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.

318 lines
13 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 .common import InfoExtractor
  9. from ..compat import (
  10. compat_parse_qs,
  11. compat_urllib_parse_urlparse,
  12. )
  13. from ..utils import (
  14. determine_ext,
  15. ExtractorError,
  16. float_or_none,
  17. int_or_none,
  18. sanitized_Request,
  19. unsmuggle_url,
  20. url_basename,
  21. xpath_with_ns,
  22. )
  23. default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  24. _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  25. class ThePlatformBaseIE(InfoExtractor):
  26. def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  27. meta = self._download_xml(smil_url, video_id, note=note)
  28. try:
  29. error_msg = next(
  30. n.attrib['abstract']
  31. for n in meta.findall(_x('.//smil:ref'))
  32. if n.attrib.get('title') == 'Geographic Restriction' or n.attrib.get('title') == 'Expired')
  33. except StopIteration:
  34. pass
  35. else:
  36. raise ExtractorError(error_msg, expected=True)
  37. formats = self._parse_smil_formats(
  38. meta, smil_url, video_id, namespace=default_ns,
  39. # the parameters are from syfy.com, other sites may use others,
  40. # they also work for nbc.com
  41. f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  42. transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  43. for _format in formats:
  44. ext = determine_ext(_format['url'])
  45. if ext == 'once':
  46. _format['ext'] = 'mp4'
  47. self._sort_formats(formats)
  48. subtitles = self._parse_smil_subtitles(meta, default_ns)
  49. return formats, subtitles
  50. def get_metadata(self, path, video_id):
  51. info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
  52. info = self._download_json(info_url, video_id)
  53. subtitles = {}
  54. captions = info.get('captions')
  55. if isinstance(captions, list):
  56. for caption in captions:
  57. lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  58. subtitles[lang] = [{
  59. 'ext': 'srt' if mime == 'text/srt' else 'ttml',
  60. 'url': src,
  61. }]
  62. return {
  63. 'title': info['title'],
  64. 'subtitles': subtitles,
  65. 'description': info['description'],
  66. 'thumbnail': info['defaultThumbnailUrl'],
  67. 'duration': int_or_none(info.get('duration'), 1000),
  68. }
  69. class ThePlatformIE(ThePlatformBaseIE):
  70. _VALID_URL = r'''(?x)
  71. (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  72. (?:(?P<media>(?:[^/]+/)+select/media/)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
  73. |theplatform:)(?P<id>[^/\?&]+)'''
  74. _TESTS = [{
  75. # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  76. 'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  77. 'info_dict': {
  78. 'id': 'e9I_cZgTgIPd',
  79. 'ext': 'flv',
  80. 'title': 'Blackberry\'s big, bold Z30',
  81. 'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
  82. 'duration': 247,
  83. },
  84. 'params': {
  85. # rtmp download
  86. 'skip_download': True,
  87. },
  88. }, {
  89. # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
  90. 'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
  91. 'info_dict': {
  92. 'id': '22d_qsQ6MIRT',
  93. 'ext': 'flv',
  94. 'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
  95. 'title': 'Tesla Model S: A second step towards a cleaner motoring future',
  96. },
  97. 'params': {
  98. # rtmp download
  99. 'skip_download': True,
  100. }
  101. }, {
  102. 'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
  103. 'info_dict': {
  104. 'id': 'yMBg9E8KFxZD',
  105. 'ext': 'mp4',
  106. 'description': 'md5:644ad9188d655b742f942bf2e06b002d',
  107. 'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
  108. }
  109. }, {
  110. 'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
  111. 'only_matching': True,
  112. }, {
  113. 'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
  114. 'md5': '734f3790fb5fc4903da391beeebc4836',
  115. 'info_dict': {
  116. 'id': 'tdy_or_siri_150701',
  117. 'ext': 'mp4',
  118. 'title': 'iPhone Siri’s sassy response to a math question has people talking',
  119. 'description': 'md5:a565d1deadd5086f3331d57298ec6333',
  120. 'duration': 83.0,
  121. 'thumbnail': 're:^https?://.*\.jpg$',
  122. 'timestamp': 1435752600,
  123. 'upload_date': '20150701',
  124. 'categories': ['Today/Shows/Orange Room', 'Today/Sections/Money', 'Today/Topics/Tech', "Today/Topics/Editor's picks"],
  125. },
  126. }, {
  127. # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
  128. # geo-restricted (US), HLS encrypted with AES-128
  129. 'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
  130. 'only_matching': True,
  131. }]
  132. @staticmethod
  133. def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
  134. flags = '10' if include_qs else '00'
  135. expiration_date = '%x' % (int(time.time()) + life)
  136. def str_to_hex(str):
  137. return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
  138. def hex_to_str(hex):
  139. return binascii.a2b_hex(hex)
  140. relative_path = url.split('http://link.theplatform.com/s/')[1].split('?')[0]
  141. clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
  142. checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
  143. sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
  144. return '%s&sig=%s' % (url, sig)
  145. def _real_extract(self, url):
  146. url, smuggled_data = unsmuggle_url(url, {})
  147. mobj = re.match(self._VALID_URL, url)
  148. provider_id = mobj.group('provider_id')
  149. video_id = mobj.group('id')
  150. if not provider_id:
  151. provider_id = 'dJ5BDC'
  152. path = provider_id
  153. if mobj.group('media'):
  154. path += '/media'
  155. path += '/' + video_id
  156. qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
  157. if 'guid' in qs_dict:
  158. webpage = self._download_webpage(url, video_id)
  159. scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
  160. feed_id = None
  161. # feed id usually locates in the last script.
  162. # Seems there's no pattern for the interested script filename, so
  163. # I try one by one
  164. for script in reversed(scripts):
  165. feed_script = self._download_webpage(
  166. self._proto_relative_url(script, 'http:'),
  167. video_id, 'Downloading feed script')
  168. feed_id = self._search_regex(
  169. r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
  170. 'default feed id', default=None)
  171. if feed_id is not None:
  172. break
  173. if feed_id is None:
  174. raise ExtractorError('Unable to find feed id')
  175. return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
  176. provider_id, feed_id, qs_dict['guid'][0]))
  177. if smuggled_data.get('force_smil_url', False):
  178. smil_url = url
  179. # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
  180. elif '/guid/' in url:
  181. headers = {}
  182. source_url = smuggled_data.get('source_url')
  183. if source_url:
  184. headers['Referer'] = source_url
  185. request = sanitized_Request(url, headers=headers)
  186. webpage = self._download_webpage(request, video_id)
  187. smil_url = self._search_regex(
  188. r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
  189. webpage, 'smil url', group='url')
  190. path = self._search_regex(
  191. r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
  192. smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4&format=SMIL'
  193. elif mobj.group('config'):
  194. config_url = url + '&form=json'
  195. config_url = config_url.replace('swf/', 'config/')
  196. config_url = config_url.replace('onsite/', 'onsite/config/')
  197. config = self._download_json(config_url, video_id, 'Downloading config')
  198. if 'releaseUrl' in config:
  199. release_url = config['releaseUrl']
  200. else:
  201. release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
  202. smil_url = release_url + '&format=SMIL&formats=MPEG4&manifest=f4m'
  203. else:
  204. smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path
  205. sig = smuggled_data.get('sig')
  206. if sig:
  207. smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
  208. formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
  209. ret = self.get_metadata(path, video_id)
  210. combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
  211. ret.update({
  212. 'id': video_id,
  213. 'formats': formats,
  214. 'subtitles': combined_subtitles,
  215. })
  216. return ret
  217. class ThePlatformFeedIE(ThePlatformBaseIE):
  218. _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s'
  219. _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[a-zA-Z0-9_]+)'
  220. _TEST = {
  221. # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
  222. 'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
  223. 'md5': '22d2b84f058d3586efcd99e57d59d314',
  224. 'info_dict': {
  225. 'id': 'n_hardball_5biden_140207',
  226. 'ext': 'mp4',
  227. 'title': 'The Biden factor: will Joe run in 2016?',
  228. 'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
  229. 'thumbnail': 're:^https?://.*\.jpg$',
  230. 'upload_date': '20140208',
  231. 'timestamp': 1391824260,
  232. 'duration': 467.0,
  233. 'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
  234. },
  235. }
  236. def _real_extract(self, url):
  237. mobj = re.match(self._VALID_URL, url)
  238. video_id = mobj.group('id')
  239. provider_id = mobj.group('provider_id')
  240. feed_id = mobj.group('feed_id')
  241. real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
  242. feed = self._download_json(real_url, video_id)
  243. entry = feed['entries'][0]
  244. formats = []
  245. subtitles = {}
  246. first_video_id = None
  247. duration = None
  248. for item in entry['media$content']:
  249. smil_url = item['plfile$url'] + '&format=SMIL&Tracking=true&Embedded=true&formats=MPEG4,F4M'
  250. cur_video_id = url_basename(smil_url)
  251. if first_video_id is None:
  252. first_video_id = cur_video_id
  253. duration = float_or_none(item.get('plfile$duration'))
  254. cur_formats, cur_subtitles = self._extract_theplatform_smil(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id)
  255. formats.extend(cur_formats)
  256. subtitles = self._merge_subtitles(subtitles, cur_subtitles)
  257. self._sort_formats(formats)
  258. thumbnails = [{
  259. 'url': thumbnail['plfile$url'],
  260. 'width': int_or_none(thumbnail.get('plfile$width')),
  261. 'height': int_or_none(thumbnail.get('plfile$height')),
  262. } for thumbnail in entry.get('media$thumbnails', [])]
  263. timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
  264. categories = [item['media$name'] for item in entry.get('media$categories', [])]
  265. ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
  266. subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
  267. ret.update({
  268. 'id': video_id,
  269. 'formats': formats,
  270. 'subtitles': subtitles,
  271. 'thumbnails': thumbnails,
  272. 'duration': duration,
  273. 'timestamp': timestamp,
  274. 'categories': categories,
  275. })
  276. return ret