Merge remote-tracking branch 'h-collector/master'
[youtube-dl] / youtube_dl / extractor / vevo.py
1 from __future__ import unicode_literals
2
3 import re
4 import xml.etree.ElementTree
5
6 from .common import InfoExtractor
7 from ..compat import (
8     compat_urllib_request,
9 )
10 from ..utils import (
11     ExtractorError,
12 )
13
14
15 class VevoIE(InfoExtractor):
16     """
17     Accepts urls from vevo.com or in the format 'vevo:{id}'
18     (currently used by MTVIE and MySpaceIE)
19     """
20     _VALID_URL = r'''(?x)
21         (?:https?://www\.vevo\.com/watch/(?:[^/]+/(?:[^/]+/)?)?|
22            https?://cache\.vevo\.com/m/html/embed\.html\?video=|
23            https?://videoplayer\.vevo\.com/embed/embedded\?videoId=|
24            vevo:)
25         (?P<id>[^&?#]+)'''
26
27     _TESTS = [{
28         'url': 'http://www.vevo.com/watch/hurts/somebody-to-die-for/GB1101300280',
29         "md5": "95ee28ee45e70130e3ab02b0f579ae23",
30         'info_dict': {
31             'id': 'GB1101300280',
32             'ext': 'mp4',
33             "upload_date": "20130624",
34             "uploader": "Hurts",
35             "title": "Somebody to Die For",
36             "duration": 230.12,
37             "width": 1920,
38             "height": 1080,
39             # timestamp and upload_date are often incorrect; seem to change randomly
40             'timestamp': int,
41         }
42     }, {
43         'note': 'v3 SMIL format',
44         'url': 'http://www.vevo.com/watch/cassadee-pope/i-wish-i-could-break-your-heart/USUV71302923',
45         'md5': 'f6ab09b034f8c22969020b042e5ac7fc',
46         'info_dict': {
47             'id': 'USUV71302923',
48             'ext': 'mp4',
49             'upload_date': '20140219',
50             'uploader': 'Cassadee Pope',
51             'title': 'I Wish I Could Break Your Heart',
52             'duration': 226.101,
53             'age_limit': 0,
54             'timestamp': int,
55         }
56     }, {
57         'note': 'Age-limited video',
58         'url': 'https://www.vevo.com/watch/justin-timberlake/tunnel-vision-explicit/USRV81300282',
59         'info_dict': {
60             'id': 'USRV81300282',
61             'ext': 'mp4',
62             'age_limit': 18,
63             'title': 'Tunnel Vision (Explicit)',
64             'uploader': 'Justin Timberlake',
65             'upload_date': 're:2013070[34]',
66             'timestamp': int,
67         },
68         'params': {
69             'skip_download': 'true',
70         }
71     }]
72     _SMIL_BASE_URL = 'http://smil.lvl3.vevo.com/'
73
74     def _real_initialize(self):
75         req = compat_urllib_request.Request(
76             'http://www.vevo.com/auth', data=b'')
77         webpage = self._download_webpage(
78             req, None,
79             note='Retrieving oauth token',
80             errnote='Unable to retrieve oauth token',
81             fatal=False)
82         if webpage is False:
83             self._oauth_token = None
84         else:
85             self._oauth_token = self._search_regex(
86                 r'access_token":\s*"([^"]+)"',
87                 webpage, 'access token', fatal=False)
88
89     def _formats_from_json(self, video_info):
90         last_version = {'version': -1}
91         for version in video_info['videoVersions']:
92             # These are the HTTP downloads, other types are for different manifests
93             if version['sourceType'] == 2:
94                 if version['version'] > last_version['version']:
95                     last_version = version
96         if last_version['version'] == -1:
97             raise ExtractorError('Unable to extract last version of the video')
98
99         renditions = xml.etree.ElementTree.fromstring(last_version['data'])
100         formats = []
101         # Already sorted from worst to best quality
102         for rend in renditions.findall('rendition'):
103             attr = rend.attrib
104             format_note = '%(videoCodec)s@%(videoBitrate)4sk, %(audioCodec)s@%(audioBitrate)3sk' % attr
105             formats.append({
106                 'url': attr['url'],
107                 'format_id': attr['name'],
108                 'format_note': format_note,
109                 'height': int(attr['frameheight']),
110                 'width': int(attr['frameWidth']),
111             })
112         return formats
113
114     def _formats_from_smil(self, smil_xml):
115         formats = []
116         smil_doc = xml.etree.ElementTree.fromstring(smil_xml.encode('utf-8'))
117         els = smil_doc.findall('.//{http://www.w3.org/2001/SMIL20/Language}video')
118         for el in els:
119             src = el.attrib['src']
120             m = re.match(r'''(?xi)
121                 (?P<ext>[a-z0-9]+):
122                 (?P<path>
123                     [/a-z0-9]+     # The directory and main part of the URL
124                     _(?P<cbr>[0-9]+)k
125                     _(?P<width>[0-9]+)x(?P<height>[0-9]+)
126                     _(?P<vcodec>[a-z0-9]+)
127                     _(?P<vbr>[0-9]+)
128                     _(?P<acodec>[a-z0-9]+)
129                     _(?P<abr>[0-9]+)
130                     \.[a-z0-9]+  # File extension
131                 )''', src)
132             if not m:
133                 continue
134
135             format_url = self._SMIL_BASE_URL + m.group('path')
136             formats.append({
137                 'url': format_url,
138                 'format_id': 'SMIL_' + m.group('cbr'),
139                 'vcodec': m.group('vcodec'),
140                 'acodec': m.group('acodec'),
141                 'vbr': int(m.group('vbr')),
142                 'abr': int(m.group('abr')),
143                 'ext': m.group('ext'),
144                 'width': int(m.group('width')),
145                 'height': int(m.group('height')),
146             })
147         return formats
148
149     def _download_api_formats(self, video_id):
150         if not self._oauth_token:
151             self._downloader.report_warning(
152                 'No oauth token available, skipping API HLS download')
153             return []
154
155         api_url = 'https://apiv2.vevo.com/video/%s/streams/hls?token=%s' % (
156             video_id, self._oauth_token)
157         api_data = self._download_json(
158             api_url, video_id,
159             note='Downloading HLS formats',
160             errnote='Failed to download HLS format list', fatal=False)
161         if api_data is None:
162             return []
163
164         m3u8_url = api_data[0]['url']
165         return self._extract_m3u8_formats(
166             m3u8_url, video_id, entry_protocol='m3u8_native', ext='mp4',
167             preference=0)
168
169     def _real_extract(self, url):
170         mobj = re.match(self._VALID_URL, url)
171         video_id = mobj.group('id')
172
173         json_url = 'http://videoplayer.vevo.com/VideoService/AuthenticateVideo?isrc=%s' % video_id
174         response = self._download_json(json_url, video_id)
175         video_info = response['video']
176
177         if not video_info:
178             if 'statusMessage' in response:
179                 raise ExtractorError('%s said: %s' % (self.IE_NAME, response['statusMessage']), expected=True)
180             raise ExtractorError('Unable to extract videos')
181
182         formats = self._formats_from_json(video_info)
183
184         is_explicit = video_info.get('isExplicit')
185         if is_explicit is True:
186             age_limit = 18
187         elif is_explicit is False:
188             age_limit = 0
189         else:
190             age_limit = None
191
192         # Download via HLS API
193         formats.extend(self._download_api_formats(video_id))
194
195         self._sort_formats(formats)
196         timestamp_ms = int(self._search_regex(
197             r'/Date\((\d+)\)/', video_info['launchDate'], 'launch date'))
198
199         return {
200             'id': video_id,
201             'title': video_info['title'],
202             'formats': formats,
203             'thumbnail': video_info['imageUrl'],
204             'timestamp': timestamp_ms // 1000,
205             'uploader': video_info['mainArtists'][0]['artistName'],
206             'duration': video_info['duration'],
207             'age_limit': age_limit,
208         }