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.

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