[brightcove:new] update embed_in_page embeds regex to match non numeric ref id
[youtube-dl] / youtube_dl / extractor / brightcove.py
1 # encoding: utf-8
2 from __future__ import unicode_literals
3
4 import re
5 import json
6
7 from .common import InfoExtractor
8 from ..compat import (
9     compat_etree_fromstring,
10     compat_parse_qs,
11     compat_str,
12     compat_urllib_parse,
13     compat_urllib_parse_urlparse,
14     compat_urlparse,
15     compat_xml_parse_error,
16     compat_HTTPError,
17 )
18 from ..utils import (
19     determine_ext,
20     ExtractorError,
21     find_xpath_attr,
22     fix_xml_ampersands,
23     float_or_none,
24     js_to_json,
25     int_or_none,
26     parse_iso8601,
27     sanitized_Request,
28     unescapeHTML,
29     unsmuggle_url,
30 )
31
32
33 class BrightcoveLegacyIE(InfoExtractor):
34     IE_NAME = 'brightcove:legacy'
35     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
36     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
37
38     _TESTS = [
39         {
40             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
41             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
42             'md5': '5423e113865d26e40624dce2e4b45d95',
43             'note': 'Test Brightcove downloads and detection in GenericIE',
44             'info_dict': {
45                 'id': '2371591881001',
46                 'ext': 'mp4',
47                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
48                 'uploader': '8TV',
49                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
50             }
51         },
52         {
53             # From http://medianetwork.oracle.com/video/player/1785452137001
54             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
55             'info_dict': {
56                 'id': '1785452137001',
57                 'ext': 'flv',
58                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
59                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
60                 'uploader': 'Oracle',
61             },
62         },
63         {
64             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
65             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
66             'info_dict': {
67                 'id': '2750934548001',
68                 'ext': 'mp4',
69                 'title': 'This Bracelet Acts as a Personal Thermostat',
70                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
71                 'uploader': 'Mashable',
72             },
73         },
74         {
75             # test that the default referer works
76             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
77             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
78             'info_dict': {
79                 'id': '2878862109001',
80                 'ext': 'mp4',
81                 'title': 'Lost in Motion II',
82                 'description': 'md5:363109c02998fee92ec02211bd8000df',
83                 'uploader': 'National Ballet of Canada',
84             },
85         },
86         {
87             # test flv videos served by akamaihd.net
88             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
89             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?%40videoPlayer=ref%3ABC2996102916001&linkBaseURL=http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fvideos%2F1331655630249%2Freplay-uci-fort-william-2014-dh&playerKey=AQ%7E%7E%2CAAAApYJ7UqE%7E%2Cxqr_zXk0I-zzNndy8NlHogrCb5QdyZRf&playerID=1398061561001#__youtubedl_smuggle=%7B%22Referer%22%3A+%22http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fstories%2F1331655643987%2Freplay-uci-dh-world-cup-2014-from-fort-william%22%7D',
90             # The md5 checksum changes on each download
91             'info_dict': {
92                 'id': '2996102916001',
93                 'ext': 'flv',
94                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
95                 'uploader': 'Red Bull TV',
96                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
97             },
98         },
99         {
100             # playlist test
101             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
102             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
103             'info_dict': {
104                 'title': 'Sealife',
105                 'id': '3550319591001',
106             },
107             'playlist_mincount': 7,
108         },
109     ]
110
111     @classmethod
112     def _build_brighcove_url(cls, object_str):
113         """
114         Build a Brightcove url from a xml string containing
115         <object class="BrightcoveExperience">{params}</object>
116         """
117
118         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
119         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
120                             lambda m: m.group(1) + '/>', object_str)
121         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
122         object_str = object_str.replace('<--', '<!--')
123         # remove namespace to simplify extraction
124         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
125         object_str = fix_xml_ampersands(object_str)
126
127         try:
128             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
129         except compat_xml_parse_error:
130             return
131
132         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
133         if fv_el is not None:
134             flashvars = dict(
135                 (k, v[0])
136                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
137         else:
138             flashvars = {}
139
140         def find_param(name):
141             if name in flashvars:
142                 return flashvars[name]
143             node = find_xpath_attr(object_doc, './param', 'name', name)
144             if node is not None:
145                 return node.attrib['value']
146             return None
147
148         params = {}
149
150         playerID = find_param('playerID')
151         if playerID is None:
152             raise ExtractorError('Cannot find player ID')
153         params['playerID'] = playerID
154
155         playerKey = find_param('playerKey')
156         # Not all pages define this value
157         if playerKey is not None:
158             params['playerKey'] = playerKey
159         # The three fields hold the id of the video
160         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
161         if videoPlayer is not None:
162             params['@videoPlayer'] = videoPlayer
163         linkBase = find_param('linkBaseURL')
164         if linkBase is not None:
165             params['linkBaseURL'] = linkBase
166         return cls._make_brightcove_url(params)
167
168     @classmethod
169     def _build_brighcove_url_from_js(cls, object_js):
170         # The layout of JS is as follows:
171         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
172         #   // build Brightcove <object /> XML
173         # }
174         m = re.search(
175             r'''(?x)customBC.\createVideo\(
176                 .*?                                                  # skipping width and height
177                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
178                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
179                                                                      # in length, however it's appended to itself
180                                                                      # in places, so truncate
181                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
182             ''', object_js)
183         if m:
184             return cls._make_brightcove_url(m.groupdict())
185
186     @classmethod
187     def _make_brightcove_url(cls, params):
188         data = compat_urllib_parse.urlencode(params)
189         return cls._FEDERATED_URL_TEMPLATE % data
190
191     @classmethod
192     def _extract_brightcove_url(cls, webpage):
193         """Try to extract the brightcove url from the webpage, returns None
194         if it can't be found
195         """
196         urls = cls._extract_brightcove_urls(webpage)
197         return urls[0] if urls else None
198
199     @classmethod
200     def _extract_brightcove_urls(cls, webpage):
201         """Return a list of all Brightcove URLs from the webpage """
202
203         url_m = re.search(
204             r'<meta\s+property=[\'"]og:video[\'"]\s+content=[\'"](https?://(?:secure|c)\.brightcove.com/[^\'"]+)[\'"]',
205             webpage)
206         if url_m:
207             url = unescapeHTML(url_m.group(1))
208             # Some sites don't add it, we can't download with this url, for example:
209             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
210             if 'playerKey' in url or 'videoId' in url:
211                 return [url]
212
213         matches = re.findall(
214             r'''(?sx)<object
215             (?:
216                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
217                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
218             ).+?>\s*</object>''',
219             webpage)
220         if matches:
221             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
222
223         return list(filter(None, [
224             cls._build_brighcove_url_from_js(custom_bc)
225             for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))
226
227     def _real_extract(self, url):
228         url, smuggled_data = unsmuggle_url(url, {})
229
230         # Change the 'videoId' and others field to '@videoPlayer'
231         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
232         # Change bckey (used by bcove.me urls) to playerKey
233         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
234         mobj = re.match(self._VALID_URL, url)
235         query_str = mobj.group('query')
236         query = compat_urlparse.parse_qs(query_str)
237
238         videoPlayer = query.get('@videoPlayer')
239         if videoPlayer:
240             # We set the original url as the default 'Referer' header
241             referer = smuggled_data.get('Referer', url)
242             return self._get_video_info(
243                 videoPlayer[0], query_str, query, referer=referer)
244         elif 'playerKey' in query:
245             player_key = query['playerKey']
246             return self._get_playlist_info(player_key[0])
247         else:
248             raise ExtractorError(
249                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
250                 expected=True)
251
252     def _get_video_info(self, video_id, query_str, query, referer=None):
253         request_url = self._FEDERATED_URL_TEMPLATE % query_str
254         req = sanitized_Request(request_url)
255         linkBase = query.get('linkBaseURL')
256         if linkBase is not None:
257             referer = linkBase[0]
258         if referer is not None:
259             req.add_header('Referer', referer)
260         webpage = self._download_webpage(req, video_id)
261
262         error_msg = self._html_search_regex(
263             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
264             'error message', default=None)
265         if error_msg is not None:
266             raise ExtractorError(
267                 'brightcove said: %s' % error_msg, expected=True)
268
269         self.report_extraction(video_id)
270         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
271         info = json.loads(info)['data']
272         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
273         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
274
275         return self._extract_video_info(video_info)
276
277     def _get_playlist_info(self, player_key):
278         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
279         playlist_info = self._download_webpage(
280             info_url, player_key, 'Downloading playlist information')
281
282         json_data = json.loads(playlist_info)
283         if 'videoList' not in json_data:
284             raise ExtractorError('Empty playlist')
285         playlist_info = json_data['videoList']
286         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
287
288         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
289                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
290
291     def _extract_video_info(self, video_info):
292         info = {
293             'id': compat_str(video_info['id']),
294             'title': video_info['displayName'].strip(),
295             'description': video_info.get('shortDescription'),
296             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
297             'uploader': video_info.get('publisherName'),
298         }
299
300         renditions = video_info.get('renditions')
301         if renditions:
302             formats = []
303             for rend in renditions:
304                 url = rend['defaultURL']
305                 if not url:
306                     continue
307                 ext = None
308                 if rend['remote']:
309                     url_comp = compat_urllib_parse_urlparse(url)
310                     if url_comp.path.endswith('.m3u8'):
311                         formats.extend(
312                             self._extract_m3u8_formats(url, info['id'], 'mp4'))
313                         continue
314                     elif 'akamaihd.net' in url_comp.netloc:
315                         # This type of renditions are served through
316                         # akamaihd.net, but they don't use f4m manifests
317                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
318                         ext = 'flv'
319                 if ext is None:
320                     ext = determine_ext(url)
321                 size = rend.get('size')
322                 formats.append({
323                     'url': url,
324                     'ext': ext,
325                     'height': rend.get('frameHeight'),
326                     'width': rend.get('frameWidth'),
327                     'filesize': size if size != 0 else None,
328                 })
329             self._sort_formats(formats)
330             info['formats'] = formats
331         elif video_info.get('FLVFullLengthURL') is not None:
332             info.update({
333                 'url': video_info['FLVFullLengthURL'],
334             })
335
336         if self._downloader.params.get('include_ads', False):
337             adServerURL = video_info.get('_youtubedl_adServerURL')
338             if adServerURL:
339                 ad_info = {
340                     '_type': 'url',
341                     'url': adServerURL,
342                 }
343                 if 'url' in info:
344                     return {
345                         '_type': 'playlist',
346                         'title': info['title'],
347                         'entries': [ad_info, info],
348                     }
349                 else:
350                     return ad_info
351
352         if 'url' not in info and not info.get('formats'):
353             raise ExtractorError('Unable to extract video url for %s' % info['id'])
354         return info
355
356
357 class BrightcoveNewIE(InfoExtractor):
358     IE_NAME = 'brightcove:new'
359     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+|ref:[^&]+)'
360     _TESTS = [{
361         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
362         'md5': 'c8100925723840d4b0d243f7025703be',
363         'info_dict': {
364             'id': '4463358922001',
365             'ext': 'mp4',
366             'title': 'Meet the man behind Popcorn Time',
367             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
368             'duration': 165.768,
369             'timestamp': 1441391203,
370             'upload_date': '20150904',
371             'uploader_id': '929656772001',
372             'formats': 'mincount:22',
373         },
374     }, {
375         # with rtmp streams
376         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
377         'info_dict': {
378             'id': '4279049078001',
379             'ext': 'mp4',
380             'title': 'Titansgrave: Chapter 0',
381             'description': 'Titansgrave: Chapter 0',
382             'duration': 1242.058,
383             'timestamp': 1433556729,
384             'upload_date': '20150606',
385             'uploader_id': '4036320279001',
386             'formats': 'mincount:41',
387         },
388         'params': {
389             'skip_download': True,
390         }
391     }, {
392         # ref: prefixed video id
393         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
394         'only_matching': True,
395     }, {
396         # non numeric ref: prefixed video id
397         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
398         'only_matching': True,
399     }]
400
401     @staticmethod
402     def _extract_url(webpage):
403         urls = BrightcoveNewIE._extract_urls(webpage)
404         return urls[0] if urls else None
405
406     @staticmethod
407     def _extract_urls(webpage):
408         # Reference:
409         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
410         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
411         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/embed-in-page.html
412         # 4. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
413
414         entries = []
415
416         # Look for iframe embeds [1]
417         for _, url in re.findall(
418                 r'<iframe[^>]+src=(["\'])((?:https?:)//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
419             entries.append(url)
420
421         # Look for embed_in_page embeds [2]
422         for video_id, account_id, player_id, embed in re.findall(
423                 # According to examples from [3] it's unclear whether video id
424                 # may be optional and what to do when it is
425                 # According to [4] data-video-id may be prefixed with ref:
426                 r'''(?sx)
427                     <video[^>]+
428                         data-video-id=["\'](\d+|ref:[^"\']+)["\'][^>]*>.*?
429                     </video>.*?
430                     <script[^>]+
431                         src=["\'](?:https?:)?//players\.brightcove\.net/
432                         (\d+)/([\da-f-]+)_([^/]+)/index(?:\.min)?\.js
433                 ''', webpage):
434             entries.append(
435                 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
436                 % (account_id, player_id, embed, video_id))
437
438         return entries
439
440     def _real_extract(self, url):
441         account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()
442
443         webpage = self._download_webpage(
444             'http://players.brightcove.net/%s/%s_%s/index.min.js'
445             % (account_id, player_id, embed), video_id)
446
447         policy_key = None
448
449         catalog = self._search_regex(
450             r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
451         if catalog:
452             catalog = self._parse_json(
453                 js_to_json(catalog), video_id, fatal=False)
454             if catalog:
455                 policy_key = catalog.get('policyKey')
456
457         if not policy_key:
458             policy_key = self._search_regex(
459                 r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
460                 webpage, 'policy key', group='pk')
461
462         req = sanitized_Request(
463             'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s'
464             % (account_id, video_id),
465             headers={'Accept': 'application/json;pk=%s' % policy_key})
466         try:
467             json_data = self._download_json(req, video_id)
468         except ExtractorError as e:
469             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
470                 json_data = self._parse_json(e.cause.read().decode(), video_id)
471                 raise ExtractorError(json_data[0]['message'], expected=True)
472             raise
473
474         title = json_data['name']
475
476         formats = []
477         for source in json_data.get('sources', []):
478             container = source.get('container')
479             source_type = source.get('type')
480             src = source.get('src')
481             if source_type == 'application/x-mpegURL' or container == 'M2TS':
482                 if not src:
483                     continue
484                 formats.extend(self._extract_m3u8_formats(
485                     src, video_id, 'mp4', entry_protocol='m3u8_native',
486                     m3u8_id='hls', fatal=False))
487             elif source_type == 'application/dash+xml':
488                 if not src:
489                     continue
490                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
491             else:
492                 streaming_src = source.get('streaming_src')
493                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
494                 if not src and not streaming_src and (not stream_name or not app_name):
495                     continue
496                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
497                 height = int_or_none(source.get('height'))
498                 width = int_or_none(source.get('width'))
499                 f = {
500                     'tbr': tbr,
501                     'filesize': int_or_none(source.get('size')),
502                     'container': container,
503                     'ext': container.lower(),
504                 }
505                 if width == 0 and height == 0:
506                     f.update({
507                         'vcodec': 'none',
508                     })
509                 else:
510                     f.update({
511                         'width': width,
512                         'height': height,
513                         'vcodec': source.get('codec'),
514                     })
515
516                 def build_format_id(kind):
517                     format_id = kind
518                     if tbr:
519                         format_id += '-%dk' % int(tbr)
520                     if height:
521                         format_id += '-%dp' % height
522                     return format_id
523
524                 if src or streaming_src:
525                     f.update({
526                         'url': src or streaming_src,
527                         'format_id': build_format_id('http' if src else 'http-streaming'),
528                         'preference': 2 if src else 1,
529                     })
530                 else:
531                     f.update({
532                         'url': app_name,
533                         'play_path': stream_name,
534                         'format_id': build_format_id('rtmp'),
535                     })
536                 formats.append(f)
537         self._sort_formats(formats)
538
539         description = json_data.get('description')
540         thumbnail = json_data.get('thumbnail')
541         timestamp = parse_iso8601(json_data.get('published_at'))
542         duration = float_or_none(json_data.get('duration'), 1000)
543         tags = json_data.get('tags', [])
544
545         return {
546             'id': video_id,
547             'title': title,
548             'description': description,
549             'thumbnail': thumbnail,
550             'duration': duration,
551             'timestamp': timestamp,
552             'uploader_id': account_id,
553             'formats': formats,
554             'tags': tags,
555         }