]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
Prepare new release.
[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 random
10 import re
11 import socket
12 import sys
13 import time
14 import math
15
16 from ..compat import (
17 compat_cookiejar,
18 compat_cookies,
19 compat_etree_fromstring,
20 compat_getpass,
21 compat_http_client,
22 compat_os_name,
23 compat_str,
24 compat_urllib_error,
25 compat_urllib_parse_unquote,
26 compat_urllib_parse_urlencode,
27 compat_urllib_request,
28 compat_urlparse,
29 )
30 from ..downloader.f4m import remove_encrypted_media
31 from ..utils import (
32 NO_DEFAULT,
33 age_restricted,
34 base_url,
35 bug_reports_message,
36 clean_html,
37 compiled_regex_type,
38 determine_ext,
39 error_to_compat_str,
40 ExtractorError,
41 fix_xml_ampersands,
42 float_or_none,
43 GeoRestrictedError,
44 GeoUtils,
45 int_or_none,
46 js_to_json,
47 parse_iso8601,
48 RegexNotFoundError,
49 sanitize_filename,
50 sanitized_Request,
51 unescapeHTML,
52 unified_strdate,
53 unified_timestamp,
54 url_basename,
55 xpath_element,
56 xpath_text,
57 xpath_with_ns,
58 determine_protocol,
59 parse_duration,
60 mimetype2ext,
61 update_Request,
62 update_url_query,
63 parse_m3u8_attributes,
64 extract_attributes,
65 parse_codecs,
66 urljoin,
67 )
68
69
70 class InfoExtractor(object):
71 """Information Extractor class.
72
73 Information extractors are the classes that, given a URL, extract
74 information about the video (or videos) the URL refers to. This
75 information includes the real video URL, the video title, author and
76 others. The information is stored in a dictionary which is then
77 passed to the YoutubeDL. The YoutubeDL processes this
78 information possibly downloading the video to the file system, among
79 other possible outcomes.
80
81 The type field determines the type of the result.
82 By far the most common value (and the default if _type is missing) is
83 "video", which indicates a single video.
84
85 For a video, the dictionaries must include the following fields:
86
87 id: Video identifier.
88 title: Video title, unescaped.
89
90 Additionally, it must contain either a formats entry or a url one:
91
92 formats: A list of dictionaries for each format available, ordered
93 from worst to best quality.
94
95 Potential fields:
96 * url Mandatory. The URL of the video file
97 * manifest_url
98 The URL of the manifest file in case of
99 fragmented media (DASH, hls, hds)
100 * ext Will be calculated from URL if missing
101 * format A human-readable description of the format
102 ("mp4 container with h264/opus").
103 Calculated from the format_id, width, height.
104 and format_note fields if missing.
105 * format_id A short description of the format
106 ("mp4_h264_opus" or "19").
107 Technically optional, but strongly recommended.
108 * format_note Additional info about the format
109 ("3D" or "DASH video")
110 * width Width of the video, if known
111 * height Height of the video, if known
112 * resolution Textual description of width and height
113 * tbr Average bitrate of audio and video in KBit/s
114 * abr Average audio bitrate in KBit/s
115 * acodec Name of the audio codec in use
116 * asr Audio sampling rate in Hertz
117 * vbr Average video bitrate in KBit/s
118 * fps Frame rate
119 * vcodec Name of the video codec in use
120 * container Name of the container format
121 * filesize The number of bytes, if known in advance
122 * filesize_approx An estimate for the number of bytes
123 * player_url SWF Player URL (used for rtmpdump).
124 * protocol The protocol that will be used for the actual
125 download, lower-case.
126 "http", "https", "rtsp", "rtmp", "rtmpe",
127 "m3u8", "m3u8_native" or "http_dash_segments".
128 * fragment_base_url
129 Base URL for fragments. Each fragment's path
130 value (if present) will be relative to
131 this URL.
132 * fragments A list of fragments of a fragmented media.
133 Each fragment entry must contain either an url
134 or a path. If an url is present it should be
135 considered by a client. Otherwise both path and
136 fragment_base_url must be present. Here is
137 the list of all potential fields:
138 * "url" - fragment's URL
139 * "path" - fragment's path relative to
140 fragment_base_url
141 * "duration" (optional, int or float)
142 * "filesize" (optional, int)
143 * preference Order number of this format. If this field is
144 present and not None, the formats get sorted
145 by this field, regardless of all other values.
146 -1 for default (order by other properties),
147 -2 or smaller for less than default.
148 < -1000 to hide the format (if there is
149 another one which is strictly better)
150 * language Language code, e.g. "de" or "en-US".
151 * language_preference Is this in the language mentioned in
152 the URL?
153 10 if it's what the URL is about,
154 -1 for default (don't know),
155 -10 otherwise, other values reserved for now.
156 * quality Order number of the video quality of this
157 format, irrespective of the file format.
158 -1 for default (order by other properties),
159 -2 or smaller for less than default.
160 * source_preference Order number for this video source
161 (quality takes higher priority)
162 -1 for default (order by other properties),
163 -2 or smaller for less than default.
164 * http_headers A dictionary of additional HTTP headers
165 to add to the request.
166 * stretched_ratio If given and not 1, indicates that the
167 video's pixels are not square.
168 width : height ratio as float.
169 * no_resume The server does not support resuming the
170 (HTTP or RTMP) download. Boolean.
171
172 url: Final video URL.
173 ext: Video filename extension.
174 format: The video format, defaults to ext (used for --get-format)
175 player_url: SWF Player URL (used for rtmpdump).
176
177 The following fields are optional:
178
179 alt_title: A secondary title of the video.
180 display_id An alternative identifier for the video, not necessarily
181 unique, but available before title. Typically, id is
182 something like "4234987", title "Dancing naked mole rats",
183 and display_id "dancing-naked-mole-rats"
184 thumbnails: A list of dictionaries, with the following entries:
185 * "id" (optional, string) - Thumbnail format ID
186 * "url"
187 * "preference" (optional, int) - quality of the image
188 * "width" (optional, int)
189 * "height" (optional, int)
190 * "resolution" (optional, string "{width}x{height"},
191 deprecated)
192 * "filesize" (optional, int)
193 thumbnail: Full URL to a video thumbnail image.
194 description: Full video description.
195 uploader: Full name of the video uploader.
196 license: License name the video is licensed under.
197 creator: The creator of the video.
198 release_date: The date (YYYYMMDD) when the video was released.
199 timestamp: UNIX timestamp of the moment the video became available.
200 upload_date: Video upload date (YYYYMMDD).
201 If not explicitly set, calculated from timestamp.
202 uploader_id: Nickname or id of the video uploader.
203 uploader_url: Full URL to a personal webpage of the video uploader.
204 location: Physical location where the video was filmed.
205 subtitles: The available subtitles as a dictionary in the format
206 {tag: subformats}. "tag" is usually a language code, and
207 "subformats" is a list sorted from lower to higher
208 preference, each element is a dictionary with the "ext"
209 entry and one of:
210 * "data": The subtitles file contents
211 * "url": A URL pointing to the subtitles file
212 "ext" will be calculated from URL if missing
213 automatic_captions: Like 'subtitles', used by the YoutubeIE for
214 automatically generated captions
215 duration: Length of the video in seconds, as an integer or float.
216 view_count: How many users have watched the video on the platform.
217 like_count: Number of positive ratings of the video
218 dislike_count: Number of negative ratings of the video
219 repost_count: Number of reposts of the video
220 average_rating: Average rating give by users, the scale used depends on the webpage
221 comment_count: Number of comments on the video
222 comments: A list of comments, each with one or more of the following
223 properties (all but one of text or html optional):
224 * "author" - human-readable name of the comment author
225 * "author_id" - user ID of the comment author
226 * "id" - Comment ID
227 * "html" - Comment as HTML
228 * "text" - Plain text of the comment
229 * "timestamp" - UNIX timestamp of comment
230 * "parent" - ID of the comment this one is replying to.
231 Set to "root" to indicate that this is a
232 comment to the original video.
233 age_limit: Age restriction for the video, as an integer (years)
234 webpage_url: The URL to the video webpage, if given to youtube-dl it
235 should allow to get the same result again. (It will be set
236 by YoutubeDL if it's missing)
237 categories: A list of categories that the video falls in, for example
238 ["Sports", "Berlin"]
239 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
240 is_live: True, False, or None (=unknown). Whether this video is a
241 live stream that goes on instead of a fixed-length video.
242 start_time: Time in seconds where the reproduction should start, as
243 specified in the URL.
244 end_time: Time in seconds where the reproduction should end, as
245 specified in the URL.
246
247 The following fields should only be used when the video belongs to some logical
248 chapter or section:
249
250 chapter: Name or title of the chapter the video belongs to.
251 chapter_number: Number of the chapter the video belongs to, as an integer.
252 chapter_id: Id of the chapter the video belongs to, as a unicode string.
253
254 The following fields should only be used when the video is an episode of some
255 series, programme or podcast:
256
257 series: Title of the series or programme the video episode belongs to.
258 season: Title of the season the video episode belongs to.
259 season_number: Number of the season the video episode belongs to, as an integer.
260 season_id: Id of the season the video episode belongs to, as a unicode string.
261 episode: Title of the video episode. Unlike mandatory video title field,
262 this field should denote the exact title of the video episode
263 without any kind of decoration.
264 episode_number: Number of the video episode within a season, as an integer.
265 episode_id: Id of the video episode, as a unicode string.
266
267 The following fields should only be used when the media is a track or a part of
268 a music album:
269
270 track: Title of the track.
271 track_number: Number of the track within an album or a disc, as an integer.
272 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
273 as a unicode string.
274 artist: Artist(s) of the track.
275 genre: Genre(s) of the track.
276 album: Title of the album the track belongs to.
277 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
278 album_artist: List of all artists appeared on the album (e.g.
279 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
280 and compilations).
281 disc_number: Number of the disc or other physical medium the track belongs to,
282 as an integer.
283 release_year: Year (YYYY) when the album was released.
284
285 Unless mentioned otherwise, the fields should be Unicode strings.
286
287 Unless mentioned otherwise, None is equivalent to absence of information.
288
289
290 _type "playlist" indicates multiple videos.
291 There must be a key "entries", which is a list, an iterable, or a PagedList
292 object, each element of which is a valid dictionary by this specification.
293
294 Additionally, playlists can have "title", "description" and "id" attributes
295 with the same semantics as videos (see above).
296
297
298 _type "multi_video" indicates that there are multiple videos that
299 form a single show, for examples multiple acts of an opera or TV episode.
300 It must have an entries key like a playlist and contain all the keys
301 required for a video at the same time.
302
303
304 _type "url" indicates that the video must be extracted from another
305 location, possibly by a different extractor. Its only required key is:
306 "url" - the next URL to extract.
307 The key "ie_key" can be set to the class name (minus the trailing "IE",
308 e.g. "Youtube") if the extractor class is known in advance.
309 Additionally, the dictionary may have any properties of the resolved entity
310 known in advance, for example "title" if the title of the referred video is
311 known ahead of time.
312
313
314 _type "url_transparent" entities have the same specification as "url", but
315 indicate that the given additional information is more precise than the one
316 associated with the resolved URL.
317 This is useful when a site employs a video service that hosts the video and
318 its technical metadata, but that video service does not embed a useful
319 title, description etc.
320
321
322 Subclasses of this one should re-define the _real_initialize() and
323 _real_extract() methods and define a _VALID_URL regexp.
324 Probably, they should also be added to the list of extractors.
325
326 _GEO_BYPASS attribute may be set to False in order to disable
327 geo restriction bypass mechanisms for a particular extractor.
328 Though it won't disable explicit geo restriction bypass based on
329 country code provided with geo_bypass_country. (experimental)
330
331 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
332 countries for this extractor. One of these countries will be used by
333 geo restriction bypass mechanism right away in order to bypass
334 geo restriction, of course, if the mechanism is not disabled. (experimental)
335
336 NB: both these geo attributes are experimental and may change in future
337 or be completely removed.
338
339 Finally, the _WORKING attribute should be set to False for broken IEs
340 in order to warn the users and skip the tests.
341 """
342
343 _ready = False
344 _downloader = None
345 _x_forwarded_for_ip = None
346 _GEO_BYPASS = True
347 _GEO_COUNTRIES = None
348 _WORKING = True
349
350 def __init__(self, downloader=None):
351 """Constructor. Receives an optional downloader."""
352 self._ready = False
353 self._x_forwarded_for_ip = None
354 self.set_downloader(downloader)
355
356 @classmethod
357 def suitable(cls, url):
358 """Receives a URL and returns True if suitable for this IE."""
359
360 # This does not use has/getattr intentionally - we want to know whether
361 # we have cached the regexp for *this* class, whereas getattr would also
362 # match the superclass
363 if '_VALID_URL_RE' not in cls.__dict__:
364 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
365 return cls._VALID_URL_RE.match(url) is not None
366
367 @classmethod
368 def _match_id(cls, url):
369 if '_VALID_URL_RE' not in cls.__dict__:
370 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
371 m = cls._VALID_URL_RE.match(url)
372 assert m
373 return m.group('id')
374
375 @classmethod
376 def working(cls):
377 """Getter method for _WORKING."""
378 return cls._WORKING
379
380 def initialize(self):
381 """Initializes an instance (authentication, etc)."""
382 self._initialize_geo_bypass(self._GEO_COUNTRIES)
383 if not self._ready:
384 self._real_initialize()
385 self._ready = True
386
387 def _initialize_geo_bypass(self, countries):
388 """
389 Initialize geo restriction bypass mechanism.
390
391 This method is used to initialize geo bypass mechanism based on faking
392 X-Forwarded-For HTTP header. A random country from provided country list
393 is selected and a random IP belonging to this country is generated. This
394 IP will be passed as X-Forwarded-For HTTP header in all subsequent
395 HTTP requests.
396
397 This method will be used for initial geo bypass mechanism initialization
398 during the instance initialization with _GEO_COUNTRIES.
399
400 You may also manually call it from extractor's code if geo countries
401 information is not available beforehand (e.g. obtained during
402 extraction) or due to some another reason.
403 """
404 if not self._x_forwarded_for_ip:
405 country_code = self._downloader.params.get('geo_bypass_country', None)
406 # If there is no explicit country for geo bypass specified and
407 # the extractor is known to be geo restricted let's fake IP
408 # as X-Forwarded-For right away.
409 if (not country_code and
410 self._GEO_BYPASS and
411 self._downloader.params.get('geo_bypass', True) and
412 countries):
413 country_code = random.choice(countries)
414 if country_code:
415 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
416 if self._downloader.params.get('verbose', False):
417 self._downloader.to_stdout(
418 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
419 % (self._x_forwarded_for_ip, country_code.upper()))
420
421 def extract(self, url):
422 """Extracts URL information and returns it in list of dicts."""
423 try:
424 for _ in range(2):
425 try:
426 self.initialize()
427 ie_result = self._real_extract(url)
428 if self._x_forwarded_for_ip:
429 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
430 return ie_result
431 except GeoRestrictedError as e:
432 if self.__maybe_fake_ip_and_retry(e.countries):
433 continue
434 raise
435 except ExtractorError:
436 raise
437 except compat_http_client.IncompleteRead as e:
438 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
439 except (KeyError, StopIteration) as e:
440 raise ExtractorError('An extractor error has occurred.', cause=e)
441
442 def __maybe_fake_ip_and_retry(self, countries):
443 if (not self._downloader.params.get('geo_bypass_country', None) and
444 self._GEO_BYPASS and
445 self._downloader.params.get('geo_bypass', True) and
446 not self._x_forwarded_for_ip and
447 countries):
448 country_code = random.choice(countries)
449 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
450 if self._x_forwarded_for_ip:
451 self.report_warning(
452 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
453 % (self._x_forwarded_for_ip, country_code.upper()))
454 return True
455 return False
456
457 def set_downloader(self, downloader):
458 """Sets the downloader for this IE."""
459 self._downloader = downloader
460
461 def _real_initialize(self):
462 """Real initialization process. Redefine in subclasses."""
463 pass
464
465 def _real_extract(self, url):
466 """Real extraction process. Redefine in subclasses."""
467 pass
468
469 @classmethod
470 def ie_key(cls):
471 """A string for getting the InfoExtractor with get_info_extractor"""
472 return compat_str(cls.__name__[:-2])
473
474 @property
475 def IE_NAME(self):
476 return compat_str(type(self).__name__[:-2])
477
478 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
479 """ Returns the response handle """
480 if note is None:
481 self.report_download_webpage(video_id)
482 elif note is not False:
483 if video_id is None:
484 self.to_screen('%s' % (note,))
485 else:
486 self.to_screen('%s: %s' % (video_id, note))
487 if isinstance(url_or_request, compat_urllib_request.Request):
488 url_or_request = update_Request(
489 url_or_request, data=data, headers=headers, query=query)
490 else:
491 if query:
492 url_or_request = update_url_query(url_or_request, query)
493 if data is not None or headers:
494 url_or_request = sanitized_Request(url_or_request, data, headers)
495 try:
496 return self._downloader.urlopen(url_or_request)
497 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
498 if errnote is False:
499 return False
500 if errnote is None:
501 errnote = 'Unable to download webpage'
502
503 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
504 if fatal:
505 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
506 else:
507 self._downloader.report_warning(errmsg)
508 return False
509
510 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
511 """ Returns a tuple (page content as string, URL handle) """
512 # Strip hashes from the URL (#1038)
513 if isinstance(url_or_request, (compat_str, str)):
514 url_or_request = url_or_request.partition('#')[0]
515
516 # Some sites check X-Forwarded-For HTTP header in order to figure out
517 # the origin of the client behind proxy. This allows bypassing geo
518 # restriction by faking this header's value to IP that belongs to some
519 # geo unrestricted country. We will do so once we encounter any
520 # geo restriction error.
521 if self._x_forwarded_for_ip:
522 if 'X-Forwarded-For' not in headers:
523 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
524
525 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
526 if urlh is False:
527 assert not fatal
528 return False
529 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
530 return (content, urlh)
531
532 @staticmethod
533 def _guess_encoding_from_content(content_type, webpage_bytes):
534 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
535 if m:
536 encoding = m.group(1)
537 else:
538 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
539 webpage_bytes[:1024])
540 if m:
541 encoding = m.group(1).decode('ascii')
542 elif webpage_bytes.startswith(b'\xff\xfe'):
543 encoding = 'utf-16'
544 else:
545 encoding = 'utf-8'
546
547 return encoding
548
549 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
550 content_type = urlh.headers.get('Content-Type', '')
551 webpage_bytes = urlh.read()
552 if prefix is not None:
553 webpage_bytes = prefix + webpage_bytes
554 if not encoding:
555 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
556 if self._downloader.params.get('dump_intermediate_pages', False):
557 try:
558 url = url_or_request.get_full_url()
559 except AttributeError:
560 url = url_or_request
561 self.to_screen('Dumping request to ' + url)
562 dump = base64.b64encode(webpage_bytes).decode('ascii')
563 self._downloader.to_screen(dump)
564 if self._downloader.params.get('write_pages', False):
565 try:
566 url = url_or_request.get_full_url()
567 except AttributeError:
568 url = url_or_request
569 basen = '%s_%s' % (video_id, url)
570 if len(basen) > 240:
571 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
572 basen = basen[:240 - len(h)] + h
573 raw_filename = basen + '.dump'
574 filename = sanitize_filename(raw_filename, restricted=True)
575 self.to_screen('Saving request to ' + filename)
576 # Working around MAX_PATH limitation on Windows (see
577 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
578 if compat_os_name == 'nt':
579 absfilepath = os.path.abspath(filename)
580 if len(absfilepath) > 259:
581 filename = '\\\\?\\' + absfilepath
582 with open(filename, 'wb') as outf:
583 outf.write(webpage_bytes)
584
585 try:
586 content = webpage_bytes.decode(encoding, 'replace')
587 except LookupError:
588 content = webpage_bytes.decode('utf-8', 'replace')
589
590 if ('<title>Access to this site is blocked</title>' in content and
591 'Websense' in content[:512]):
592 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
593 blocked_iframe = self._html_search_regex(
594 r'<iframe src="([^"]+)"', content,
595 'Websense information URL', default=None)
596 if blocked_iframe:
597 msg += ' Visit %s for more details' % blocked_iframe
598 raise ExtractorError(msg, expected=True)
599 if '<title>The URL you requested has been blocked</title>' in content[:512]:
600 msg = (
601 'Access to this webpage has been blocked by Indian censorship. '
602 'Use a VPN or proxy server (with --proxy) to route around it.')
603 block_msg = self._html_search_regex(
604 r'</h1><p>(.*?)</p>',
605 content, 'block message', default=None)
606 if block_msg:
607 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
608 raise ExtractorError(msg, expected=True)
609
610 return content
611
612 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}):
613 """ Returns the data of the page as a string """
614 success = False
615 try_count = 0
616 while success is False:
617 try:
618 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
619 success = True
620 except compat_http_client.IncompleteRead as e:
621 try_count += 1
622 if try_count >= tries:
623 raise e
624 self._sleep(timeout, video_id)
625 if res is False:
626 return res
627 else:
628 content, _ = res
629 return content
630
631 def _download_xml(self, url_or_request, video_id,
632 note='Downloading XML', errnote='Unable to download XML',
633 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
634 """Return the xml as an xml.etree.ElementTree.Element"""
635 xml_string = self._download_webpage(
636 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
637 if xml_string is False:
638 return xml_string
639 if transform_source:
640 xml_string = transform_source(xml_string)
641 return compat_etree_fromstring(xml_string.encode('utf-8'))
642
643 def _download_json(self, url_or_request, video_id,
644 note='Downloading JSON metadata',
645 errnote='Unable to download JSON metadata',
646 transform_source=None,
647 fatal=True, encoding=None, data=None, headers={}, query={}):
648 json_string = self._download_webpage(
649 url_or_request, video_id, note, errnote, fatal=fatal,
650 encoding=encoding, data=data, headers=headers, query=query)
651 if (not fatal) and json_string is False:
652 return None
653 return self._parse_json(
654 json_string, video_id, transform_source=transform_source, fatal=fatal)
655
656 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
657 if transform_source:
658 json_string = transform_source(json_string)
659 try:
660 return json.loads(json_string)
661 except ValueError as ve:
662 errmsg = '%s: Failed to parse JSON ' % video_id
663 if fatal:
664 raise ExtractorError(errmsg, cause=ve)
665 else:
666 self.report_warning(errmsg + str(ve))
667
668 def report_warning(self, msg, video_id=None):
669 idstr = '' if video_id is None else '%s: ' % video_id
670 self._downloader.report_warning(
671 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
672
673 def to_screen(self, msg):
674 """Print msg to screen, prefixing it with '[ie_name]'"""
675 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
676
677 def report_extraction(self, id_or_name):
678 """Report information extraction."""
679 self.to_screen('%s: Extracting information' % id_or_name)
680
681 def report_download_webpage(self, video_id):
682 """Report webpage download."""
683 self.to_screen('%s: Downloading webpage' % video_id)
684
685 def report_age_confirmation(self):
686 """Report attempt to confirm age."""
687 self.to_screen('Confirming age')
688
689 def report_login(self):
690 """Report attempt to log in."""
691 self.to_screen('Logging in')
692
693 @staticmethod
694 def raise_login_required(msg='This video is only available for registered users'):
695 raise ExtractorError(
696 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
697 expected=True)
698
699 @staticmethod
700 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
701 raise GeoRestrictedError(msg, countries=countries)
702
703 # Methods for following #608
704 @staticmethod
705 def url_result(url, ie=None, video_id=None, video_title=None):
706 """Returns a URL that points to a page that should be processed"""
707 # TODO: ie should be the class used for getting the info
708 video_info = {'_type': 'url',
709 'url': url,
710 'ie_key': ie}
711 if video_id is not None:
712 video_info['id'] = video_id
713 if video_title is not None:
714 video_info['title'] = video_title
715 return video_info
716
717 @staticmethod
718 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
719 """Returns a playlist"""
720 video_info = {'_type': 'playlist',
721 'entries': entries}
722 if playlist_id:
723 video_info['id'] = playlist_id
724 if playlist_title:
725 video_info['title'] = playlist_title
726 if playlist_description:
727 video_info['description'] = playlist_description
728 return video_info
729
730 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
731 """
732 Perform a regex search on the given string, using a single or a list of
733 patterns returning the first matching group.
734 In case of failure return a default value or raise a WARNING or a
735 RegexNotFoundError, depending on fatal, specifying the field name.
736 """
737 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
738 mobj = re.search(pattern, string, flags)
739 else:
740 for p in pattern:
741 mobj = re.search(p, string, flags)
742 if mobj:
743 break
744
745 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
746 _name = '\033[0;34m%s\033[0m' % name
747 else:
748 _name = name
749
750 if mobj:
751 if group is None:
752 # return the first matching group
753 return next(g for g in mobj.groups() if g is not None)
754 else:
755 return mobj.group(group)
756 elif default is not NO_DEFAULT:
757 return default
758 elif fatal:
759 raise RegexNotFoundError('Unable to extract %s' % _name)
760 else:
761 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
762 return None
763
764 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
765 """
766 Like _search_regex, but strips HTML tags and unescapes entities.
767 """
768 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
769 if res:
770 return clean_html(res).strip()
771 else:
772 return res
773
774 def _get_netrc_login_info(self, netrc_machine=None):
775 username = None
776 password = None
777 netrc_machine = netrc_machine or self._NETRC_MACHINE
778
779 if self._downloader.params.get('usenetrc', False):
780 try:
781 info = netrc.netrc().authenticators(netrc_machine)
782 if info is not None:
783 username = info[0]
784 password = info[2]
785 else:
786 raise netrc.NetrcParseError(
787 'No authenticators for %s' % netrc_machine)
788 except (IOError, netrc.NetrcParseError) as err:
789 self._downloader.report_warning(
790 'parsing .netrc: %s' % error_to_compat_str(err))
791
792 return username, password
793
794 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
795 """
796 Get the login info as (username, password)
797 First look for the manually specified credentials using username_option
798 and password_option as keys in params dictionary. If no such credentials
799 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
800 value.
801 If there's no info available, return (None, None)
802 """
803 if self._downloader is None:
804 return (None, None)
805
806 downloader_params = self._downloader.params
807
808 # Attempt to use provided username and password or .netrc data
809 if downloader_params.get(username_option) is not None:
810 username = downloader_params[username_option]
811 password = downloader_params[password_option]
812 else:
813 username, password = self._get_netrc_login_info(netrc_machine)
814
815 return username, password
816
817 def _get_tfa_info(self, note='two-factor verification code'):
818 """
819 Get the two-factor authentication info
820 TODO - asking the user will be required for sms/phone verify
821 currently just uses the command line option
822 If there's no info available, return None
823 """
824 if self._downloader is None:
825 return None
826 downloader_params = self._downloader.params
827
828 if downloader_params.get('twofactor') is not None:
829 return downloader_params['twofactor']
830
831 return compat_getpass('Type %s and press [Return]: ' % note)
832
833 # Helper functions for extracting OpenGraph info
834 @staticmethod
835 def _og_regexes(prop):
836 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
837 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
838 % {'prop': re.escape(prop)})
839 template = r'<meta[^>]+?%s[^>]+?%s'
840 return [
841 template % (property_re, content_re),
842 template % (content_re, property_re),
843 ]
844
845 @staticmethod
846 def _meta_regex(prop):
847 return r'''(?isx)<meta
848 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
849 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
850
851 def _og_search_property(self, prop, html, name=None, **kargs):
852 if not isinstance(prop, (list, tuple)):
853 prop = [prop]
854 if name is None:
855 name = 'OpenGraph %s' % prop[0]
856 og_regexes = []
857 for p in prop:
858 og_regexes.extend(self._og_regexes(p))
859 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
860 if escaped is None:
861 return None
862 return unescapeHTML(escaped)
863
864 def _og_search_thumbnail(self, html, **kargs):
865 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
866
867 def _og_search_description(self, html, **kargs):
868 return self._og_search_property('description', html, fatal=False, **kargs)
869
870 def _og_search_title(self, html, **kargs):
871 return self._og_search_property('title', html, **kargs)
872
873 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
874 regexes = self._og_regexes('video') + self._og_regexes('video:url')
875 if secure:
876 regexes = self._og_regexes('video:secure_url') + regexes
877 return self._html_search_regex(regexes, html, name, **kargs)
878
879 def _og_search_url(self, html, **kargs):
880 return self._og_search_property('url', html, **kargs)
881
882 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
883 if not isinstance(name, (list, tuple)):
884 name = [name]
885 if display_name is None:
886 display_name = name[0]
887 return self._html_search_regex(
888 [self._meta_regex(n) for n in name],
889 html, display_name, fatal=fatal, group='content', **kwargs)
890
891 def _dc_search_uploader(self, html):
892 return self._html_search_meta('dc.creator', html, 'uploader')
893
894 def _rta_search(self, html):
895 # See http://www.rtalabel.org/index.php?content=howtofaq#single
896 if re.search(r'(?ix)<meta\s+name="rating"\s+'
897 r' content="RTA-5042-1996-1400-1577-RTA"',
898 html):
899 return 18
900 return 0
901
902 def _media_rating_search(self, html):
903 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
904 rating = self._html_search_meta('rating', html)
905
906 if not rating:
907 return None
908
909 RATING_TABLE = {
910 'safe for kids': 0,
911 'general': 8,
912 '14 years': 14,
913 'mature': 17,
914 'restricted': 19,
915 }
916 return RATING_TABLE.get(rating.lower())
917
918 def _family_friendly_search(self, html):
919 # See http://schema.org/VideoObject
920 family_friendly = self._html_search_meta('isFamilyFriendly', html)
921
922 if not family_friendly:
923 return None
924
925 RATING_TABLE = {
926 '1': 0,
927 'true': 0,
928 '0': 18,
929 'false': 18,
930 }
931 return RATING_TABLE.get(family_friendly.lower())
932
933 def _twitter_search_player(self, html):
934 return self._html_search_meta('twitter:player', html,
935 'twitter card player')
936
937 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
938 json_ld = self._search_regex(
939 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
940 html, 'JSON-LD', group='json_ld', **kwargs)
941 default = kwargs.get('default', NO_DEFAULT)
942 if not json_ld:
943 return default if default is not NO_DEFAULT else {}
944 # JSON-LD may be malformed and thus `fatal` should be respected.
945 # At the same time `default` may be passed that assumes `fatal=False`
946 # for _search_regex. Let's simulate the same behavior here as well.
947 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
948 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
949
950 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
951 if isinstance(json_ld, compat_str):
952 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
953 if not json_ld:
954 return {}
955 info = {}
956 if not isinstance(json_ld, (list, tuple, dict)):
957 return info
958 if isinstance(json_ld, dict):
959 json_ld = [json_ld]
960 for e in json_ld:
961 if e.get('@context') == 'http://schema.org':
962 item_type = e.get('@type')
963 if expected_type is not None and expected_type != item_type:
964 return info
965 if item_type == 'TVEpisode':
966 info.update({
967 'episode': unescapeHTML(e.get('name')),
968 'episode_number': int_or_none(e.get('episodeNumber')),
969 'description': unescapeHTML(e.get('description')),
970 })
971 part_of_season = e.get('partOfSeason')
972 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
973 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
974 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
975 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
976 info['series'] = unescapeHTML(part_of_series.get('name'))
977 elif item_type == 'Article':
978 info.update({
979 'timestamp': parse_iso8601(e.get('datePublished')),
980 'title': unescapeHTML(e.get('headline')),
981 'description': unescapeHTML(e.get('articleBody')),
982 })
983 elif item_type == 'VideoObject':
984 info.update({
985 'url': e.get('contentUrl'),
986 'title': unescapeHTML(e.get('name')),
987 'description': unescapeHTML(e.get('description')),
988 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
989 'duration': parse_duration(e.get('duration')),
990 'timestamp': unified_timestamp(e.get('uploadDate')),
991 'filesize': float_or_none(e.get('contentSize')),
992 'tbr': int_or_none(e.get('bitrate')),
993 'width': int_or_none(e.get('width')),
994 'height': int_or_none(e.get('height')),
995 })
996 break
997 return dict((k, v) for k, v in info.items() if v is not None)
998
999 @staticmethod
1000 def _hidden_inputs(html):
1001 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1002 hidden_inputs = {}
1003 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1004 attrs = extract_attributes(input)
1005 if not input:
1006 continue
1007 if attrs.get('type') not in ('hidden', 'submit'):
1008 continue
1009 name = attrs.get('name') or attrs.get('id')
1010 value = attrs.get('value')
1011 if name and value is not None:
1012 hidden_inputs[name] = value
1013 return hidden_inputs
1014
1015 def _form_hidden_inputs(self, form_id, html):
1016 form = self._search_regex(
1017 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1018 html, '%s form' % form_id, group='form')
1019 return self._hidden_inputs(form)
1020
1021 def _sort_formats(self, formats, field_preference=None):
1022 if not formats:
1023 raise ExtractorError('No video formats found')
1024
1025 for f in formats:
1026 # Automatically determine tbr when missing based on abr and vbr (improves
1027 # formats sorting in some cases)
1028 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1029 f['tbr'] = f['abr'] + f['vbr']
1030
1031 def _formats_key(f):
1032 # TODO remove the following workaround
1033 from ..utils import determine_ext
1034 if not f.get('ext') and 'url' in f:
1035 f['ext'] = determine_ext(f['url'])
1036
1037 if isinstance(field_preference, (list, tuple)):
1038 return tuple(
1039 f.get(field)
1040 if f.get(field) is not None
1041 else ('' if field == 'format_id' else -1)
1042 for field in field_preference)
1043
1044 preference = f.get('preference')
1045 if preference is None:
1046 preference = 0
1047 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1048 preference -= 0.5
1049
1050 protocol = f.get('protocol') or determine_protocol(f)
1051 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1052
1053 if f.get('vcodec') == 'none': # audio only
1054 preference -= 50
1055 if self._downloader.params.get('prefer_free_formats'):
1056 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1057 else:
1058 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1059 ext_preference = 0
1060 try:
1061 audio_ext_preference = ORDER.index(f['ext'])
1062 except ValueError:
1063 audio_ext_preference = -1
1064 else:
1065 if f.get('acodec') == 'none': # video only
1066 preference -= 40
1067 if self._downloader.params.get('prefer_free_formats'):
1068 ORDER = ['flv', 'mp4', 'webm']
1069 else:
1070 ORDER = ['webm', 'flv', 'mp4']
1071 try:
1072 ext_preference = ORDER.index(f['ext'])
1073 except ValueError:
1074 ext_preference = -1
1075 audio_ext_preference = 0
1076
1077 return (
1078 preference,
1079 f.get('language_preference') if f.get('language_preference') is not None else -1,
1080 f.get('quality') if f.get('quality') is not None else -1,
1081 f.get('tbr') if f.get('tbr') is not None else -1,
1082 f.get('filesize') if f.get('filesize') is not None else -1,
1083 f.get('vbr') if f.get('vbr') is not None else -1,
1084 f.get('height') if f.get('height') is not None else -1,
1085 f.get('width') if f.get('width') is not None else -1,
1086 proto_preference,
1087 ext_preference,
1088 f.get('abr') if f.get('abr') is not None else -1,
1089 audio_ext_preference,
1090 f.get('fps') if f.get('fps') is not None else -1,
1091 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1092 f.get('source_preference') if f.get('source_preference') is not None else -1,
1093 f.get('format_id') if f.get('format_id') is not None else '',
1094 )
1095 formats.sort(key=_formats_key)
1096
1097 def _check_formats(self, formats, video_id):
1098 if formats:
1099 formats[:] = filter(
1100 lambda f: self._is_valid_url(
1101 f['url'], video_id,
1102 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1103 formats)
1104
1105 @staticmethod
1106 def _remove_duplicate_formats(formats):
1107 format_urls = set()
1108 unique_formats = []
1109 for f in formats:
1110 if f['url'] not in format_urls:
1111 format_urls.add(f['url'])
1112 unique_formats.append(f)
1113 formats[:] = unique_formats
1114
1115 def _is_valid_url(self, url, video_id, item='video', headers={}):
1116 url = self._proto_relative_url(url, scheme='http:')
1117 # For now assume non HTTP(S) URLs always valid
1118 if not (url.startswith('http://') or url.startswith('https://')):
1119 return True
1120 try:
1121 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1122 return True
1123 except ExtractorError as e:
1124 if isinstance(e.cause, compat_urllib_error.URLError):
1125 self.to_screen(
1126 '%s: %s URL is invalid, skipping' % (video_id, item))
1127 return False
1128 raise
1129
1130 def http_scheme(self):
1131 """ Either "http:" or "https:", depending on the user's preferences """
1132 return (
1133 'http:'
1134 if self._downloader.params.get('prefer_insecure', False)
1135 else 'https:')
1136
1137 def _proto_relative_url(self, url, scheme=None):
1138 if url is None:
1139 return url
1140 if url.startswith('//'):
1141 if scheme is None:
1142 scheme = self.http_scheme()
1143 return scheme + url
1144 else:
1145 return url
1146
1147 def _sleep(self, timeout, video_id, msg_template=None):
1148 if msg_template is None:
1149 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1150 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1151 self.to_screen(msg)
1152 time.sleep(timeout)
1153
1154 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1155 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1156 fatal=True, m3u8_id=None):
1157 manifest = self._download_xml(
1158 manifest_url, video_id, 'Downloading f4m manifest',
1159 'Unable to download f4m manifest',
1160 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1161 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1162 transform_source=transform_source,
1163 fatal=fatal)
1164
1165 if manifest is False:
1166 return []
1167
1168 return self._parse_f4m_formats(
1169 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1170 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1171
1172 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1173 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1174 fatal=True, m3u8_id=None):
1175 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1176 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1177 if akamai_pv is not None and ';' in akamai_pv.text:
1178 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1179 if playerVerificationChallenge.strip() != '':
1180 return []
1181
1182 formats = []
1183 manifest_version = '1.0'
1184 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1185 if not media_nodes:
1186 manifest_version = '2.0'
1187 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1188 # Remove unsupported DRM protected media from final formats
1189 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1190 media_nodes = remove_encrypted_media(media_nodes)
1191 if not media_nodes:
1192 return formats
1193 base_url = xpath_text(
1194 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1195 'base URL', default=None)
1196 if base_url:
1197 base_url = base_url.strip()
1198
1199 bootstrap_info = xpath_element(
1200 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1201 'bootstrap info', default=None)
1202
1203 vcodec = None
1204 mime_type = xpath_text(
1205 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1206 'base URL', default=None)
1207 if mime_type and mime_type.startswith('audio/'):
1208 vcodec = 'none'
1209
1210 for i, media_el in enumerate(media_nodes):
1211 tbr = int_or_none(media_el.attrib.get('bitrate'))
1212 width = int_or_none(media_el.attrib.get('width'))
1213 height = int_or_none(media_el.attrib.get('height'))
1214 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1215 # If <bootstrapInfo> is present, the specified f4m is a
1216 # stream-level manifest, and only set-level manifests may refer to
1217 # external resources. See section 11.4 and section 4 of F4M spec
1218 if bootstrap_info is None:
1219 media_url = None
1220 # @href is introduced in 2.0, see section 11.6 of F4M spec
1221 if manifest_version == '2.0':
1222 media_url = media_el.attrib.get('href')
1223 if media_url is None:
1224 media_url = media_el.attrib.get('url')
1225 if not media_url:
1226 continue
1227 manifest_url = (
1228 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1229 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1230 # If media_url is itself a f4m manifest do the recursive extraction
1231 # since bitrates in parent manifest (this one) and media_url manifest
1232 # may differ leading to inability to resolve the format by requested
1233 # bitrate in f4m downloader
1234 ext = determine_ext(manifest_url)
1235 if ext == 'f4m':
1236 f4m_formats = self._extract_f4m_formats(
1237 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1238 transform_source=transform_source, fatal=fatal)
1239 # Sometimes stream-level manifest contains single media entry that
1240 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1241 # At the same time parent's media entry in set-level manifest may
1242 # contain it. We will copy it from parent in such cases.
1243 if len(f4m_formats) == 1:
1244 f = f4m_formats[0]
1245 f.update({
1246 'tbr': f.get('tbr') or tbr,
1247 'width': f.get('width') or width,
1248 'height': f.get('height') or height,
1249 'format_id': f.get('format_id') if not tbr else format_id,
1250 'vcodec': vcodec,
1251 })
1252 formats.extend(f4m_formats)
1253 continue
1254 elif ext == 'm3u8':
1255 formats.extend(self._extract_m3u8_formats(
1256 manifest_url, video_id, 'mp4', preference=preference,
1257 m3u8_id=m3u8_id, fatal=fatal))
1258 continue
1259 formats.append({
1260 'format_id': format_id,
1261 'url': manifest_url,
1262 'manifest_url': manifest_url,
1263 'ext': 'flv' if bootstrap_info is not None else None,
1264 'tbr': tbr,
1265 'width': width,
1266 'height': height,
1267 'vcodec': vcodec,
1268 'preference': preference,
1269 })
1270 return formats
1271
1272 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1273 return {
1274 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1275 'url': m3u8_url,
1276 'ext': ext,
1277 'protocol': 'm3u8',
1278 'preference': preference - 100 if preference else -100,
1279 'resolution': 'multiple',
1280 'format_note': 'Quality selection URL',
1281 }
1282
1283 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1284 entry_protocol='m3u8', preference=None,
1285 m3u8_id=None, note=None, errnote=None,
1286 fatal=True, live=False):
1287
1288 res = self._download_webpage_handle(
1289 m3u8_url, video_id,
1290 note=note or 'Downloading m3u8 information',
1291 errnote=errnote or 'Failed to download m3u8 information',
1292 fatal=fatal)
1293 if res is False:
1294 return []
1295 m3u8_doc, urlh = res
1296 m3u8_url = urlh.geturl()
1297
1298 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1299 return []
1300
1301 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1302
1303 format_url = lambda u: (
1304 u
1305 if re.match(r'^https?://', u)
1306 else compat_urlparse.urljoin(m3u8_url, u))
1307
1308 # We should try extracting formats only from master playlists [1], i.e.
1309 # playlists that describe available qualities. On the other hand media
1310 # playlists [2] should be returned as is since they contain just the media
1311 # without qualities renditions.
1312 # Fortunately, master playlist can be easily distinguished from media
1313 # playlist based on particular tags availability. As of [1, 2] master
1314 # playlist tags MUST NOT appear in a media playist and vice versa.
1315 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1316 # and MUST NOT appear in master playlist thus we can clearly detect media
1317 # playlist with this criterion.
1318 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1319 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1320 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1321 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1322 return [{
1323 'url': m3u8_url,
1324 'format_id': m3u8_id,
1325 'ext': ext,
1326 'protocol': entry_protocol,
1327 'preference': preference,
1328 }]
1329 audio_in_video_stream = {}
1330 last_info = {}
1331 last_media = {}
1332 for line in m3u8_doc.splitlines():
1333 if line.startswith('#EXT-X-STREAM-INF:'):
1334 last_info = parse_m3u8_attributes(line)
1335 elif line.startswith('#EXT-X-MEDIA:'):
1336 media = parse_m3u8_attributes(line)
1337 media_type = media.get('TYPE')
1338 if media_type in ('VIDEO', 'AUDIO'):
1339 group_id = media.get('GROUP-ID')
1340 media_url = media.get('URI')
1341 if media_url:
1342 format_id = []
1343 for v in (group_id, media.get('NAME')):
1344 if v:
1345 format_id.append(v)
1346 f = {
1347 'format_id': '-'.join(format_id),
1348 'url': format_url(media_url),
1349 'language': media.get('LANGUAGE'),
1350 'ext': ext,
1351 'protocol': entry_protocol,
1352 'preference': preference,
1353 }
1354 if media_type == 'AUDIO':
1355 f['vcodec'] = 'none'
1356 if group_id and not audio_in_video_stream.get(group_id):
1357 audio_in_video_stream[group_id] = False
1358 formats.append(f)
1359 else:
1360 # When there is no URI in EXT-X-MEDIA let this tag's
1361 # data be used by regular URI lines below
1362 last_media = media
1363 if media_type == 'AUDIO' and group_id:
1364 audio_in_video_stream[group_id] = True
1365 elif line.startswith('#') or not line.strip():
1366 continue
1367 else:
1368 tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000)
1369 format_id = []
1370 if m3u8_id:
1371 format_id.append(m3u8_id)
1372 # Despite specification does not mention NAME attribute for
1373 # EXT-X-STREAM-INF it still sometimes may be present
1374 stream_name = last_info.get('NAME') or last_media.get('NAME')
1375 # Bandwidth of live streams may differ over time thus making
1376 # format_id unpredictable. So it's better to keep provided
1377 # format_id intact.
1378 if not live:
1379 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1380 manifest_url = format_url(line.strip())
1381 f = {
1382 'format_id': '-'.join(format_id),
1383 'url': manifest_url,
1384 'manifest_url': manifest_url,
1385 'tbr': tbr,
1386 'ext': ext,
1387 'fps': float_or_none(last_info.get('FRAME-RATE')),
1388 'protocol': entry_protocol,
1389 'preference': preference,
1390 }
1391 resolution = last_info.get('RESOLUTION')
1392 if resolution:
1393 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1394 if mobj:
1395 f['width'] = int(mobj.group('width'))
1396 f['height'] = int(mobj.group('height'))
1397 # Unified Streaming Platform
1398 mobj = re.search(
1399 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1400 if mobj:
1401 abr, vbr = mobj.groups()
1402 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1403 f.update({
1404 'vbr': vbr,
1405 'abr': abr,
1406 })
1407 f.update(parse_codecs(last_info.get('CODECS')))
1408 if audio_in_video_stream.get(last_info.get('AUDIO')) is False and f['vcodec'] != 'none':
1409 # TODO: update acodec for audio only formats with the same GROUP-ID
1410 f['acodec'] = 'none'
1411 formats.append(f)
1412 last_info = {}
1413 last_media = {}
1414 return formats
1415
1416 @staticmethod
1417 def _xpath_ns(path, namespace=None):
1418 if not namespace:
1419 return path
1420 out = []
1421 for c in path.split('/'):
1422 if not c or c == '.':
1423 out.append(c)
1424 else:
1425 out.append('{%s}%s' % (namespace, c))
1426 return '/'.join(out)
1427
1428 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1429 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1430
1431 if smil is False:
1432 assert not fatal
1433 return []
1434
1435 namespace = self._parse_smil_namespace(smil)
1436
1437 return self._parse_smil_formats(
1438 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1439
1440 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1441 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1442 if smil is False:
1443 return {}
1444 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1445
1446 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1447 return self._download_xml(
1448 smil_url, video_id, 'Downloading SMIL file',
1449 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1450
1451 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1452 namespace = self._parse_smil_namespace(smil)
1453
1454 formats = self._parse_smil_formats(
1455 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1456 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1457
1458 video_id = os.path.splitext(url_basename(smil_url))[0]
1459 title = None
1460 description = None
1461 upload_date = None
1462 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1463 name = meta.attrib.get('name')
1464 content = meta.attrib.get('content')
1465 if not name or not content:
1466 continue
1467 if not title and name == 'title':
1468 title = content
1469 elif not description and name in ('description', 'abstract'):
1470 description = content
1471 elif not upload_date and name == 'date':
1472 upload_date = unified_strdate(content)
1473
1474 thumbnails = [{
1475 'id': image.get('type'),
1476 'url': image.get('src'),
1477 'width': int_or_none(image.get('width')),
1478 'height': int_or_none(image.get('height')),
1479 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1480
1481 return {
1482 'id': video_id,
1483 'title': title or video_id,
1484 'description': description,
1485 'upload_date': upload_date,
1486 'thumbnails': thumbnails,
1487 'formats': formats,
1488 'subtitles': subtitles,
1489 }
1490
1491 def _parse_smil_namespace(self, smil):
1492 return self._search_regex(
1493 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1494
1495 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1496 base = smil_url
1497 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1498 b = meta.get('base') or meta.get('httpBase')
1499 if b:
1500 base = b
1501 break
1502
1503 formats = []
1504 rtmp_count = 0
1505 http_count = 0
1506 m3u8_count = 0
1507
1508 srcs = []
1509 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1510 for medium in media:
1511 src = medium.get('src')
1512 if not src or src in srcs:
1513 continue
1514 srcs.append(src)
1515
1516 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1517 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1518 width = int_or_none(medium.get('width'))
1519 height = int_or_none(medium.get('height'))
1520 proto = medium.get('proto')
1521 ext = medium.get('ext')
1522 src_ext = determine_ext(src)
1523 streamer = medium.get('streamer') or base
1524
1525 if proto == 'rtmp' or streamer.startswith('rtmp'):
1526 rtmp_count += 1
1527 formats.append({
1528 'url': streamer,
1529 'play_path': src,
1530 'ext': 'flv',
1531 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1532 'tbr': bitrate,
1533 'filesize': filesize,
1534 'width': width,
1535 'height': height,
1536 })
1537 if transform_rtmp_url:
1538 streamer, src = transform_rtmp_url(streamer, src)
1539 formats[-1].update({
1540 'url': streamer,
1541 'play_path': src,
1542 })
1543 continue
1544
1545 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1546 src_url = src_url.strip()
1547
1548 if proto == 'm3u8' or src_ext == 'm3u8':
1549 m3u8_formats = self._extract_m3u8_formats(
1550 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1551 if len(m3u8_formats) == 1:
1552 m3u8_count += 1
1553 m3u8_formats[0].update({
1554 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1555 'tbr': bitrate,
1556 'width': width,
1557 'height': height,
1558 })
1559 formats.extend(m3u8_formats)
1560 continue
1561
1562 if src_ext == 'f4m':
1563 f4m_url = src_url
1564 if not f4m_params:
1565 f4m_params = {
1566 'hdcore': '3.2.0',
1567 'plugin': 'flowplayer-3.2.0.1',
1568 }
1569 f4m_url += '&' if '?' in f4m_url else '?'
1570 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1571 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1572 continue
1573
1574 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1575 http_count += 1
1576 formats.append({
1577 'url': src_url,
1578 'ext': ext or src_ext or 'flv',
1579 'format_id': 'http-%d' % (bitrate or http_count),
1580 'tbr': bitrate,
1581 'filesize': filesize,
1582 'width': width,
1583 'height': height,
1584 })
1585 continue
1586
1587 return formats
1588
1589 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1590 urls = []
1591 subtitles = {}
1592 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1593 src = textstream.get('src')
1594 if not src or src in urls:
1595 continue
1596 urls.append(src)
1597 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1598 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1599 subtitles.setdefault(lang, []).append({
1600 'url': src,
1601 'ext': ext,
1602 })
1603 return subtitles
1604
1605 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1606 xspf = self._download_xml(
1607 playlist_url, playlist_id, 'Downloading xpsf playlist',
1608 'Unable to download xspf manifest', fatal=fatal)
1609 if xspf is False:
1610 return []
1611 return self._parse_xspf(xspf, playlist_id)
1612
1613 def _parse_xspf(self, playlist, playlist_id):
1614 NS_MAP = {
1615 'xspf': 'http://xspf.org/ns/0/',
1616 's1': 'http://static.streamone.nl/player/ns/0',
1617 }
1618
1619 entries = []
1620 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1621 title = xpath_text(
1622 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1623 description = xpath_text(
1624 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1625 thumbnail = xpath_text(
1626 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1627 duration = float_or_none(
1628 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1629
1630 formats = [{
1631 'url': location.text,
1632 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1633 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1634 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1635 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1636 self._sort_formats(formats)
1637
1638 entries.append({
1639 'id': playlist_id,
1640 'title': title,
1641 'description': description,
1642 'thumbnail': thumbnail,
1643 'duration': duration,
1644 'formats': formats,
1645 })
1646 return entries
1647
1648 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1649 res = self._download_webpage_handle(
1650 mpd_url, video_id,
1651 note=note or 'Downloading MPD manifest',
1652 errnote=errnote or 'Failed to download MPD manifest',
1653 fatal=fatal)
1654 if res is False:
1655 return []
1656 mpd, urlh = res
1657 mpd_base_url = base_url(urlh.geturl())
1658
1659 return self._parse_mpd_formats(
1660 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1661 formats_dict=formats_dict, mpd_url=mpd_url)
1662
1663 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1664 """
1665 Parse formats from MPD manifest.
1666 References:
1667 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1668 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1669 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1670 """
1671 if mpd_doc.get('type') == 'dynamic':
1672 return []
1673
1674 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1675
1676 def _add_ns(path):
1677 return self._xpath_ns(path, namespace)
1678
1679 def is_drm_protected(element):
1680 return element.find(_add_ns('ContentProtection')) is not None
1681
1682 def extract_multisegment_info(element, ms_parent_info):
1683 ms_info = ms_parent_info.copy()
1684
1685 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1686 # common attributes and elements. We will only extract relevant
1687 # for us.
1688 def extract_common(source):
1689 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1690 if segment_timeline is not None:
1691 s_e = segment_timeline.findall(_add_ns('S'))
1692 if s_e:
1693 ms_info['total_number'] = 0
1694 ms_info['s'] = []
1695 for s in s_e:
1696 r = int(s.get('r', 0))
1697 ms_info['total_number'] += 1 + r
1698 ms_info['s'].append({
1699 't': int(s.get('t', 0)),
1700 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1701 'd': int(s.attrib['d']),
1702 'r': r,
1703 })
1704 start_number = source.get('startNumber')
1705 if start_number:
1706 ms_info['start_number'] = int(start_number)
1707 timescale = source.get('timescale')
1708 if timescale:
1709 ms_info['timescale'] = int(timescale)
1710 segment_duration = source.get('duration')
1711 if segment_duration:
1712 ms_info['segment_duration'] = int(segment_duration)
1713
1714 def extract_Initialization(source):
1715 initialization = source.find(_add_ns('Initialization'))
1716 if initialization is not None:
1717 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1718
1719 segment_list = element.find(_add_ns('SegmentList'))
1720 if segment_list is not None:
1721 extract_common(segment_list)
1722 extract_Initialization(segment_list)
1723 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1724 if segment_urls_e:
1725 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1726 else:
1727 segment_template = element.find(_add_ns('SegmentTemplate'))
1728 if segment_template is not None:
1729 extract_common(segment_template)
1730 media = segment_template.get('media')
1731 if media:
1732 ms_info['media'] = media
1733 initialization = segment_template.get('initialization')
1734 if initialization:
1735 ms_info['initialization'] = initialization
1736 else:
1737 extract_Initialization(segment_template)
1738 return ms_info
1739
1740 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1741 formats = []
1742 for period in mpd_doc.findall(_add_ns('Period')):
1743 period_duration = parse_duration(period.get('duration')) or mpd_duration
1744 period_ms_info = extract_multisegment_info(period, {
1745 'start_number': 1,
1746 'timescale': 1,
1747 })
1748 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1749 if is_drm_protected(adaptation_set):
1750 continue
1751 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1752 for representation in adaptation_set.findall(_add_ns('Representation')):
1753 if is_drm_protected(representation):
1754 continue
1755 representation_attrib = adaptation_set.attrib.copy()
1756 representation_attrib.update(representation.attrib)
1757 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1758 mime_type = representation_attrib['mimeType']
1759 content_type = mime_type.split('/')[0]
1760 if content_type == 'text':
1761 # TODO implement WebVTT downloading
1762 pass
1763 elif content_type == 'video' or content_type == 'audio':
1764 base_url = ''
1765 for element in (representation, adaptation_set, period, mpd_doc):
1766 base_url_e = element.find(_add_ns('BaseURL'))
1767 if base_url_e is not None:
1768 base_url = base_url_e.text + base_url
1769 if re.match(r'^https?://', base_url):
1770 break
1771 if mpd_base_url and not re.match(r'^https?://', base_url):
1772 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1773 mpd_base_url += '/'
1774 base_url = mpd_base_url + base_url
1775 representation_id = representation_attrib.get('id')
1776 lang = representation_attrib.get('lang')
1777 url_el = representation.find(_add_ns('BaseURL'))
1778 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1779 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1780 f = {
1781 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1782 'url': base_url,
1783 'manifest_url': mpd_url,
1784 'ext': mimetype2ext(mime_type),
1785 'width': int_or_none(representation_attrib.get('width')),
1786 'height': int_or_none(representation_attrib.get('height')),
1787 'tbr': int_or_none(bandwidth, 1000),
1788 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1789 'fps': int_or_none(representation_attrib.get('frameRate')),
1790 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1791 'format_note': 'DASH %s' % content_type,
1792 'filesize': filesize,
1793 }
1794 f.update(parse_codecs(representation_attrib.get('codecs')))
1795 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1796
1797 def prepare_template(template_name, identifiers):
1798 t = representation_ms_info[template_name]
1799 t = t.replace('$RepresentationID$', representation_id)
1800 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1801 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1802 t.replace('$$', '$')
1803 return t
1804
1805 # @initialization is a regular template like @media one
1806 # so it should be handled just the same way (see
1807 # https://github.com/rg3/youtube-dl/issues/11605)
1808 if 'initialization' in representation_ms_info:
1809 initialization_template = prepare_template(
1810 'initialization',
1811 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1812 # $Time$ shall not be included for @initialization thus
1813 # only $Bandwidth$ remains
1814 ('Bandwidth', ))
1815 representation_ms_info['initialization_url'] = initialization_template % {
1816 'Bandwidth': bandwidth,
1817 }
1818
1819 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1820
1821 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1822
1823 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1824 # can't be used at the same time
1825 if '%(Number' in media_template and 's' not in representation_ms_info:
1826 segment_duration = None
1827 if 'total_number' not in representation_ms_info and 'segment_duration':
1828 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1829 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1830 representation_ms_info['fragments'] = [{
1831 'url': media_template % {
1832 'Number': segment_number,
1833 'Bandwidth': bandwidth,
1834 },
1835 'duration': segment_duration,
1836 } for segment_number in range(
1837 representation_ms_info['start_number'],
1838 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1839 else:
1840 # $Number*$ or $Time$ in media template with S list available
1841 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1842 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1843 representation_ms_info['fragments'] = []
1844 segment_time = 0
1845 segment_d = None
1846 segment_number = representation_ms_info['start_number']
1847
1848 def add_segment_url():
1849 segment_url = media_template % {
1850 'Time': segment_time,
1851 'Bandwidth': bandwidth,
1852 'Number': segment_number,
1853 }
1854 representation_ms_info['fragments'].append({
1855 'url': segment_url,
1856 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1857 })
1858
1859 for num, s in enumerate(representation_ms_info['s']):
1860 segment_time = s.get('t') or segment_time
1861 segment_d = s['d']
1862 add_segment_url()
1863 segment_number += 1
1864 for r in range(s.get('r', 0)):
1865 segment_time += segment_d
1866 add_segment_url()
1867 segment_number += 1
1868 segment_time += segment_d
1869 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1870 # No media template
1871 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1872 # or any YouTube dashsegments video
1873 fragments = []
1874 segment_index = 0
1875 timescale = representation_ms_info['timescale']
1876 for s in representation_ms_info['s']:
1877 duration = float_or_none(s['d'], timescale)
1878 for r in range(s.get('r', 0) + 1):
1879 fragments.append({
1880 'url': representation_ms_info['segment_urls'][segment_index],
1881 'duration': duration,
1882 })
1883 segment_index += 1
1884 representation_ms_info['fragments'] = fragments
1885 # NB: MPD manifest may contain direct URLs to unfragmented media.
1886 # No fragments key is present in this case.
1887 if 'fragments' in representation_ms_info:
1888 f.update({
1889 'fragments': [],
1890 'protocol': 'http_dash_segments',
1891 })
1892 if 'initialization_url' in representation_ms_info:
1893 initialization_url = representation_ms_info['initialization_url']
1894 if not f.get('url'):
1895 f['url'] = initialization_url
1896 f['fragments'].append({'url': initialization_url})
1897 f['fragments'].extend(representation_ms_info['fragments'])
1898 for fragment in f['fragments']:
1899 fragment['url'] = urljoin(base_url, fragment['url'])
1900 try:
1901 existing_format = next(
1902 fo for fo in formats
1903 if fo['format_id'] == representation_id)
1904 except StopIteration:
1905 full_info = formats_dict.get(representation_id, {}).copy()
1906 full_info.update(f)
1907 formats.append(full_info)
1908 else:
1909 existing_format.update(f)
1910 else:
1911 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1912 return formats
1913
1914 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1915 res = self._download_webpage_handle(
1916 ism_url, video_id,
1917 note=note or 'Downloading ISM manifest',
1918 errnote=errnote or 'Failed to download ISM manifest',
1919 fatal=fatal)
1920 if res is False:
1921 return []
1922 ism, urlh = res
1923
1924 return self._parse_ism_formats(
1925 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
1926
1927 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
1928 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
1929 return []
1930
1931 duration = int(ism_doc.attrib['Duration'])
1932 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
1933
1934 formats = []
1935 for stream in ism_doc.findall('StreamIndex'):
1936 stream_type = stream.get('Type')
1937 if stream_type not in ('video', 'audio'):
1938 continue
1939 url_pattern = stream.attrib['Url']
1940 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
1941 stream_name = stream.get('Name')
1942 for track in stream.findall('QualityLevel'):
1943 fourcc = track.get('FourCC')
1944 # TODO: add support for WVC1 and WMAP
1945 if fourcc not in ('H264', 'AVC1', 'AACL'):
1946 self.report_warning('%s is not a supported codec' % fourcc)
1947 continue
1948 tbr = int(track.attrib['Bitrate']) // 1000
1949 width = int_or_none(track.get('MaxWidth'))
1950 height = int_or_none(track.get('MaxHeight'))
1951 sampling_rate = int_or_none(track.get('SamplingRate'))
1952
1953 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
1954 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
1955
1956 fragments = []
1957 fragment_ctx = {
1958 'time': 0,
1959 }
1960 stream_fragments = stream.findall('c')
1961 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
1962 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
1963 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
1964 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
1965 if not fragment_ctx['duration']:
1966 try:
1967 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
1968 except IndexError:
1969 next_fragment_time = duration
1970 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
1971 for _ in range(fragment_repeat):
1972 fragments.append({
1973 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
1974 'duration': fragment_ctx['duration'] / stream_timescale,
1975 })
1976 fragment_ctx['time'] += fragment_ctx['duration']
1977
1978 format_id = []
1979 if ism_id:
1980 format_id.append(ism_id)
1981 if stream_name:
1982 format_id.append(stream_name)
1983 format_id.append(compat_str(tbr))
1984
1985 formats.append({
1986 'format_id': '-'.join(format_id),
1987 'url': ism_url,
1988 'manifest_url': ism_url,
1989 'ext': 'ismv' if stream_type == 'video' else 'isma',
1990 'width': width,
1991 'height': height,
1992 'tbr': tbr,
1993 'asr': sampling_rate,
1994 'vcodec': 'none' if stream_type == 'audio' else fourcc,
1995 'acodec': 'none' if stream_type == 'video' else fourcc,
1996 'protocol': 'ism',
1997 'fragments': fragments,
1998 '_download_params': {
1999 'duration': duration,
2000 'timescale': stream_timescale,
2001 'width': width or 0,
2002 'height': height or 0,
2003 'fourcc': fourcc,
2004 'codec_private_data': track.get('CodecPrivateData'),
2005 'sampling_rate': sampling_rate,
2006 'channels': int_or_none(track.get('Channels', 2)),
2007 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2008 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2009 },
2010 })
2011 return formats
2012
2013 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2014 def absolute_url(video_url):
2015 return compat_urlparse.urljoin(base_url, video_url)
2016
2017 def parse_content_type(content_type):
2018 if not content_type:
2019 return {}
2020 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2021 if ctr:
2022 mimetype, codecs = ctr.groups()
2023 f = parse_codecs(codecs)
2024 f['ext'] = mimetype2ext(mimetype)
2025 return f
2026 return {}
2027
2028 def _media_formats(src, cur_media_type):
2029 full_url = absolute_url(src)
2030 ext = determine_ext(full_url)
2031 if ext == 'm3u8':
2032 is_plain_url = False
2033 formats = self._extract_m3u8_formats(
2034 full_url, video_id, ext='mp4',
2035 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2036 preference=preference)
2037 elif ext == 'mpd':
2038 is_plain_url = False
2039 formats = self._extract_mpd_formats(
2040 full_url, video_id, mpd_id=mpd_id)
2041 else:
2042 is_plain_url = True
2043 formats = [{
2044 'url': full_url,
2045 'vcodec': 'none' if cur_media_type == 'audio' else None,
2046 }]
2047 return is_plain_url, formats
2048
2049 entries = []
2050 media_tags = [(media_tag, media_type, '')
2051 for media_tag, media_type
2052 in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
2053 media_tags.extend(re.findall(
2054 # We only allow video|audio followed by a whitespace or '>'.
2055 # Allowing more characters may end up in significant slow down (see
2056 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2057 # http://www.porntrex.com/maps/videositemap.xml).
2058 r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2059 for media_tag, media_type, media_content in media_tags:
2060 media_info = {
2061 'formats': [],
2062 'subtitles': {},
2063 }
2064 media_attributes = extract_attributes(media_tag)
2065 src = media_attributes.get('src')
2066 if src:
2067 _, formats = _media_formats(src, media_type)
2068 media_info['formats'].extend(formats)
2069 media_info['thumbnail'] = media_attributes.get('poster')
2070 if media_content:
2071 for source_tag in re.findall(r'<source[^>]+>', media_content):
2072 source_attributes = extract_attributes(source_tag)
2073 src = source_attributes.get('src')
2074 if not src:
2075 continue
2076 is_plain_url, formats = _media_formats(src, media_type)
2077 if is_plain_url:
2078 f = parse_content_type(source_attributes.get('type'))
2079 f.update(formats[0])
2080 media_info['formats'].append(f)
2081 else:
2082 media_info['formats'].extend(formats)
2083 for track_tag in re.findall(r'<track[^>]+>', media_content):
2084 track_attributes = extract_attributes(track_tag)
2085 kind = track_attributes.get('kind')
2086 if not kind or kind in ('subtitles', 'captions'):
2087 src = track_attributes.get('src')
2088 if not src:
2089 continue
2090 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2091 media_info['subtitles'].setdefault(lang, []).append({
2092 'url': absolute_url(src),
2093 })
2094 if media_info['formats'] or media_info['subtitles']:
2095 entries.append(media_info)
2096 return entries
2097
2098 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2099 formats = []
2100 hdcore_sign = 'hdcore=3.7.0'
2101 f4m_url = re.sub(r'(https?://[^/+])/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2102 hds_host = hosts.get('hds')
2103 if hds_host:
2104 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2105 if 'hdcore=' not in f4m_url:
2106 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2107 f4m_formats = self._extract_f4m_formats(
2108 f4m_url, video_id, f4m_id='hds', fatal=False)
2109 for entry in f4m_formats:
2110 entry.update({'extra_param_to_segment_url': hdcore_sign})
2111 formats.extend(f4m_formats)
2112 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2113 hls_host = hosts.get('hls')
2114 if hls_host:
2115 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2116 formats.extend(self._extract_m3u8_formats(
2117 m3u8_url, video_id, 'mp4', 'm3u8_native',
2118 m3u8_id='hls', fatal=False))
2119 return formats
2120
2121 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2122 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2123 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
2124 http_base_url = 'http' + url_base
2125 formats = []
2126 if 'm3u8' not in skip_protocols:
2127 formats.extend(self._extract_m3u8_formats(
2128 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2129 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2130 if 'f4m' not in skip_protocols:
2131 formats.extend(self._extract_f4m_formats(
2132 http_base_url + '/manifest.f4m',
2133 video_id, f4m_id='hds', fatal=False))
2134 if 'dash' not in skip_protocols:
2135 formats.extend(self._extract_mpd_formats(
2136 http_base_url + '/manifest.mpd',
2137 video_id, mpd_id='dash', fatal=False))
2138 if re.search(r'(?:/smil:|\.smil)', url_base):
2139 if 'smil' not in skip_protocols:
2140 rtmp_formats = self._extract_smil_formats(
2141 http_base_url + '/jwplayer.smil',
2142 video_id, fatal=False)
2143 for rtmp_format in rtmp_formats:
2144 rtsp_format = rtmp_format.copy()
2145 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2146 del rtsp_format['play_path']
2147 del rtsp_format['ext']
2148 rtsp_format.update({
2149 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2150 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2151 'protocol': 'rtsp',
2152 })
2153 formats.extend([rtmp_format, rtsp_format])
2154 else:
2155 for protocol in ('rtmp', 'rtsp'):
2156 if protocol not in skip_protocols:
2157 formats.append({
2158 'url': protocol + url_base,
2159 'format_id': protocol,
2160 'protocol': protocol,
2161 })
2162 return formats
2163
2164 @staticmethod
2165 def _find_jwplayer_data(webpage):
2166 mobj = re.search(
2167 r'jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)\.setup\s*\((?P<options>[^)]+)\)',
2168 webpage)
2169 if mobj:
2170 return mobj.group('options')
2171
2172 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2173 jwplayer_data = self._parse_json(
2174 self._find_jwplayer_data(webpage), video_id,
2175 transform_source=js_to_json)
2176 return self._parse_jwplayer_data(
2177 jwplayer_data, video_id, *args, **kwargs)
2178
2179 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2180 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2181 # JWPlayer backward compatibility: flattened playlists
2182 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2183 if 'playlist' not in jwplayer_data:
2184 jwplayer_data = {'playlist': [jwplayer_data]}
2185
2186 entries = []
2187
2188 # JWPlayer backward compatibility: single playlist item
2189 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2190 if not isinstance(jwplayer_data['playlist'], list):
2191 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2192
2193 for video_data in jwplayer_data['playlist']:
2194 # JWPlayer backward compatibility: flattened sources
2195 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2196 if 'sources' not in video_data:
2197 video_data['sources'] = [video_data]
2198
2199 this_video_id = video_id or video_data['mediaid']
2200
2201 formats = self._parse_jwplayer_formats(
2202 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2203 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2204 self._sort_formats(formats)
2205
2206 subtitles = {}
2207 tracks = video_data.get('tracks')
2208 if tracks and isinstance(tracks, list):
2209 for track in tracks:
2210 if track.get('kind') != 'captions':
2211 continue
2212 track_url = urljoin(base_url, track.get('file'))
2213 if not track_url:
2214 continue
2215 subtitles.setdefault(track.get('label') or 'en', []).append({
2216 'url': self._proto_relative_url(track_url)
2217 })
2218
2219 entries.append({
2220 'id': this_video_id,
2221 'title': video_data['title'] if require_title else video_data.get('title'),
2222 'description': video_data.get('description'),
2223 'thumbnail': self._proto_relative_url(video_data.get('image')),
2224 'timestamp': int_or_none(video_data.get('pubdate')),
2225 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2226 'subtitles': subtitles,
2227 'formats': formats,
2228 })
2229 if len(entries) == 1:
2230 return entries[0]
2231 else:
2232 return self.playlist_result(entries)
2233
2234 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2235 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2236 formats = []
2237 for source in jwplayer_sources_data:
2238 source_url = self._proto_relative_url(source['file'])
2239 if base_url:
2240 source_url = compat_urlparse.urljoin(base_url, source_url)
2241 source_type = source.get('type') or ''
2242 ext = mimetype2ext(source_type) or determine_ext(source_url)
2243 if source_type == 'hls' or ext == 'm3u8':
2244 formats.extend(self._extract_m3u8_formats(
2245 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2246 m3u8_id=m3u8_id, fatal=False))
2247 elif ext == 'mpd':
2248 formats.extend(self._extract_mpd_formats(
2249 source_url, video_id, mpd_id=mpd_id, fatal=False))
2250 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2251 elif source_type.startswith('audio') or ext in (
2252 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2253 formats.append({
2254 'url': source_url,
2255 'vcodec': 'none',
2256 'ext': ext,
2257 })
2258 else:
2259 height = int_or_none(source.get('height'))
2260 if height is None:
2261 # Often no height is provided but there is a label in
2262 # format like "1080p", "720p SD", or 1080.
2263 height = int_or_none(self._search_regex(
2264 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2265 'height', default=None))
2266 a_format = {
2267 'url': source_url,
2268 'width': int_or_none(source.get('width')),
2269 'height': height,
2270 'tbr': int_or_none(source.get('bitrate')),
2271 'ext': ext,
2272 }
2273 if source_url.startswith('rtmp'):
2274 a_format['ext'] = 'flv'
2275 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2276 # of jwplayer.flash.swf
2277 rtmp_url_parts = re.split(
2278 r'((?:mp4|mp3|flv):)', source_url, 1)
2279 if len(rtmp_url_parts) == 3:
2280 rtmp_url, prefix, play_path = rtmp_url_parts
2281 a_format.update({
2282 'url': rtmp_url,
2283 'play_path': prefix + play_path,
2284 })
2285 if rtmp_params:
2286 a_format.update(rtmp_params)
2287 formats.append(a_format)
2288 return formats
2289
2290 def _live_title(self, name):
2291 """ Generate the title for a live video """
2292 now = datetime.datetime.now()
2293 now_str = now.strftime('%Y-%m-%d %H:%M')
2294 return name + ' ' + now_str
2295
2296 def _int(self, v, name, fatal=False, **kwargs):
2297 res = int_or_none(v, **kwargs)
2298 if 'get_attr' in kwargs:
2299 print(getattr(v, kwargs['get_attr']))
2300 if res is None:
2301 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2302 if fatal:
2303 raise ExtractorError(msg)
2304 else:
2305 self._downloader.report_warning(msg)
2306 return res
2307
2308 def _float(self, v, name, fatal=False, **kwargs):
2309 res = float_or_none(v, **kwargs)
2310 if res is None:
2311 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2312 if fatal:
2313 raise ExtractorError(msg)
2314 else:
2315 self._downloader.report_warning(msg)
2316 return res
2317
2318 def _set_cookie(self, domain, name, value, expire_time=None):
2319 cookie = compat_cookiejar.Cookie(
2320 0, name, value, None, None, domain, None,
2321 None, '/', True, False, expire_time, '', None, None, None)
2322 self._downloader.cookiejar.set_cookie(cookie)
2323
2324 def _get_cookies(self, url):
2325 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2326 req = sanitized_Request(url)
2327 self._downloader.cookiejar.add_cookie_header(req)
2328 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2329
2330 def get_testcases(self, include_onlymatching=False):
2331 t = getattr(self, '_TEST', None)
2332 if t:
2333 assert not hasattr(self, '_TESTS'), \
2334 '%s has _TEST and _TESTS' % type(self).__name__
2335 tests = [t]
2336 else:
2337 tests = getattr(self, '_TESTS', [])
2338 for t in tests:
2339 if not include_onlymatching and t.get('only_matching', False):
2340 continue
2341 t['name'] = type(self).__name__[:-len('IE')]
2342 yield t
2343
2344 def is_suitable(self, age_limit):
2345 """ Test whether the extractor is generally suitable for the given
2346 age limit (i.e. pornographic sites are not, all others usually are) """
2347
2348 any_restricted = False
2349 for tc in self.get_testcases(include_onlymatching=False):
2350 if tc.get('playlist', []):
2351 tc = tc['playlist'][0]
2352 is_restricted = age_restricted(
2353 tc.get('info_dict', {}).get('age_limit'), age_limit)
2354 if not is_restricted:
2355 return True
2356 any_restricted = any_restricted or is_restricted
2357 return not any_restricted
2358
2359 def extract_subtitles(self, *args, **kwargs):
2360 if (self._downloader.params.get('writesubtitles', False) or
2361 self._downloader.params.get('listsubtitles')):
2362 return self._get_subtitles(*args, **kwargs)
2363 return {}
2364
2365 def _get_subtitles(self, *args, **kwargs):
2366 raise NotImplementedError('This method must be implemented by subclasses')
2367
2368 @staticmethod
2369 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2370 """ Merge subtitle items for one language. Items with duplicated URLs
2371 will be dropped. """
2372 list1_urls = set([item['url'] for item in subtitle_list1])
2373 ret = list(subtitle_list1)
2374 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2375 return ret
2376
2377 @classmethod
2378 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2379 """ Merge two subtitle dictionaries, language by language. """
2380 ret = dict(subtitle_dict1)
2381 for lang in subtitle_dict2:
2382 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2383 return ret
2384
2385 def extract_automatic_captions(self, *args, **kwargs):
2386 if (self._downloader.params.get('writeautomaticsub', False) or
2387 self._downloader.params.get('listsubtitles')):
2388 return self._get_automatic_captions(*args, **kwargs)
2389 return {}
2390
2391 def _get_automatic_captions(self, *args, **kwargs):
2392 raise NotImplementedError('This method must be implemented by subclasses')
2393
2394 def mark_watched(self, *args, **kwargs):
2395 if (self._downloader.params.get('mark_watched', False) and
2396 (self._get_login_info()[0] is not None or
2397 self._downloader.params.get('cookiefile') is not None)):
2398 self._mark_watched(*args, **kwargs)
2399
2400 def _mark_watched(self, *args, **kwargs):
2401 raise NotImplementedError('This method must be implemented by subclasses')
2402
2403 def geo_verification_headers(self):
2404 headers = {}
2405 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2406 if geo_verification_proxy:
2407 headers['Ytdl-request-proxy'] = geo_verification_proxy
2408 return headers
2409
2410 def _generic_id(self, url):
2411 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2412
2413 def _generic_title(self, url):
2414 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2415
2416
2417 class SearchInfoExtractor(InfoExtractor):
2418 """
2419 Base class for paged search queries extractors.
2420 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2421 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2422 """
2423
2424 @classmethod
2425 def _make_valid_url(cls):
2426 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2427
2428 @classmethod
2429 def suitable(cls, url):
2430 return re.match(cls._make_valid_url(), url) is not None
2431
2432 def _real_extract(self, query):
2433 mobj = re.match(self._make_valid_url(), query)
2434 if mobj is None:
2435 raise ExtractorError('Invalid search query "%s"' % query)
2436
2437 prefix = mobj.group('prefix')
2438 query = mobj.group('query')
2439 if prefix == '':
2440 return self._get_n_results(query, 1)
2441 elif prefix == 'all':
2442 return self._get_n_results(query, self._MAX_RESULTS)
2443 else:
2444 n = int(prefix)
2445 if n <= 0:
2446 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2447 elif n > self._MAX_RESULTS:
2448 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2449 n = self._MAX_RESULTS
2450 return self._get_n_results(query, n)
2451
2452 def _get_n_results(self, query, n):
2453 """Get a specified number of results for a query"""
2454 raise NotImplementedError('This method must be implemented by subclasses')
2455
2456 @property
2457 def SEARCH_KEY(self):
2458 return self._SEARCH_KEY