('data-time'); // extract tweet timestamp $item->timestamp = $tweet->find('span.js-short-timestamp', 0)->getAttribute('data-time'); // extract tweet timestamp
$item->content = str_replace('href="/', 'href="https://twitter.com/', strip_tags($tweet->find('p.tweet-text', 0)->innertext, '<a>')); // extract tweet text $item->content = str_replace('href="/', 'href="https://twitter.com/', strip_tags($tweet->find('p.js-tweet-text', 0)->innertext, '<a>')); // extract tweet text
$item->title = $item->fullname . ' (@'. $item->username . ') | ' . $item->content; $item->title = $item->fullname . ' (@'. $item->username . ') | ' . $item->content;
$this->items[] = $item; $this->items[] = $item;
} }

18
bridges/YoutubeBridge.php
View file

@ -8,6 +8,10 @@
* @use1(u="username") * @use1(u="username")
* @use2(p="playlist id") * @use2(p="playlist id")
* @use3(s="search keyword",pa="page") * @use3(s="search keyword",pa="page")
*
* WARNING: to parse big playlists (over ~90 videos), you need to edit simple_html_dom.php:
* change: define('MAX_FILE_SIZE', 600000);
* into: define('MAX_FILE_SIZE', 900000); (or more)
*/ */
class YoutubeBridge extends BridgeAbstract{ class YoutubeBridge extends BridgeAbstract{
@ -32,21 +36,21 @@ class YoutubeBridge extends BridgeAbstract{
$this->request = $param['p']; $this->request = $param['p'];
$html = file_get_html('https://www.youtube.com/playlist?list='.urlencode($this->request).'') or $this->returnError('Could not request Youtube.', 404); $html = file_get_html('https://www.youtube.com/playlist?list='.urlencode($this->request).'') or $this->returnError('Could not request Youtube.', 404);
foreach($html->find('li.playlist-video-item') as $element) { foreach($html->find('tr.pl-video') as $element) {
$item = new \Item(); $item = new \Item();
$item->uri = 'https://www.youtube.com'.$element->find('a',0)->href; $item->uri = 'https://www.youtube.com'.$element->find('.pl-video-title a',0)->href;
$item->thumbnailUri = 'https:'.$element->find('img',0)->src; $item->thumbnailUri = 'https:'.str_replace('/default.','/mqdefault.',$element->find('.pl-video-thumbnail img',0)->src);
$item->title = trim($element->find('h3',0)->plaintext); $item->title = trim($element->find('.pl-video-title a',0)->plaintext);
$item->content = '<a href="' . $item->uri . '"><img src="' . $item->thumbnailUri . '" /></a><br><a href="' . $item->uri . '">' . $item->title . '</a>'; $item->content = '<a href="' . $item->uri . '"><img src="' . $item->thumbnailUri . '" /></a><br><a href="' . $item->uri . '">' . $item->title . '</a>';
$this->items[] = $item; $this->items[] = $item;
} }
$this->request = 'Playlist '.str_replace(' - YouTube', '', $html->find('title', 0)->plaintext).', by '.$html->find('h1', 0)->plaintext; $this->request = 'Playlist '.trim(str_replace(' - YouTube', '', $html->find('title', 0)->plaintext)).', by '.$html->find('h1', 0)->plaintext;
} }
else if (isset($param['s'])) { /* search mode */ else if (isset($param['s'])) { /* search mode */
$this->request = $param['s']; $page = 1; if (isset($param['pa'])) $page = (int)preg_replace("/[^0-9]/",'', $param['pa']); $this->request = $param['s']; $page = 1; if (isset($param['pa'])) $page = (int)preg_replace("/[^0-9]/",'', $param['pa']);
$html = file_get_html('https://www.youtube.com/results?search_query='.urlencode($this->request).'&page='.$page.'&filters=video&search_sort=video_date_uploaded') or $this->returnError('Could not request Youtube.', 404); $html = file_get_html('https://www.youtube.com/results?search_query='.urlencode($this->request).'&&page='.$page.'&filters=video&search_sort=video_date_uploaded') or $this->returnError('Could not request Youtube.', 404);
foreach($html->find('li.context-data-item') as $element) { foreach($html->find('li.yt-lockup') as $element) {
$item = new \Item(); $item = new \Item();
$item->uri = 'https://www.youtube.com'.$element->find('a',0)->href; $item->uri = 'https://www.youtube.com'.$element->find('a',0)->href;
$checkthumb = $element->find('img', 0)->getAttribute('data-thumb'); $checkthumb = $element->find('img', 0)->getAttribute('data-thumb');