aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/extractor/brightcove.py
blob: c172bad2d1814937e0b488fc5f81c913aef60087 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
# encoding: utf-8
from __future__ import unicode_literals

import re
import json

from .common import InfoExtractor
from ..compat import (
    compat_etree_fromstring,
    compat_parse_qs,
    compat_str,
    compat_urllib_parse_urlparse,
    compat_urlparse,
    compat_xml_parse_error,
    compat_HTTPError,
)
from ..utils import (
    determine_ext,
    ExtractorError,
    find_xpath_attr,
    fix_xml_ampersands,
    float_or_none,
    js_to_json,
    int_or_none,
    parse_iso8601,
    unescapeHTML,
    unsmuggle_url,
    update_url_query,
    clean_html,
)


class BrightcoveLegacyIE(InfoExtractor):
    IE_NAME = 'brightcove:legacy'
    _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
    _FEDERATED_URL = 'http://c.brightcove.com/services/viewer/htmlFederated'

    _TESTS = [
        {
            # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
            'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
            'md5': '5423e113865d26e40624dce2e4b45d95',
            'note': 'Test Brightcove downloads and detection in GenericIE',
            'info_dict': {
                'id': '2371591881001',
                'ext': 'mp4',
                'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
                'uploader': '8TV',
                'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
                'timestamp': 1368213670,
                'upload_date': '20130510',
                'uploader_id': '1589608506001',
            }
        },
        {
            # From http://medianetwork.oracle.com/video/player/1785452137001
            'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
            'info_dict': {
                'id': '1785452137001',
                'ext': 'flv',
                'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
                'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
                'uploader': 'Oracle',
                'timestamp': 1344975024,
                'upload_date': '20120814',
                'uploader_id': '1460825906',
            },
        },
        {
            # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
            'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
            'info_dict': {
                'id': '2750934548001',
                'ext': 'mp4',
                'title': 'This Bracelet Acts as a Personal Thermostat',
                'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
                'uploader': 'Mashable',
                'timestamp': 1382041798,
                'upload_date': '20131017',
                'uploader_id': '1130468786001',
            },
        },
        {
            # test that the default referer works
            # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
            'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
            'info_dict': {
                'id': '2878862109001',
                'ext': 'mp4',
                'title': 'Lost in Motion II',
                'description': 'md5:363109c02998fee92ec02211bd8000df',
                'uploader': 'National Ballet of Canada',
            },
            'skip': 'Video gone',
        },
        {
            # test flv videos served by akamaihd.net
            # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
            'url': 'http://c.brightcove.com/services/viewer/htmlFederated?%40videoPlayer=ref%3Aevent-stream-356&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',
            # The md5 checksum changes on each download
            'info_dict': {
                'id': '3750436379001',
                'ext': 'flv',
                'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
                'uploader': 'RBTV Old (do not use)',
                'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
                'timestamp': 1409122195,
                'upload_date': '20140827',
                'uploader_id': '710858724001',
            },
        },
        {
            # playlist with 'videoList'
            # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
            'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
            'info_dict': {
                'title': 'Sealife',
                'id': '3550319591001',
            },
            'playlist_mincount': 7,
        },
        {
            # playlist with 'playlistTab' (https://github.com/rg3/youtube-dl/issues/9965)
            'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
            'info_dict': {
                'id': '1522758701001',
                'title': 'Lesson 08',
            },
            'playlist_mincount': 10,
        },
    ]
    FLV_VCODECS = {
        1: 'SORENSON',
        2: 'ON2',
        3: 'H264',
        4: 'VP8',
    }

    @classmethod
    def _build_brighcove_url(cls, object_str):
        """
        Build a Brightcove url from a xml string containing
        <object class="BrightcoveExperience">{params}</object>
        """

        # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
        object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
                            lambda m: m.group(1) + '/>', object_str)
        # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
        object_str = object_str.replace('<--', '<!--')
        # remove namespace to simplify extraction
        object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
        object_str = fix_xml_ampersands(object_str)

        try:
            object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
        except compat_xml_parse_error:
            return

        fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
        if fv_el is not None:
            flashvars = dict(
                (k, v[0])
                for k, v in compat_parse_qs(fv_el.attrib['value']).items())
        else:
            flashvars = {}

        data_url = object_doc.attrib.get('data', '')
        data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)

        def find_param(name):
            if name in flashvars:
                return flashvars[name]
            node = find_xpath_attr(object_doc, './param', 'name', name)
            if node is not None:
                return node.attrib['value']
            return data_url_params.get(name)

        params = {}

        playerID = find_param('playerID')
        if playerID is None:
            raise ExtractorError('Cannot find player ID')
        params['playerID'] = playerID

        playerKey = find_param('playerKey')
        # Not all pages define this value
        if playerKey is not None:
            params['playerKey'] = playerKey
        # These fields hold the id of the video
        videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
        if videoPlayer is not None:
            params['@videoPlayer'] = videoPlayer
        linkBase = find_param('linkBaseURL')
        if linkBase is not None:
            params['linkBaseURL'] = linkBase
        return cls._make_brightcove_url(params)

    @classmethod
    def _build_brighcove_url_from_js(cls, object_js):
        # The layout of JS is as follows:
        # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
        #   // build Brightcove <object /> XML
        # }
        m = re.search(
            r'''(?x)customBC.\createVideo\(
                .*?                                                  # skipping width and height
                ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
                ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
                                                                     # in length, however it's appended to itself
                                                                     # in places, so truncate
                ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
            ''', object_js)
        if m:
            return cls._make_brightcove_url(m.groupdict())

    @classmethod
    def _make_brightcove_url(cls, params):
        return update_url_query(cls._FEDERATED_URL, params)

    @classmethod
    def _extract_brightcove_url(cls, webpage):
        """Try to extract the brightcove url from the webpage, returns None
        if it can't be found
        """
        urls = cls._extract_brightcove_urls(webpage)
        return urls[0] if urls else None

    @classmethod
    def _extract_brightcove_urls(cls, webpage):
        """Return a list of all Brightcove URLs from the webpage """

        url_m = re.search(
            r'<meta\s+property=[\'"]og:video[\'"]\s+content=[\'"](https?://(?:secure|c)\.brightcove.com/[^\'"]+)[\'"]',
            webpage)
        if url_m:
            url = unescapeHTML(url_m.group(1))
            # Some sites don't add it, we can't download with this url, for example:
            # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
            if 'playerKey' in url or 'videoId' in url:
                return [url]

        matches = re.findall(
            r'''(?sx)<object
            (?:
                [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
                [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
            ).+?>\s*</object>''',
            webpage)
        if matches:
            return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))

        return list(filter(None, [
            cls._build_brighcove_url_from_js(custom_bc)
            for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))

    def _real_extract(self, url):
        url, smuggled_data = unsmuggle_url(url, {})

        # Change the 'videoId' and others field to '@videoPlayer'
        url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
        # Change bckey (used by bcove.me urls) to playerKey
        url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
        mobj = re.match(self._VALID_URL, url)
        query_str = mobj.group('query')
        query = compat_urlparse.parse_qs(query_str)

        videoPlayer = query.get('@videoPlayer')
        if videoPlayer:
            # We set the original url as the default 'Referer' header
            referer = smuggled_data.get('Referer', url)
            return self._get_video_info(
                videoPlayer[0], query, referer=referer)
        elif 'playerKey' in query:
            player_key = query['playerKey']
            return self._get_playlist_info(player_key[0])
        else:
            raise ExtractorError(
                'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
                expected=True)

    def _get_video_info(self, video_id, query, referer=None):
        headers = {}
        linkBase = query.get('linkBaseURL')
        if linkBase is not None:
            referer = linkBase[0]
        if referer is not None:
            headers['Referer'] = referer
        webpage = self._download_webpage(self._FEDERATED_URL, video_id, headers=headers, query=query)

        error_msg = self._html_search_regex(
            r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
            'error message', default=None)
        if error_msg is not None:
            raise ExtractorError(
                'brightcove said: %s' % error_msg, expected=True)

        self.report_extraction(video_id)
        info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
        info = json.loads(info)['data']
        video_info = info['programmedContent']['videoPlayer']['mediaDTO']
        video_info['_youtubedl_adServerURL'] = info.get('adServerURL')

        return self._extract_video_info(video_info)

    def _get_playlist_info(self, player_key):
        info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
        playlist_info = self._download_webpage(
            info_url, player_key, 'Downloading playlist information')

        json_data = json.loads(playlist_info)
        if 'videoList' in json_data:
            playlist_info = json_data['videoList']
            playlist_dto = playlist_info['mediaCollectionDTO']
        elif 'playlistTabs' in json_data:
            playlist_info = json_data['playlistTabs']
            playlist_dto = playlist_info['lineupListDTO']['playlistDTOs'][0]
        else:
            raise ExtractorError('Empty playlist')

        videos = [self._extract_video_info(video_info) for video_info in playlist_dto['videoDTOs']]

        return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
                                    playlist_title=playlist_dto['displayName'])

    def _extract_video_info(self, video_info):
        video_id = compat_str(video_info['id'])
        publisher_id = video_info.get('publisherId')
        info = {
            'id': video_id,
            'title': video_info['displayName'].strip(),
            'description': video_info.get('shortDescription'),
            'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
            'uploader': video_info.get('publisherName'),
            'uploader_id': compat_str(publisher_id) if publisher_id else None,
            'duration': float_or_none(video_info.get('length'), 1000),
            'timestamp': int_or_none(video_info.get('creationDate'), 1000),
        }

        renditions = video_info.get('renditions', []) + video_info.get('IOSRenditions', [])
        if renditions:
            formats = []
            for rend in renditions:
                url = rend['defaultURL']
                if not url:
                    continue
                ext = None
                if rend['remote']:
                    url_comp = compat_urllib_parse_urlparse(url)
                    if url_comp.path.endswith('.m3u8'):
                        formats.extend(
                            self._extract_m3u8_formats(
                                url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
                        continue
                    elif 'akamaihd.net' in url_comp.netloc:
                        # This type of renditions are served through
                        # akamaihd.net, but they don't use f4m manifests
                        url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
                        ext = 'flv'
                if ext is None:
                    ext = determine_ext(url)
                tbr = int_or_none(rend.get('encodingRate'), 1000)
                a_format = {
                    'format_id': 'http%s' % ('-%s' % tbr if tbr else ''),
                    'url': url,
                    'ext': ext,
                    'filesize': int_or_none(rend.get('size')) or None,
                    'tbr': tbr,
                }
                if rend.get('audioOnly'):
                    a_format.update({
                        'vcodec': 'none',
                    })
                else:
                    a_format.update({
                        'height': int_or_none(rend.get('frameHeight')),
                        'width': int_or_none(rend.get('frameWidth')),
                        'vcodec': rend.get('videoCodec'),
                    })

                # m3u8 manifests with remote == false are media playlists
                # Not calling _extract_m3u8_formats here to save network traffic
                if ext == 'm3u8':
                    a_format.update({
                        'format_id': 'hls%s' % ('-%s' % tbr if tbr else ''),
                        'ext': 'mp4',
                        'protocol': 'm3u8_native',
                    })

                formats.append(a_format)
            self._sort_formats(formats)
            info['formats'] = formats
        elif video_info.get('FLVFullLengthURL') is not None:
            info.update({
                'url': video_info['FLVFullLengthURL'],
                'vcodec': self.FLV_VCODECS.get(video_info.get('FLVFullCodec')),
                'filesize': int_or_none(video_info.get('FLVFullSize')),
            })

        if self._downloader.params.get('include_ads', False):
            adServerURL = video_info.get('_youtubedl_adServerURL')
            if adServerURL:
                ad_info = {
                    '_type': 'url',
                    'url': adServerURL,
                }
                if 'url' in info:
                    return {
                        '_type': 'playlist',
                        'title': info['title'],
                        'entries': [ad_info, info],
                    }
                else:
                    return ad_info

        if 'url' not in info and not info.get('formats'):
            raise ExtractorError('Unable to extract video url for %s' % video_id)
        return info


class BrightcoveNewIE(InfoExtractor):
    IE_NAME = 'brightcove:new'
    _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+|ref:[^&]+)'
    _TESTS = [{
        'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
        'md5': 'c8100925723840d4b0d243f7025703be',
        'info_dict': {
            'id': '4463358922001',
            'ext': 'mp4',
            'title': 'Meet the man behind Popcorn Time',
            'description': 'md5:eac376a4fe366edc70279bfb681aea16',
            'duration': 165.768,
            'timestamp': 1441391203,
            'upload_date': '20150904',
            'uploader_id': '929656772001',
            'formats': 'mincount:22',
        },
    }, {
        # with rtmp streams
        'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
        'info_dict': {
            'id': '4279049078001',
            'ext': 'mp4',
            'title': 'Titansgrave: Chapter 0',
            'description': 'Titansgrave: Chapter 0',
            'duration': 1242.058,
            'timestamp': 1433556729,
            'upload_date': '20150606',
            'uploader_id': '4036320279001',
            'formats': 'mincount:41',
        },
        'params': {
            # m3u8 download
            'skip_download': True,
        }
    }, {
        # ref: prefixed video id
        'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
        'only_matching': True,
    }, {
        # non numeric ref: prefixed video id
        'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
        'only_matching': True,
    }, {
        # unavailable video without message but with error_code
        'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
        'only_matching': True,
    }]

    @staticmethod
    def _extract_url(webpage):
        urls = BrightcoveNewIE._extract_urls(webpage)
        return urls[0] if urls else None

    @staticmethod
    def _extract_urls(webpage):
        # Reference:
        # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
        # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
        # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/embed-in-page.html
        # 4. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player

        entries = []

        # Look for iframe embeds [1]
        for _, url in re.findall(
                r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
            entries.append(url if url.startswith('http') else 'http:' + url)

        # Look for embed_in_page embeds [2]
        for video_id, account_id, player_id, embed in re.findall(
                # According to examples from [3] it's unclear whether video id
                # may be optional and what to do when it is
                # According to [4] data-video-id may be prefixed with ref:
                r'''(?sx)
                    <video[^>]+
                        data-video-id=["\'](\d+|ref:[^"\']+)["\'][^>]*>.*?
                    </video>.*?
                    <script[^>]+
                        src=["\'](?:https?:)?//players\.brightcove\.net/
                        (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
                ''', webpage):
            entries.append(
                'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
                % (account_id, player_id, embed, video_id))

        return entries

    def _real_extract(self, url):
        account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()

        webpage = self._download_webpage(
            'http://players.brightcove.net/%s/%s_%s/index.min.js'
            % (account_id, player_id, embed), video_id)

        policy_key = None

        catalog = self._search_regex(
            r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
        if catalog:
            catalog = self._parse_json(
                js_to_json(catalog), video_id, fatal=False)
            if catalog:
                policy_key = catalog.get('policyKey')

        if not policy_key:
            policy_key = self._search_regex(
                r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
                webpage, 'policy key', group='pk')

        api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s' % (account_id, video_id)
        try:
            json_data = self._download_json(api_url, video_id, headers={
                'Accept': 'application/json;pk=%s' % policy_key
            })
        except ExtractorError as e:
            if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
                json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
                raise ExtractorError(
                    json_data.get('message') or json_data['error_code'], expected=True)
            raise

        title = json_data['name'].strip()

        formats = []
        for source in json_data.get('sources', []):
            container = source.get('container')
            source_type = source.get('type')
            src = source.get('src')
            if source_type == 'application/x-mpegURL' or container == 'M2TS':
                if not src:
                    continue
                formats.extend(self._extract_m3u8_formats(
                    src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
            elif source_type == 'application/dash+xml':
                if not src:
                    continue
                formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
            else:
                streaming_src = source.get('streaming_src')
                stream_name, app_name = source.get('stream_name'), source.get('app_name')
                if not src and not streaming_src and (not stream_name or not app_name):
                    continue
                tbr = float_or_none(source.get('avg_bitrate'), 1000)
                height = int_or_none(source.get('height'))
                width = int_or_none(source.get('width'))
                f = {
                    'tbr': tbr,
                    'filesize': int_or_none(source.get('size')),
                    'container': container,
                    'ext': container.lower(),
                }
                if width == 0 and height == 0:
                    f.update({
                        'vcodec': 'none',
                    })
                else:
                    f.update({
                        'width': width,
                        'height': height,
                        'vcodec': source.get('codec'),
                    })

                def build_format_id(kind):
                    format_id = kind
                    if tbr:
                        format_id += '-%dk' % int(tbr)
                    if height:
                        format_id += '-%dp' % height
                    return format_id

                if src or streaming_src:
                    f.update({
                        'url': src or streaming_src,
                        'format_id': build_format_id('http' if src else 'http-streaming'),
                        'source_preference': 0 if src else -1,
                    })
                else:
                    f.update({
                        'url': app_name,
                        'play_path': stream_name,
                        'format_id': build_format_id('rtmp'),
                    })
                formats.append(f)

        errors = json_data.get('errors')
        if not formats and errors:
            error = errors[0]
            raise ExtractorError(
                error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)

        self._sort_formats(formats)

        subtitles = {}
        for text_track in json_data.get('text_tracks', []):
            if text_track.get('src'):
                subtitles.setdefault(text_track.get('srclang'), []).append({
                    'url': text_track['src'],
                })

        return {
            'id': video_id,
            'title': title,
            'description': clean_html(json_data.get('description')),
            'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
            'duration': float_or_none(json_data.get('duration'), 1000),
            'timestamp': parse_iso8601(json_data.get('published_at')),
            'uploader_id': account_id,
            'formats': formats,
            'subtitles': subtitles,
            'tags': json_data.get('tags', []),
        }