You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

411 lines
17 KiB

10 years ago
10 years ago
  1. # coding: utf-8
  2. from __future__ import unicode_literals
  3. import re
  4. import time
  5. import hmac
  6. import binascii
  7. import hashlib
  8. from .once import OnceIE
  9. from .adobepass import AdobePassIE
  10. from ..compat import (
  11. compat_parse_qs,
  12. compat_urllib_parse_urlparse,
  13. )
  14. from ..utils import (
  15. determine_ext,
  16. ExtractorError,
  17. float_or_none,
  18. int_or_none,
  19. sanitized_Request,
  20. unsmuggle_url,
  21. update_url_query,
  22. xpath_with_ns,
  23. mimetype2ext,
  24. find_xpath_attr,
  25. )
  26. default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  27. _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  28. class ThePlatformBaseIE(OnceIE):
  29. _TP_TLD = 'com'
  30. def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  31. meta = self._download_xml(
  32. smil_url, video_id, note=note, query={'format': 'SMIL'},
  33. headers=self.geo_verification_headers())
  34. error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  35. if error_element is not None:
  36. exception = find_xpath_attr(
  37. error_element, _x('.//smil:param'), 'name', 'exception')
  38. if exception is not None:
  39. if exception.get('value') == 'GeoLocationBlocked':
  40. self.raise_geo_restricted(error_element.attrib['abstract'])
  41. elif error_element.attrib['src'].startswith(
  42. 'http://link.theplatform.%s/s/errorFiles/Unavailable.'
  43. % self._TP_TLD):
  44. raise ExtractorError(
  45. error_element.attrib['abstract'], expected=True)
  46. smil_formats = self._parse_smil_formats(
  47. meta, smil_url, video_id, namespace=default_ns,
  48. # the parameters are from syfy.com, other sites may use others,
  49. # they also work for nbc.com
  50. f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  51. transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  52. formats = []
  53. for _format in smil_formats:
  54. if OnceIE.suitable(_format['url']):
  55. formats.extend(self._extract_once_formats(_format['url']))
  56. else:
  57. media_url = _format['url']
  58. if determine_ext(media_url) == 'm3u8':
  59. hdnea2 = self._get_cookies(media_url).get('hdnea2')
  60. if hdnea2:
  61. _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  62. formats.append(_format)
  63. subtitles = self._parse_smil_subtitles(meta, default_ns)
  64. return formats, subtitles
  65. def _download_theplatform_metadata(self, path, video_id):
  66. info_url = 'http://link.theplatform.%s/s/%s?format=preview' % (self._TP_TLD, path)
  67. return self._download_json(info_url, video_id)
  68. def _parse_theplatform_metadata(self, info):
  69. subtitles = {}
  70. captions = info.get('captions')
  71. if isinstance(captions, list):
  72. for caption in captions:
  73. lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  74. subtitles.setdefault(lang, []).append({
  75. 'ext': mimetype2ext(mime),
  76. 'url': src,
  77. })
  78. duration = info.get('duration')
  79. tp_chapters = info.get('chapters', [])
  80. chapters = []
  81. if tp_chapters:
  82. def _add_chapter(start_time, end_time):
  83. start_time = float_or_none(start_time, 1000)
  84. end_time = float_or_none(end_time, 1000)
  85. if start_time is None or end_time is None:
  86. return
  87. chapters.append({
  88. 'start_time': start_time,
  89. 'end_time': end_time,
  90. })
  91. for chapter in tp_chapters[:-1]:
  92. _add_chapter(chapter.get('startTime'), chapter.get('endTime'))
  93. _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration)
  94. return {
  95. 'title': info['title'],
  96. 'subtitles': subtitles,
  97. 'description': info['description'],
  98. 'thumbnail': info['defaultThumbnailUrl'],
  99. 'duration': float_or_none(duration, 1000),
  100. 'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
  101. 'uploader': info.get('billingCode'),
  102. 'chapters': chapters,
  103. }
  104. def _extract_theplatform_metadata(self, path, video_id):
  105. info = self._download_theplatform_metadata(path, video_id)
  106. return self._parse_theplatform_metadata(info)
  107. class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
  108. _VALID_URL = r'''(?x)
  109. (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  110. (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
  111. |theplatform:)(?P<id>[^/\?&]+)'''
  112. _TESTS = [{
  113. # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  114. 'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  115. 'info_dict': {
  116. 'id': 'e9I_cZgTgIPd',
  117. 'ext': 'flv',
  118. 'title': 'Blackberry\'s big, bold Z30',
  119. 'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
  120. 'duration': 247,
  121. 'timestamp': 1383239700,
  122. 'upload_date': '20131031',
  123. 'uploader': 'CBSI-NEW',
  124. },
  125. 'params': {
  126. # rtmp download
  127. 'skip_download': True,
  128. },
  129. 'skip': '404 Not Found',
  130. }, {
  131. # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
  132. 'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
  133. 'info_dict': {
  134. 'id': '22d_qsQ6MIRT',
  135. 'ext': 'flv',
  136. 'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
  137. 'title': 'Tesla Model S: A second step towards a cleaner motoring future',
  138. 'timestamp': 1426176191,
  139. 'upload_date': '20150312',
  140. 'uploader': 'CBSI-NEW',
  141. },
  142. 'params': {
  143. # rtmp download
  144. 'skip_download': True,
  145. }
  146. }, {
  147. 'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
  148. 'info_dict': {
  149. 'id': 'yMBg9E8KFxZD',
  150. 'ext': 'mp4',
  151. 'description': 'md5:644ad9188d655b742f942bf2e06b002d',
  152. 'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
  153. 'uploader': 'EGSM',
  154. }
  155. }, {
  156. 'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
  157. 'only_matching': True,
  158. }, {
  159. 'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
  160. 'md5': 'fb96bb3d85118930a5b055783a3bd992',
  161. 'info_dict': {
  162. 'id': 'tdy_or_siri_150701',
  163. 'ext': 'mp4',
  164. 'title': 'iPhone Siri’s sassy response to a math question has people talking',
  165. 'description': 'md5:a565d1deadd5086f3331d57298ec6333',
  166. 'duration': 83.0,
  167. 'thumbnail': r're:^https?://.*\.jpg$',
  168. 'timestamp': 1435752600,
  169. 'upload_date': '20150701',
  170. 'uploader': 'NBCU-NEWS',
  171. },
  172. }, {
  173. # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
  174. # geo-restricted (US), HLS encrypted with AES-128
  175. 'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
  176. 'only_matching': True,
  177. }]
  178. @classmethod
  179. def _extract_urls(cls, webpage):
  180. m = re.search(
  181. r'''(?x)
  182. <meta\s+
  183. property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
  184. content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
  185. ''', webpage)
  186. if m:
  187. return [m.group('url')]
  188. # Are whitesapces ignored in URLs?
  189. # https://github.com/ytdl-org/youtube-dl/issues/12044
  190. matches = re.findall(
  191. r'(?s)<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
  192. if matches:
  193. return [re.sub(r'\s', '', list(zip(*matches))[1][0])]
  194. @staticmethod
  195. def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
  196. flags = '10' if include_qs else '00'
  197. expiration_date = '%x' % (int(time.time()) + life)
  198. def str_to_hex(str):
  199. return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
  200. def hex_to_bytes(hex):
  201. return binascii.a2b_hex(hex.encode('ascii'))
  202. relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1)
  203. clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
  204. checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
  205. sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
  206. return '%s&sig=%s' % (url, sig)
  207. def _real_extract(self, url):
  208. url, smuggled_data = unsmuggle_url(url, {})
  209. mobj = re.match(self._VALID_URL, url)
  210. provider_id = mobj.group('provider_id')
  211. video_id = mobj.group('id')
  212. if not provider_id:
  213. provider_id = 'dJ5BDC'
  214. path = provider_id + '/'
  215. if mobj.group('media'):
  216. path += mobj.group('media')
  217. path += video_id
  218. qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
  219. if 'guid' in qs_dict:
  220. webpage = self._download_webpage(url, video_id)
  221. scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
  222. feed_id = None
  223. # feed id usually locates in the last script.
  224. # Seems there's no pattern for the interested script filename, so
  225. # I try one by one
  226. for script in reversed(scripts):
  227. feed_script = self._download_webpage(
  228. self._proto_relative_url(script, 'http:'),
  229. video_id, 'Downloading feed script')
  230. feed_id = self._search_regex(
  231. r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
  232. 'default feed id', default=None)
  233. if feed_id is not None:
  234. break
  235. if feed_id is None:
  236. raise ExtractorError('Unable to find feed id')
  237. return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
  238. provider_id, feed_id, qs_dict['guid'][0]))
  239. if smuggled_data.get('force_smil_url', False):
  240. smil_url = url
  241. # Explicitly specified SMIL (see https://github.com/ytdl-org/youtube-dl/issues/7385)
  242. elif '/guid/' in url:
  243. headers = {}
  244. source_url = smuggled_data.get('source_url')
  245. if source_url:
  246. headers['Referer'] = source_url
  247. request = sanitized_Request(url, headers=headers)
  248. webpage = self._download_webpage(request, video_id)
  249. smil_url = self._search_regex(
  250. r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
  251. webpage, 'smil url', group='url')
  252. path = self._search_regex(
  253. r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
  254. smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
  255. elif mobj.group('config'):
  256. config_url = url + '&form=json'
  257. config_url = config_url.replace('swf/', 'config/')
  258. config_url = config_url.replace('onsite/', 'onsite/config/')
  259. config = self._download_json(config_url, video_id, 'Downloading config')
  260. if 'releaseUrl' in config:
  261. release_url = config['releaseUrl']
  262. else:
  263. release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
  264. smil_url = release_url + '&formats=MPEG4&manifest=f4m'
  265. else:
  266. smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
  267. sig = smuggled_data.get('sig')
  268. if sig:
  269. smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
  270. formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
  271. self._sort_formats(formats)
  272. ret = self._extract_theplatform_metadata(path, video_id)
  273. combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
  274. ret.update({
  275. 'id': video_id,
  276. 'formats': formats,
  277. 'subtitles': combined_subtitles,
  278. })
  279. return ret
  280. class ThePlatformFeedIE(ThePlatformBaseIE):
  281. _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
  282. _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[^&]+))'
  283. _TESTS = [{
  284. # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
  285. 'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
  286. 'md5': '6e32495b5073ab414471b615c5ded394',
  287. 'info_dict': {
  288. 'id': 'n_hardball_5biden_140207',
  289. 'ext': 'mp4',
  290. 'title': 'The Biden factor: will Joe run in 2016?',
  291. 'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
  292. 'thumbnail': r're:^https?://.*\.jpg$',
  293. 'upload_date': '20140208',
  294. 'timestamp': 1391824260,
  295. 'duration': 467.0,
  296. 'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
  297. 'uploader': 'NBCU-NEWS',
  298. },
  299. }, {
  300. 'url': 'http://feed.theplatform.com/f/2E2eJC/nnd_NBCNews?byGuid=nn_netcast_180306.Copy.01',
  301. 'only_matching': True,
  302. }]
  303. def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
  304. real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
  305. entry = self._download_json(real_url, video_id)['entries'][0]
  306. main_smil_url = 'http://link.theplatform.com/s/%s/media/guid/%d/%s' % (provider_id, account_id, entry['guid']) if account_id else entry.get('plmedia$publicUrl')
  307. formats = []
  308. subtitles = {}
  309. first_video_id = None
  310. duration = None
  311. asset_types = []
  312. for item in entry['media$content']:
  313. smil_url = item['plfile$url']
  314. cur_video_id = ThePlatformIE._match_id(smil_url)
  315. if first_video_id is None:
  316. first_video_id = cur_video_id
  317. duration = float_or_none(item.get('plfile$duration'))
  318. file_asset_types = item.get('plfile$assetTypes') or compat_parse_qs(compat_urllib_parse_urlparse(smil_url).query)['assetTypes']
  319. for asset_type in file_asset_types:
  320. if asset_type in asset_types:
  321. continue
  322. asset_types.append(asset_type)
  323. query = {
  324. 'mbr': 'true',
  325. 'formats': item['plfile$format'],
  326. 'assetTypes': asset_type,
  327. }
  328. if asset_type in asset_types_query:
  329. query.update(asset_types_query[asset_type])
  330. cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
  331. main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
  332. formats.extend(cur_formats)
  333. subtitles = self._merge_subtitles(subtitles, cur_subtitles)
  334. self._sort_formats(formats)
  335. thumbnails = [{
  336. 'url': thumbnail['plfile$url'],
  337. 'width': int_or_none(thumbnail.get('plfile$width')),
  338. 'height': int_or_none(thumbnail.get('plfile$height')),
  339. } for thumbnail in entry.get('media$thumbnails', [])]
  340. timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
  341. categories = [item['media$name'] for item in entry.get('media$categories', [])]
  342. ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
  343. subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
  344. ret.update({
  345. 'id': video_id,
  346. 'formats': formats,
  347. 'subtitles': subtitles,
  348. 'thumbnails': thumbnails,
  349. 'duration': duration,
  350. 'timestamp': timestamp,
  351. 'categories': categories,
  352. })
  353. if custom_fields:
  354. ret.update(custom_fields(entry))
  355. return ret
  356. def _real_extract(self, url):
  357. mobj = re.match(self._VALID_URL, url)
  358. video_id = mobj.group('id')
  359. provider_id = mobj.group('provider_id')
  360. feed_id = mobj.group('feed_id')
  361. filter_query = mobj.group('filter')
  362. return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)