]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
Update README.md
[youtubedl] / youtube_dl / extractor / common.py
1 # coding: utf-8
2 from __future__ import unicode_literals
3
4 import base64
5 import datetime
6 import hashlib
7 import json
8 import netrc
9 import os
10 import random
11 import re
12 import socket
13 import sys
14 import time
15 import math
16
17 from ..compat import (
18 compat_cookiejar_Cookie,
19 compat_cookies,
20 compat_etree_Element,
21 compat_etree_fromstring,
22 compat_getpass,
23 compat_integer_types,
24 compat_http_client,
25 compat_os_name,
26 compat_str,
27 compat_urllib_error,
28 compat_urllib_parse_unquote,
29 compat_urllib_parse_urlencode,
30 compat_urllib_request,
31 compat_urlparse,
32 compat_xml_parse_error,
33 )
34 from ..downloader.f4m import (
35 get_base_url,
36 remove_encrypted_media,
37 )
38 from ..utils import (
39 NO_DEFAULT,
40 age_restricted,
41 base_url,
42 bug_reports_message,
43 clean_html,
44 compiled_regex_type,
45 determine_ext,
46 determine_protocol,
47 dict_get,
48 error_to_compat_str,
49 ExtractorError,
50 extract_attributes,
51 fix_xml_ampersands,
52 float_or_none,
53 GeoRestrictedError,
54 GeoUtils,
55 int_or_none,
56 js_to_json,
57 JSON_LD_RE,
58 mimetype2ext,
59 orderedSet,
60 parse_bitrate,
61 parse_codecs,
62 parse_duration,
63 parse_iso8601,
64 parse_m3u8_attributes,
65 parse_resolution,
66 RegexNotFoundError,
67 sanitized_Request,
68 sanitize_filename,
69 str_or_none,
70 strip_or_none,
71 unescapeHTML,
72 unified_strdate,
73 unified_timestamp,
74 update_Request,
75 update_url_query,
76 urljoin,
77 url_basename,
78 url_or_none,
79 xpath_element,
80 xpath_text,
81 xpath_with_ns,
82 )
83
84
85 class InfoExtractor(object):
86 """Information Extractor class.
87
88 Information extractors are the classes that, given a URL, extract
89 information about the video (or videos) the URL refers to. This
90 information includes the real video URL, the video title, author and
91 others. The information is stored in a dictionary which is then
92 passed to the YoutubeDL. The YoutubeDL processes this
93 information possibly downloading the video to the file system, among
94 other possible outcomes.
95
96 The type field determines the type of the result.
97 By far the most common value (and the default if _type is missing) is
98 "video", which indicates a single video.
99
100 For a video, the dictionaries must include the following fields:
101
102 id: Video identifier.
103 title: Video title, unescaped.
104
105 Additionally, it must contain either a formats entry or a url one:
106
107 formats: A list of dictionaries for each format available, ordered
108 from worst to best quality.
109
110 Potential fields:
111 * url The mandatory URL representing the media:
112 for plain file media - HTTP URL of this file,
113 for RTMP - RTMP URL,
114 for HLS - URL of the M3U8 media playlist,
115 for HDS - URL of the F4M manifest,
116 for DASH
117 - HTTP URL to plain file media (in case of
118 unfragmented media)
119 - URL of the MPD manifest or base URL
120 representing the media if MPD manifest
121 is parsed from a string (in case of
122 fragmented media)
123 for MSS - URL of the ISM manifest.
124 * manifest_url
125 The URL of the manifest file in case of
126 fragmented media:
127 for HLS - URL of the M3U8 master playlist,
128 for HDS - URL of the F4M manifest,
129 for DASH - URL of the MPD manifest,
130 for MSS - URL of the ISM manifest.
131 * ext Will be calculated from URL if missing
132 * format A human-readable description of the format
133 ("mp4 container with h264/opus").
134 Calculated from the format_id, width, height.
135 and format_note fields if missing.
136 * format_id A short description of the format
137 ("mp4_h264_opus" or "19").
138 Technically optional, but strongly recommended.
139 * format_note Additional info about the format
140 ("3D" or "DASH video")
141 * width Width of the video, if known
142 * height Height of the video, if known
143 * resolution Textual description of width and height
144 * tbr Average bitrate of audio and video in KBit/s
145 * abr Average audio bitrate in KBit/s
146 * acodec Name of the audio codec in use
147 * asr Audio sampling rate in Hertz
148 * vbr Average video bitrate in KBit/s
149 * fps Frame rate
150 * vcodec Name of the video codec in use
151 * container Name of the container format
152 * filesize The number of bytes, if known in advance
153 * filesize_approx An estimate for the number of bytes
154 * player_url SWF Player URL (used for rtmpdump).
155 * protocol The protocol that will be used for the actual
156 download, lower-case.
157 "http", "https", "rtsp", "rtmp", "rtmpe",
158 "m3u8", "m3u8_native" or "http_dash_segments".
159 * fragment_base_url
160 Base URL for fragments. Each fragment's path
161 value (if present) will be relative to
162 this URL.
163 * fragments A list of fragments of a fragmented media.
164 Each fragment entry must contain either an url
165 or a path. If an url is present it should be
166 considered by a client. Otherwise both path and
167 fragment_base_url must be present. Here is
168 the list of all potential fields:
169 * "url" - fragment's URL
170 * "path" - fragment's path relative to
171 fragment_base_url
172 * "duration" (optional, int or float)
173 * "filesize" (optional, int)
174 * preference Order number of this format. If this field is
175 present and not None, the formats get sorted
176 by this field, regardless of all other values.
177 -1 for default (order by other properties),
178 -2 or smaller for less than default.
179 < -1000 to hide the format (if there is
180 another one which is strictly better)
181 * language Language code, e.g. "de" or "en-US".
182 * language_preference Is this in the language mentioned in
183 the URL?
184 10 if it's what the URL is about,
185 -1 for default (don't know),
186 -10 otherwise, other values reserved for now.
187 * quality Order number of the video quality of this
188 format, irrespective of the file format.
189 -1 for default (order by other properties),
190 -2 or smaller for less than default.
191 * source_preference Order number for this video source
192 (quality takes higher priority)
193 -1 for default (order by other properties),
194 -2 or smaller for less than default.
195 * http_headers A dictionary of additional HTTP headers
196 to add to the request.
197 * stretched_ratio If given and not 1, indicates that the
198 video's pixels are not square.
199 width : height ratio as float.
200 * no_resume The server does not support resuming the
201 (HTTP or RTMP) download. Boolean.
202 * downloader_options A dictionary of downloader options as
203 described in FileDownloader
204
205 url: Final video URL.
206 ext: Video filename extension.
207 format: The video format, defaults to ext (used for --get-format)
208 player_url: SWF Player URL (used for rtmpdump).
209
210 The following fields are optional:
211
212 alt_title: A secondary title of the video.
213 display_id An alternative identifier for the video, not necessarily
214 unique, but available before title. Typically, id is
215 something like "4234987", title "Dancing naked mole rats",
216 and display_id "dancing-naked-mole-rats"
217 thumbnails: A list of dictionaries, with the following entries:
218 * "id" (optional, string) - Thumbnail format ID
219 * "url"
220 * "preference" (optional, int) - quality of the image
221 * "width" (optional, int)
222 * "height" (optional, int)
223 * "resolution" (optional, string "{width}x{height}",
224 deprecated)
225 * "filesize" (optional, int)
226 thumbnail: Full URL to a video thumbnail image.
227 description: Full video description.
228 uploader: Full name of the video uploader.
229 license: License name the video is licensed under.
230 creator: The creator of the video.
231 release_date: The date (YYYYMMDD) when the video was released.
232 timestamp: UNIX timestamp of the moment the video became available.
233 upload_date: Video upload date (YYYYMMDD).
234 If not explicitly set, calculated from timestamp.
235 uploader_id: Nickname or id of the video uploader.
236 uploader_url: Full URL to a personal webpage of the video uploader.
237 channel: Full name of the channel the video is uploaded on.
238 Note that channel fields may or may not repeat uploader
239 fields. This depends on a particular extractor.
240 channel_id: Id of the channel.
241 channel_url: Full URL to a channel webpage.
242 location: Physical location where the video was filmed.
243 subtitles: The available subtitles as a dictionary in the format
244 {tag: subformats}. "tag" is usually a language code, and
245 "subformats" is a list sorted from lower to higher
246 preference, each element is a dictionary with the "ext"
247 entry and one of:
248 * "data": The subtitles file contents
249 * "url": A URL pointing to the subtitles file
250 "ext" will be calculated from URL if missing
251 automatic_captions: Like 'subtitles', used by the YoutubeIE for
252 automatically generated captions
253 duration: Length of the video in seconds, as an integer or float.
254 view_count: How many users have watched the video on the platform.
255 like_count: Number of positive ratings of the video
256 dislike_count: Number of negative ratings of the video
257 repost_count: Number of reposts of the video
258 average_rating: Average rating give by users, the scale used depends on the webpage
259 comment_count: Number of comments on the video
260 comments: A list of comments, each with one or more of the following
261 properties (all but one of text or html optional):
262 * "author" - human-readable name of the comment author
263 * "author_id" - user ID of the comment author
264 * "id" - Comment ID
265 * "html" - Comment as HTML
266 * "text" - Plain text of the comment
267 * "timestamp" - UNIX timestamp of comment
268 * "parent" - ID of the comment this one is replying to.
269 Set to "root" to indicate that this is a
270 comment to the original video.
271 age_limit: Age restriction for the video, as an integer (years)
272 webpage_url: The URL to the video webpage, if given to youtube-dl it
273 should allow to get the same result again. (It will be set
274 by YoutubeDL if it's missing)
275 categories: A list of categories that the video falls in, for example
276 ["Sports", "Berlin"]
277 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
278 is_live: True, False, or None (=unknown). Whether this video is a
279 live stream that goes on instead of a fixed-length video.
280 start_time: Time in seconds where the reproduction should start, as
281 specified in the URL.
282 end_time: Time in seconds where the reproduction should end, as
283 specified in the URL.
284 chapters: A list of dictionaries, with the following entries:
285 * "start_time" - The start time of the chapter in seconds
286 * "end_time" - The end time of the chapter in seconds
287 * "title" (optional, string)
288
289 The following fields should only be used when the video belongs to some logical
290 chapter or section:
291
292 chapter: Name or title of the chapter the video belongs to.
293 chapter_number: Number of the chapter the video belongs to, as an integer.
294 chapter_id: Id of the chapter the video belongs to, as a unicode string.
295
296 The following fields should only be used when the video is an episode of some
297 series, programme or podcast:
298
299 series: Title of the series or programme the video episode belongs to.
300 season: Title of the season the video episode belongs to.
301 season_number: Number of the season the video episode belongs to, as an integer.
302 season_id: Id of the season the video episode belongs to, as a unicode string.
303 episode: Title of the video episode. Unlike mandatory video title field,
304 this field should denote the exact title of the video episode
305 without any kind of decoration.
306 episode_number: Number of the video episode within a season, as an integer.
307 episode_id: Id of the video episode, as a unicode string.
308
309 The following fields should only be used when the media is a track or a part of
310 a music album:
311
312 track: Title of the track.
313 track_number: Number of the track within an album or a disc, as an integer.
314 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
315 as a unicode string.
316 artist: Artist(s) of the track.
317 genre: Genre(s) of the track.
318 album: Title of the album the track belongs to.
319 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
320 album_artist: List of all artists appeared on the album (e.g.
321 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
322 and compilations).
323 disc_number: Number of the disc or other physical medium the track belongs to,
324 as an integer.
325 release_year: Year (YYYY) when the album was released.
326
327 Unless mentioned otherwise, the fields should be Unicode strings.
328
329 Unless mentioned otherwise, None is equivalent to absence of information.
330
331
332 _type "playlist" indicates multiple videos.
333 There must be a key "entries", which is a list, an iterable, or a PagedList
334 object, each element of which is a valid dictionary by this specification.
335
336 Additionally, playlists can have "id", "title", "description", "uploader",
337 "uploader_id", "uploader_url" attributes with the same semantics as videos
338 (see above).
339
340
341 _type "multi_video" indicates that there are multiple videos that
342 form a single show, for examples multiple acts of an opera or TV episode.
343 It must have an entries key like a playlist and contain all the keys
344 required for a video at the same time.
345
346
347 _type "url" indicates that the video must be extracted from another
348 location, possibly by a different extractor. Its only required key is:
349 "url" - the next URL to extract.
350 The key "ie_key" can be set to the class name (minus the trailing "IE",
351 e.g. "Youtube") if the extractor class is known in advance.
352 Additionally, the dictionary may have any properties of the resolved entity
353 known in advance, for example "title" if the title of the referred video is
354 known ahead of time.
355
356
357 _type "url_transparent" entities have the same specification as "url", but
358 indicate that the given additional information is more precise than the one
359 associated with the resolved URL.
360 This is useful when a site employs a video service that hosts the video and
361 its technical metadata, but that video service does not embed a useful
362 title, description etc.
363
364
365 Subclasses of this one should re-define the _real_initialize() and
366 _real_extract() methods and define a _VALID_URL regexp.
367 Probably, they should also be added to the list of extractors.
368
369 _GEO_BYPASS attribute may be set to False in order to disable
370 geo restriction bypass mechanisms for a particular extractor.
371 Though it won't disable explicit geo restriction bypass based on
372 country code provided with geo_bypass_country.
373
374 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
375 countries for this extractor. One of these countries will be used by
376 geo restriction bypass mechanism right away in order to bypass
377 geo restriction, of course, if the mechanism is not disabled.
378
379 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
380 IP blocks in CIDR notation for this extractor. One of these IP blocks
381 will be used by geo restriction bypass mechanism similarly
382 to _GEO_COUNTRIES.
383
384 Finally, the _WORKING attribute should be set to False for broken IEs
385 in order to warn the users and skip the tests.
386 """
387
388 _ready = False
389 _downloader = None
390 _x_forwarded_for_ip = None
391 _GEO_BYPASS = True
392 _GEO_COUNTRIES = None
393 _GEO_IP_BLOCKS = None
394 _WORKING = True
395
396 def __init__(self, downloader=None):
397 """Constructor. Receives an optional downloader."""
398 self._ready = False
399 self._x_forwarded_for_ip = None
400 self.set_downloader(downloader)
401
402 @classmethod
403 def suitable(cls, url):
404 """Receives a URL and returns True if suitable for this IE."""
405
406 # This does not use has/getattr intentionally - we want to know whether
407 # we have cached the regexp for *this* class, whereas getattr would also
408 # match the superclass
409 if '_VALID_URL_RE' not in cls.__dict__:
410 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
411 return cls._VALID_URL_RE.match(url) is not None
412
413 @classmethod
414 def _match_id(cls, url):
415 if '_VALID_URL_RE' not in cls.__dict__:
416 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
417 m = cls._VALID_URL_RE.match(url)
418 assert m
419 return compat_str(m.group('id'))
420
421 @classmethod
422 def working(cls):
423 """Getter method for _WORKING."""
424 return cls._WORKING
425
426 def initialize(self):
427 """Initializes an instance (authentication, etc)."""
428 self._initialize_geo_bypass({
429 'countries': self._GEO_COUNTRIES,
430 'ip_blocks': self._GEO_IP_BLOCKS,
431 })
432 if not self._ready:
433 self._real_initialize()
434 self._ready = True
435
436 def _initialize_geo_bypass(self, geo_bypass_context):
437 """
438 Initialize geo restriction bypass mechanism.
439
440 This method is used to initialize geo bypass mechanism based on faking
441 X-Forwarded-For HTTP header. A random country from provided country list
442 is selected and a random IP belonging to this country is generated. This
443 IP will be passed as X-Forwarded-For HTTP header in all subsequent
444 HTTP requests.
445
446 This method will be used for initial geo bypass mechanism initialization
447 during the instance initialization with _GEO_COUNTRIES and
448 _GEO_IP_BLOCKS.
449
450 You may also manually call it from extractor's code if geo bypass
451 information is not available beforehand (e.g. obtained during
452 extraction) or due to some other reason. In this case you should pass
453 this information in geo bypass context passed as first argument. It may
454 contain following fields:
455
456 countries: List of geo unrestricted countries (similar
457 to _GEO_COUNTRIES)
458 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
459 (similar to _GEO_IP_BLOCKS)
460
461 """
462 if not self._x_forwarded_for_ip:
463
464 # Geo bypass mechanism is explicitly disabled by user
465 if not self._downloader.params.get('geo_bypass', True):
466 return
467
468 if not geo_bypass_context:
469 geo_bypass_context = {}
470
471 # Backward compatibility: previously _initialize_geo_bypass
472 # expected a list of countries, some 3rd party code may still use
473 # it this way
474 if isinstance(geo_bypass_context, (list, tuple)):
475 geo_bypass_context = {
476 'countries': geo_bypass_context,
477 }
478
479 # The whole point of geo bypass mechanism is to fake IP
480 # as X-Forwarded-For HTTP header based on some IP block or
481 # country code.
482
483 # Path 1: bypassing based on IP block in CIDR notation
484
485 # Explicit IP block specified by user, use it right away
486 # regardless of whether extractor is geo bypassable or not
487 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
488
489 # Otherwise use random IP block from geo bypass context but only
490 # if extractor is known as geo bypassable
491 if not ip_block:
492 ip_blocks = geo_bypass_context.get('ip_blocks')
493 if self._GEO_BYPASS and ip_blocks:
494 ip_block = random.choice(ip_blocks)
495
496 if ip_block:
497 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
498 if self._downloader.params.get('verbose', False):
499 self._downloader.to_screen(
500 '[debug] Using fake IP %s as X-Forwarded-For.'
501 % self._x_forwarded_for_ip)
502 return
503
504 # Path 2: bypassing based on country code
505
506 # Explicit country code specified by user, use it right away
507 # regardless of whether extractor is geo bypassable or not
508 country = self._downloader.params.get('geo_bypass_country', None)
509
510 # Otherwise use random country code from geo bypass context but
511 # only if extractor is known as geo bypassable
512 if not country:
513 countries = geo_bypass_context.get('countries')
514 if self._GEO_BYPASS and countries:
515 country = random.choice(countries)
516
517 if country:
518 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
519 if self._downloader.params.get('verbose', False):
520 self._downloader.to_screen(
521 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
522 % (self._x_forwarded_for_ip, country.upper()))
523
524 def extract(self, url):
525 """Extracts URL information and returns it in list of dicts."""
526 try:
527 for _ in range(2):
528 try:
529 self.initialize()
530 ie_result = self._real_extract(url)
531 if self._x_forwarded_for_ip:
532 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
533 return ie_result
534 except GeoRestrictedError as e:
535 if self.__maybe_fake_ip_and_retry(e.countries):
536 continue
537 raise
538 except ExtractorError:
539 raise
540 except compat_http_client.IncompleteRead as e:
541 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
542 except (KeyError, StopIteration) as e:
543 raise ExtractorError('An extractor error has occurred.', cause=e)
544
545 def __maybe_fake_ip_and_retry(self, countries):
546 if (not self._downloader.params.get('geo_bypass_country', None)
547 and self._GEO_BYPASS
548 and self._downloader.params.get('geo_bypass', True)
549 and not self._x_forwarded_for_ip
550 and countries):
551 country_code = random.choice(countries)
552 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
553 if self._x_forwarded_for_ip:
554 self.report_warning(
555 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
556 % (self._x_forwarded_for_ip, country_code.upper()))
557 return True
558 return False
559
560 def set_downloader(self, downloader):
561 """Sets the downloader for this IE."""
562 self._downloader = downloader
563
564 def _real_initialize(self):
565 """Real initialization process. Redefine in subclasses."""
566 pass
567
568 def _real_extract(self, url):
569 """Real extraction process. Redefine in subclasses."""
570 pass
571
572 @classmethod
573 def ie_key(cls):
574 """A string for getting the InfoExtractor with get_info_extractor"""
575 return compat_str(cls.__name__[:-2])
576
577 @property
578 def IE_NAME(self):
579 return compat_str(type(self).__name__[:-2])
580
581 @staticmethod
582 def __can_accept_status_code(err, expected_status):
583 assert isinstance(err, compat_urllib_error.HTTPError)
584 if expected_status is None:
585 return False
586 if isinstance(expected_status, compat_integer_types):
587 return err.code == expected_status
588 elif isinstance(expected_status, (list, tuple)):
589 return err.code in expected_status
590 elif callable(expected_status):
591 return expected_status(err.code) is True
592 else:
593 assert False
594
595 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
596 """
597 Return the response handle.
598
599 See _download_webpage docstring for arguments specification.
600 """
601 if note is None:
602 self.report_download_webpage(video_id)
603 elif note is not False:
604 if video_id is None:
605 self.to_screen('%s' % (note,))
606 else:
607 self.to_screen('%s: %s' % (video_id, note))
608
609 # Some sites check X-Forwarded-For HTTP header in order to figure out
610 # the origin of the client behind proxy. This allows bypassing geo
611 # restriction by faking this header's value to IP that belongs to some
612 # geo unrestricted country. We will do so once we encounter any
613 # geo restriction error.
614 if self._x_forwarded_for_ip:
615 if 'X-Forwarded-For' not in headers:
616 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
617
618 if isinstance(url_or_request, compat_urllib_request.Request):
619 url_or_request = update_Request(
620 url_or_request, data=data, headers=headers, query=query)
621 else:
622 if query:
623 url_or_request = update_url_query(url_or_request, query)
624 if data is not None or headers:
625 url_or_request = sanitized_Request(url_or_request, data, headers)
626 try:
627 return self._downloader.urlopen(url_or_request)
628 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
629 if isinstance(err, compat_urllib_error.HTTPError):
630 if self.__can_accept_status_code(err, expected_status):
631 # Retain reference to error to prevent file object from
632 # being closed before it can be read. Works around the
633 # effects of <https://bugs.python.org/issue15002>
634 # introduced in Python 3.4.1.
635 err.fp._error = err
636 return err.fp
637
638 if errnote is False:
639 return False
640 if errnote is None:
641 errnote = 'Unable to download webpage'
642
643 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
644 if fatal:
645 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
646 else:
647 self._downloader.report_warning(errmsg)
648 return False
649
650 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
651 """
652 Return a tuple (page content as string, URL handle).
653
654 See _download_webpage docstring for arguments specification.
655 """
656 # Strip hashes from the URL (#1038)
657 if isinstance(url_or_request, (compat_str, str)):
658 url_or_request = url_or_request.partition('#')[0]
659
660 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
661 if urlh is False:
662 assert not fatal
663 return False
664 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
665 return (content, urlh)
666
667 @staticmethod
668 def _guess_encoding_from_content(content_type, webpage_bytes):
669 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
670 if m:
671 encoding = m.group(1)
672 else:
673 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
674 webpage_bytes[:1024])
675 if m:
676 encoding = m.group(1).decode('ascii')
677 elif webpage_bytes.startswith(b'\xff\xfe'):
678 encoding = 'utf-16'
679 else:
680 encoding = 'utf-8'
681
682 return encoding
683
684 def __check_blocked(self, content):
685 first_block = content[:512]
686 if ('<title>Access to this site is blocked</title>' in content
687 and 'Websense' in first_block):
688 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
689 blocked_iframe = self._html_search_regex(
690 r'<iframe src="([^"]+)"', content,
691 'Websense information URL', default=None)
692 if blocked_iframe:
693 msg += ' Visit %s for more details' % blocked_iframe
694 raise ExtractorError(msg, expected=True)
695 if '<title>The URL you requested has been blocked</title>' in first_block:
696 msg = (
697 'Access to this webpage has been blocked by Indian censorship. '
698 'Use a VPN or proxy server (with --proxy) to route around it.')
699 block_msg = self._html_search_regex(
700 r'</h1><p>(.*?)</p>',
701 content, 'block message', default=None)
702 if block_msg:
703 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
704 raise ExtractorError(msg, expected=True)
705 if ('<title>TTK :: Š”Š¾ŃŃ‚ŃƒŠæ Šŗ рŠµŃŃƒŃ€ŃŃƒ Š¾Š³Ń€Š°Š½ŠøчŠµŠ½</title>' in content
706 and 'blocklist.rkn.gov.ru' in content):
707 raise ExtractorError(
708 'Access to this webpage has been blocked by decision of the Russian government. '
709 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
710 expected=True)
711
712 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
713 content_type = urlh.headers.get('Content-Type', '')
714 webpage_bytes = urlh.read()
715 if prefix is not None:
716 webpage_bytes = prefix + webpage_bytes
717 if not encoding:
718 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
719 if self._downloader.params.get('dump_intermediate_pages', False):
720 self.to_screen('Dumping request to ' + urlh.geturl())
721 dump = base64.b64encode(webpage_bytes).decode('ascii')
722 self._downloader.to_screen(dump)
723 if self._downloader.params.get('write_pages', False):
724 basen = '%s_%s' % (video_id, urlh.geturl())
725 if len(basen) > 240:
726 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
727 basen = basen[:240 - len(h)] + h
728 raw_filename = basen + '.dump'
729 filename = sanitize_filename(raw_filename, restricted=True)
730 self.to_screen('Saving request to ' + filename)
731 # Working around MAX_PATH limitation on Windows (see
732 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
733 if compat_os_name == 'nt':
734 absfilepath = os.path.abspath(filename)
735 if len(absfilepath) > 259:
736 filename = '\\\\?\\' + absfilepath
737 with open(filename, 'wb') as outf:
738 outf.write(webpage_bytes)
739
740 try:
741 content = webpage_bytes.decode(encoding, 'replace')
742 except LookupError:
743 content = webpage_bytes.decode('utf-8', 'replace')
744
745 self.__check_blocked(content)
746
747 return content
748
749 def _download_webpage(
750 self, url_or_request, video_id, note=None, errnote=None,
751 fatal=True, tries=1, timeout=5, encoding=None, data=None,
752 headers={}, query={}, expected_status=None):
753 """
754 Return the data of the page as a string.
755
756 Arguments:
757 url_or_request -- plain text URL as a string or
758 a compat_urllib_request.Requestobject
759 video_id -- Video/playlist/item identifier (string)
760
761 Keyword arguments:
762 note -- note printed before downloading (string)
763 errnote -- note printed in case of an error (string)
764 fatal -- flag denoting whether error should be considered fatal,
765 i.e. whether it should cause ExtractionError to be raised,
766 otherwise a warning will be reported and extraction continued
767 tries -- number of tries
768 timeout -- sleep interval between tries
769 encoding -- encoding for a page content decoding, guessed automatically
770 when not explicitly specified
771 data -- POST data (bytes)
772 headers -- HTTP headers (dict)
773 query -- URL query (dict)
774 expected_status -- allows to accept failed HTTP requests (non 2xx
775 status code) by explicitly specifying a set of accepted status
776 codes. Can be any of the following entities:
777 - an integer type specifying an exact failed status code to
778 accept
779 - a list or a tuple of integer types specifying a list of
780 failed status codes to accept
781 - a callable accepting an actual failed status code and
782 returning True if it should be accepted
783 Note that this argument does not affect success status codes (2xx)
784 which are always accepted.
785 """
786
787 success = False
788 try_count = 0
789 while success is False:
790 try:
791 res = self._download_webpage_handle(
792 url_or_request, video_id, note, errnote, fatal,
793 encoding=encoding, data=data, headers=headers, query=query,
794 expected_status=expected_status)
795 success = True
796 except compat_http_client.IncompleteRead as e:
797 try_count += 1
798 if try_count >= tries:
799 raise e
800 self._sleep(timeout, video_id)
801 if res is False:
802 return res
803 else:
804 content, _ = res
805 return content
806
807 def _download_xml_handle(
808 self, url_or_request, video_id, note='Downloading XML',
809 errnote='Unable to download XML', transform_source=None,
810 fatal=True, encoding=None, data=None, headers={}, query={},
811 expected_status=None):
812 """
813 Return a tuple (xml as an compat_etree_Element, URL handle).
814
815 See _download_webpage docstring for arguments specification.
816 """
817 res = self._download_webpage_handle(
818 url_or_request, video_id, note, errnote, fatal=fatal,
819 encoding=encoding, data=data, headers=headers, query=query,
820 expected_status=expected_status)
821 if res is False:
822 return res
823 xml_string, urlh = res
824 return self._parse_xml(
825 xml_string, video_id, transform_source=transform_source,
826 fatal=fatal), urlh
827
828 def _download_xml(
829 self, url_or_request, video_id,
830 note='Downloading XML', errnote='Unable to download XML',
831 transform_source=None, fatal=True, encoding=None,
832 data=None, headers={}, query={}, expected_status=None):
833 """
834 Return the xml as an compat_etree_Element.
835
836 See _download_webpage docstring for arguments specification.
837 """
838 res = self._download_xml_handle(
839 url_or_request, video_id, note=note, errnote=errnote,
840 transform_source=transform_source, fatal=fatal, encoding=encoding,
841 data=data, headers=headers, query=query,
842 expected_status=expected_status)
843 return res if res is False else res[0]
844
845 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
846 if transform_source:
847 xml_string = transform_source(xml_string)
848 try:
849 return compat_etree_fromstring(xml_string.encode('utf-8'))
850 except compat_xml_parse_error as ve:
851 errmsg = '%s: Failed to parse XML ' % video_id
852 if fatal:
853 raise ExtractorError(errmsg, cause=ve)
854 else:
855 self.report_warning(errmsg + str(ve))
856
857 def _download_json_handle(
858 self, url_or_request, video_id, note='Downloading JSON metadata',
859 errnote='Unable to download JSON metadata', transform_source=None,
860 fatal=True, encoding=None, data=None, headers={}, query={},
861 expected_status=None):
862 """
863 Return a tuple (JSON object, URL handle).
864
865 See _download_webpage docstring for arguments specification.
866 """
867 res = self._download_webpage_handle(
868 url_or_request, video_id, note, errnote, fatal=fatal,
869 encoding=encoding, data=data, headers=headers, query=query,
870 expected_status=expected_status)
871 if res is False:
872 return res
873 json_string, urlh = res
874 return self._parse_json(
875 json_string, video_id, transform_source=transform_source,
876 fatal=fatal), urlh
877
878 def _download_json(
879 self, url_or_request, video_id, note='Downloading JSON metadata',
880 errnote='Unable to download JSON metadata', transform_source=None,
881 fatal=True, encoding=None, data=None, headers={}, query={},
882 expected_status=None):
883 """
884 Return the JSON object as a dict.
885
886 See _download_webpage docstring for arguments specification.
887 """
888 res = self._download_json_handle(
889 url_or_request, video_id, note=note, errnote=errnote,
890 transform_source=transform_source, fatal=fatal, encoding=encoding,
891 data=data, headers=headers, query=query,
892 expected_status=expected_status)
893 return res if res is False else res[0]
894
895 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
896 if transform_source:
897 json_string = transform_source(json_string)
898 try:
899 return json.loads(json_string)
900 except ValueError as ve:
901 errmsg = '%s: Failed to parse JSON ' % video_id
902 if fatal:
903 raise ExtractorError(errmsg, cause=ve)
904 else:
905 self.report_warning(errmsg + str(ve))
906
907 def report_warning(self, msg, video_id=None):
908 idstr = '' if video_id is None else '%s: ' % video_id
909 self._downloader.report_warning(
910 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
911
912 def to_screen(self, msg):
913 """Print msg to screen, prefixing it with '[ie_name]'"""
914 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
915
916 def report_extraction(self, id_or_name):
917 """Report information extraction."""
918 self.to_screen('%s: Extracting information' % id_or_name)
919
920 def report_download_webpage(self, video_id):
921 """Report webpage download."""
922 self.to_screen('%s: Downloading webpage' % video_id)
923
924 def report_age_confirmation(self):
925 """Report attempt to confirm age."""
926 self.to_screen('Confirming age')
927
928 def report_login(self):
929 """Report attempt to log in."""
930 self.to_screen('Logging in')
931
932 @staticmethod
933 def raise_login_required(msg='This video is only available for registered users'):
934 raise ExtractorError(
935 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
936 expected=True)
937
938 @staticmethod
939 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
940 raise GeoRestrictedError(msg, countries=countries)
941
942 # Methods for following #608
943 @staticmethod
944 def url_result(url, ie=None, video_id=None, video_title=None):
945 """Returns a URL that points to a page that should be processed"""
946 # TODO: ie should be the class used for getting the info
947 video_info = {'_type': 'url',
948 'url': url,
949 'ie_key': ie}
950 if video_id is not None:
951 video_info['id'] = video_id
952 if video_title is not None:
953 video_info['title'] = video_title
954 return video_info
955
956 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
957 urls = orderedSet(
958 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
959 for m in matches)
960 return self.playlist_result(
961 urls, playlist_id=playlist_id, playlist_title=playlist_title)
962
963 @staticmethod
964 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
965 """Returns a playlist"""
966 video_info = {'_type': 'playlist',
967 'entries': entries}
968 if playlist_id:
969 video_info['id'] = playlist_id
970 if playlist_title:
971 video_info['title'] = playlist_title
972 if playlist_description:
973 video_info['description'] = playlist_description
974 return video_info
975
976 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
977 """
978 Perform a regex search on the given string, using a single or a list of
979 patterns returning the first matching group.
980 In case of failure return a default value or raise a WARNING or a
981 RegexNotFoundError, depending on fatal, specifying the field name.
982 """
983 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
984 mobj = re.search(pattern, string, flags)
985 else:
986 for p in pattern:
987 mobj = re.search(p, string, flags)
988 if mobj:
989 break
990
991 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
992 _name = '\033[0;34m%s\033[0m' % name
993 else:
994 _name = name
995
996 if mobj:
997 if group is None:
998 # return the first matching group
999 return next(g for g in mobj.groups() if g is not None)
1000 else:
1001 return mobj.group(group)
1002 elif default is not NO_DEFAULT:
1003 return default
1004 elif fatal:
1005 raise RegexNotFoundError('Unable to extract %s' % _name)
1006 else:
1007 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
1008 return None
1009
1010 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1011 """
1012 Like _search_regex, but strips HTML tags and unescapes entities.
1013 """
1014 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1015 if res:
1016 return clean_html(res).strip()
1017 else:
1018 return res
1019
1020 def _get_netrc_login_info(self, netrc_machine=None):
1021 username = None
1022 password = None
1023 netrc_machine = netrc_machine or self._NETRC_MACHINE
1024
1025 if self._downloader.params.get('usenetrc', False):
1026 try:
1027 info = netrc.netrc().authenticators(netrc_machine)
1028 if info is not None:
1029 username = info[0]
1030 password = info[2]
1031 else:
1032 raise netrc.NetrcParseError(
1033 'No authenticators for %s' % netrc_machine)
1034 except (IOError, netrc.NetrcParseError) as err:
1035 self._downloader.report_warning(
1036 'parsing .netrc: %s' % error_to_compat_str(err))
1037
1038 return username, password
1039
1040 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1041 """
1042 Get the login info as (username, password)
1043 First look for the manually specified credentials using username_option
1044 and password_option as keys in params dictionary. If no such credentials
1045 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1046 value.
1047 If there's no info available, return (None, None)
1048 """
1049 if self._downloader is None:
1050 return (None, None)
1051
1052 downloader_params = self._downloader.params
1053
1054 # Attempt to use provided username and password or .netrc data
1055 if downloader_params.get(username_option) is not None:
1056 username = downloader_params[username_option]
1057 password = downloader_params[password_option]
1058 else:
1059 username, password = self._get_netrc_login_info(netrc_machine)
1060
1061 return username, password
1062
1063 def _get_tfa_info(self, note='two-factor verification code'):
1064 """
1065 Get the two-factor authentication info
1066 TODO - asking the user will be required for sms/phone verify
1067 currently just uses the command line option
1068 If there's no info available, return None
1069 """
1070 if self._downloader is None:
1071 return None
1072 downloader_params = self._downloader.params
1073
1074 if downloader_params.get('twofactor') is not None:
1075 return downloader_params['twofactor']
1076
1077 return compat_getpass('Type %s and press [Return]: ' % note)
1078
1079 # Helper functions for extracting OpenGraph info
1080 @staticmethod
1081 def _og_regexes(prop):
1082 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1083 property_re = (r'(?:name|property)=(?:\'og[:-]%(prop)s\'|"og[:-]%(prop)s"|\s*og[:-]%(prop)s\b)'
1084 % {'prop': re.escape(prop)})
1085 template = r'<meta[^>]+?%s[^>]+?%s'
1086 return [
1087 template % (property_re, content_re),
1088 template % (content_re, property_re),
1089 ]
1090
1091 @staticmethod
1092 def _meta_regex(prop):
1093 return r'''(?isx)<meta
1094 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1095 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1096
1097 def _og_search_property(self, prop, html, name=None, **kargs):
1098 if not isinstance(prop, (list, tuple)):
1099 prop = [prop]
1100 if name is None:
1101 name = 'OpenGraph %s' % prop[0]
1102 og_regexes = []
1103 for p in prop:
1104 og_regexes.extend(self._og_regexes(p))
1105 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1106 if escaped is None:
1107 return None
1108 return unescapeHTML(escaped)
1109
1110 def _og_search_thumbnail(self, html, **kargs):
1111 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1112
1113 def _og_search_description(self, html, **kargs):
1114 return self._og_search_property('description', html, fatal=False, **kargs)
1115
1116 def _og_search_title(self, html, **kargs):
1117 return self._og_search_property('title', html, **kargs)
1118
1119 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1120 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1121 if secure:
1122 regexes = self._og_regexes('video:secure_url') + regexes
1123 return self._html_search_regex(regexes, html, name, **kargs)
1124
1125 def _og_search_url(self, html, **kargs):
1126 return self._og_search_property('url', html, **kargs)
1127
1128 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1129 if not isinstance(name, (list, tuple)):
1130 name = [name]
1131 if display_name is None:
1132 display_name = name[0]
1133 return self._html_search_regex(
1134 [self._meta_regex(n) for n in name],
1135 html, display_name, fatal=fatal, group='content', **kwargs)
1136
1137 def _dc_search_uploader(self, html):
1138 return self._html_search_meta('dc.creator', html, 'uploader')
1139
1140 def _rta_search(self, html):
1141 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1142 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1143 r' content="RTA-5042-1996-1400-1577-RTA"',
1144 html):
1145 return 18
1146 return 0
1147
1148 def _media_rating_search(self, html):
1149 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1150 rating = self._html_search_meta('rating', html)
1151
1152 if not rating:
1153 return None
1154
1155 RATING_TABLE = {
1156 'safe for kids': 0,
1157 'general': 8,
1158 '14 years': 14,
1159 'mature': 17,
1160 'restricted': 19,
1161 }
1162 return RATING_TABLE.get(rating.lower())
1163
1164 def _family_friendly_search(self, html):
1165 # See http://schema.org/VideoObject
1166 family_friendly = self._html_search_meta(
1167 'isFamilyFriendly', html, default=None)
1168
1169 if not family_friendly:
1170 return None
1171
1172 RATING_TABLE = {
1173 '1': 0,
1174 'true': 0,
1175 '0': 18,
1176 'false': 18,
1177 }
1178 return RATING_TABLE.get(family_friendly.lower())
1179
1180 def _twitter_search_player(self, html):
1181 return self._html_search_meta('twitter:player', html,
1182 'twitter card player')
1183
1184 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1185 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1186 default = kwargs.get('default', NO_DEFAULT)
1187 # JSON-LD may be malformed and thus `fatal` should be respected.
1188 # At the same time `default` may be passed that assumes `fatal=False`
1189 # for _search_regex. Let's simulate the same behavior here as well.
1190 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1191 json_ld = []
1192 for mobj in json_ld_list:
1193 json_ld_item = self._parse_json(
1194 mobj.group('json_ld'), video_id, fatal=fatal)
1195 if not json_ld_item:
1196 continue
1197 if isinstance(json_ld_item, dict):
1198 json_ld.append(json_ld_item)
1199 elif isinstance(json_ld_item, (list, tuple)):
1200 json_ld.extend(json_ld_item)
1201 if json_ld:
1202 json_ld = self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1203 if json_ld:
1204 return json_ld
1205 if default is not NO_DEFAULT:
1206 return default
1207 elif fatal:
1208 raise RegexNotFoundError('Unable to extract JSON-LD')
1209 else:
1210 self._downloader.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1211 return {}
1212
1213 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1214 if isinstance(json_ld, compat_str):
1215 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1216 if not json_ld:
1217 return {}
1218 info = {}
1219 if not isinstance(json_ld, (list, tuple, dict)):
1220 return info
1221 if isinstance(json_ld, dict):
1222 json_ld = [json_ld]
1223
1224 INTERACTION_TYPE_MAP = {
1225 'CommentAction': 'comment',
1226 'AgreeAction': 'like',
1227 'DisagreeAction': 'dislike',
1228 'LikeAction': 'like',
1229 'DislikeAction': 'dislike',
1230 'ListenAction': 'view',
1231 'WatchAction': 'view',
1232 'ViewAction': 'view',
1233 }
1234
1235 def extract_interaction_statistic(e):
1236 interaction_statistic = e.get('interactionStatistic')
1237 if not isinstance(interaction_statistic, list):
1238 return
1239 for is_e in interaction_statistic:
1240 if not isinstance(is_e, dict):
1241 continue
1242 if is_e.get('@type') != 'InteractionCounter':
1243 continue
1244 interaction_type = is_e.get('interactionType')
1245 if not isinstance(interaction_type, compat_str):
1246 continue
1247 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1248 if interaction_count is None:
1249 continue
1250 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1251 if not count_kind:
1252 continue
1253 count_key = '%s_count' % count_kind
1254 if info.get(count_key) is not None:
1255 continue
1256 info[count_key] = interaction_count
1257
1258 def extract_video_object(e):
1259 assert e['@type'] == 'VideoObject'
1260 info.update({
1261 'url': url_or_none(e.get('contentUrl')),
1262 'title': unescapeHTML(e.get('name')),
1263 'description': unescapeHTML(e.get('description')),
1264 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
1265 'duration': parse_duration(e.get('duration')),
1266 'timestamp': unified_timestamp(e.get('uploadDate')),
1267 'filesize': float_or_none(e.get('contentSize')),
1268 'tbr': int_or_none(e.get('bitrate')),
1269 'width': int_or_none(e.get('width')),
1270 'height': int_or_none(e.get('height')),
1271 'view_count': int_or_none(e.get('interactionCount')),
1272 })
1273 extract_interaction_statistic(e)
1274
1275 for e in json_ld:
1276 if '@context' in e:
1277 item_type = e.get('@type')
1278 if expected_type is not None and expected_type != item_type:
1279 continue
1280 if item_type in ('TVEpisode', 'Episode'):
1281 episode_name = unescapeHTML(e.get('name'))
1282 info.update({
1283 'episode': episode_name,
1284 'episode_number': int_or_none(e.get('episodeNumber')),
1285 'description': unescapeHTML(e.get('description')),
1286 })
1287 if not info.get('title') and episode_name:
1288 info['title'] = episode_name
1289 part_of_season = e.get('partOfSeason')
1290 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1291 info.update({
1292 'season': unescapeHTML(part_of_season.get('name')),
1293 'season_number': int_or_none(part_of_season.get('seasonNumber')),
1294 })
1295 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1296 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1297 info['series'] = unescapeHTML(part_of_series.get('name'))
1298 elif item_type == 'Movie':
1299 info.update({
1300 'title': unescapeHTML(e.get('name')),
1301 'description': unescapeHTML(e.get('description')),
1302 'duration': parse_duration(e.get('duration')),
1303 'timestamp': unified_timestamp(e.get('dateCreated')),
1304 })
1305 elif item_type in ('Article', 'NewsArticle'):
1306 info.update({
1307 'timestamp': parse_iso8601(e.get('datePublished')),
1308 'title': unescapeHTML(e.get('headline')),
1309 'description': unescapeHTML(e.get('articleBody')),
1310 })
1311 elif item_type == 'VideoObject':
1312 extract_video_object(e)
1313 if expected_type is None:
1314 continue
1315 else:
1316 break
1317 video = e.get('video')
1318 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1319 extract_video_object(video)
1320 if expected_type is None:
1321 continue
1322 else:
1323 break
1324 return dict((k, v) for k, v in info.items() if v is not None)
1325
1326 @staticmethod
1327 def _hidden_inputs(html):
1328 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1329 hidden_inputs = {}
1330 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1331 attrs = extract_attributes(input)
1332 if not input:
1333 continue
1334 if attrs.get('type') not in ('hidden', 'submit'):
1335 continue
1336 name = attrs.get('name') or attrs.get('id')
1337 value = attrs.get('value')
1338 if name and value is not None:
1339 hidden_inputs[name] = value
1340 return hidden_inputs
1341
1342 def _form_hidden_inputs(self, form_id, html):
1343 form = self._search_regex(
1344 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1345 html, '%s form' % form_id, group='form')
1346 return self._hidden_inputs(form)
1347
1348 def _sort_formats(self, formats, field_preference=None):
1349 if not formats:
1350 raise ExtractorError('No video formats found')
1351
1352 for f in formats:
1353 # Automatically determine tbr when missing based on abr and vbr (improves
1354 # formats sorting in some cases)
1355 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1356 f['tbr'] = f['abr'] + f['vbr']
1357
1358 def _formats_key(f):
1359 # TODO remove the following workaround
1360 from ..utils import determine_ext
1361 if not f.get('ext') and 'url' in f:
1362 f['ext'] = determine_ext(f['url'])
1363
1364 if isinstance(field_preference, (list, tuple)):
1365 return tuple(
1366 f.get(field)
1367 if f.get(field) is not None
1368 else ('' if field == 'format_id' else -1)
1369 for field in field_preference)
1370
1371 preference = f.get('preference')
1372 if preference is None:
1373 preference = 0
1374 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1375 preference -= 0.5
1376
1377 protocol = f.get('protocol') or determine_protocol(f)
1378 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1379
1380 if f.get('vcodec') == 'none': # audio only
1381 preference -= 50
1382 if self._downloader.params.get('prefer_free_formats'):
1383 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1384 else:
1385 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1386 ext_preference = 0
1387 try:
1388 audio_ext_preference = ORDER.index(f['ext'])
1389 except ValueError:
1390 audio_ext_preference = -1
1391 else:
1392 if f.get('acodec') == 'none': # video only
1393 preference -= 40
1394 if self._downloader.params.get('prefer_free_formats'):
1395 ORDER = ['flv', 'mp4', 'webm']
1396 else:
1397 ORDER = ['webm', 'flv', 'mp4']
1398 try:
1399 ext_preference = ORDER.index(f['ext'])
1400 except ValueError:
1401 ext_preference = -1
1402 audio_ext_preference = 0
1403
1404 return (
1405 preference,
1406 f.get('language_preference') if f.get('language_preference') is not None else -1,
1407 f.get('quality') if f.get('quality') is not None else -1,
1408 f.get('tbr') if f.get('tbr') is not None else -1,
1409 f.get('filesize') if f.get('filesize') is not None else -1,
1410 f.get('vbr') if f.get('vbr') is not None else -1,
1411 f.get('height') if f.get('height') is not None else -1,
1412 f.get('width') if f.get('width') is not None else -1,
1413 proto_preference,
1414 ext_preference,
1415 f.get('abr') if f.get('abr') is not None else -1,
1416 audio_ext_preference,
1417 f.get('fps') if f.get('fps') is not None else -1,
1418 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1419 f.get('source_preference') if f.get('source_preference') is not None else -1,
1420 f.get('format_id') if f.get('format_id') is not None else '',
1421 )
1422 formats.sort(key=_formats_key)
1423
1424 def _check_formats(self, formats, video_id):
1425 if formats:
1426 formats[:] = filter(
1427 lambda f: self._is_valid_url(
1428 f['url'], video_id,
1429 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1430 formats)
1431
1432 @staticmethod
1433 def _remove_duplicate_formats(formats):
1434 format_urls = set()
1435 unique_formats = []
1436 for f in formats:
1437 if f['url'] not in format_urls:
1438 format_urls.add(f['url'])
1439 unique_formats.append(f)
1440 formats[:] = unique_formats
1441
1442 def _is_valid_url(self, url, video_id, item='video', headers={}):
1443 url = self._proto_relative_url(url, scheme='http:')
1444 # For now assume non HTTP(S) URLs always valid
1445 if not (url.startswith('http://') or url.startswith('https://')):
1446 return True
1447 try:
1448 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1449 return True
1450 except ExtractorError:
1451 self.to_screen(
1452 '%s: %s URL is invalid, skipping' % (video_id, item))
1453 return False
1454
1455 def http_scheme(self):
1456 """ Either "http:" or "https:", depending on the user's preferences """
1457 return (
1458 'http:'
1459 if self._downloader.params.get('prefer_insecure', False)
1460 else 'https:')
1461
1462 def _proto_relative_url(self, url, scheme=None):
1463 if url is None:
1464 return url
1465 if url.startswith('//'):
1466 if scheme is None:
1467 scheme = self.http_scheme()
1468 return scheme + url
1469 else:
1470 return url
1471
1472 def _sleep(self, timeout, video_id, msg_template=None):
1473 if msg_template is None:
1474 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1475 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1476 self.to_screen(msg)
1477 time.sleep(timeout)
1478
1479 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1480 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1481 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1482 manifest = self._download_xml(
1483 manifest_url, video_id, 'Downloading f4m manifest',
1484 'Unable to download f4m manifest',
1485 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1486 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1487 transform_source=transform_source,
1488 fatal=fatal, data=data, headers=headers, query=query)
1489
1490 if manifest is False:
1491 return []
1492
1493 return self._parse_f4m_formats(
1494 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1495 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1496
1497 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1498 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1499 fatal=True, m3u8_id=None):
1500 if not isinstance(manifest, compat_etree_Element) and not fatal:
1501 return []
1502
1503 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1504 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1505 if akamai_pv is not None and ';' in akamai_pv.text:
1506 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1507 if playerVerificationChallenge.strip() != '':
1508 return []
1509
1510 formats = []
1511 manifest_version = '1.0'
1512 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1513 if not media_nodes:
1514 manifest_version = '2.0'
1515 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1516 # Remove unsupported DRM protected media from final formats
1517 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1518 media_nodes = remove_encrypted_media(media_nodes)
1519 if not media_nodes:
1520 return formats
1521
1522 manifest_base_url = get_base_url(manifest)
1523
1524 bootstrap_info = xpath_element(
1525 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1526 'bootstrap info', default=None)
1527
1528 vcodec = None
1529 mime_type = xpath_text(
1530 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1531 'base URL', default=None)
1532 if mime_type and mime_type.startswith('audio/'):
1533 vcodec = 'none'
1534
1535 for i, media_el in enumerate(media_nodes):
1536 tbr = int_or_none(media_el.attrib.get('bitrate'))
1537 width = int_or_none(media_el.attrib.get('width'))
1538 height = int_or_none(media_el.attrib.get('height'))
1539 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1540 # If <bootstrapInfo> is present, the specified f4m is a
1541 # stream-level manifest, and only set-level manifests may refer to
1542 # external resources. See section 11.4 and section 4 of F4M spec
1543 if bootstrap_info is None:
1544 media_url = None
1545 # @href is introduced in 2.0, see section 11.6 of F4M spec
1546 if manifest_version == '2.0':
1547 media_url = media_el.attrib.get('href')
1548 if media_url is None:
1549 media_url = media_el.attrib.get('url')
1550 if not media_url:
1551 continue
1552 manifest_url = (
1553 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1554 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1555 # If media_url is itself a f4m manifest do the recursive extraction
1556 # since bitrates in parent manifest (this one) and media_url manifest
1557 # may differ leading to inability to resolve the format by requested
1558 # bitrate in f4m downloader
1559 ext = determine_ext(manifest_url)
1560 if ext == 'f4m':
1561 f4m_formats = self._extract_f4m_formats(
1562 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1563 transform_source=transform_source, fatal=fatal)
1564 # Sometimes stream-level manifest contains single media entry that
1565 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1566 # At the same time parent's media entry in set-level manifest may
1567 # contain it. We will copy it from parent in such cases.
1568 if len(f4m_formats) == 1:
1569 f = f4m_formats[0]
1570 f.update({
1571 'tbr': f.get('tbr') or tbr,
1572 'width': f.get('width') or width,
1573 'height': f.get('height') or height,
1574 'format_id': f.get('format_id') if not tbr else format_id,
1575 'vcodec': vcodec,
1576 })
1577 formats.extend(f4m_formats)
1578 continue
1579 elif ext == 'm3u8':
1580 formats.extend(self._extract_m3u8_formats(
1581 manifest_url, video_id, 'mp4', preference=preference,
1582 m3u8_id=m3u8_id, fatal=fatal))
1583 continue
1584 formats.append({
1585 'format_id': format_id,
1586 'url': manifest_url,
1587 'manifest_url': manifest_url,
1588 'ext': 'flv' if bootstrap_info is not None else None,
1589 'protocol': 'f4m',
1590 'tbr': tbr,
1591 'width': width,
1592 'height': height,
1593 'vcodec': vcodec,
1594 'preference': preference,
1595 })
1596 return formats
1597
1598 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1599 return {
1600 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1601 'url': m3u8_url,
1602 'ext': ext,
1603 'protocol': 'm3u8',
1604 'preference': preference - 100 if preference else -100,
1605 'resolution': 'multiple',
1606 'format_note': 'Quality selection URL',
1607 }
1608
1609 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1610 entry_protocol='m3u8', preference=None,
1611 m3u8_id=None, note=None, errnote=None,
1612 fatal=True, live=False, data=None, headers={},
1613 query={}):
1614 res = self._download_webpage_handle(
1615 m3u8_url, video_id,
1616 note=note or 'Downloading m3u8 information',
1617 errnote=errnote or 'Failed to download m3u8 information',
1618 fatal=fatal, data=data, headers=headers, query=query)
1619
1620 if res is False:
1621 return []
1622
1623 m3u8_doc, urlh = res
1624 m3u8_url = urlh.geturl()
1625
1626 return self._parse_m3u8_formats(
1627 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1628 preference=preference, m3u8_id=m3u8_id, live=live)
1629
1630 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1631 entry_protocol='m3u8', preference=None,
1632 m3u8_id=None, live=False):
1633 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1634 return []
1635
1636 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1637 return []
1638
1639 formats = []
1640
1641 format_url = lambda u: (
1642 u
1643 if re.match(r'^https?://', u)
1644 else compat_urlparse.urljoin(m3u8_url, u))
1645
1646 # References:
1647 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1648 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1649 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1650
1651 # We should try extracting formats only from master playlists [1, 4.3.4],
1652 # i.e. playlists that describe available qualities. On the other hand
1653 # media playlists [1, 4.3.3] should be returned as is since they contain
1654 # just the media without qualities renditions.
1655 # Fortunately, master playlist can be easily distinguished from media
1656 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1657 # master playlist tags MUST NOT appear in a media playist and vice versa.
1658 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1659 # media playlist and MUST NOT appear in master playlist thus we can
1660 # clearly detect media playlist with this criterion.
1661
1662 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1663 return [{
1664 'url': m3u8_url,
1665 'format_id': m3u8_id,
1666 'ext': ext,
1667 'protocol': entry_protocol,
1668 'preference': preference,
1669 }]
1670
1671 groups = {}
1672 last_stream_inf = {}
1673
1674 def extract_media(x_media_line):
1675 media = parse_m3u8_attributes(x_media_line)
1676 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1677 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1678 if not (media_type and group_id and name):
1679 return
1680 groups.setdefault(group_id, []).append(media)
1681 if media_type not in ('VIDEO', 'AUDIO'):
1682 return
1683 media_url = media.get('URI')
1684 if media_url:
1685 format_id = []
1686 for v in (m3u8_id, group_id, name):
1687 if v:
1688 format_id.append(v)
1689 f = {
1690 'format_id': '-'.join(format_id),
1691 'url': format_url(media_url),
1692 'manifest_url': m3u8_url,
1693 'language': media.get('LANGUAGE'),
1694 'ext': ext,
1695 'protocol': entry_protocol,
1696 'preference': preference,
1697 }
1698 if media_type == 'AUDIO':
1699 f['vcodec'] = 'none'
1700 formats.append(f)
1701
1702 def build_stream_name():
1703 # Despite specification does not mention NAME attribute for
1704 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1705 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1706 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1707 stream_name = last_stream_inf.get('NAME')
1708 if stream_name:
1709 return stream_name
1710 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1711 # from corresponding rendition group
1712 stream_group_id = last_stream_inf.get('VIDEO')
1713 if not stream_group_id:
1714 return
1715 stream_group = groups.get(stream_group_id)
1716 if not stream_group:
1717 return stream_group_id
1718 rendition = stream_group[0]
1719 return rendition.get('NAME') or stream_group_id
1720
1721 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1722 # chance to detect video only formats when EXT-X-STREAM-INF tags
1723 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1724 for line in m3u8_doc.splitlines():
1725 if line.startswith('#EXT-X-MEDIA:'):
1726 extract_media(line)
1727
1728 for line in m3u8_doc.splitlines():
1729 if line.startswith('#EXT-X-STREAM-INF:'):
1730 last_stream_inf = parse_m3u8_attributes(line)
1731 elif line.startswith('#') or not line.strip():
1732 continue
1733 else:
1734 tbr = float_or_none(
1735 last_stream_inf.get('AVERAGE-BANDWIDTH')
1736 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1737 format_id = []
1738 if m3u8_id:
1739 format_id.append(m3u8_id)
1740 stream_name = build_stream_name()
1741 # Bandwidth of live streams may differ over time thus making
1742 # format_id unpredictable. So it's better to keep provided
1743 # format_id intact.
1744 if not live:
1745 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1746 manifest_url = format_url(line.strip())
1747 f = {
1748 'format_id': '-'.join(format_id),
1749 'url': manifest_url,
1750 'manifest_url': m3u8_url,
1751 'tbr': tbr,
1752 'ext': ext,
1753 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1754 'protocol': entry_protocol,
1755 'preference': preference,
1756 }
1757 resolution = last_stream_inf.get('RESOLUTION')
1758 if resolution:
1759 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1760 if mobj:
1761 f['width'] = int(mobj.group('width'))
1762 f['height'] = int(mobj.group('height'))
1763 # Unified Streaming Platform
1764 mobj = re.search(
1765 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1766 if mobj:
1767 abr, vbr = mobj.groups()
1768 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1769 f.update({
1770 'vbr': vbr,
1771 'abr': abr,
1772 })
1773 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1774 f.update(codecs)
1775 audio_group_id = last_stream_inf.get('AUDIO')
1776 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1777 # references a rendition group MUST have a CODECS attribute.
1778 # However, this is not always respected, for example, [2]
1779 # contains EXT-X-STREAM-INF tag which references AUDIO
1780 # rendition group but does not have CODECS and despite
1781 # referencing an audio group it represents a complete
1782 # (with audio and video) format. So, for such cases we will
1783 # ignore references to rendition groups and treat them
1784 # as complete formats.
1785 if audio_group_id and codecs and f.get('vcodec') != 'none':
1786 audio_group = groups.get(audio_group_id)
1787 if audio_group and audio_group[0].get('URI'):
1788 # TODO: update acodec for audio only formats with
1789 # the same GROUP-ID
1790 f['acodec'] = 'none'
1791 formats.append(f)
1792
1793 # for DailyMotion
1794 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
1795 if progressive_uri:
1796 http_f = f.copy()
1797 del http_f['manifest_url']
1798 http_f.update({
1799 'format_id': f['format_id'].replace('hls-', 'http-'),
1800 'protocol': 'http',
1801 'url': progressive_uri,
1802 })
1803 formats.append(http_f)
1804
1805 last_stream_inf = {}
1806 return formats
1807
1808 @staticmethod
1809 def _xpath_ns(path, namespace=None):
1810 if not namespace:
1811 return path
1812 out = []
1813 for c in path.split('/'):
1814 if not c or c == '.':
1815 out.append(c)
1816 else:
1817 out.append('{%s}%s' % (namespace, c))
1818 return '/'.join(out)
1819
1820 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1821 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1822
1823 if smil is False:
1824 assert not fatal
1825 return []
1826
1827 namespace = self._parse_smil_namespace(smil)
1828
1829 return self._parse_smil_formats(
1830 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1831
1832 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1833 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1834 if smil is False:
1835 return {}
1836 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1837
1838 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1839 return self._download_xml(
1840 smil_url, video_id, 'Downloading SMIL file',
1841 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1842
1843 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1844 namespace = self._parse_smil_namespace(smil)
1845
1846 formats = self._parse_smil_formats(
1847 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1848 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1849
1850 video_id = os.path.splitext(url_basename(smil_url))[0]
1851 title = None
1852 description = None
1853 upload_date = None
1854 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1855 name = meta.attrib.get('name')
1856 content = meta.attrib.get('content')
1857 if not name or not content:
1858 continue
1859 if not title and name == 'title':
1860 title = content
1861 elif not description and name in ('description', 'abstract'):
1862 description = content
1863 elif not upload_date and name == 'date':
1864 upload_date = unified_strdate(content)
1865
1866 thumbnails = [{
1867 'id': image.get('type'),
1868 'url': image.get('src'),
1869 'width': int_or_none(image.get('width')),
1870 'height': int_or_none(image.get('height')),
1871 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1872
1873 return {
1874 'id': video_id,
1875 'title': title or video_id,
1876 'description': description,
1877 'upload_date': upload_date,
1878 'thumbnails': thumbnails,
1879 'formats': formats,
1880 'subtitles': subtitles,
1881 }
1882
1883 def _parse_smil_namespace(self, smil):
1884 return self._search_regex(
1885 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1886
1887 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1888 base = smil_url
1889 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1890 b = meta.get('base') or meta.get('httpBase')
1891 if b:
1892 base = b
1893 break
1894
1895 formats = []
1896 rtmp_count = 0
1897 http_count = 0
1898 m3u8_count = 0
1899
1900 srcs = []
1901 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1902 for medium in media:
1903 src = medium.get('src')
1904 if not src or src in srcs:
1905 continue
1906 srcs.append(src)
1907
1908 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1909 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1910 width = int_or_none(medium.get('width'))
1911 height = int_or_none(medium.get('height'))
1912 proto = medium.get('proto')
1913 ext = medium.get('ext')
1914 src_ext = determine_ext(src)
1915 streamer = medium.get('streamer') or base
1916
1917 if proto == 'rtmp' or streamer.startswith('rtmp'):
1918 rtmp_count += 1
1919 formats.append({
1920 'url': streamer,
1921 'play_path': src,
1922 'ext': 'flv',
1923 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1924 'tbr': bitrate,
1925 'filesize': filesize,
1926 'width': width,
1927 'height': height,
1928 })
1929 if transform_rtmp_url:
1930 streamer, src = transform_rtmp_url(streamer, src)
1931 formats[-1].update({
1932 'url': streamer,
1933 'play_path': src,
1934 })
1935 continue
1936
1937 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1938 src_url = src_url.strip()
1939
1940 if proto == 'm3u8' or src_ext == 'm3u8':
1941 m3u8_formats = self._extract_m3u8_formats(
1942 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1943 if len(m3u8_formats) == 1:
1944 m3u8_count += 1
1945 m3u8_formats[0].update({
1946 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1947 'tbr': bitrate,
1948 'width': width,
1949 'height': height,
1950 })
1951 formats.extend(m3u8_formats)
1952 elif src_ext == 'f4m':
1953 f4m_url = src_url
1954 if not f4m_params:
1955 f4m_params = {
1956 'hdcore': '3.2.0',
1957 'plugin': 'flowplayer-3.2.0.1',
1958 }
1959 f4m_url += '&' if '?' in f4m_url else '?'
1960 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1961 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1962 elif src_ext == 'mpd':
1963 formats.extend(self._extract_mpd_formats(
1964 src_url, video_id, mpd_id='dash', fatal=False))
1965 elif re.search(r'\.ism/[Mm]anifest', src_url):
1966 formats.extend(self._extract_ism_formats(
1967 src_url, video_id, ism_id='mss', fatal=False))
1968 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1969 http_count += 1
1970 formats.append({
1971 'url': src_url,
1972 'ext': ext or src_ext or 'flv',
1973 'format_id': 'http-%d' % (bitrate or http_count),
1974 'tbr': bitrate,
1975 'filesize': filesize,
1976 'width': width,
1977 'height': height,
1978 })
1979
1980 return formats
1981
1982 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1983 urls = []
1984 subtitles = {}
1985 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1986 src = textstream.get('src')
1987 if not src or src in urls:
1988 continue
1989 urls.append(src)
1990 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1991 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1992 subtitles.setdefault(lang, []).append({
1993 'url': src,
1994 'ext': ext,
1995 })
1996 return subtitles
1997
1998 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1999 xspf = self._download_xml(
2000 xspf_url, playlist_id, 'Downloading xpsf playlist',
2001 'Unable to download xspf manifest', fatal=fatal)
2002 if xspf is False:
2003 return []
2004 return self._parse_xspf(
2005 xspf, playlist_id, xspf_url=xspf_url,
2006 xspf_base_url=base_url(xspf_url))
2007
2008 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2009 NS_MAP = {
2010 'xspf': 'http://xspf.org/ns/0/',
2011 's1': 'http://static.streamone.nl/player/ns/0',
2012 }
2013
2014 entries = []
2015 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2016 title = xpath_text(
2017 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2018 description = xpath_text(
2019 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2020 thumbnail = xpath_text(
2021 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2022 duration = float_or_none(
2023 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2024
2025 formats = []
2026 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2027 format_url = urljoin(xspf_base_url, location.text)
2028 if not format_url:
2029 continue
2030 formats.append({
2031 'url': format_url,
2032 'manifest_url': xspf_url,
2033 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2034 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2035 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2036 })
2037 self._sort_formats(formats)
2038
2039 entries.append({
2040 'id': playlist_id,
2041 'title': title,
2042 'description': description,
2043 'thumbnail': thumbnail,
2044 'duration': duration,
2045 'formats': formats,
2046 })
2047 return entries
2048
2049 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
2050 res = self._download_xml_handle(
2051 mpd_url, video_id,
2052 note=note or 'Downloading MPD manifest',
2053 errnote=errnote or 'Failed to download MPD manifest',
2054 fatal=fatal, data=data, headers=headers, query=query)
2055 if res is False:
2056 return []
2057 mpd_doc, urlh = res
2058 if mpd_doc is None:
2059 return []
2060 mpd_base_url = base_url(urlh.geturl())
2061
2062 return self._parse_mpd_formats(
2063 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2064 formats_dict=formats_dict, mpd_url=mpd_url)
2065
2066 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2067 """
2068 Parse formats from MPD manifest.
2069 References:
2070 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2071 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2072 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2073 """
2074 if mpd_doc.get('type') == 'dynamic':
2075 return []
2076
2077 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2078
2079 def _add_ns(path):
2080 return self._xpath_ns(path, namespace)
2081
2082 def is_drm_protected(element):
2083 return element.find(_add_ns('ContentProtection')) is not None
2084
2085 def extract_multisegment_info(element, ms_parent_info):
2086 ms_info = ms_parent_info.copy()
2087
2088 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2089 # common attributes and elements. We will only extract relevant
2090 # for us.
2091 def extract_common(source):
2092 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2093 if segment_timeline is not None:
2094 s_e = segment_timeline.findall(_add_ns('S'))
2095 if s_e:
2096 ms_info['total_number'] = 0
2097 ms_info['s'] = []
2098 for s in s_e:
2099 r = int(s.get('r', 0))
2100 ms_info['total_number'] += 1 + r
2101 ms_info['s'].append({
2102 't': int(s.get('t', 0)),
2103 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2104 'd': int(s.attrib['d']),
2105 'r': r,
2106 })
2107 start_number = source.get('startNumber')
2108 if start_number:
2109 ms_info['start_number'] = int(start_number)
2110 timescale = source.get('timescale')
2111 if timescale:
2112 ms_info['timescale'] = int(timescale)
2113 segment_duration = source.get('duration')
2114 if segment_duration:
2115 ms_info['segment_duration'] = float(segment_duration)
2116
2117 def extract_Initialization(source):
2118 initialization = source.find(_add_ns('Initialization'))
2119 if initialization is not None:
2120 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2121
2122 segment_list = element.find(_add_ns('SegmentList'))
2123 if segment_list is not None:
2124 extract_common(segment_list)
2125 extract_Initialization(segment_list)
2126 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2127 if segment_urls_e:
2128 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2129 else:
2130 segment_template = element.find(_add_ns('SegmentTemplate'))
2131 if segment_template is not None:
2132 extract_common(segment_template)
2133 media = segment_template.get('media')
2134 if media:
2135 ms_info['media'] = media
2136 initialization = segment_template.get('initialization')
2137 if initialization:
2138 ms_info['initialization'] = initialization
2139 else:
2140 extract_Initialization(segment_template)
2141 return ms_info
2142
2143 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2144 formats = []
2145 for period in mpd_doc.findall(_add_ns('Period')):
2146 period_duration = parse_duration(period.get('duration')) or mpd_duration
2147 period_ms_info = extract_multisegment_info(period, {
2148 'start_number': 1,
2149 'timescale': 1,
2150 })
2151 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2152 if is_drm_protected(adaptation_set):
2153 continue
2154 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2155 for representation in adaptation_set.findall(_add_ns('Representation')):
2156 if is_drm_protected(representation):
2157 continue
2158 representation_attrib = adaptation_set.attrib.copy()
2159 representation_attrib.update(representation.attrib)
2160 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2161 mime_type = representation_attrib['mimeType']
2162 content_type = mime_type.split('/')[0]
2163 if content_type == 'text':
2164 # TODO implement WebVTT downloading
2165 pass
2166 elif content_type in ('video', 'audio'):
2167 base_url = ''
2168 for element in (representation, adaptation_set, period, mpd_doc):
2169 base_url_e = element.find(_add_ns('BaseURL'))
2170 if base_url_e is not None:
2171 base_url = base_url_e.text + base_url
2172 if re.match(r'^https?://', base_url):
2173 break
2174 if mpd_base_url and not re.match(r'^https?://', base_url):
2175 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2176 mpd_base_url += '/'
2177 base_url = mpd_base_url + base_url
2178 representation_id = representation_attrib.get('id')
2179 lang = representation_attrib.get('lang')
2180 url_el = representation.find(_add_ns('BaseURL'))
2181 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2182 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2183 f = {
2184 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2185 'manifest_url': mpd_url,
2186 'ext': mimetype2ext(mime_type),
2187 'width': int_or_none(representation_attrib.get('width')),
2188 'height': int_or_none(representation_attrib.get('height')),
2189 'tbr': float_or_none(bandwidth, 1000),
2190 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2191 'fps': int_or_none(representation_attrib.get('frameRate')),
2192 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2193 'format_note': 'DASH %s' % content_type,
2194 'filesize': filesize,
2195 'container': mimetype2ext(mime_type) + '_dash',
2196 }
2197 f.update(parse_codecs(representation_attrib.get('codecs')))
2198 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2199
2200 def prepare_template(template_name, identifiers):
2201 tmpl = representation_ms_info[template_name]
2202 # First of, % characters outside $...$ templates
2203 # must be escaped by doubling for proper processing
2204 # by % operator string formatting used further (see
2205 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2206 t = ''
2207 in_template = False
2208 for c in tmpl:
2209 t += c
2210 if c == '$':
2211 in_template = not in_template
2212 elif c == '%' and not in_template:
2213 t += c
2214 # Next, $...$ templates are translated to their
2215 # %(...) counterparts to be used with % operator
2216 t = t.replace('$RepresentationID$', representation_id)
2217 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2218 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2219 t.replace('$$', '$')
2220 return t
2221
2222 # @initialization is a regular template like @media one
2223 # so it should be handled just the same way (see
2224 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2225 if 'initialization' in representation_ms_info:
2226 initialization_template = prepare_template(
2227 'initialization',
2228 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2229 # $Time$ shall not be included for @initialization thus
2230 # only $Bandwidth$ remains
2231 ('Bandwidth', ))
2232 representation_ms_info['initialization_url'] = initialization_template % {
2233 'Bandwidth': bandwidth,
2234 }
2235
2236 def location_key(location):
2237 return 'url' if re.match(r'^https?://', location) else 'path'
2238
2239 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2240
2241 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2242 media_location_key = location_key(media_template)
2243
2244 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2245 # can't be used at the same time
2246 if '%(Number' in media_template and 's' not in representation_ms_info:
2247 segment_duration = None
2248 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2249 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2250 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2251 representation_ms_info['fragments'] = [{
2252 media_location_key: media_template % {
2253 'Number': segment_number,
2254 'Bandwidth': bandwidth,
2255 },
2256 'duration': segment_duration,
2257 } for segment_number in range(
2258 representation_ms_info['start_number'],
2259 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2260 else:
2261 # $Number*$ or $Time$ in media template with S list available
2262 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2263 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2264 representation_ms_info['fragments'] = []
2265 segment_time = 0
2266 segment_d = None
2267 segment_number = representation_ms_info['start_number']
2268
2269 def add_segment_url():
2270 segment_url = media_template % {
2271 'Time': segment_time,
2272 'Bandwidth': bandwidth,
2273 'Number': segment_number,
2274 }
2275 representation_ms_info['fragments'].append({
2276 media_location_key: segment_url,
2277 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2278 })
2279
2280 for num, s in enumerate(representation_ms_info['s']):
2281 segment_time = s.get('t') or segment_time
2282 segment_d = s['d']
2283 add_segment_url()
2284 segment_number += 1
2285 for r in range(s.get('r', 0)):
2286 segment_time += segment_d
2287 add_segment_url()
2288 segment_number += 1
2289 segment_time += segment_d
2290 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2291 # No media template
2292 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2293 # or any YouTube dashsegments video
2294 fragments = []
2295 segment_index = 0
2296 timescale = representation_ms_info['timescale']
2297 for s in representation_ms_info['s']:
2298 duration = float_or_none(s['d'], timescale)
2299 for r in range(s.get('r', 0) + 1):
2300 segment_uri = representation_ms_info['segment_urls'][segment_index]
2301 fragments.append({
2302 location_key(segment_uri): segment_uri,
2303 'duration': duration,
2304 })
2305 segment_index += 1
2306 representation_ms_info['fragments'] = fragments
2307 elif 'segment_urls' in representation_ms_info:
2308 # Segment URLs with no SegmentTimeline
2309 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2310 # https://github.com/ytdl-org/youtube-dl/pull/14844
2311 fragments = []
2312 segment_duration = float_or_none(
2313 representation_ms_info['segment_duration'],
2314 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2315 for segment_url in representation_ms_info['segment_urls']:
2316 fragment = {
2317 location_key(segment_url): segment_url,
2318 }
2319 if segment_duration:
2320 fragment['duration'] = segment_duration
2321 fragments.append(fragment)
2322 representation_ms_info['fragments'] = fragments
2323 # If there is a fragments key available then we correctly recognized fragmented media.
2324 # Otherwise we will assume unfragmented media with direct access. Technically, such
2325 # assumption is not necessarily correct since we may simply have no support for
2326 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2327 if 'fragments' in representation_ms_info:
2328 f.update({
2329 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2330 'url': mpd_url or base_url,
2331 'fragment_base_url': base_url,
2332 'fragments': [],
2333 'protocol': 'http_dash_segments',
2334 })
2335 if 'initialization_url' in representation_ms_info:
2336 initialization_url = representation_ms_info['initialization_url']
2337 if not f.get('url'):
2338 f['url'] = initialization_url
2339 f['fragments'].append({location_key(initialization_url): initialization_url})
2340 f['fragments'].extend(representation_ms_info['fragments'])
2341 else:
2342 # Assuming direct URL to unfragmented media.
2343 f['url'] = base_url
2344
2345 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2346 # is not necessarily unique within a Period thus formats with
2347 # the same `format_id` are quite possible. There are numerous examples
2348 # of such manifests (see https://github.com/ytdl-org/youtube-dl/issues/15111,
2349 # https://github.com/ytdl-org/youtube-dl/issues/13919)
2350 full_info = formats_dict.get(representation_id, {}).copy()
2351 full_info.update(f)
2352 formats.append(full_info)
2353 else:
2354 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2355 return formats
2356
2357 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2358 res = self._download_xml_handle(
2359 ism_url, video_id,
2360 note=note or 'Downloading ISM manifest',
2361 errnote=errnote or 'Failed to download ISM manifest',
2362 fatal=fatal, data=data, headers=headers, query=query)
2363 if res is False:
2364 return []
2365 ism_doc, urlh = res
2366 if ism_doc is None:
2367 return []
2368
2369 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2370
2371 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2372 """
2373 Parse formats from ISM manifest.
2374 References:
2375 1. [MS-SSTR]: Smooth Streaming Protocol,
2376 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2377 """
2378 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2379 return []
2380
2381 duration = int(ism_doc.attrib['Duration'])
2382 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2383
2384 formats = []
2385 for stream in ism_doc.findall('StreamIndex'):
2386 stream_type = stream.get('Type')
2387 if stream_type not in ('video', 'audio'):
2388 continue
2389 url_pattern = stream.attrib['Url']
2390 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2391 stream_name = stream.get('Name')
2392 for track in stream.findall('QualityLevel'):
2393 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2394 # TODO: add support for WVC1 and WMAP
2395 if fourcc not in ('H264', 'AVC1', 'AACL'):
2396 self.report_warning('%s is not a supported codec' % fourcc)
2397 continue
2398 tbr = int(track.attrib['Bitrate']) // 1000
2399 # [1] does not mention Width and Height attributes. However,
2400 # they're often present while MaxWidth and MaxHeight are
2401 # missing, so should be used as fallbacks
2402 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2403 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2404 sampling_rate = int_or_none(track.get('SamplingRate'))
2405
2406 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2407 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2408
2409 fragments = []
2410 fragment_ctx = {
2411 'time': 0,
2412 }
2413 stream_fragments = stream.findall('c')
2414 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2415 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2416 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2417 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2418 if not fragment_ctx['duration']:
2419 try:
2420 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2421 except IndexError:
2422 next_fragment_time = duration
2423 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2424 for _ in range(fragment_repeat):
2425 fragments.append({
2426 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2427 'duration': fragment_ctx['duration'] / stream_timescale,
2428 })
2429 fragment_ctx['time'] += fragment_ctx['duration']
2430
2431 format_id = []
2432 if ism_id:
2433 format_id.append(ism_id)
2434 if stream_name:
2435 format_id.append(stream_name)
2436 format_id.append(compat_str(tbr))
2437
2438 formats.append({
2439 'format_id': '-'.join(format_id),
2440 'url': ism_url,
2441 'manifest_url': ism_url,
2442 'ext': 'ismv' if stream_type == 'video' else 'isma',
2443 'width': width,
2444 'height': height,
2445 'tbr': tbr,
2446 'asr': sampling_rate,
2447 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2448 'acodec': 'none' if stream_type == 'video' else fourcc,
2449 'protocol': 'ism',
2450 'fragments': fragments,
2451 '_download_params': {
2452 'duration': duration,
2453 'timescale': stream_timescale,
2454 'width': width or 0,
2455 'height': height or 0,
2456 'fourcc': fourcc,
2457 'codec_private_data': track.get('CodecPrivateData'),
2458 'sampling_rate': sampling_rate,
2459 'channels': int_or_none(track.get('Channels', 2)),
2460 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2461 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2462 },
2463 })
2464 return formats
2465
2466 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2467 def absolute_url(item_url):
2468 return urljoin(base_url, item_url)
2469
2470 def parse_content_type(content_type):
2471 if not content_type:
2472 return {}
2473 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2474 if ctr:
2475 mimetype, codecs = ctr.groups()
2476 f = parse_codecs(codecs)
2477 f['ext'] = mimetype2ext(mimetype)
2478 return f
2479 return {}
2480
2481 def _media_formats(src, cur_media_type, type_info={}):
2482 full_url = absolute_url(src)
2483 ext = type_info.get('ext') or determine_ext(full_url)
2484 if ext == 'm3u8':
2485 is_plain_url = False
2486 formats = self._extract_m3u8_formats(
2487 full_url, video_id, ext='mp4',
2488 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2489 preference=preference, fatal=False)
2490 elif ext == 'mpd':
2491 is_plain_url = False
2492 formats = self._extract_mpd_formats(
2493 full_url, video_id, mpd_id=mpd_id, fatal=False)
2494 else:
2495 is_plain_url = True
2496 formats = [{
2497 'url': full_url,
2498 'vcodec': 'none' if cur_media_type == 'audio' else None,
2499 }]
2500 return is_plain_url, formats
2501
2502 entries = []
2503 # amp-video and amp-audio are very similar to their HTML5 counterparts
2504 # so we wll include them right here (see
2505 # https://www.ampproject.org/docs/reference/components/amp-video)
2506 media_tags = [(media_tag, media_type, '')
2507 for media_tag, media_type
2508 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2509 media_tags.extend(re.findall(
2510 # We only allow video|audio followed by a whitespace or '>'.
2511 # Allowing more characters may end up in significant slow down (see
2512 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2513 # http://www.porntrex.com/maps/videositemap.xml).
2514 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2515 for media_tag, media_type, media_content in media_tags:
2516 media_info = {
2517 'formats': [],
2518 'subtitles': {},
2519 }
2520 media_attributes = extract_attributes(media_tag)
2521 src = strip_or_none(media_attributes.get('src'))
2522 if src:
2523 _, formats = _media_formats(src, media_type)
2524 media_info['formats'].extend(formats)
2525 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2526 if media_content:
2527 for source_tag in re.findall(r'<source[^>]+>', media_content):
2528 s_attr = extract_attributes(source_tag)
2529 # data-video-src and data-src are non standard but seen
2530 # several times in the wild
2531 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2532 if not src:
2533 continue
2534 f = parse_content_type(s_attr.get('type'))
2535 is_plain_url, formats = _media_formats(src, media_type, f)
2536 if is_plain_url:
2537 # width, height, res, label and title attributes are
2538 # all not standard but seen several times in the wild
2539 labels = [
2540 s_attr.get(lbl)
2541 for lbl in ('label', 'title')
2542 if str_or_none(s_attr.get(lbl))
2543 ]
2544 width = int_or_none(s_attr.get('width'))
2545 height = (int_or_none(s_attr.get('height'))
2546 or int_or_none(s_attr.get('res')))
2547 if not width or not height:
2548 for lbl in labels:
2549 resolution = parse_resolution(lbl)
2550 if not resolution:
2551 continue
2552 width = width or resolution.get('width')
2553 height = height or resolution.get('height')
2554 for lbl in labels:
2555 tbr = parse_bitrate(lbl)
2556 if tbr:
2557 break
2558 else:
2559 tbr = None
2560 f.update({
2561 'width': width,
2562 'height': height,
2563 'tbr': tbr,
2564 'format_id': s_attr.get('label') or s_attr.get('title'),
2565 })
2566 f.update(formats[0])
2567 media_info['formats'].append(f)
2568 else:
2569 media_info['formats'].extend(formats)
2570 for track_tag in re.findall(r'<track[^>]+>', media_content):
2571 track_attributes = extract_attributes(track_tag)
2572 kind = track_attributes.get('kind')
2573 if not kind or kind in ('subtitles', 'captions'):
2574 src = strip_or_none(track_attributes.get('src'))
2575 if not src:
2576 continue
2577 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2578 media_info['subtitles'].setdefault(lang, []).append({
2579 'url': absolute_url(src),
2580 })
2581 for f in media_info['formats']:
2582 f.setdefault('http_headers', {})['Referer'] = base_url
2583 if media_info['formats'] or media_info['subtitles']:
2584 entries.append(media_info)
2585 return entries
2586
2587 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2588 formats = []
2589 hdcore_sign = 'hdcore=3.7.0'
2590 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2591 hds_host = hosts.get('hds')
2592 if hds_host:
2593 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2594 if 'hdcore=' not in f4m_url:
2595 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2596 f4m_formats = self._extract_f4m_formats(
2597 f4m_url, video_id, f4m_id='hds', fatal=False)
2598 for entry in f4m_formats:
2599 entry.update({'extra_param_to_segment_url': hdcore_sign})
2600 formats.extend(f4m_formats)
2601 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2602 hls_host = hosts.get('hls')
2603 if hls_host:
2604 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2605 formats.extend(self._extract_m3u8_formats(
2606 m3u8_url, video_id, 'mp4', 'm3u8_native',
2607 m3u8_id='hls', fatal=False))
2608 return formats
2609
2610 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2611 query = compat_urlparse.urlparse(url).query
2612 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2613 mobj = re.search(
2614 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2615 url_base = mobj.group('url')
2616 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2617 formats = []
2618
2619 def manifest_url(manifest):
2620 m_url = '%s/%s' % (http_base_url, manifest)
2621 if query:
2622 m_url += '?%s' % query
2623 return m_url
2624
2625 if 'm3u8' not in skip_protocols:
2626 formats.extend(self._extract_m3u8_formats(
2627 manifest_url('playlist.m3u8'), video_id, 'mp4',
2628 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2629 if 'f4m' not in skip_protocols:
2630 formats.extend(self._extract_f4m_formats(
2631 manifest_url('manifest.f4m'),
2632 video_id, f4m_id='hds', fatal=False))
2633 if 'dash' not in skip_protocols:
2634 formats.extend(self._extract_mpd_formats(
2635 manifest_url('manifest.mpd'),
2636 video_id, mpd_id='dash', fatal=False))
2637 if re.search(r'(?:/smil:|\.smil)', url_base):
2638 if 'smil' not in skip_protocols:
2639 rtmp_formats = self._extract_smil_formats(
2640 manifest_url('jwplayer.smil'),
2641 video_id, fatal=False)
2642 for rtmp_format in rtmp_formats:
2643 rtsp_format = rtmp_format.copy()
2644 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2645 del rtsp_format['play_path']
2646 del rtsp_format['ext']
2647 rtsp_format.update({
2648 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2649 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2650 'protocol': 'rtsp',
2651 })
2652 formats.extend([rtmp_format, rtsp_format])
2653 else:
2654 for protocol in ('rtmp', 'rtsp'):
2655 if protocol not in skip_protocols:
2656 formats.append({
2657 'url': '%s:%s' % (protocol, url_base),
2658 'format_id': protocol,
2659 'protocol': protocol,
2660 })
2661 return formats
2662
2663 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2664 mobj = re.search(
2665 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2666 webpage)
2667 if mobj:
2668 try:
2669 jwplayer_data = self._parse_json(mobj.group('options'),
2670 video_id=video_id,
2671 transform_source=transform_source)
2672 except ExtractorError:
2673 pass
2674 else:
2675 if isinstance(jwplayer_data, dict):
2676 return jwplayer_data
2677
2678 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2679 jwplayer_data = self._find_jwplayer_data(
2680 webpage, video_id, transform_source=js_to_json)
2681 return self._parse_jwplayer_data(
2682 jwplayer_data, video_id, *args, **kwargs)
2683
2684 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2685 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2686 # JWPlayer backward compatibility: flattened playlists
2687 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2688 if 'playlist' not in jwplayer_data:
2689 jwplayer_data = {'playlist': [jwplayer_data]}
2690
2691 entries = []
2692
2693 # JWPlayer backward compatibility: single playlist item
2694 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2695 if not isinstance(jwplayer_data['playlist'], list):
2696 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2697
2698 for video_data in jwplayer_data['playlist']:
2699 # JWPlayer backward compatibility: flattened sources
2700 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2701 if 'sources' not in video_data:
2702 video_data['sources'] = [video_data]
2703
2704 this_video_id = video_id or video_data['mediaid']
2705
2706 formats = self._parse_jwplayer_formats(
2707 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2708 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2709
2710 subtitles = {}
2711 tracks = video_data.get('tracks')
2712 if tracks and isinstance(tracks, list):
2713 for track in tracks:
2714 if not isinstance(track, dict):
2715 continue
2716 track_kind = track.get('kind')
2717 if not track_kind or not isinstance(track_kind, compat_str):
2718 continue
2719 if track_kind.lower() not in ('captions', 'subtitles'):
2720 continue
2721 track_url = urljoin(base_url, track.get('file'))
2722 if not track_url:
2723 continue
2724 subtitles.setdefault(track.get('label') or 'en', []).append({
2725 'url': self._proto_relative_url(track_url)
2726 })
2727
2728 entry = {
2729 'id': this_video_id,
2730 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2731 'description': clean_html(video_data.get('description')),
2732 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2733 'timestamp': int_or_none(video_data.get('pubdate')),
2734 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2735 'subtitles': subtitles,
2736 }
2737 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2738 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2739 entry.update({
2740 '_type': 'url_transparent',
2741 'url': formats[0]['url'],
2742 })
2743 else:
2744 self._sort_formats(formats)
2745 entry['formats'] = formats
2746 entries.append(entry)
2747 if len(entries) == 1:
2748 return entries[0]
2749 else:
2750 return self.playlist_result(entries)
2751
2752 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2753 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2754 urls = []
2755 formats = []
2756 for source in jwplayer_sources_data:
2757 if not isinstance(source, dict):
2758 continue
2759 source_url = urljoin(
2760 base_url, self._proto_relative_url(source.get('file')))
2761 if not source_url or source_url in urls:
2762 continue
2763 urls.append(source_url)
2764 source_type = source.get('type') or ''
2765 ext = mimetype2ext(source_type) or determine_ext(source_url)
2766 if source_type == 'hls' or ext == 'm3u8':
2767 formats.extend(self._extract_m3u8_formats(
2768 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2769 m3u8_id=m3u8_id, fatal=False))
2770 elif source_type == 'dash' or ext == 'mpd':
2771 formats.extend(self._extract_mpd_formats(
2772 source_url, video_id, mpd_id=mpd_id, fatal=False))
2773 elif ext == 'smil':
2774 formats.extend(self._extract_smil_formats(
2775 source_url, video_id, fatal=False))
2776 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2777 elif source_type.startswith('audio') or ext in (
2778 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2779 formats.append({
2780 'url': source_url,
2781 'vcodec': 'none',
2782 'ext': ext,
2783 })
2784 else:
2785 height = int_or_none(source.get('height'))
2786 if height is None:
2787 # Often no height is provided but there is a label in
2788 # format like "1080p", "720p SD", or 1080.
2789 height = int_or_none(self._search_regex(
2790 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2791 'height', default=None))
2792 a_format = {
2793 'url': source_url,
2794 'width': int_or_none(source.get('width')),
2795 'height': height,
2796 'tbr': int_or_none(source.get('bitrate')),
2797 'ext': ext,
2798 }
2799 if source_url.startswith('rtmp'):
2800 a_format['ext'] = 'flv'
2801 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2802 # of jwplayer.flash.swf
2803 rtmp_url_parts = re.split(
2804 r'((?:mp4|mp3|flv):)', source_url, 1)
2805 if len(rtmp_url_parts) == 3:
2806 rtmp_url, prefix, play_path = rtmp_url_parts
2807 a_format.update({
2808 'url': rtmp_url,
2809 'play_path': prefix + play_path,
2810 })
2811 if rtmp_params:
2812 a_format.update(rtmp_params)
2813 formats.append(a_format)
2814 return formats
2815
2816 def _live_title(self, name):
2817 """ Generate the title for a live video """
2818 now = datetime.datetime.now()
2819 now_str = now.strftime('%Y-%m-%d %H:%M')
2820 return name + ' ' + now_str
2821
2822 def _int(self, v, name, fatal=False, **kwargs):
2823 res = int_or_none(v, **kwargs)
2824 if 'get_attr' in kwargs:
2825 print(getattr(v, kwargs['get_attr']))
2826 if res is None:
2827 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2828 if fatal:
2829 raise ExtractorError(msg)
2830 else:
2831 self._downloader.report_warning(msg)
2832 return res
2833
2834 def _float(self, v, name, fatal=False, **kwargs):
2835 res = float_or_none(v, **kwargs)
2836 if res is None:
2837 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2838 if fatal:
2839 raise ExtractorError(msg)
2840 else:
2841 self._downloader.report_warning(msg)
2842 return res
2843
2844 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2845 path='/', secure=False, discard=False, rest={}, **kwargs):
2846 cookie = compat_cookiejar_Cookie(
2847 0, name, value, port, port is not None, domain, True,
2848 domain.startswith('.'), path, True, secure, expire_time,
2849 discard, None, None, rest)
2850 self._downloader.cookiejar.set_cookie(cookie)
2851
2852 def _get_cookies(self, url):
2853 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2854 req = sanitized_Request(url)
2855 self._downloader.cookiejar.add_cookie_header(req)
2856 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2857
2858 def _apply_first_set_cookie_header(self, url_handle, cookie):
2859 """
2860 Apply first Set-Cookie header instead of the last. Experimental.
2861
2862 Some sites (e.g. [1-3]) may serve two cookies under the same name
2863 in Set-Cookie header and expect the first (old) one to be set rather
2864 than second (new). However, as of RFC6265 the newer one cookie
2865 should be set into cookie store what actually happens.
2866 We will workaround this issue by resetting the cookie to
2867 the first one manually.
2868 1. https://new.vk.com/
2869 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
2870 3. https://learning.oreilly.com/
2871 """
2872 for header, cookies in url_handle.headers.items():
2873 if header.lower() != 'set-cookie':
2874 continue
2875 if sys.version_info[0] >= 3:
2876 cookies = cookies.encode('iso-8859-1')
2877 cookies = cookies.decode('utf-8')
2878 cookie_value = re.search(
2879 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
2880 if cookie_value:
2881 value, domain = cookie_value.groups()
2882 self._set_cookie(domain, cookie, value)
2883 break
2884
2885 def get_testcases(self, include_onlymatching=False):
2886 t = getattr(self, '_TEST', None)
2887 if t:
2888 assert not hasattr(self, '_TESTS'), \
2889 '%s has _TEST and _TESTS' % type(self).__name__
2890 tests = [t]
2891 else:
2892 tests = getattr(self, '_TESTS', [])
2893 for t in tests:
2894 if not include_onlymatching and t.get('only_matching', False):
2895 continue
2896 t['name'] = type(self).__name__[:-len('IE')]
2897 yield t
2898
2899 def is_suitable(self, age_limit):
2900 """ Test whether the extractor is generally suitable for the given
2901 age limit (i.e. pornographic sites are not, all others usually are) """
2902
2903 any_restricted = False
2904 for tc in self.get_testcases(include_onlymatching=False):
2905 if tc.get('playlist', []):
2906 tc = tc['playlist'][0]
2907 is_restricted = age_restricted(
2908 tc.get('info_dict', {}).get('age_limit'), age_limit)
2909 if not is_restricted:
2910 return True
2911 any_restricted = any_restricted or is_restricted
2912 return not any_restricted
2913
2914 def extract_subtitles(self, *args, **kwargs):
2915 if (self._downloader.params.get('writesubtitles', False)
2916 or self._downloader.params.get('listsubtitles')):
2917 return self._get_subtitles(*args, **kwargs)
2918 return {}
2919
2920 def _get_subtitles(self, *args, **kwargs):
2921 raise NotImplementedError('This method must be implemented by subclasses')
2922
2923 @staticmethod
2924 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2925 """ Merge subtitle items for one language. Items with duplicated URLs
2926 will be dropped. """
2927 list1_urls = set([item['url'] for item in subtitle_list1])
2928 ret = list(subtitle_list1)
2929 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2930 return ret
2931
2932 @classmethod
2933 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2934 """ Merge two subtitle dictionaries, language by language. """
2935 ret = dict(subtitle_dict1)
2936 for lang in subtitle_dict2:
2937 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2938 return ret
2939
2940 def extract_automatic_captions(self, *args, **kwargs):
2941 if (self._downloader.params.get('writeautomaticsub', False)
2942 or self._downloader.params.get('listsubtitles')):
2943 return self._get_automatic_captions(*args, **kwargs)
2944 return {}
2945
2946 def _get_automatic_captions(self, *args, **kwargs):
2947 raise NotImplementedError('This method must be implemented by subclasses')
2948
2949 def mark_watched(self, *args, **kwargs):
2950 if (self._downloader.params.get('mark_watched', False)
2951 and (self._get_login_info()[0] is not None
2952 or self._downloader.params.get('cookiefile') is not None)):
2953 self._mark_watched(*args, **kwargs)
2954
2955 def _mark_watched(self, *args, **kwargs):
2956 raise NotImplementedError('This method must be implemented by subclasses')
2957
2958 def geo_verification_headers(self):
2959 headers = {}
2960 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2961 if geo_verification_proxy:
2962 headers['Ytdl-request-proxy'] = geo_verification_proxy
2963 return headers
2964
2965 def _generic_id(self, url):
2966 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2967
2968 def _generic_title(self, url):
2969 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2970
2971
2972 class SearchInfoExtractor(InfoExtractor):
2973 """
2974 Base class for paged search queries extractors.
2975 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2976 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2977 """
2978
2979 @classmethod
2980 def _make_valid_url(cls):
2981 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2982
2983 @classmethod
2984 def suitable(cls, url):
2985 return re.match(cls._make_valid_url(), url) is not None
2986
2987 def _real_extract(self, query):
2988 mobj = re.match(self._make_valid_url(), query)
2989 if mobj is None:
2990 raise ExtractorError('Invalid search query "%s"' % query)
2991
2992 prefix = mobj.group('prefix')
2993 query = mobj.group('query')
2994 if prefix == '':
2995 return self._get_n_results(query, 1)
2996 elif prefix == 'all':
2997 return self._get_n_results(query, self._MAX_RESULTS)
2998 else:
2999 n = int(prefix)
3000 if n <= 0:
3001 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3002 elif n > self._MAX_RESULTS:
3003 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3004 n = self._MAX_RESULTS
3005 return self._get_n_results(query, n)
3006
3007 def _get_n_results(self, query, n):
3008 """Get a specified number of results for a query"""
3009 raise NotImplementedError('This method must be implemented by subclasses')
3010
3011 @property
3012 def SEARCH_KEY(self):
3013 return self._SEARCH_KEY