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.

573 lines
23 KiB

11 years ago
11 years ago
  1. import base64
  2. import hashlib
  3. import json
  4. import os
  5. import re
  6. import socket
  7. import sys
  8. import netrc
  9. import xml.etree.ElementTree
  10. from ..utils import (
  11. compat_http_client,
  12. compat_urllib_error,
  13. compat_urllib_parse_urlparse,
  14. compat_str,
  15. clean_html,
  16. compiled_regex_type,
  17. ExtractorError,
  18. RegexNotFoundError,
  19. sanitize_filename,
  20. unescapeHTML,
  21. )
  22. _NO_DEFAULT = object()
  23. class InfoExtractor(object):
  24. """Information Extractor class.
  25. Information extractors are the classes that, given a URL, extract
  26. information about the video (or videos) the URL refers to. This
  27. information includes the real video URL, the video title, author and
  28. others. The information is stored in a dictionary which is then
  29. passed to the FileDownloader. The FileDownloader processes this
  30. information possibly downloading the video to the file system, among
  31. other possible outcomes.
  32. The dictionaries must include the following fields:
  33. id: Video identifier.
  34. title: Video title, unescaped.
  35. Additionally, it must contain either a formats entry or a url one:
  36. formats: A list of dictionaries for each format available, ordered
  37. from worst to best quality.
  38. Potential fields:
  39. * url Mandatory. The URL of the video file
  40. * ext Will be calculated from url if missing
  41. * format A human-readable description of the format
  42. ("mp4 container with h264/opus").
  43. Calculated from the format_id, width, height.
  44. and format_note fields if missing.
  45. * format_id A short description of the format
  46. ("mp4_h264_opus" or "19").
  47. Technically optional, but strongly recommended.
  48. * format_note Additional info about the format
  49. ("3D" or "DASH video")
  50. * width Width of the video, if known
  51. * height Height of the video, if known
  52. * resolution Textual description of width and height
  53. * tbr Average bitrate of audio and video in KBit/s
  54. * abr Average audio bitrate in KBit/s
  55. * acodec Name of the audio codec in use
  56. * asr Audio sampling rate in Hertz
  57. * vbr Average video bitrate in KBit/s
  58. * vcodec Name of the video codec in use
  59. * container Name of the container format
  60. * filesize The number of bytes, if known in advance
  61. * player_url SWF Player URL (used for rtmpdump).
  62. * protocol The protocol that will be used for the actual
  63. download, lower-case.
  64. "http", "https", "rtsp", "rtmp", "m3u8" or so.
  65. * preference Order number of this format. If this field is
  66. present and not None, the formats get sorted
  67. by this field.
  68. -1 for default (order by other properties),
  69. -2 or smaller for less than default.
  70. * quality Order number of the video quality of this
  71. format, irrespective of the file format.
  72. -1 for default (order by other properties),
  73. -2 or smaller for less than default.
  74. url: Final video URL.
  75. ext: Video filename extension.
  76. format: The video format, defaults to ext (used for --get-format)
  77. player_url: SWF Player URL (used for rtmpdump).
  78. The following fields are optional:
  79. thumbnails: A list of dictionaries (with the entries "resolution" and
  80. "url") for the varying thumbnails
  81. thumbnail: Full URL to a video thumbnail image.
  82. description: One-line video description.
  83. uploader: Full name of the video uploader.
  84. upload_date: Video upload date (YYYYMMDD).
  85. uploader_id: Nickname or id of the video uploader.
  86. location: Physical location of the video.
  87. subtitles: The subtitle file contents as a dictionary in the format
  88. {language: subtitles}.
  89. duration: Length of the video in seconds, as an integer.
  90. view_count: How many users have watched the video on the platform.
  91. like_count: Number of positive ratings of the video
  92. dislike_count: Number of negative ratings of the video
  93. comment_count: Number of comments on the video
  94. age_limit: Age restriction for the video, as an integer (years)
  95. webpage_url: The url to the video webpage, if given to youtube-dl it
  96. should allow to get the same result again. (It will be set
  97. by YoutubeDL if it's missing)
  98. Unless mentioned otherwise, the fields should be Unicode strings.
  99. Subclasses of this one should re-define the _real_initialize() and
  100. _real_extract() methods and define a _VALID_URL regexp.
  101. Probably, they should also be added to the list of extractors.
  102. _real_extract() must return a *list* of information dictionaries as
  103. described above.
  104. Finally, the _WORKING attribute should be set to False for broken IEs
  105. in order to warn the users and skip the tests.
  106. """
  107. _ready = False
  108. _downloader = None
  109. _WORKING = True
  110. def __init__(self, downloader=None):
  111. """Constructor. Receives an optional downloader."""
  112. self._ready = False
  113. self.set_downloader(downloader)
  114. @classmethod
  115. def suitable(cls, url):
  116. """Receives a URL and returns True if suitable for this IE."""
  117. # This does not use has/getattr intentionally - we want to know whether
  118. # we have cached the regexp for *this* class, whereas getattr would also
  119. # match the superclass
  120. if '_VALID_URL_RE' not in cls.__dict__:
  121. cls._VALID_URL_RE = re.compile(cls._VALID_URL)
  122. return cls._VALID_URL_RE.match(url) is not None
  123. @classmethod
  124. def working(cls):
  125. """Getter method for _WORKING."""
  126. return cls._WORKING
  127. def initialize(self):
  128. """Initializes an instance (authentication, etc)."""
  129. if not self._ready:
  130. self._real_initialize()
  131. self._ready = True
  132. def extract(self, url):
  133. """Extracts URL information and returns it in list of dicts."""
  134. self.initialize()
  135. return self._real_extract(url)
  136. def set_downloader(self, downloader):
  137. """Sets the downloader for this IE."""
  138. self._downloader = downloader
  139. def _real_initialize(self):
  140. """Real initialization process. Redefine in subclasses."""
  141. pass
  142. def _real_extract(self, url):
  143. """Real extraction process. Redefine in subclasses."""
  144. pass
  145. @classmethod
  146. def ie_key(cls):
  147. """A string for getting the InfoExtractor with get_info_extractor"""
  148. return cls.__name__[:-2]
  149. @property
  150. def IE_NAME(self):
  151. return type(self).__name__[:-2]
  152. def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
  153. """ Returns the response handle """
  154. if note is None:
  155. self.report_download_webpage(video_id)
  156. elif note is not False:
  157. if video_id is None:
  158. self.to_screen(u'%s' % (note,))
  159. else:
  160. self.to_screen(u'%s: %s' % (video_id, note))
  161. try:
  162. return self._downloader.urlopen(url_or_request)
  163. except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
  164. if errnote is False:
  165. return False
  166. if errnote is None:
  167. errnote = u'Unable to download webpage'
  168. errmsg = u'%s: %s' % (errnote, compat_str(err))
  169. if fatal:
  170. raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
  171. else:
  172. self._downloader.report_warning(errmsg)
  173. return False
  174. def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
  175. """ Returns a tuple (page content as string, URL handle) """
  176. # Strip hashes from the URL (#1038)
  177. if isinstance(url_or_request, (compat_str, str)):
  178. url_or_request = url_or_request.partition('#')[0]
  179. urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
  180. if urlh is False:
  181. assert not fatal
  182. return False
  183. content_type = urlh.headers.get('Content-Type', '')
  184. webpage_bytes = urlh.read()
  185. m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
  186. if m:
  187. encoding = m.group(1)
  188. else:
  189. m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
  190. webpage_bytes[:1024])
  191. if m:
  192. encoding = m.group(1).decode('ascii')
  193. elif webpage_bytes.startswith(b'\xff\xfe'):
  194. encoding = 'utf-16'
  195. else:
  196. encoding = 'utf-8'
  197. if self._downloader.params.get('dump_intermediate_pages', False):
  198. try:
  199. url = url_or_request.get_full_url()
  200. except AttributeError:
  201. url = url_or_request
  202. self.to_screen(u'Dumping request to ' + url)
  203. dump = base64.b64encode(webpage_bytes).decode('ascii')
  204. self._downloader.to_screen(dump)
  205. if self._downloader.params.get('write_pages', False):
  206. try:
  207. url = url_or_request.get_full_url()
  208. except AttributeError:
  209. url = url_or_request
  210. if len(url) > 200:
  211. h = u'___' + hashlib.md5(url.encode('utf-8')).hexdigest()
  212. url = url[:200 - len(h)] + h
  213. raw_filename = ('%s_%s.dump' % (video_id, url))
  214. filename = sanitize_filename(raw_filename, restricted=True)
  215. self.to_screen(u'Saving request to ' + filename)
  216. with open(filename, 'wb') as outf:
  217. outf.write(webpage_bytes)
  218. content = webpage_bytes.decode(encoding, 'replace')
  219. return (content, urlh)
  220. def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
  221. """ Returns the data of the page as a string """
  222. res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
  223. if res is False:
  224. return res
  225. else:
  226. content, _ = res
  227. return content
  228. def _download_xml(self, url_or_request, video_id,
  229. note=u'Downloading XML', errnote=u'Unable to download XML',
  230. transform_source=None):
  231. """Return the xml as an xml.etree.ElementTree.Element"""
  232. xml_string = self._download_webpage(url_or_request, video_id, note, errnote)
  233. if transform_source:
  234. xml_string = transform_source(xml_string)
  235. return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
  236. def _download_json(self, url_or_request, video_id,
  237. note=u'Downloading JSON metadata',
  238. errnote=u'Unable to download JSON metadata',
  239. transform_source=None):
  240. json_string = self._download_webpage(url_or_request, video_id, note, errnote)
  241. if transform_source:
  242. json_string = transform_source(json_string)
  243. try:
  244. return json.loads(json_string)
  245. except ValueError as ve:
  246. raise ExtractorError('Failed to download JSON', cause=ve)
  247. def report_warning(self, msg, video_id=None):
  248. idstr = u'' if video_id is None else u'%s: ' % video_id
  249. self._downloader.report_warning(
  250. u'[%s] %s%s' % (self.IE_NAME, idstr, msg))
  251. def to_screen(self, msg):
  252. """Print msg to screen, prefixing it with '[ie_name]'"""
  253. self._downloader.to_screen(u'[%s] %s' % (self.IE_NAME, msg))
  254. def report_extraction(self, id_or_name):
  255. """Report information extraction."""
  256. self.to_screen(u'%s: Extracting information' % id_or_name)
  257. def report_download_webpage(self, video_id):
  258. """Report webpage download."""
  259. self.to_screen(u'%s: Downloading webpage' % video_id)
  260. def report_age_confirmation(self):
  261. """Report attempt to confirm age."""
  262. self.to_screen(u'Confirming age')
  263. def report_login(self):
  264. """Report attempt to log in."""
  265. self.to_screen(u'Logging in')
  266. #Methods for following #608
  267. @staticmethod
  268. def url_result(url, ie=None, video_id=None):
  269. """Returns a url that points to a page that should be processed"""
  270. #TODO: ie should be the class used for getting the info
  271. video_info = {'_type': 'url',
  272. 'url': url,
  273. 'ie_key': ie}
  274. if video_id is not None:
  275. video_info['id'] = video_id
  276. return video_info
  277. @staticmethod
  278. def playlist_result(entries, playlist_id=None, playlist_title=None):
  279. """Returns a playlist"""
  280. video_info = {'_type': 'playlist',
  281. 'entries': entries}
  282. if playlist_id:
  283. video_info['id'] = playlist_id
  284. if playlist_title:
  285. video_info['title'] = playlist_title
  286. return video_info
  287. def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0):
  288. """
  289. Perform a regex search on the given string, using a single or a list of
  290. patterns returning the first matching group.
  291. In case of failure return a default value or raise a WARNING or a
  292. RegexNotFoundError, depending on fatal, specifying the field name.
  293. """
  294. if isinstance(pattern, (str, compat_str, compiled_regex_type)):
  295. mobj = re.search(pattern, string, flags)
  296. else:
  297. for p in pattern:
  298. mobj = re.search(p, string, flags)
  299. if mobj: break
  300. if os.name != 'nt' and sys.stderr.isatty():
  301. _name = u'\033[0;34m%s\033[0m' % name
  302. else:
  303. _name = name
  304. if mobj:
  305. # return the first matching group
  306. return next(g for g in mobj.groups() if g is not None)
  307. elif default is not _NO_DEFAULT:
  308. return default
  309. elif fatal:
  310. raise RegexNotFoundError(u'Unable to extract %s' % _name)
  311. else:
  312. self._downloader.report_warning(u'unable to extract %s; '
  313. u'please report this issue on http://yt-dl.org/bug' % _name)
  314. return None
  315. def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0):
  316. """
  317. Like _search_regex, but strips HTML tags and unescapes entities.
  318. """
  319. res = self._search_regex(pattern, string, name, default, fatal, flags)
  320. if res:
  321. return clean_html(res).strip()
  322. else:
  323. return res
  324. def _get_login_info(self):
  325. """
  326. Get the the login info as (username, password)
  327. It will look in the netrc file using the _NETRC_MACHINE value
  328. If there's no info available, return (None, None)
  329. """
  330. if self._downloader is None:
  331. return (None, None)
  332. username = None
  333. password = None
  334. downloader_params = self._downloader.params
  335. # Attempt to use provided username and password or .netrc data
  336. if downloader_params.get('username', None) is not None:
  337. username = downloader_params['username']
  338. password = downloader_params['password']
  339. elif downloader_params.get('usenetrc', False):
  340. try:
  341. info = netrc.netrc().authenticators(self._NETRC_MACHINE)
  342. if info is not None:
  343. username = info[0]
  344. password = info[2]
  345. else:
  346. raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
  347. except (IOError, netrc.NetrcParseError) as err:
  348. self._downloader.report_warning(u'parsing .netrc: %s' % compat_str(err))
  349. return (username, password)
  350. # Helper functions for extracting OpenGraph info
  351. @staticmethod
  352. def _og_regexes(prop):
  353. content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
  354. property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
  355. template = r'<meta[^>]+?%s[^>]+?%s'
  356. return [
  357. template % (property_re, content_re),
  358. template % (content_re, property_re),
  359. ]
  360. def _og_search_property(self, prop, html, name=None, **kargs):
  361. if name is None:
  362. name = 'OpenGraph %s' % prop
  363. escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
  364. if escaped is None:
  365. return None
  366. return unescapeHTML(escaped)
  367. def _og_search_thumbnail(self, html, **kargs):
  368. return self._og_search_property('image', html, u'thumbnail url', fatal=False, **kargs)
  369. def _og_search_description(self, html, **kargs):
  370. return self._og_search_property('description', html, fatal=False, **kargs)
  371. def _og_search_title(self, html, **kargs):
  372. return self._og_search_property('title', html, **kargs)
  373. def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
  374. regexes = self._og_regexes('video')
  375. if secure: regexes = self._og_regexes('video:secure_url') + regexes
  376. return self._html_search_regex(regexes, html, name, **kargs)
  377. def _html_search_meta(self, name, html, display_name=None):
  378. if display_name is None:
  379. display_name = name
  380. return self._html_search_regex(
  381. r'''(?ix)<meta
  382. (?=[^>]+(?:itemprop|name|property)=["\']%s["\'])
  383. [^>]+content=["\']([^"\']+)["\']''' % re.escape(name),
  384. html, display_name, fatal=False)
  385. def _dc_search_uploader(self, html):
  386. return self._html_search_meta('dc.creator', html, 'uploader')
  387. def _rta_search(self, html):
  388. # See http://www.rtalabel.org/index.php?content=howtofaq#single
  389. if re.search(r'(?ix)<meta\s+name="rating"\s+'
  390. r' content="RTA-5042-1996-1400-1577-RTA"',
  391. html):
  392. return 18
  393. return 0
  394. def _media_rating_search(self, html):
  395. # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
  396. rating = self._html_search_meta('rating', html)
  397. if not rating:
  398. return None
  399. RATING_TABLE = {
  400. 'safe for kids': 0,
  401. 'general': 8,
  402. '14 years': 14,
  403. 'mature': 17,
  404. 'restricted': 19,
  405. }
  406. return RATING_TABLE.get(rating.lower(), None)
  407. def _twitter_search_player(self, html):
  408. return self._html_search_meta('twitter:player', html,
  409. 'twitter card player')
  410. def _sort_formats(self, formats):
  411. if not formats:
  412. raise ExtractorError(u'No video formats found')
  413. def _formats_key(f):
  414. # TODO remove the following workaround
  415. from ..utils import determine_ext
  416. if not f.get('ext') and 'url' in f:
  417. f['ext'] = determine_ext(f['url'])
  418. preference = f.get('preference')
  419. if preference is None:
  420. proto = f.get('protocol')
  421. if proto is None:
  422. proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
  423. preference = 0 if proto in ['http', 'https'] else -0.1
  424. if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
  425. preference -= 0.5
  426. if f.get('vcodec') == 'none': # audio only
  427. if self._downloader.params.get('prefer_free_formats'):
  428. ORDER = [u'aac', u'mp3', u'm4a', u'webm', u'ogg', u'opus']
  429. else:
  430. ORDER = [u'webm', u'opus', u'ogg', u'mp3', u'aac', u'm4a']
  431. ext_preference = 0
  432. try:
  433. audio_ext_preference = ORDER.index(f['ext'])
  434. except ValueError:
  435. audio_ext_preference = -1
  436. else:
  437. if self._downloader.params.get('prefer_free_formats'):
  438. ORDER = [u'flv', u'mp4', u'webm']
  439. else:
  440. ORDER = [u'webm', u'flv', u'mp4']
  441. try:
  442. ext_preference = ORDER.index(f['ext'])
  443. except ValueError:
  444. ext_preference = -1
  445. audio_ext_preference = 0
  446. return (
  447. preference,
  448. f.get('quality') if f.get('quality') is not None else -1,
  449. f.get('height') if f.get('height') is not None else -1,
  450. f.get('width') if f.get('width') is not None else -1,
  451. ext_preference,
  452. f.get('tbr') if f.get('tbr') is not None else -1,
  453. f.get('vbr') if f.get('vbr') is not None else -1,
  454. f.get('abr') if f.get('abr') is not None else -1,
  455. audio_ext_preference,
  456. f.get('filesize') if f.get('filesize') is not None else -1,
  457. f.get('format_id'),
  458. )
  459. formats.sort(key=_formats_key)
  460. class SearchInfoExtractor(InfoExtractor):
  461. """
  462. Base class for paged search queries extractors.
  463. They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
  464. Instances should define _SEARCH_KEY and _MAX_RESULTS.
  465. """
  466. @classmethod
  467. def _make_valid_url(cls):
  468. return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
  469. @classmethod
  470. def suitable(cls, url):
  471. return re.match(cls._make_valid_url(), url) is not None
  472. def _real_extract(self, query):
  473. mobj = re.match(self._make_valid_url(), query)
  474. if mobj is None:
  475. raise ExtractorError(u'Invalid search query "%s"' % query)
  476. prefix = mobj.group('prefix')
  477. query = mobj.group('query')
  478. if prefix == '':
  479. return self._get_n_results(query, 1)
  480. elif prefix == 'all':
  481. return self._get_n_results(query, self._MAX_RESULTS)
  482. else:
  483. n = int(prefix)
  484. if n <= 0:
  485. raise ExtractorError(u'invalid download number %s for query "%s"' % (n, query))
  486. elif n > self._MAX_RESULTS:
  487. self._downloader.report_warning(u'%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
  488. n = self._MAX_RESULTS
  489. return self._get_n_results(query, n)
  490. def _get_n_results(self, query, n):
  491. """Get a specified number of results for a query"""
  492. raise NotImplementedError("This method must be implemented by subclasses")
  493. @property
  494. def SEARCH_KEY(self):
  495. return self._SEARCH_KEY