]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
03f3f18c83012cdced0e305fe1cc02d69a85bb7c
[youtubedl] / youtube_dl / extractor / common.py
1 from __future__ import unicode_literals
2
3 import base64
4 import datetime
5 import hashlib
6 import json
7 import netrc
8 import os
9 import re
10 import socket
11 import sys
12 import time
13 import xml.etree.ElementTree
14
15 from ..compat import (
16 compat_cookiejar,
17 compat_http_client,
18 compat_urllib_error,
19 compat_urllib_parse_urlparse,
20 compat_urlparse,
21 compat_str,
22 )
23 from ..utils import (
24 age_restricted,
25 clean_html,
26 compiled_regex_type,
27 ExtractorError,
28 float_or_none,
29 int_or_none,
30 RegexNotFoundError,
31 sanitize_filename,
32 unescapeHTML,
33 )
34 _NO_DEFAULT = object()
35
36
37 class InfoExtractor(object):
38 """Information Extractor class.
39
40 Information extractors are the classes that, given a URL, extract
41 information about the video (or videos) the URL refers to. This
42 information includes the real video URL, the video title, author and
43 others. The information is stored in a dictionary which is then
44 passed to the YoutubeDL. The YoutubeDL processes this
45 information possibly downloading the video to the file system, among
46 other possible outcomes.
47
48 The type field determines the the type of the result.
49 By far the most common value (and the default if _type is missing) is
50 "video", which indicates a single video.
51
52 For a video, the dictionaries must include the following fields:
53
54 id: Video identifier.
55 title: Video title, unescaped.
56
57 Additionally, it must contain either a formats entry or a url one:
58
59 formats: A list of dictionaries for each format available, ordered
60 from worst to best quality.
61
62 Potential fields:
63 * url Mandatory. The URL of the video file
64 * ext Will be calculated from url if missing
65 * format A human-readable description of the format
66 ("mp4 container with h264/opus").
67 Calculated from the format_id, width, height.
68 and format_note fields if missing.
69 * format_id A short description of the format
70 ("mp4_h264_opus" or "19").
71 Technically optional, but strongly recommended.
72 * format_note Additional info about the format
73 ("3D" or "DASH video")
74 * width Width of the video, if known
75 * height Height of the video, if known
76 * resolution Textual description of width and height
77 * tbr Average bitrate of audio and video in KBit/s
78 * abr Average audio bitrate in KBit/s
79 * acodec Name of the audio codec in use
80 * asr Audio sampling rate in Hertz
81 * vbr Average video bitrate in KBit/s
82 * fps Frame rate
83 * vcodec Name of the video codec in use
84 * container Name of the container format
85 * filesize The number of bytes, if known in advance
86 * filesize_approx An estimate for the number of bytes
87 * player_url SWF Player URL (used for rtmpdump).
88 * protocol The protocol that will be used for the actual
89 download, lower-case.
90 "http", "https", "rtsp", "rtmp", "m3u8" or so.
91 * preference Order number of this format. If this field is
92 present and not None, the formats get sorted
93 by this field, regardless of all other values.
94 -1 for default (order by other properties),
95 -2 or smaller for less than default.
96 < -1000 to hide the format (if there is
97 another one which is strictly better)
98 * language_preference Is this in the correct requested
99 language?
100 10 if it's what the URL is about,
101 -1 for default (don't know),
102 -10 otherwise, other values reserved for now.
103 * quality Order number of the video quality of this
104 format, irrespective of the file format.
105 -1 for default (order by other properties),
106 -2 or smaller for less than default.
107 * source_preference Order number for this video source
108 (quality takes higher priority)
109 -1 for default (order by other properties),
110 -2 or smaller for less than default.
111 * http_referer HTTP Referer header value to set.
112 * http_method HTTP method to use for the download.
113 * http_headers A dictionary of additional HTTP headers
114 to add to the request.
115 * http_post_data Additional data to send with a POST
116 request.
117 * stretched_ratio If given and not 1, indicates that the
118 video's pixels are not square.
119 width : height ratio as float.
120 url: Final video URL.
121 ext: Video filename extension.
122 format: The video format, defaults to ext (used for --get-format)
123 player_url: SWF Player URL (used for rtmpdump).
124
125 The following fields are optional:
126
127 alt_title: A secondary title of the video.
128 display_id An alternative identifier for the video, not necessarily
129 unique, but available before title. Typically, id is
130 something like "4234987", title "Dancing naked mole rats",
131 and display_id "dancing-naked-mole-rats"
132 thumbnails: A list of dictionaries, with the following entries:
133 * "url"
134 * "width" (optional, int)
135 * "height" (optional, int)
136 * "resolution" (optional, string "{width}x{height"},
137 deprecated)
138 thumbnail: Full URL to a video thumbnail image.
139 description: Full video description.
140 uploader: Full name of the video uploader.
141 timestamp: UNIX timestamp of the moment the video became available.
142 upload_date: Video upload date (YYYYMMDD).
143 If not explicitly set, calculated from timestamp.
144 uploader_id: Nickname or id of the video uploader.
145 location: Physical location where the video was filmed.
146 subtitles: The subtitle file contents as a dictionary in the format
147 {language: subtitles}.
148 duration: Length of the video in seconds, as an integer.
149 view_count: How many users have watched the video on the platform.
150 like_count: Number of positive ratings of the video
151 dislike_count: Number of negative ratings of the video
152 comment_count: Number of comments on the video
153 comments: A list of comments, each with one or more of the following
154 properties (all but one of text or html optional):
155 * "author" - human-readable name of the comment author
156 * "author_id" - user ID of the comment author
157 * "id" - Comment ID
158 * "html" - Comment as HTML
159 * "text" - Plain text of the comment
160 * "timestamp" - UNIX timestamp of comment
161 * "parent" - ID of the comment this one is replying to.
162 Set to "root" to indicate that this is a
163 comment to the original video.
164 age_limit: Age restriction for the video, as an integer (years)
165 webpage_url: The url to the video webpage, if given to youtube-dl it
166 should allow to get the same result again. (It will be set
167 by YoutubeDL if it's missing)
168 categories: A list of categories that the video falls in, for example
169 ["Sports", "Berlin"]
170 is_live: True, False, or None (=unknown). Whether this video is a
171 live stream that goes on instead of a fixed-length video.
172
173 Unless mentioned otherwise, the fields should be Unicode strings.
174
175 Unless mentioned otherwise, None is equivalent to absence of information.
176
177
178 _type "playlist" indicates multiple videos.
179 There must be a key "entries", which is a list, an iterable, or a PagedList
180 object, each element of which is a valid dictionary by this specification.
181
182 Additionally, playlists can have "title" and "id" attributes with the same
183 semantics as videos (see above).
184
185
186 _type "multi_video" indicates that there are multiple videos that
187 form a single show, for examples multiple acts of an opera or TV episode.
188 It must have an entries key like a playlist and contain all the keys
189 required for a video at the same time.
190
191
192 _type "url" indicates that the video must be extracted from another
193 location, possibly by a different extractor. Its only required key is:
194 "url" - the next URL to extract.
195 The key "ie_key" can be set to the class name (minus the trailing "IE",
196 e.g. "Youtube") if the extractor class is known in advance.
197 Additionally, the dictionary may have any properties of the resolved entity
198 known in advance, for example "title" if the title of the referred video is
199 known ahead of time.
200
201
202 _type "url_transparent" entities have the same specification as "url", but
203 indicate that the given additional information is more precise than the one
204 associated with the resolved URL.
205 This is useful when a site employs a video service that hosts the video and
206 its technical metadata, but that video service does not embed a useful
207 title, description etc.
208
209
210 Subclasses of this one should re-define the _real_initialize() and
211 _real_extract() methods and define a _VALID_URL regexp.
212 Probably, they should also be added to the list of extractors.
213
214 Finally, the _WORKING attribute should be set to False for broken IEs
215 in order to warn the users and skip the tests.
216 """
217
218 _ready = False
219 _downloader = None
220 _WORKING = True
221
222 def __init__(self, downloader=None):
223 """Constructor. Receives an optional downloader."""
224 self._ready = False
225 self.set_downloader(downloader)
226
227 @classmethod
228 def suitable(cls, url):
229 """Receives a URL and returns True if suitable for this IE."""
230
231 # This does not use has/getattr intentionally - we want to know whether
232 # we have cached the regexp for *this* class, whereas getattr would also
233 # match the superclass
234 if '_VALID_URL_RE' not in cls.__dict__:
235 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
236 return cls._VALID_URL_RE.match(url) is not None
237
238 @classmethod
239 def _match_id(cls, url):
240 if '_VALID_URL_RE' not in cls.__dict__:
241 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
242 m = cls._VALID_URL_RE.match(url)
243 assert m
244 return m.group('id')
245
246 @classmethod
247 def working(cls):
248 """Getter method for _WORKING."""
249 return cls._WORKING
250
251 def initialize(self):
252 """Initializes an instance (authentication, etc)."""
253 if not self._ready:
254 self._real_initialize()
255 self._ready = True
256
257 def extract(self, url):
258 """Extracts URL information and returns it in list of dicts."""
259 self.initialize()
260 return self._real_extract(url)
261
262 def set_downloader(self, downloader):
263 """Sets the downloader for this IE."""
264 self._downloader = downloader
265
266 def _real_initialize(self):
267 """Real initialization process. Redefine in subclasses."""
268 pass
269
270 def _real_extract(self, url):
271 """Real extraction process. Redefine in subclasses."""
272 pass
273
274 @classmethod
275 def ie_key(cls):
276 """A string for getting the InfoExtractor with get_info_extractor"""
277 return cls.__name__[:-2]
278
279 @property
280 def IE_NAME(self):
281 return type(self).__name__[:-2]
282
283 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
284 """ Returns the response handle """
285 if note is None:
286 self.report_download_webpage(video_id)
287 elif note is not False:
288 if video_id is None:
289 self.to_screen('%s' % (note,))
290 else:
291 self.to_screen('%s: %s' % (video_id, note))
292 try:
293 return self._downloader.urlopen(url_or_request)
294 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
295 if errnote is False:
296 return False
297 if errnote is None:
298 errnote = 'Unable to download webpage'
299 errmsg = '%s: %s' % (errnote, compat_str(err))
300 if fatal:
301 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
302 else:
303 self._downloader.report_warning(errmsg)
304 return False
305
306 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
307 """ Returns a tuple (page content as string, URL handle) """
308 # Strip hashes from the URL (#1038)
309 if isinstance(url_or_request, (compat_str, str)):
310 url_or_request = url_or_request.partition('#')[0]
311
312 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
313 if urlh is False:
314 assert not fatal
315 return False
316 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal)
317 return (content, urlh)
318
319 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None):
320 content_type = urlh.headers.get('Content-Type', '')
321 webpage_bytes = urlh.read()
322 if prefix is not None:
323 webpage_bytes = prefix + webpage_bytes
324 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
325 if m:
326 encoding = m.group(1)
327 else:
328 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
329 webpage_bytes[:1024])
330 if m:
331 encoding = m.group(1).decode('ascii')
332 elif webpage_bytes.startswith(b'\xff\xfe'):
333 encoding = 'utf-16'
334 else:
335 encoding = 'utf-8'
336 if self._downloader.params.get('dump_intermediate_pages', False):
337 try:
338 url = url_or_request.get_full_url()
339 except AttributeError:
340 url = url_or_request
341 self.to_screen('Dumping request to ' + url)
342 dump = base64.b64encode(webpage_bytes).decode('ascii')
343 self._downloader.to_screen(dump)
344 if self._downloader.params.get('write_pages', False):
345 try:
346 url = url_or_request.get_full_url()
347 except AttributeError:
348 url = url_or_request
349 basen = '%s_%s' % (video_id, url)
350 if len(basen) > 240:
351 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
352 basen = basen[:240 - len(h)] + h
353 raw_filename = basen + '.dump'
354 filename = sanitize_filename(raw_filename, restricted=True)
355 self.to_screen('Saving request to ' + filename)
356 # Working around MAX_PATH limitation on Windows (see
357 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
358 if os.name == 'nt':
359 absfilepath = os.path.abspath(filename)
360 if len(absfilepath) > 259:
361 filename = '\\\\?\\' + absfilepath
362 with open(filename, 'wb') as outf:
363 outf.write(webpage_bytes)
364
365 try:
366 content = webpage_bytes.decode(encoding, 'replace')
367 except LookupError:
368 content = webpage_bytes.decode('utf-8', 'replace')
369
370 if ('<title>Access to this site is blocked</title>' in content and
371 'Websense' in content[:512]):
372 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
373 blocked_iframe = self._html_search_regex(
374 r'<iframe src="([^"]+)"', content,
375 'Websense information URL', default=None)
376 if blocked_iframe:
377 msg += ' Visit %s for more details' % blocked_iframe
378 raise ExtractorError(msg, expected=True)
379
380 return content
381
382 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5):
383 """ Returns the data of the page as a string """
384 success = False
385 try_count = 0
386 while success is False:
387 try:
388 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
389 success = True
390 except compat_http_client.IncompleteRead as e:
391 try_count += 1
392 if try_count >= tries:
393 raise e
394 self._sleep(timeout, video_id)
395 if res is False:
396 return res
397 else:
398 content, _ = res
399 return content
400
401 def _download_xml(self, url_or_request, video_id,
402 note='Downloading XML', errnote='Unable to download XML',
403 transform_source=None, fatal=True):
404 """Return the xml as an xml.etree.ElementTree.Element"""
405 xml_string = self._download_webpage(
406 url_or_request, video_id, note, errnote, fatal=fatal)
407 if xml_string is False:
408 return xml_string
409 if transform_source:
410 xml_string = transform_source(xml_string)
411 return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
412
413 def _download_json(self, url_or_request, video_id,
414 note='Downloading JSON metadata',
415 errnote='Unable to download JSON metadata',
416 transform_source=None,
417 fatal=True):
418 json_string = self._download_webpage(
419 url_or_request, video_id, note, errnote, fatal=fatal)
420 if (not fatal) and json_string is False:
421 return None
422 return self._parse_json(
423 json_string, video_id, transform_source=transform_source, fatal=fatal)
424
425 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
426 if transform_source:
427 json_string = transform_source(json_string)
428 try:
429 return json.loads(json_string)
430 except ValueError as ve:
431 errmsg = '%s: Failed to parse JSON ' % video_id
432 if fatal:
433 raise ExtractorError(errmsg, cause=ve)
434 else:
435 self.report_warning(errmsg + str(ve))
436
437 def report_warning(self, msg, video_id=None):
438 idstr = '' if video_id is None else '%s: ' % video_id
439 self._downloader.report_warning(
440 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
441
442 def to_screen(self, msg):
443 """Print msg to screen, prefixing it with '[ie_name]'"""
444 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
445
446 def report_extraction(self, id_or_name):
447 """Report information extraction."""
448 self.to_screen('%s: Extracting information' % id_or_name)
449
450 def report_download_webpage(self, video_id):
451 """Report webpage download."""
452 self.to_screen('%s: Downloading webpage' % video_id)
453
454 def report_age_confirmation(self):
455 """Report attempt to confirm age."""
456 self.to_screen('Confirming age')
457
458 def report_login(self):
459 """Report attempt to log in."""
460 self.to_screen('Logging in')
461
462 # Methods for following #608
463 @staticmethod
464 def url_result(url, ie=None, video_id=None):
465 """Returns a url that points to a page that should be processed"""
466 # TODO: ie should be the class used for getting the info
467 video_info = {'_type': 'url',
468 'url': url,
469 'ie_key': ie}
470 if video_id is not None:
471 video_info['id'] = video_id
472 return video_info
473
474 @staticmethod
475 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
476 """Returns a playlist"""
477 video_info = {'_type': 'playlist',
478 'entries': entries}
479 if playlist_id:
480 video_info['id'] = playlist_id
481 if playlist_title:
482 video_info['title'] = playlist_title
483 if playlist_description:
484 video_info['description'] = playlist_description
485 return video_info
486
487 def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
488 """
489 Perform a regex search on the given string, using a single or a list of
490 patterns returning the first matching group.
491 In case of failure return a default value or raise a WARNING or a
492 RegexNotFoundError, depending on fatal, specifying the field name.
493 """
494 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
495 mobj = re.search(pattern, string, flags)
496 else:
497 for p in pattern:
498 mobj = re.search(p, string, flags)
499 if mobj:
500 break
501
502 if os.name != 'nt' and sys.stderr.isatty():
503 _name = '\033[0;34m%s\033[0m' % name
504 else:
505 _name = name
506
507 if mobj:
508 if group is None:
509 # return the first matching group
510 return next(g for g in mobj.groups() if g is not None)
511 else:
512 return mobj.group(group)
513 elif default is not _NO_DEFAULT:
514 return default
515 elif fatal:
516 raise RegexNotFoundError('Unable to extract %s' % _name)
517 else:
518 self._downloader.report_warning('unable to extract %s; '
519 'please report this issue on http://yt-dl.org/bug' % _name)
520 return None
521
522 def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0, group=None):
523 """
524 Like _search_regex, but strips HTML tags and unescapes entities.
525 """
526 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
527 if res:
528 return clean_html(res).strip()
529 else:
530 return res
531
532 def _get_login_info(self):
533 """
534 Get the the login info as (username, password)
535 It will look in the netrc file using the _NETRC_MACHINE value
536 If there's no info available, return (None, None)
537 """
538 if self._downloader is None:
539 return (None, None)
540
541 username = None
542 password = None
543 downloader_params = self._downloader.params
544
545 # Attempt to use provided username and password or .netrc data
546 if downloader_params.get('username', None) is not None:
547 username = downloader_params['username']
548 password = downloader_params['password']
549 elif downloader_params.get('usenetrc', False):
550 try:
551 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
552 if info is not None:
553 username = info[0]
554 password = info[2]
555 else:
556 raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
557 except (IOError, netrc.NetrcParseError) as err:
558 self._downloader.report_warning('parsing .netrc: %s' % compat_str(err))
559
560 return (username, password)
561
562 def _get_tfa_info(self):
563 """
564 Get the two-factor authentication info
565 TODO - asking the user will be required for sms/phone verify
566 currently just uses the command line option
567 If there's no info available, return None
568 """
569 if self._downloader is None:
570 return None
571 downloader_params = self._downloader.params
572
573 if downloader_params.get('twofactor', None) is not None:
574 return downloader_params['twofactor']
575
576 return None
577
578 # Helper functions for extracting OpenGraph info
579 @staticmethod
580 def _og_regexes(prop):
581 content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
582 property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
583 template = r'<meta[^>]+?%s[^>]+?%s'
584 return [
585 template % (property_re, content_re),
586 template % (content_re, property_re),
587 ]
588
589 def _og_search_property(self, prop, html, name=None, **kargs):
590 if name is None:
591 name = 'OpenGraph %s' % prop
592 escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
593 if escaped is None:
594 return None
595 return unescapeHTML(escaped)
596
597 def _og_search_thumbnail(self, html, **kargs):
598 return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
599
600 def _og_search_description(self, html, **kargs):
601 return self._og_search_property('description', html, fatal=False, **kargs)
602
603 def _og_search_title(self, html, **kargs):
604 return self._og_search_property('title', html, **kargs)
605
606 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
607 regexes = self._og_regexes('video') + self._og_regexes('video:url')
608 if secure:
609 regexes = self._og_regexes('video:secure_url') + regexes
610 return self._html_search_regex(regexes, html, name, **kargs)
611
612 def _og_search_url(self, html, **kargs):
613 return self._og_search_property('url', html, **kargs)
614
615 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
616 if display_name is None:
617 display_name = name
618 return self._html_search_regex(
619 r'''(?isx)<meta
620 (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1)
621 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name),
622 html, display_name, fatal=fatal, group='content', **kwargs)
623
624 def _dc_search_uploader(self, html):
625 return self._html_search_meta('dc.creator', html, 'uploader')
626
627 def _rta_search(self, html):
628 # See http://www.rtalabel.org/index.php?content=howtofaq#single
629 if re.search(r'(?ix)<meta\s+name="rating"\s+'
630 r' content="RTA-5042-1996-1400-1577-RTA"',
631 html):
632 return 18
633 return 0
634
635 def _media_rating_search(self, html):
636 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
637 rating = self._html_search_meta('rating', html)
638
639 if not rating:
640 return None
641
642 RATING_TABLE = {
643 'safe for kids': 0,
644 'general': 8,
645 '14 years': 14,
646 'mature': 17,
647 'restricted': 19,
648 }
649 return RATING_TABLE.get(rating.lower(), None)
650
651 def _twitter_search_player(self, html):
652 return self._html_search_meta('twitter:player', html,
653 'twitter card player')
654
655 def _sort_formats(self, formats):
656 if not formats:
657 raise ExtractorError('No video formats found')
658
659 def _formats_key(f):
660 # TODO remove the following workaround
661 from ..utils import determine_ext
662 if not f.get('ext') and 'url' in f:
663 f['ext'] = determine_ext(f['url'])
664
665 preference = f.get('preference')
666 if preference is None:
667 proto = f.get('protocol')
668 if proto is None:
669 proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
670
671 preference = 0 if proto in ['http', 'https'] else -0.1
672 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
673 preference -= 0.5
674
675 if f.get('vcodec') == 'none': # audio only
676 if self._downloader.params.get('prefer_free_formats'):
677 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
678 else:
679 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
680 ext_preference = 0
681 try:
682 audio_ext_preference = ORDER.index(f['ext'])
683 except ValueError:
684 audio_ext_preference = -1
685 else:
686 if self._downloader.params.get('prefer_free_formats'):
687 ORDER = ['flv', 'mp4', 'webm']
688 else:
689 ORDER = ['webm', 'flv', 'mp4']
690 try:
691 ext_preference = ORDER.index(f['ext'])
692 except ValueError:
693 ext_preference = -1
694 audio_ext_preference = 0
695
696 return (
697 preference,
698 f.get('language_preference') if f.get('language_preference') is not None else -1,
699 f.get('quality') if f.get('quality') is not None else -1,
700 f.get('height') if f.get('height') is not None else -1,
701 f.get('width') if f.get('width') is not None else -1,
702 ext_preference,
703 f.get('tbr') if f.get('tbr') is not None else -1,
704 f.get('vbr') if f.get('vbr') is not None else -1,
705 f.get('abr') if f.get('abr') is not None else -1,
706 audio_ext_preference,
707 f.get('fps') if f.get('fps') is not None else -1,
708 f.get('filesize') if f.get('filesize') is not None else -1,
709 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
710 f.get('source_preference') if f.get('source_preference') is not None else -1,
711 f.get('format_id'),
712 )
713 formats.sort(key=_formats_key)
714
715 def http_scheme(self):
716 """ Either "http:" or "https:", depending on the user's preferences """
717 return (
718 'http:'
719 if self._downloader.params.get('prefer_insecure', False)
720 else 'https:')
721
722 def _proto_relative_url(self, url, scheme=None):
723 if url is None:
724 return url
725 if url.startswith('//'):
726 if scheme is None:
727 scheme = self.http_scheme()
728 return scheme + url
729 else:
730 return url
731
732 def _sleep(self, timeout, video_id, msg_template=None):
733 if msg_template is None:
734 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
735 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
736 self.to_screen(msg)
737 time.sleep(timeout)
738
739 def _extract_f4m_formats(self, manifest_url, video_id):
740 manifest = self._download_xml(
741 manifest_url, video_id, 'Downloading f4m manifest',
742 'Unable to download f4m manifest')
743
744 formats = []
745 manifest_version = '1.0'
746 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
747 if not media_nodes:
748 manifest_version = '2.0'
749 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
750 for i, media_el in enumerate(media_nodes):
751 if manifest_version == '2.0':
752 manifest_url = '/'.join(manifest_url.split('/')[:-1]) + '/' + media_el.attrib.get('href')
753 tbr = int_or_none(media_el.attrib.get('bitrate'))
754 format_id = 'f4m-%d' % (i if tbr is None else tbr)
755 formats.append({
756 'format_id': format_id,
757 'url': manifest_url,
758 'ext': 'flv',
759 'tbr': tbr,
760 'width': int_or_none(media_el.attrib.get('width')),
761 'height': int_or_none(media_el.attrib.get('height')),
762 })
763 self._sort_formats(formats)
764
765 return formats
766
767 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
768 entry_protocol='m3u8', preference=None):
769
770 formats = [{
771 'format_id': 'm3u8-meta',
772 'url': m3u8_url,
773 'ext': ext,
774 'protocol': 'm3u8',
775 'preference': -1,
776 'resolution': 'multiple',
777 'format_note': 'Quality selection URL',
778 }]
779
780 format_url = lambda u: (
781 u
782 if re.match(r'^https?://', u)
783 else compat_urlparse.urljoin(m3u8_url, u))
784
785 m3u8_doc = self._download_webpage(
786 m3u8_url, video_id,
787 note='Downloading m3u8 information',
788 errnote='Failed to download m3u8 information')
789 last_info = None
790 kv_rex = re.compile(
791 r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
792 for line in m3u8_doc.splitlines():
793 if line.startswith('#EXT-X-STREAM-INF:'):
794 last_info = {}
795 for m in kv_rex.finditer(line):
796 v = m.group('val')
797 if v.startswith('"'):
798 v = v[1:-1]
799 last_info[m.group('key')] = v
800 elif line.startswith('#') or not line.strip():
801 continue
802 else:
803 if last_info is None:
804 formats.append({'url': format_url(line)})
805 continue
806 tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
807
808 f = {
809 'format_id': 'm3u8-%d' % (tbr if tbr else len(formats)),
810 'url': format_url(line.strip()),
811 'tbr': tbr,
812 'ext': ext,
813 'protocol': entry_protocol,
814 'preference': preference,
815 }
816 codecs = last_info.get('CODECS')
817 if codecs:
818 # TODO: looks like video codec is not always necessarily goes first
819 va_codecs = codecs.split(',')
820 if va_codecs[0]:
821 f['vcodec'] = va_codecs[0].partition('.')[0]
822 if len(va_codecs) > 1 and va_codecs[1]:
823 f['acodec'] = va_codecs[1].partition('.')[0]
824 resolution = last_info.get('RESOLUTION')
825 if resolution:
826 width_str, height_str = resolution.split('x')
827 f['width'] = int(width_str)
828 f['height'] = int(height_str)
829 formats.append(f)
830 last_info = {}
831 self._sort_formats(formats)
832 return formats
833
834 # TODO: improve extraction
835 def _extract_smil_formats(self, smil_url, video_id):
836 smil = self._download_xml(
837 smil_url, video_id, 'Downloading SMIL file',
838 'Unable to download SMIL file')
839
840 base = smil.find('./head/meta').get('base')
841
842 formats = []
843 rtmp_count = 0
844 for video in smil.findall('./body/switch/video'):
845 src = video.get('src')
846 if not src:
847 continue
848 bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
849 width = int_or_none(video.get('width'))
850 height = int_or_none(video.get('height'))
851 proto = video.get('proto')
852 if not proto:
853 if base:
854 if base.startswith('rtmp'):
855 proto = 'rtmp'
856 elif base.startswith('http'):
857 proto = 'http'
858 ext = video.get('ext')
859 if proto == 'm3u8':
860 formats.extend(self._extract_m3u8_formats(src, video_id, ext))
861 elif proto == 'rtmp':
862 rtmp_count += 1
863 streamer = video.get('streamer') or base
864 formats.append({
865 'url': streamer,
866 'play_path': src,
867 'ext': 'flv',
868 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
869 'tbr': bitrate,
870 'width': width,
871 'height': height,
872 })
873 self._sort_formats(formats)
874
875 return formats
876
877 def _live_title(self, name):
878 """ Generate the title for a live video """
879 now = datetime.datetime.now()
880 now_str = now.strftime("%Y-%m-%d %H:%M")
881 return name + ' ' + now_str
882
883 def _int(self, v, name, fatal=False, **kwargs):
884 res = int_or_none(v, **kwargs)
885 if 'get_attr' in kwargs:
886 print(getattr(v, kwargs['get_attr']))
887 if res is None:
888 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
889 if fatal:
890 raise ExtractorError(msg)
891 else:
892 self._downloader.report_warning(msg)
893 return res
894
895 def _float(self, v, name, fatal=False, **kwargs):
896 res = float_or_none(v, **kwargs)
897 if res is None:
898 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
899 if fatal:
900 raise ExtractorError(msg)
901 else:
902 self._downloader.report_warning(msg)
903 return res
904
905 def _set_cookie(self, domain, name, value, expire_time=None):
906 cookie = compat_cookiejar.Cookie(
907 0, name, value, None, None, domain, None,
908 None, '/', True, False, expire_time, '', None, None, None)
909 self._downloader.cookiejar.set_cookie(cookie)
910
911 def get_testcases(self, include_onlymatching=False):
912 t = getattr(self, '_TEST', None)
913 if t:
914 assert not hasattr(self, '_TESTS'), \
915 '%s has _TEST and _TESTS' % type(self).__name__
916 tests = [t]
917 else:
918 tests = getattr(self, '_TESTS', [])
919 for t in tests:
920 if not include_onlymatching and t.get('only_matching', False):
921 continue
922 t['name'] = type(self).__name__[:-len('IE')]
923 yield t
924
925 def is_suitable(self, age_limit):
926 """ Test whether the extractor is generally suitable for the given
927 age limit (i.e. pornographic sites are not, all others usually are) """
928
929 any_restricted = False
930 for tc in self.get_testcases(include_onlymatching=False):
931 if 'playlist' in tc:
932 tc = tc['playlist'][0]
933 is_restricted = age_restricted(
934 tc.get('info_dict', {}).get('age_limit'), age_limit)
935 if not is_restricted:
936 return True
937 any_restricted = any_restricted or is_restricted
938 return not any_restricted
939
940
941 class SearchInfoExtractor(InfoExtractor):
942 """
943 Base class for paged search queries extractors.
944 They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
945 Instances should define _SEARCH_KEY and _MAX_RESULTS.
946 """
947
948 @classmethod
949 def _make_valid_url(cls):
950 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
951
952 @classmethod
953 def suitable(cls, url):
954 return re.match(cls._make_valid_url(), url) is not None
955
956 def _real_extract(self, query):
957 mobj = re.match(self._make_valid_url(), query)
958 if mobj is None:
959 raise ExtractorError('Invalid search query "%s"' % query)
960
961 prefix = mobj.group('prefix')
962 query = mobj.group('query')
963 if prefix == '':
964 return self._get_n_results(query, 1)
965 elif prefix == 'all':
966 return self._get_n_results(query, self._MAX_RESULTS)
967 else:
968 n = int(prefix)
969 if n <= 0:
970 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
971 elif n > self._MAX_RESULTS:
972 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
973 n = self._MAX_RESULTS
974 return self._get_n_results(query, n)
975
976 def _get_n_results(self, query, n):
977 """Get a specified number of results for a query"""
978 raise NotImplementedError("This method must be implemented by subclasses")
979
980 @property
981 def SEARCH_KEY(self):
982 return self._SEARCH_KEY