|
1
|
|
|
<?php |
|
2
|
|
|
namespace Tartana\Host; |
|
3
|
|
|
|
|
4
|
|
|
use GuzzleHttp\Psr7\Request; |
|
5
|
|
|
use GuzzleHttp\RequestOptions; |
|
6
|
|
|
use Symfony\Component\DomCrawler\Crawler; |
|
7
|
|
|
use Tartana\Entity\Download; |
|
8
|
|
|
use Tartana\Host\Common\Http; |
|
9
|
|
|
use Tartana\Util; |
|
10
|
|
|
|
|
11
|
|
|
class Youtubecom extends Http |
|
12
|
|
|
{ |
|
13
|
|
|
|
|
14
|
3 |
|
public function fetchLinkList($link) |
|
15
|
|
|
{ |
|
16
|
3 |
|
if (!Util::startsWith($link, 'https://www.youtube.com/playlist?')) { |
|
17
|
1 |
|
return parent::fetchLinkList($link); |
|
18
|
|
|
} |
|
19
|
|
|
|
|
20
|
2 |
|
$resp = $this->getClient()->request('get', $link); |
|
21
|
|
|
|
|
22
|
2 |
|
$crawler = new Crawler($resp->getBody()->getContents()); |
|
23
|
|
|
|
|
24
|
|
|
// Extracting the links which belong to the playlist |
|
25
|
2 |
|
$links = $crawler->filter('a.pl-video-title-link')->extract([ |
|
26
|
|
|
'href' |
|
27
|
2 |
|
]); |
|
28
|
|
|
|
|
29
|
2 |
|
$data = []; |
|
30
|
2 |
|
foreach ($links as $link) { |
|
31
|
2 |
|
$uri = parse_url($link); |
|
32
|
|
|
|
|
33
|
2 |
|
if (!key_exists('path', $uri) || $uri['path'] != '/watch') { |
|
34
|
1 |
|
continue; |
|
35
|
|
|
} |
|
36
|
|
|
|
|
37
|
1 |
|
$params = []; |
|
38
|
1 |
|
parse_str($uri['query'], $params); |
|
39
|
1 |
|
if (key_exists('v', $params)) { |
|
40
|
1 |
|
$data[] = 'https://www.youtube.com/watch?v=' . $params['v']; |
|
41
|
|
|
} |
|
42
|
|
|
} |
|
43
|
2 |
|
return array_unique($data); |
|
44
|
|
|
} |
|
45
|
|
|
|
|
46
|
4 |
|
public function fetchDownloadInfo(array $downloads) |
|
47
|
|
|
{ |
|
48
|
4 |
|
foreach ($downloads as $download) { |
|
49
|
|
|
// Connection check |
|
50
|
|
|
try { |
|
51
|
4 |
|
$data = $this->getStreamData($download); |
|
52
|
3 |
|
if (!empty($data['title']) && empty($download->getFileName())) { |
|
53
|
1 |
|
$download->setFileName($this->makeSafe($data['title'] . '.mp4')); |
|
54
|
|
|
} |
|
55
|
|
|
|
|
56
|
3 |
|
if (key_exists('errorcode', $data) && $data['errorcode'] > 0 && key_exists('reason', $data)) { |
|
57
|
1 |
|
$download->setMessage($data['reason']); |
|
58
|
3 |
|
$download->setState(Download::STATE_DOWNLOADING_ERROR); |
|
59
|
|
|
} |
|
60
|
1 |
|
} catch (\Exception $e) { |
|
61
|
1 |
|
$this->log('Exception fetching head for connection test: ' . $e->getMessage()); |
|
62
|
1 |
|
$download->setMessage('TARTANA_DOWNLOAD_MESSAGE_INVALID_URL'); |
|
63
|
4 |
|
$download->setState(Download::STATE_DOWNLOADING_ERROR); |
|
64
|
|
|
} |
|
65
|
|
|
} |
|
66
|
4 |
|
} |
|
67
|
|
|
|
|
68
|
3 |
|
protected function getUrlToDownload(Download $download) |
|
69
|
|
|
{ |
|
70
|
3 |
|
$data = $this->getStreamData($download); |
|
71
|
|
|
|
|
72
|
3 |
|
if (!key_exists('url_encoded_fmt_stream_map', $data)) { |
|
73
|
1 |
|
return null; |
|
74
|
|
|
} |
|
75
|
|
|
|
|
76
|
2 |
|
foreach (explode(',', $data['url_encoded_fmt_stream_map']) as $streamData) { |
|
77
|
2 |
|
parse_str($streamData, $streamData); |
|
78
|
|
|
|
|
79
|
2 |
|
if (key_exists('url', $streamData)) { |
|
80
|
1 |
|
$url = urldecode($streamData['url']); |
|
81
|
1 |
|
$this->log('Youtube real url: ' . $url); |
|
82
|
2 |
|
return $url; |
|
83
|
|
|
} |
|
84
|
|
|
} |
|
85
|
1 |
|
return null; |
|
86
|
|
|
} |
|
87
|
|
|
|
|
88
|
7 |
|
private function getStreamData(Download $download) |
|
89
|
|
|
{ |
|
90
|
7 |
|
if (preg_match( |
|
91
|
7 |
|
'%(?:youtube(?:-nocookie)?\.com/(?:[^/]+/.+/|(?:v|e(?:mbed)?)/|.*[?&]v=)|youtu\.be/)([^"&?/ ]{11})%i', |
|
92
|
7 |
|
$download->getLink(), |
|
93
|
|
|
$match |
|
94
|
|
|
)) { |
|
95
|
|
|
$headers = [ |
|
96
|
6 |
|
'User-Agent' => 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/44.0 (Chrome)', |
|
97
|
6 |
|
'Referer' => $download->getLink(), |
|
98
|
6 |
|
'Accept-Language' => 'en-us,en;q=0.5', |
|
99
|
6 |
|
'Accept-Encoding' => 'gzip, deflate', |
|
100
|
6 |
|
'Accept-Charset' => 'ISO-8859-1,utf-8;q=0.7,*;q=0.7', |
|
101
|
6 |
|
'Accept' => 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', |
|
102
|
6 |
|
'Connection' => 'close', |
|
103
|
6 |
|
'Origin' => 'https://www.youtube.com', |
|
104
|
6 |
|
'Host' => 'www.youtube.com' |
|
105
|
|
|
]; |
|
106
|
|
|
// Getting the link information initializing the cookie |
|
107
|
6 |
|
$this->getClient()->request( |
|
108
|
6 |
|
'get', |
|
109
|
6 |
|
$download->getLink() . '&gl=US&hl=en&has_verified=1&bpctr=9999999999', |
|
110
|
|
|
[ |
|
111
|
6 |
|
RequestOptions::HEADERS => $headers |
|
112
|
|
|
] |
|
113
|
|
|
); |
|
114
|
|
|
|
|
115
|
|
|
// Fetching the video information |
|
116
|
6 |
|
$res = $this->getClient()->request( |
|
117
|
6 |
|
'get', |
|
118
|
6 |
|
'http://www.youtube.com/get_video_info?video_id=' . $match[1] . '&el=info&ps=default&eurl=&gl=US&hl=en', |
|
119
|
|
|
[ |
|
120
|
6 |
|
RequestOptions::HEADERS => $headers |
|
121
|
|
|
] |
|
122
|
|
|
); |
|
123
|
|
|
|
|
124
|
5 |
|
parse_str($res->getBody()->getContents(), $videoData); |
|
125
|
5 |
|
return $videoData; |
|
126
|
|
|
} |
|
127
|
|
|
|
|
128
|
|
|
return [ |
|
129
|
1 |
|
'title' => '', |
|
130
|
|
|
'url_encoded_fmt_stream_map' => '' |
|
131
|
|
|
]; |
|
132
|
|
|
} |
|
133
|
|
|
|
|
134
|
|
|
/** |
|
135
|
|
|
* |
|
136
|
|
|
* @param string $file |
|
137
|
|
|
* @return string |
|
138
|
|
|
* @see http://stackoverflow.com/questions/2668854/sanitizing-strings-to-make-them-url-and-filename-safe |
|
139
|
|
|
* @see http://stackoverflow.com/a/24812812/356375 |
|
140
|
|
|
*/ |
|
141
|
1 |
|
private function makeSafe($file) |
|
142
|
|
|
{ |
|
143
|
|
|
// Remove any trailing dots, as those aren't ever valid file names. |
|
144
|
1 |
|
$file = rtrim($file, '.'); |
|
145
|
|
|
|
|
146
|
|
|
$regex = array( |
|
147
|
1 |
|
'#(\.){2,}#', |
|
148
|
|
|
'#[^A-Za-z0-9\.\_\- ]#', |
|
149
|
|
|
'#^\.#' |
|
150
|
|
|
); |
|
151
|
|
|
|
|
152
|
1 |
|
return trim(preg_replace($regex, '', $file)); |
|
153
|
|
|
} |
|
154
|
|
|
} |
|
155
|
|
|
|