2 from __future__
import unicode_literals
18 from ..compat
import (
19 compat_cookiejar_Cookie
,
22 compat_etree_fromstring
,
29 compat_urllib_parse_unquote
,
30 compat_urllib_parse_urlencode
,
31 compat_urllib_request
,
33 compat_xml_parse_error
,
35 from ..downloader
import FileDownloader
36 from ..downloader
.f4m
import (
38 remove_encrypted_media
,
66 parse_m3u8_attributes
,
88 class InfoExtractor(object):
89 """Information Extractor class.
91 Information extractors are the classes that, given a URL, extract
92 information about the video (or videos) the URL refers to. This
93 information includes the real video URL, the video title, author and
94 others. The information is stored in a dictionary which is then
95 passed to the YoutubeDL. The YoutubeDL processes this
96 information possibly downloading the video to the file system, among
97 other possible outcomes.
99 The type field determines the type of the result.
100 By far the most common value (and the default if _type is missing) is
101 "video", which indicates a single video.
103 For a video, the dictionaries must include the following fields:
105 id: Video identifier.
106 title: Video title, unescaped.
108 Additionally, it must contain either a formats entry or a url one:
110 formats: A list of dictionaries for each format available, ordered
111 from worst to best quality.
114 * url The mandatory URL representing the media:
115 for plain file media - HTTP URL of this file,
117 for HLS - URL of the M3U8 media playlist,
118 for HDS - URL of the F4M manifest,
120 - HTTP URL to plain file media (in case of
122 - URL of the MPD manifest or base URL
123 representing the media if MPD manifest
124 is parsed from a string (in case of
126 for MSS - URL of the ISM manifest.
128 The URL of the manifest file in case of
130 for HLS - URL of the M3U8 master playlist,
131 for HDS - URL of the F4M manifest,
132 for DASH - URL of the MPD manifest,
133 for MSS - URL of the ISM manifest.
134 * ext Will be calculated from URL if missing
135 * format A human-readable description of the format
136 ("mp4 container with h264/opus").
137 Calculated from the format_id, width, height.
138 and format_note fields if missing.
139 * format_id A short description of the format
140 ("mp4_h264_opus" or "19").
141 Technically optional, but strongly recommended.
142 * format_note Additional info about the format
143 ("3D" or "DASH video")
144 * width Width of the video, if known
145 * height Height of the video, if known
146 * resolution Textual description of width and height
147 * tbr Average bitrate of audio and video in KBit/s
148 * abr Average audio bitrate in KBit/s
149 * acodec Name of the audio codec in use
150 * asr Audio sampling rate in Hertz
151 * vbr Average video bitrate in KBit/s
153 * vcodec Name of the video codec in use
154 * container Name of the container format
155 * filesize The number of bytes, if known in advance
156 * filesize_approx An estimate for the number of bytes
157 * player_url SWF Player URL (used for rtmpdump).
158 * protocol The protocol that will be used for the actual
159 download, lower-case.
160 "http", "https", "rtsp", "rtmp", "rtmpe",
161 "m3u8", "m3u8_native" or "http_dash_segments".
163 Base URL for fragments. Each fragment's path
164 value (if present) will be relative to
166 * fragments A list of fragments of a fragmented media.
167 Each fragment entry must contain either an url
168 or a path. If an url is present it should be
169 considered by a client. Otherwise both path and
170 fragment_base_url must be present. Here is
171 the list of all potential fields:
172 * "url" - fragment's URL
173 * "path" - fragment's path relative to
175 * "duration" (optional, int or float)
176 * "filesize" (optional, int)
177 * preference Order number of this format. If this field is
178 present and not None, the formats get sorted
179 by this field, regardless of all other values.
180 -1 for default (order by other properties),
181 -2 or smaller for less than default.
182 < -1000 to hide the format (if there is
183 another one which is strictly better)
184 * language Language code, e.g. "de" or "en-US".
185 * language_preference Is this in the language mentioned in
187 10 if it's what the URL is about,
188 -1 for default (don't know),
189 -10 otherwise, other values reserved for now.
190 * quality Order number of the video quality of this
191 format, irrespective of the file format.
192 -1 for default (order by other properties),
193 -2 or smaller for less than default.
194 * source_preference Order number for this video source
195 (quality takes higher priority)
196 -1 for default (order by other properties),
197 -2 or smaller for less than default.
198 * http_headers A dictionary of additional HTTP headers
199 to add to the request.
200 * stretched_ratio If given and not 1, indicates that the
201 video's pixels are not square.
202 width : height ratio as float.
203 * no_resume The server does not support resuming the
204 (HTTP or RTMP) download. Boolean.
205 * downloader_options A dictionary of downloader options as
206 described in FileDownloader
208 url: Final video URL.
209 ext: Video filename extension.
210 format: The video format, defaults to ext (used for --get-format)
211 player_url: SWF Player URL (used for rtmpdump).
213 The following fields are optional:
215 alt_title: A secondary title of the video.
216 display_id An alternative identifier for the video, not necessarily
217 unique, but available before title. Typically, id is
218 something like "4234987", title "Dancing naked mole rats",
219 and display_id "dancing-naked-mole-rats"
220 thumbnails: A list of dictionaries, with the following entries:
221 * "id" (optional, string) - Thumbnail format ID
223 * "preference" (optional, int) - quality of the image
224 * "width" (optional, int)
225 * "height" (optional, int)
226 * "resolution" (optional, string "{width}x{height}",
228 * "filesize" (optional, int)
229 thumbnail: Full URL to a video thumbnail image.
230 description: Full video description.
231 uploader: Full name of the video uploader.
232 license: License name the video is licensed under.
233 creator: The creator of the video.
234 release_date: The date (YYYYMMDD) when the video was released.
235 timestamp: UNIX timestamp of the moment the video became available.
236 upload_date: Video upload date (YYYYMMDD).
237 If not explicitly set, calculated from timestamp.
238 uploader_id: Nickname or id of the video uploader.
239 uploader_url: Full URL to a personal webpage of the video uploader.
240 channel: Full name of the channel the video is uploaded on.
241 Note that channel fields may or may not repeat uploader
242 fields. This depends on a particular extractor.
243 channel_id: Id of the channel.
244 channel_url: Full URL to a channel webpage.
245 location: Physical location where the video was filmed.
246 subtitles: The available subtitles as a dictionary in the format
247 {tag: subformats}. "tag" is usually a language code, and
248 "subformats" is a list sorted from lower to higher
249 preference, each element is a dictionary with the "ext"
251 * "data": The subtitles file contents
252 * "url": A URL pointing to the subtitles file
253 "ext" will be calculated from URL if missing
254 automatic_captions: Like 'subtitles', used by the YoutubeIE for
255 automatically generated captions
256 duration: Length of the video in seconds, as an integer or float.
257 view_count: How many users have watched the video on the platform.
258 like_count: Number of positive ratings of the video
259 dislike_count: Number of negative ratings of the video
260 repost_count: Number of reposts of the video
261 average_rating: Average rating give by users, the scale used depends on the webpage
262 comment_count: Number of comments on the video
263 comments: A list of comments, each with one or more of the following
264 properties (all but one of text or html optional):
265 * "author" - human-readable name of the comment author
266 * "author_id" - user ID of the comment author
268 * "html" - Comment as HTML
269 * "text" - Plain text of the comment
270 * "timestamp" - UNIX timestamp of comment
271 * "parent" - ID of the comment this one is replying to.
272 Set to "root" to indicate that this is a
273 comment to the original video.
274 age_limit: Age restriction for the video, as an integer (years)
275 webpage_url: The URL to the video webpage, if given to yt-dlp it
276 should allow to get the same result again. (It will be set
277 by YoutubeDL if it's missing)
278 categories: A list of categories that the video falls in, for example
280 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
281 is_live: True, False, or None (=unknown). Whether this video is a
282 live stream that goes on instead of a fixed-length video.
283 was_live: True, False, or None (=unknown). Whether this video was
284 originally a live stream.
285 start_time: Time in seconds where the reproduction should start, as
286 specified in the URL.
287 end_time: Time in seconds where the reproduction should end, as
288 specified in the URL.
289 chapters: A list of dictionaries, with the following entries:
290 * "start_time" - The start time of the chapter in seconds
291 * "end_time" - The end time of the chapter in seconds
292 * "title" (optional, string)
293 playable_in_embed: Whether this video is allowed to play in embedded
294 players on other sites. Can be True (=always allowed),
295 False (=never allowed), None (=unknown), or a string
296 specifying the criteria for embedability (Eg: 'whitelist').
298 The following fields should only be used when the video belongs to some logical
301 chapter: Name or title of the chapter the video belongs to.
302 chapter_number: Number of the chapter the video belongs to, as an integer.
303 chapter_id: Id of the chapter the video belongs to, as a unicode string.
305 The following fields should only be used when the video is an episode of some
306 series, programme or podcast:
308 series: Title of the series or programme the video episode belongs to.
309 season: Title of the season the video episode belongs to.
310 season_number: Number of the season the video episode belongs to, as an integer.
311 season_id: Id of the season the video episode belongs to, as a unicode string.
312 episode: Title of the video episode. Unlike mandatory video title field,
313 this field should denote the exact title of the video episode
314 without any kind of decoration.
315 episode_number: Number of the video episode within a season, as an integer.
316 episode_id: Id of the video episode, as a unicode string.
318 The following fields should only be used when the media is a track or a part of
321 track: Title of the track.
322 track_number: Number of the track within an album or a disc, as an integer.
323 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
325 artist: Artist(s) of the track.
326 genre: Genre(s) of the track.
327 album: Title of the album the track belongs to.
328 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
329 album_artist: List of all artists appeared on the album (e.g.
330 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
332 disc_number: Number of the disc or other physical medium the track belongs to,
334 release_year: Year (YYYY) when the album was released.
336 Unless mentioned otherwise, the fields should be Unicode strings.
338 Unless mentioned otherwise, None is equivalent to absence of information.
341 _type "playlist" indicates multiple videos.
342 There must be a key "entries", which is a list, an iterable, or a PagedList
343 object, each element of which is a valid dictionary by this specification.
345 Additionally, playlists can have "id", "title", and any other relevent
346 attributes with the same semantics as videos (see above).
349 _type "multi_video" indicates that there are multiple videos that
350 form a single show, for examples multiple acts of an opera or TV episode.
351 It must have an entries key like a playlist and contain all the keys
352 required for a video at the same time.
355 _type "url" indicates that the video must be extracted from another
356 location, possibly by a different extractor. Its only required key is:
357 "url" - the next URL to extract.
358 The key "ie_key" can be set to the class name (minus the trailing "IE",
359 e.g. "Youtube") if the extractor class is known in advance.
360 Additionally, the dictionary may have any properties of the resolved entity
361 known in advance, for example "title" if the title of the referred video is
365 _type "url_transparent" entities have the same specification as "url", but
366 indicate that the given additional information is more precise than the one
367 associated with the resolved URL.
368 This is useful when a site employs a video service that hosts the video and
369 its technical metadata, but that video service does not embed a useful
370 title, description etc.
373 Subclasses of this one should re-define the _real_initialize() and
374 _real_extract() methods and define a _VALID_URL regexp.
375 Probably, they should also be added to the list of extractors.
377 _GEO_BYPASS attribute may be set to False in order to disable
378 geo restriction bypass mechanisms for a particular extractor.
379 Though it won't disable explicit geo restriction bypass based on
380 country code provided with geo_bypass_country.
382 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
383 countries for this extractor. One of these countries will be used by
384 geo restriction bypass mechanism right away in order to bypass
385 geo restriction, of course, if the mechanism is not disabled.
387 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
388 IP blocks in CIDR notation for this extractor. One of these IP blocks
389 will be used by geo restriction bypass mechanism similarly
392 Finally, the _WORKING attribute should be set to False for broken IEs
393 in order to warn the users and skip the tests.
398 _x_forwarded_for_ip
= None
400 _GEO_COUNTRIES
= None
401 _GEO_IP_BLOCKS
= None
404 def __init__(self
, downloader
=None):
405 """Constructor. Receives an optional downloader."""
407 self
._x
_forwarded
_for
_ip
= None
408 self
.set_downloader(downloader
)
411 def suitable(cls
, url
):
412 """Receives a URL and returns True if suitable for this IE."""
414 # This does not use has/getattr intentionally - we want to know whether
415 # we have cached the regexp for *this* class, whereas getattr would also
416 # match the superclass
417 if '_VALID_URL_RE' not in cls
.__dict
__:
418 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
419 return cls
._VALID
_URL
_RE
.match(url
) is not None
422 def _match_id(cls
, url
):
423 if '_VALID_URL_RE' not in cls
.__dict
__:
424 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
425 m
= cls
._VALID
_URL
_RE
.match(url
)
427 return compat_str(m
.group('id'))
431 """Getter method for _WORKING."""
434 def initialize(self
):
435 """Initializes an instance (authentication, etc)."""
436 self
._initialize
_geo
_bypass
({
437 'countries': self
._GEO
_COUNTRIES
,
438 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
441 self
._real
_initialize
()
444 def _initialize_geo_bypass(self
, geo_bypass_context
):
446 Initialize geo restriction bypass mechanism.
448 This method is used to initialize geo bypass mechanism based on faking
449 X-Forwarded-For HTTP header. A random country from provided country list
450 is selected and a random IP belonging to this country is generated. This
451 IP will be passed as X-Forwarded-For HTTP header in all subsequent
454 This method will be used for initial geo bypass mechanism initialization
455 during the instance initialization with _GEO_COUNTRIES and
458 You may also manually call it from extractor's code if geo bypass
459 information is not available beforehand (e.g. obtained during
460 extraction) or due to some other reason. In this case you should pass
461 this information in geo bypass context passed as first argument. It may
462 contain following fields:
464 countries: List of geo unrestricted countries (similar
466 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
467 (similar to _GEO_IP_BLOCKS)
470 if not self
._x
_forwarded
_for
_ip
:
472 # Geo bypass mechanism is explicitly disabled by user
473 if not self
._downloader
.params
.get('geo_bypass', True):
476 if not geo_bypass_context
:
477 geo_bypass_context
= {}
479 # Backward compatibility: previously _initialize_geo_bypass
480 # expected a list of countries, some 3rd party code may still use
482 if isinstance(geo_bypass_context
, (list, tuple)):
483 geo_bypass_context
= {
484 'countries': geo_bypass_context
,
487 # The whole point of geo bypass mechanism is to fake IP
488 # as X-Forwarded-For HTTP header based on some IP block or
491 # Path 1: bypassing based on IP block in CIDR notation
493 # Explicit IP block specified by user, use it right away
494 # regardless of whether extractor is geo bypassable or not
495 ip_block
= self
._downloader
.params
.get('geo_bypass_ip_block', None)
497 # Otherwise use random IP block from geo bypass context but only
498 # if extractor is known as geo bypassable
500 ip_blocks
= geo_bypass_context
.get('ip_blocks')
501 if self
._GEO
_BYPASS
and ip_blocks
:
502 ip_block
= random
.choice(ip_blocks
)
505 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
506 if self
._downloader
.params
.get('verbose', False):
507 self
._downloader
.to_screen(
508 '[debug] Using fake IP %s as X-Forwarded-For.'
509 % self
._x
_forwarded
_for
_ip
)
512 # Path 2: bypassing based on country code
514 # Explicit country code specified by user, use it right away
515 # regardless of whether extractor is geo bypassable or not
516 country
= self
._downloader
.params
.get('geo_bypass_country', None)
518 # Otherwise use random country code from geo bypass context but
519 # only if extractor is known as geo bypassable
521 countries
= geo_bypass_context
.get('countries')
522 if self
._GEO
_BYPASS
and countries
:
523 country
= random
.choice(countries
)
526 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
527 if self
._downloader
.params
.get('verbose', False):
528 self
._downloader
.to_screen(
529 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
530 % (self
._x
_forwarded
_for
_ip
, country
.upper()))
532 def extract(self
, url
):
533 """Extracts URL information and returns it in list of dicts."""
538 ie_result
= self
._real
_extract
(url
)
539 if self
._x
_forwarded
_for
_ip
:
540 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
542 except GeoRestrictedError
as e
:
543 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
546 except ExtractorError
:
548 except compat_http_client
.IncompleteRead
as e
:
549 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
550 except (KeyError, StopIteration) as e
:
551 raise ExtractorError('An extractor error has occurred.', cause
=e
)
553 def __maybe_fake_ip_and_retry(self
, countries
):
554 if (not self
._downloader
.params
.get('geo_bypass_country', None)
556 and self
._downloader
.params
.get('geo_bypass', True)
557 and not self
._x
_forwarded
_for
_ip
559 country_code
= random
.choice(countries
)
560 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
561 if self
._x
_forwarded
_for
_ip
:
563 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
564 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
568 def set_downloader(self
, downloader
):
569 """Sets the downloader for this IE."""
570 self
._downloader
= downloader
572 def _real_initialize(self
):
573 """Real initialization process. Redefine in subclasses."""
576 def _real_extract(self
, url
):
577 """Real extraction process. Redefine in subclasses."""
582 """A string for getting the InfoExtractor with get_info_extractor"""
583 return compat_str(cls
.__name
__[:-2])
587 return compat_str(type(self
).__name
__[:-2])
590 def __can_accept_status_code(err
, expected_status
):
591 assert isinstance(err
, compat_urllib_error
.HTTPError
)
592 if expected_status
is None:
594 if isinstance(expected_status
, compat_integer_types
):
595 return err
.code
== expected_status
596 elif isinstance(expected_status
, (list, tuple)):
597 return err
.code
in expected_status
598 elif callable(expected_status
):
599 return expected_status(err
.code
) is True
603 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
605 Return the response handle.
607 See _download_webpage docstring for arguments specification.
610 self
.report_download_webpage(video_id
)
611 elif note
is not False:
613 self
.to_screen('%s' % (note
,))
615 self
.to_screen('%s: %s' % (video_id
, note
))
617 # Some sites check X-Forwarded-For HTTP header in order to figure out
618 # the origin of the client behind proxy. This allows bypassing geo
619 # restriction by faking this header's value to IP that belongs to some
620 # geo unrestricted country. We will do so once we encounter any
621 # geo restriction error.
622 if self
._x
_forwarded
_for
_ip
:
623 if 'X-Forwarded-For' not in headers
:
624 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
626 if isinstance(url_or_request
, compat_urllib_request
.Request
):
627 url_or_request
= update_Request(
628 url_or_request
, data
=data
, headers
=headers
, query
=query
)
631 url_or_request
= update_url_query(url_or_request
, query
)
632 if data
is not None or headers
:
633 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
634 exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
635 if hasattr(ssl
, 'CertificateError'):
636 exceptions
.append(ssl
.CertificateError
)
638 return self
._downloader
.urlopen(url_or_request
)
639 except tuple(exceptions
) as err
:
640 if isinstance(err
, compat_urllib_error
.HTTPError
):
641 if self
.__can
_accept
_status
_code
(err
, expected_status
):
642 # Retain reference to error to prevent file object from
643 # being closed before it can be read. Works around the
644 # effects of <https://bugs.python.org/issue15002>
645 # introduced in Python 3.4.1.
652 errnote
= 'Unable to download webpage'
654 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
656 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
658 self
._downloader
.report_warning(errmsg
)
661 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):
663 Return a tuple (page content as string, URL handle).
665 See _download_webpage docstring for arguments specification.
667 # Strip hashes from the URL (#1038)
668 if isinstance(url_or_request
, (compat_str
, str)):
669 url_or_request
= url_or_request
.partition('#')[0]
671 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
675 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
676 return (content
, urlh
)
679 def _guess_encoding_from_content(content_type
, webpage_bytes
):
680 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
682 encoding
= m
.group(1)
684 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
685 webpage_bytes[:1024])
687 encoding = m.group(1).decode('ascii')
688 elif webpage_bytes.startswith(b'\xff\xfe'):
695 def __check_blocked(self, content):
696 first_block = content[:512]
697 if ('<title>Access to this site is blocked</title>' in content
698 and 'Websense' in first_block):
699 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
700 blocked_iframe = self._html_search_regex(
701 r'<iframe src="([^
"]+)"', content,
702 'Websense information URL
', default=None)
704 msg += ' Visit
%s for more details
' % blocked_iframe
705 raise ExtractorError(msg, expected=True)
706 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
708 'Access to this webpage has been blocked by Indian censorship
. '
709 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
710 block_msg = self._html_search_regex(
711 r'</h1
><p
>(.*?
)</p
>',
712 content, 'block message
', default=None)
714 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
715 raise ExtractorError(msg, expected=True)
716 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
717 and 'blocklist
.rkn
.gov
.ru
' in content):
718 raise ExtractorError(
719 'Access to this webpage has been blocked by decision of the Russian government
. '
720 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
723 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
724 content_type = urlh.headers.get('Content
-Type
', '')
725 webpage_bytes = urlh.read()
726 if prefix is not None:
727 webpage_bytes = prefix + webpage_bytes
729 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
730 if self._downloader.params.get('dump_intermediate_pages
', False):
731 self.to_screen('Dumping request to
' + urlh.geturl())
732 dump = base64.b64encode(webpage_bytes).decode('ascii
')
733 self._downloader.to_screen(dump)
734 if self._downloader.params.get('write_pages
', False):
735 basen = '%s_%s' % (video_id, urlh.geturl())
737 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
738 basen = basen[:240 - len(h)] + h
739 raw_filename = basen + '.dump
'
740 filename = sanitize_filename(raw_filename, restricted=True)
741 self.to_screen('Saving request to
' + filename)
742 # Working around MAX_PATH limitation on Windows (see
743 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
744 if compat_os_name == 'nt
':
745 absfilepath = os.path.abspath(filename)
746 if len(absfilepath) > 259:
747 filename = '\\\\?
\\' + absfilepath
748 with open(filename, 'wb
') as outf:
749 outf.write(webpage_bytes)
752 content = webpage_bytes.decode(encoding, 'replace
')
754 content = webpage_bytes.decode('utf
-8', 'replace
')
756 self.__check_blocked(content)
760 def _download_webpage(
761 self, url_or_request, video_id, note=None, errnote=None,
762 fatal=True, tries=1, timeout=5, encoding=None, data=None,
763 headers={}, query={}, expected_status=None):
765 Return the data of the page as a string.
768 url_or_request -- plain text URL as a string or
769 a compat_urllib_request.Requestobject
770 video_id -- Video/playlist/item identifier (string)
773 note -- note printed before downloading (string)
774 errnote -- note printed in case of an error (string)
775 fatal -- flag denoting whether error should be considered fatal,
776 i.e. whether it should cause ExtractionError to be raised,
777 otherwise a warning will be reported and extraction continued
778 tries -- number of tries
779 timeout -- sleep interval between tries
780 encoding -- encoding for a page content decoding, guessed automatically
781 when not explicitly specified
782 data -- POST data (bytes)
783 headers -- HTTP headers (dict)
784 query -- URL query (dict)
785 expected_status -- allows to accept failed HTTP requests (non 2xx
786 status code) by explicitly specifying a set of accepted status
787 codes. Can be any of the following entities:
788 - an integer type specifying an exact failed status code to
790 - a list or a tuple of integer types specifying a list of
791 failed status codes to accept
792 - a callable accepting an actual failed status code and
793 returning True if it should be accepted
794 Note that this argument does not affect success status codes (2xx)
795 which are always accepted.
800 while success is False:
802 res = self._download_webpage_handle(
803 url_or_request, video_id, note, errnote, fatal,
804 encoding=encoding, data=data, headers=headers, query=query,
805 expected_status=expected_status)
807 except compat_http_client.IncompleteRead as e:
809 if try_count >= tries:
811 self._sleep(timeout, video_id)
818 def _download_xml_handle(
819 self, url_or_request, video_id, note='Downloading XML
',
820 errnote='Unable to download XML
', transform_source=None,
821 fatal=True, encoding=None, data=None, headers={}, query={},
822 expected_status=None):
824 Return a tuple (xml as an compat_etree_Element, URL handle).
826 See _download_webpage docstring for arguments specification.
828 res = self._download_webpage_handle(
829 url_or_request, video_id, note, errnote, fatal=fatal,
830 encoding=encoding, data=data, headers=headers, query=query,
831 expected_status=expected_status)
834 xml_string, urlh = res
835 return self._parse_xml(
836 xml_string, video_id, transform_source=transform_source,
840 self, url_or_request, video_id,
841 note='Downloading XML
', errnote='Unable to download XML
',
842 transform_source=None, fatal=True, encoding=None,
843 data=None, headers={}, query={}, expected_status=None):
845 Return the xml as an compat_etree_Element.
847 See _download_webpage docstring for arguments specification.
849 res = self._download_xml_handle(
850 url_or_request, video_id, note=note, errnote=errnote,
851 transform_source=transform_source, fatal=fatal, encoding=encoding,
852 data=data, headers=headers, query=query,
853 expected_status=expected_status)
854 return res if res is False else res[0]
856 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
858 xml_string = transform_source(xml_string)
860 return compat_etree_fromstring(xml_string.encode('utf
-8'))
861 except compat_xml_parse_error as ve:
862 errmsg = '%s: Failed to parse XML
' % video_id
864 raise ExtractorError(errmsg, cause=ve)
866 self.report_warning(errmsg + str(ve))
868 def _download_json_handle(
869 self, url_or_request, video_id, note='Downloading JSON metadata
',
870 errnote='Unable to download JSON metadata
', transform_source=None,
871 fatal=True, encoding=None, data=None, headers={}, query={},
872 expected_status=None):
874 Return a tuple (JSON object, URL handle).
876 See _download_webpage docstring for arguments specification.
878 res = self._download_webpage_handle(
879 url_or_request, video_id, note, errnote, fatal=fatal,
880 encoding=encoding, data=data, headers=headers, query=query,
881 expected_status=expected_status)
884 json_string, urlh = res
885 return self._parse_json(
886 json_string, video_id, transform_source=transform_source,
890 self, url_or_request, video_id, note='Downloading JSON metadata
',
891 errnote='Unable to download JSON metadata
', transform_source=None,
892 fatal=True, encoding=None, data=None, headers={}, query={},
893 expected_status=None):
895 Return the JSON object as a dict.
897 See _download_webpage docstring for arguments specification.
899 res = self._download_json_handle(
900 url_or_request, video_id, note=note, errnote=errnote,
901 transform_source=transform_source, fatal=fatal, encoding=encoding,
902 data=data, headers=headers, query=query,
903 expected_status=expected_status)
904 return res if res is False else res[0]
906 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
908 json_string = transform_source(json_string)
910 return json.loads(json_string)
911 except ValueError as ve:
912 errmsg = '%s: Failed to parse JSON
' % video_id
914 raise ExtractorError(errmsg, cause=ve)
916 self.report_warning(errmsg + str(ve))
918 def report_warning(self, msg, video_id=None):
919 idstr = '' if video_id is None else '%s: ' % video_id
920 self._downloader.report_warning(
921 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
923 def to_screen(self, msg):
924 """Print msg to screen, prefixing it with '[ie_name
]'"""
925 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
927 def report_extraction(self, id_or_name):
928 """Report information extraction."""
929 self.to_screen('%s: Extracting information
' % id_or_name)
931 def report_download_webpage(self, video_id):
932 """Report webpage download."""
933 self.to_screen('%s: Downloading webpage
' % video_id)
935 def report_age_confirmation(self):
936 """Report attempt to confirm age."""
937 self.to_screen('Confirming age
')
939 def report_login(self):
940 """Report attempt to log in."""
941 self.to_screen('Logging
in')
944 def raise_login_required(msg='This video
is only available
for registered users
'):
945 raise ExtractorError(
946 '%s. Use
--username
and --password
or --netrc to provide account credentials
.' % msg,
950 def raise_geo_restricted(msg='This video
is not available
from your location due to geo restriction
', countries=None):
951 raise GeoRestrictedError(msg, countries=countries)
953 # Methods for following #608
955 def url_result(url, ie=None, video_id=None, video_title=None):
956 """Returns a URL that points to a page that should be processed"""
957 # TODO: ie should be the class used for getting the info
958 video_info = {'_type
': 'url
',
961 if video_id is not None:
962 video_info['id'] = video_id
963 if video_title is not None:
964 video_info['title
'] = video_title
967 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
969 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
971 return self.playlist_result(
972 urls, playlist_id=playlist_id, playlist_title=playlist_title)
975 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
976 """Returns a playlist"""
977 video_info = {'_type
': 'playlist
',
979 video_info.update(kwargs)
981 video_info['id'] = playlist_id
983 video_info['title
'] = playlist_title
984 if playlist_description is not None:
985 video_info['description
'] = playlist_description
988 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
990 Perform a regex search on the given string, using a single or a list of
991 patterns returning the first matching group.
992 In case of failure return a default value or raise a WARNING or a
993 RegexNotFoundError, depending on fatal, specifying the field name.
995 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
996 mobj = re.search(pattern, string, flags)
999 mobj = re.search(p, string, flags)
1003 if not self._downloader.params.get('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
1004 _name = '\033[0;34m
%s\033[0m
' % name
1010 # return the first matching group
1011 return next(g for g in mobj.groups() if g is not None)
1013 return mobj.group(group)
1014 elif default is not NO_DEFAULT:
1017 raise RegexNotFoundError('Unable to extract
%s' % _name)
1019 self._downloader.report_warning('unable to extract
%s' % _name + bug_reports_message())
1022 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1024 Like _search_regex, but strips HTML tags and unescapes entities.
1026 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1028 return clean_html(res).strip()
1032 def _get_netrc_login_info(self, netrc_machine=None):
1035 netrc_machine = netrc_machine or self._NETRC_MACHINE
1037 if self._downloader.params.get('usenetrc
', False):
1039 info = netrc.netrc().authenticators(netrc_machine)
1040 if info is not None:
1044 raise netrc.NetrcParseError(
1045 'No authenticators
for %s' % netrc_machine)
1046 except (IOError, netrc.NetrcParseError) as err:
1047 self._downloader.report_warning(
1048 'parsing
.netrc
: %s' % error_to_compat_str(err))
1050 return username, password
1052 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1054 Get the login info as (username, password)
1055 First look for the manually specified credentials using username_option
1056 and password_option as keys in params dictionary. If no such credentials
1057 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1059 If there's no info available
, return (None, None)
1061 if self._downloader is None:
1064 downloader_params = self._downloader.params
1066 # Attempt to use provided username and password or .netrc data
1067 if downloader_params.get(username_option) is not None:
1068 username = downloader_params[username_option]
1069 password = downloader_params[password_option]
1071 username, password = self._get_netrc_login_info(netrc_machine)
1073 return username, password
1075 def _get_tfa_info(self, note='two-factor verification code'):
1077 Get the two
-factor authentication info
1078 TODO
- asking the user will be required
for sms
/phone verify
1079 currently just uses the command line option
1080 If there
's no info available, return None
1082 if self._downloader is None:
1084 downloader_params = self._downloader.params
1086 if downloader_params.get('twofactor
') is not None:
1087 return downloader_params['twofactor
']
1089 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1091 # Helper functions for extracting OpenGraph info
1093 def _og_regexes(prop):
1094 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1095 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1096 % {'prop': re.escape(prop)})
1097 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1099 template % (property_re, content_re),
1100 template % (content_re, property_re),
1104 def _meta_regex(prop):
1105 return r'''(?isx)<meta
1106 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1107 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1109 def _og_search_property(self, prop, html, name=None, **kargs):
1110 if not isinstance(prop, (list, tuple)):
1113 name = 'OpenGraph
%s' % prop[0]
1116 og_regexes.extend(self._og_regexes(p))
1117 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1120 return unescapeHTML(escaped)
1122 def _og_search_thumbnail(self, html, **kargs):
1123 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1125 def _og_search_description(self, html, **kargs):
1126 return self._og_search_property('description
', html, fatal=False, **kargs)
1128 def _og_search_title(self, html, **kargs):
1129 return self._og_search_property('title
', html, **kargs)
1131 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1132 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1134 regexes = self._og_regexes('video
:secure_url
') + regexes
1135 return self._html_search_regex(regexes, html, name, **kargs)
1137 def _og_search_url(self, html, **kargs):
1138 return self._og_search_property('url
', html, **kargs)
1140 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1141 if not isinstance(name, (list, tuple)):
1143 if display_name is None:
1144 display_name = name[0]
1145 return self._html_search_regex(
1146 [self._meta_regex(n) for n in name],
1147 html, display_name, fatal=fatal, group='content
', **kwargs)
1149 def _dc_search_uploader(self, html):
1150 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1152 def _rta_search(self, html):
1153 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1154 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1155 r' content
="RTA-5042-1996-1400-1577-RTA"',
1160 def _media_rating_search(self, html):
1161 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1162 rating = self._html_search_meta('rating
', html)
1174 return RATING_TABLE.get(rating.lower())
1176 def _family_friendly_search(self, html):
1177 # See http://schema.org/VideoObject
1178 family_friendly = self._html_search_meta(
1179 'isFamilyFriendly
', html, default=None)
1181 if not family_friendly:
1190 return RATING_TABLE.get(family_friendly.lower())
1192 def _twitter_search_player(self, html):
1193 return self._html_search_meta('twitter
:player
', html,
1194 'twitter card player
')
1196 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1197 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1198 default = kwargs.get('default
', NO_DEFAULT)
1199 # JSON-LD may be malformed and thus `fatal` should be respected.
1200 # At the same time `default` may be passed that assumes `fatal=False`
1201 # for _search_regex. Let's simulate the same behavior here
as well
.
1202 fatal
= kwargs
.get('fatal', True) if default
== NO_DEFAULT
else False
1204 for mobj
in json_ld_list
:
1205 json_ld_item
= self
._parse
_json
(
1206 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1207 if not json_ld_item
:
1209 if isinstance(json_ld_item
, dict):
1210 json_ld
.append(json_ld_item
)
1211 elif isinstance(json_ld_item
, (list, tuple)):
1212 json_ld
.extend(json_ld_item
)
1214 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1217 if default
is not NO_DEFAULT
:
1220 raise RegexNotFoundError('Unable to extract JSON-LD')
1222 self
._downloader
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1225 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1226 if isinstance(json_ld
, compat_str
):
1227 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1231 if not isinstance(json_ld
, (list, tuple, dict)):
1233 if isinstance(json_ld
, dict):
1236 INTERACTION_TYPE_MAP
= {
1237 'CommentAction': 'comment',
1238 'AgreeAction': 'like',
1239 'DisagreeAction': 'dislike',
1240 'LikeAction': 'like',
1241 'DislikeAction': 'dislike',
1242 'ListenAction': 'view',
1243 'WatchAction': 'view',
1244 'ViewAction': 'view',
1247 def extract_interaction_type(e
):
1248 interaction_type
= e
.get('interactionType')
1249 if isinstance(interaction_type
, dict):
1250 interaction_type
= interaction_type
.get('@type')
1251 return str_or_none(interaction_type
)
1253 def extract_interaction_statistic(e
):
1254 interaction_statistic
= e
.get('interactionStatistic')
1255 if isinstance(interaction_statistic
, dict):
1256 interaction_statistic
= [interaction_statistic
]
1257 if not isinstance(interaction_statistic
, list):
1259 for is_e
in interaction_statistic
:
1260 if not isinstance(is_e
, dict):
1262 if is_e
.get('@type') != 'InteractionCounter':
1264 interaction_type
= extract_interaction_type(is_e
)
1265 if not interaction_type
:
1267 # For interaction count some sites provide string instead of
1268 # an integer (as per spec) with non digit characters (e.g. ",")
1269 # so extracting count with more relaxed str_to_int
1270 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1271 if interaction_count
is None:
1273 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1276 count_key
= '%s_count' % count_kind
1277 if info
.get(count_key
) is not None:
1279 info
[count_key
] = interaction_count
1281 def extract_video_object(e
):
1282 assert e
['@type'] == 'VideoObject'
1284 'url': url_or_none(e
.get('contentUrl')),
1285 'title': unescapeHTML(e
.get('name')),
1286 'description': unescapeHTML(e
.get('description')),
1287 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1288 'duration': parse_duration(e
.get('duration')),
1289 'timestamp': unified_timestamp(e
.get('uploadDate')),
1290 'uploader': str_or_none(e
.get('author')),
1291 'filesize': float_or_none(e
.get('contentSize')),
1292 'tbr': int_or_none(e
.get('bitrate')),
1293 'width': int_or_none(e
.get('width')),
1294 'height': int_or_none(e
.get('height')),
1295 'view_count': int_or_none(e
.get('interactionCount')),
1297 extract_interaction_statistic(e
)
1301 item_type
= e
.get('@type')
1302 if expected_type
is not None and expected_type
!= item_type
:
1304 if item_type
in ('TVEpisode', 'Episode'):
1305 episode_name
= unescapeHTML(e
.get('name'))
1307 'episode': episode_name
,
1308 'episode_number': int_or_none(e
.get('episodeNumber')),
1309 'description': unescapeHTML(e
.get('description')),
1311 if not info
.get('title') and episode_name
:
1312 info
['title'] = episode_name
1313 part_of_season
= e
.get('partOfSeason')
1314 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1316 'season': unescapeHTML(part_of_season
.get('name')),
1317 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1319 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1320 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1321 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1322 elif item_type
== 'Movie':
1324 'title': unescapeHTML(e
.get('name')),
1325 'description': unescapeHTML(e
.get('description')),
1326 'duration': parse_duration(e
.get('duration')),
1327 'timestamp': unified_timestamp(e
.get('dateCreated')),
1329 elif item_type
in ('Article', 'NewsArticle'):
1331 'timestamp': parse_iso8601(e
.get('datePublished')),
1332 'title': unescapeHTML(e
.get('headline')),
1333 'description': unescapeHTML(e
.get('articleBody')),
1335 elif item_type
== 'VideoObject':
1336 extract_video_object(e
)
1337 if expected_type
is None:
1341 video
= e
.get('video')
1342 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1343 extract_video_object(video
)
1344 if expected_type
is None:
1348 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1351 def _hidden_inputs(html
):
1352 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1354 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1355 attrs
= extract_attributes(input)
1358 if attrs
.get('type') not in ('hidden', 'submit'):
1360 name
= attrs
.get('name') or attrs
.get('id')
1361 value
= attrs
.get('value')
1362 if name
and value
is not None:
1363 hidden_inputs
[name
] = value
1364 return hidden_inputs
1366 def _form_hidden_inputs(self
, form_id
, html
):
1367 form
= self
._search
_regex
(
1368 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1369 html, '%s form
' % form_id, group='form
')
1370 return self._hidden_inputs(form)
1373 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<seperator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1375 default = ('hidden
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1376 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1377 'proto
', 'ext
', 'has_audio
', 'source
', 'format_id
') # These must not be aliases
1380 'vcodec
': {'type': 'ordered
', 'regex
': True,
1381 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1382 'acodec
': {'type': 'ordered
', 'regex
': True,
1383 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1384 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1385 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', 'm3u8
', '.*dash
', '', 'mms|rtsp
', 'none
', 'f4
']},
1386 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1387 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1388 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1389 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1390 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1391 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1392 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1393 'ie_pref
': {'priority': True, 'type': 'extractor'},
1394 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1395 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1396 'lang
': {'priority': True, 'convert': 'ignore', 'type': 'extractor', 'field': 'language_preference'},
1397 'quality
': {'convert': 'float_none', 'type': 'extractor'},
1398 'filesize
': {'convert': 'bytes'},
1399 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1400 'id': {'convert': 'string', 'field': 'format_id'},
1401 'height
': {'convert': 'float_none'},
1402 'width
': {'convert': 'float_none'},
1403 'fps
': {'convert': 'float_none'},
1404 'tbr
': {'convert': 'float_none'},
1405 'vbr
': {'convert': 'float_none'},
1406 'abr
': {'convert': 'float_none'},
1407 'asr
': {'convert': 'float_none'},
1408 'source
': {'convert': 'ignore', 'type': 'extractor', 'field': 'source_preference'},
1410 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1411 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1412 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1413 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1414 'res
': {'type': 'multiple', 'field': ('height', 'width'), 'function': min},
1416 # Most of these exist only for compatibility reasons
1417 'dimension
': {'type': 'alias', 'field': 'res'},
1418 'resolution
': {'type': 'alias', 'field': 'res'},
1419 'extension
': {'type': 'alias', 'field': 'ext'},
1420 'bitrate
': {'type': 'alias', 'field': 'br'},
1421 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1422 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1423 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1424 'framerate
': {'type': 'alias', 'field': 'fps'},
1425 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1426 'protocol
': {'type': 'alias', 'field': 'proto'},
1427 'source_preference
': {'type': 'alias', 'field': 'source'},
1428 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1429 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1430 'samplerate
': {'type': 'alias', 'field': 'asr'},
1431 'video_ext
': {'type': 'alias', 'field': 'vext'},
1432 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1433 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1434 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1435 'video
': {'type': 'alias', 'field': 'hasvid'},
1436 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1437 'audio
': {'type': 'alias', 'field': 'hasaud'},
1438 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1439 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1440 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1441 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1442 'format_id
': {'type': 'alias', 'field': 'id'},
1447 def _get_field_setting(self, field, key):
1448 if field not in self.settings:
1449 self.settings[field] = {}
1450 propObj = self.settings[field]
1451 if key not in propObj:
1452 type = propObj.get('type')
1454 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1455 elif key == 'convert
':
1456 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1458 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,), 'function': max}.get(key, None)
1459 propObj[key] = default
1462 def _resolve_field_value(self, field, value, convertNone=False):
1467 value = value.lower()
1468 conversion = self._get_field_setting(field, 'convert
')
1469 if conversion == 'ignore
':
1471 if conversion == 'string
':
1473 elif conversion == 'float_none
':
1474 return float_or_none(value)
1475 elif conversion == 'bytes':
1476 return FileDownloader.parse_bytes(value)
1477 elif conversion == 'order
':
1478 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1479 use_regex = self._get_field_setting(field, 'regex
')
1480 list_length = len(order_list)
1481 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1482 if use_regex and value is not None:
1483 for i, regex in enumerate(order_list):
1484 if regex and re.match(regex, value):
1485 return list_length - i
1486 return list_length - empty_pos # not in list
1487 else: # not regex or value = None
1488 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1490 if value.isnumeric():
1493 self.settings[field]['convert
'] = 'string
'
1496 def evaluate_params(self, params, sort_extractor):
1497 self._use_free_order = params.get('prefer_free_formats
', False)
1498 self._sort_user = params.get('format_sort
', [])
1499 self._sort_extractor = sort_extractor
1501 def add_item(field, reverse, closest, limit_text):
1502 field = field.lower()
1503 if field in self._order:
1505 self._order.append(field)
1506 limit = self._resolve_field_value(field, limit_text)
1509 'closest
': False if limit is None else closest,
1510 'limit_text
': limit_text,
1512 if field in self.settings:
1513 self.settings[field].update(data)
1515 self.settings[field] = data
1518 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1519 + (tuple() if params.get('format_sort_force
', False)
1520 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1521 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1523 for item in sort_list:
1524 match = re.match(self.regex, item)
1526 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1527 field = match.group('field
')
1530 if self._get_field_setting(field, 'type') == 'alias
':
1531 field = self._get_field_setting(field, 'field
')
1532 reverse = match.group('reverse
') is not None
1533 closest = match.group('seperator
') == '~
'
1534 limit_text = match.group('limit
')
1536 has_limit = limit_text is not None
1537 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1538 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1540 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1541 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1542 limit_count = len(limits)
1543 for (i, f) in enumerate(fields):
1544 add_item(f, reverse, closest,
1545 limits[i] if i < limit_count
1546 else limits[0] if has_limit and not has_multiple_limits
1549 def print_verbose_info(self, to_screen):
1550 to_screen('[debug
] Sort order given by user
: %s' % ','.join(self._sort_user))
1551 if self._sort_extractor:
1552 to_screen('[debug
] Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1553 to_screen('[debug
] Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1554 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1555 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1556 self._get_field_setting(field, 'limit_text
'),
1557 self._get_field_setting(field, 'limit
'))
1558 if self._get_field_setting(field, 'limit_text
') is not None else '')
1559 for field in self._order if self._get_field_setting(field, 'visible
')]))
1561 def _calculate_field_preference_from_value(self, format, field, type, value):
1562 reverse = self._get_field_setting(field, 'reverse
')
1563 closest = self._get_field_setting(field, 'closest
')
1564 limit = self._get_field_setting(field, 'limit
')
1566 if type == 'extractor
':
1567 maximum = self._get_field_setting(field, 'max')
1568 if value is None or (maximum is not None and value >= maximum):
1570 elif type == 'boolean
':
1571 in_list = self._get_field_setting(field, 'in_list
')
1572 not_in_list = self._get_field_setting(field, 'not_in_list
')
1573 value = 0 if ((in_list is None or value in in_list) and (not_in_list is None or value not in not_in_list)) else -1
1574 elif type == 'ordered
':
1575 value = self._resolve_field_value(field, value, True)
1577 # try to convert to number
1578 val_num = float_or_none(value)
1579 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1583 return ((-10, 0) if value is None
1584 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1585 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1586 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1587 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1588 else (-1, value, 0))
1590 def _calculate_field_preference(self, format, field):
1591 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1592 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1593 if type == 'multiple
':
1594 type = 'field
' # Only 'field
' is allowed in multiple for now
1595 actual_fields = self._get_field_setting(field, 'field
')
1597 def wrapped_function(values):
1598 values = tuple(filter(lambda x: x is not None, values))
1599 return (self._get_field_setting(field, 'function
')(*values) if len(values) > 1
1600 else values[0] if values
1603 value = wrapped_function((get_value(f) for f in actual_fields))
1605 value = get_value(field)
1606 return self._calculate_field_preference_from_value(format, field, type, value)
1608 def calculate_preference(self, format):
1609 # Determine missing protocol
1610 if not format.get('protocol
'):
1611 format['protocol
'] = determine_protocol(format)
1613 # Determine missing ext
1614 if not format.get('ext
') and 'url
' in format:
1615 format['ext
'] = determine_ext(format['url
'])
1616 if format.get('vcodec
') == 'none
':
1617 format['audio_ext
'] = format['ext
']
1618 format['video_ext
'] = 'none
'
1620 format['video_ext
'] = format['ext
']
1621 format['audio_ext
'] = 'none
'
1622 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1623 # format['preference
'] = -1000
1625 # Determine missing bitrates
1626 if format.get('tbr
') is None:
1627 if format.get('vbr
') is not None and format.get('abr
') is not None:
1628 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1630 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1631 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1632 if format.get('acodec
') != "none" and format.get('abr
') is None:
1633 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1635 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1637 def _sort_formats(self, formats, field_preference=[]):
1639 raise ExtractorError('No video formats found
')
1640 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1641 format_sort.evaluate_params(self._downloader.params, field_preference)
1642 if self._downloader.params.get('verbose
', False):
1643 format_sort.print_verbose_info(self._downloader.to_screen)
1644 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1646 def _check_formats(self, formats, video_id):
1648 formats[:] = filter(
1649 lambda f: self._is_valid_url(
1651 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1655 def _remove_duplicate_formats(formats):
1659 if f['url
'] not in format_urls:
1660 format_urls.add(f['url
'])
1661 unique_formats.append(f)
1662 formats[:] = unique_formats
1664 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1665 url = self._proto_relative_url(url, scheme='http
:')
1666 # For now assume non HTTP(S) URLs always valid
1667 if not (url.startswith('http
://') or url.startswith('https
://')):
1670 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1672 except ExtractorError as e:
1674 '%s: %s URL
is invalid
, skipping
: %s'
1675 % (video_id, item, error_to_compat_str(e.cause)))
1678 def http_scheme(self):
1679 """ Either "http:" or "https:", depending on the user's preferences
"""
1682 if self._downloader.params.get('prefer_insecure', False)
1685 def _proto_relative_url(self, url, scheme=None):
1688 if url.startswith('//'):
1690 scheme = self.http_scheme()
1695 def _sleep(self, timeout, video_id, msg_template=None):
1696 if msg_template is None:
1697 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1698 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1702 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1703 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1704 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1705 manifest = self._download_xml(
1706 manifest_url, video_id, 'Downloading f4m manifest',
1707 'Unable to download f4m manifest',
1708 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1709 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1710 transform_source=transform_source,
1711 fatal=fatal, data=data, headers=headers, query=query)
1713 if manifest is False:
1716 return self._parse_f4m_formats(
1717 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1718 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1720 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1721 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1722 fatal=True, m3u8_id=None):
1723 if not isinstance(manifest, compat_etree_Element) and not fatal:
1726 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1727 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1728 if akamai_pv is not None and ';' in akamai_pv.text:
1729 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1730 if playerVerificationChallenge.strip() != '':
1734 manifest_version = '1.0'
1735 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1737 manifest_version = '2.0'
1738 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1739 # Remove unsupported DRM protected media from final formats
1740 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1741 media_nodes = remove_encrypted_media(media_nodes)
1745 manifest_base_url = get_base_url(manifest)
1747 bootstrap_info = xpath_element(
1748 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1749 'bootstrap info', default=None)
1752 mime_type = xpath_text(
1753 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1754 'base URL', default=None)
1755 if mime_type and mime_type.startswith('audio/'):
1758 for i, media_el in enumerate(media_nodes):
1759 tbr = int_or_none(media_el.attrib.get('bitrate'))
1760 width = int_or_none(media_el.attrib.get('width'))
1761 height = int_or_none(media_el.attrib.get('height'))
1762 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1763 # If <bootstrapInfo> is present, the specified f4m is a
1764 # stream-level manifest, and only set-level manifests may refer to
1765 # external resources. See section 11.4 and section 4 of F4M spec
1766 if bootstrap_info is None:
1768 # @href is introduced in 2.0, see section 11.6 of F4M spec
1769 if manifest_version == '2.0':
1770 media_url = media_el.attrib.get('href')
1771 if media_url is None:
1772 media_url = media_el.attrib.get('url')
1776 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1777 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1778 # If media_url is itself a f4m manifest do the recursive extraction
1779 # since bitrates in parent manifest (this one) and media_url manifest
1780 # may differ leading to inability to resolve the format by requested
1781 # bitrate in f4m downloader
1782 ext = determine_ext(manifest_url)
1784 f4m_formats = self._extract_f4m_formats(
1785 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1786 transform_source=transform_source, fatal=fatal)
1787 # Sometimes stream-level manifest contains single media entry that
1788 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1789 # At the same time parent's media entry in set-level manifest may
1790 # contain it. We will copy it from parent in such cases.
1791 if len(f4m_formats) == 1:
1794 'tbr': f.get('tbr') or tbr,
1795 'width': f.get('width') or width,
1796 'height': f.get('height') or height,
1797 'format_id': f.get('format_id') if not tbr else format_id,
1800 formats.extend(f4m_formats)
1803 formats.extend(self._extract_m3u8_formats(
1804 manifest_url, video_id, 'mp4', preference=preference,
1805 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1808 'format_id': format_id,
1809 'url': manifest_url,
1810 'manifest_url': manifest_url,
1811 'ext': 'flv' if bootstrap_info is not None else None,
1817 'preference': preference,
1822 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1824 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1828 'preference': preference - 100 if preference else -100,
1830 'resolution': 'multiple',
1831 'format_note': 'Quality selection URL',
1834 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1835 entry_protocol='m3u8', preference=None, quality=None,
1836 m3u8_id=None, live=False, note=None, errnote=None,
1837 fatal=True, data=None, headers={}, query={}):
1838 res = self._download_webpage_handle(
1840 note=note or 'Downloading m3u8 information',
1841 errnote=errnote or 'Failed to download m3u8 information',
1842 fatal=fatal, data=data, headers=headers, query=query)
1847 m3u8_doc, urlh = res
1848 m3u8_url = urlh.geturl()
1850 return self._parse_m3u8_formats(
1851 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1852 preference=preference, quality=quality, m3u8_id=m3u8_id,
1853 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1854 headers=headers, query=query, video_id=video_id)
1856 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1857 entry_protocol='m3u8', preference=None, quality=None,
1858 m3u8_id=None, live=False, note=None, errnote=None,
1859 fatal=True, data=None, headers={}, query={}, video_id=None):
1860 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1863 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1868 format_url = lambda u: (
1870 if re.match(r'^https?://', u)
1871 else compat_urlparse.urljoin(m3u8_url, u))
1873 split_discontinuity = self._downloader.params.get('hls_split_discontinuity', False)
1876 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1877 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1878 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1880 # We should try extracting formats only from master playlists [1, 4.3.4],
1881 # i.e. playlists that describe available qualities. On the other hand
1882 # media playlists [1, 4.3.3] should be returned as is since they contain
1883 # just the media without qualities renditions.
1884 # Fortunately, master playlist can be easily distinguished from media
1885 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1886 # master playlist tags MUST NOT appear in a media playlist and vice versa.
1887 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1888 # media playlist and MUST NOT appear in master playlist thus we can
1889 # clearly detect media playlist with this criterion.
1891 def _extract_m3u8_playlist_formats(format_url, m3u8_doc=None):
1893 res = self._download_webpage_handle(
1894 format_url, video_id,
1896 errnote=errnote or 'Failed to download m3u8 playlist information',
1897 fatal=fatal, data=data, headers=headers, query=query)
1902 m3u8_doc, urlh = res
1903 format_url = urlh.geturl()
1905 playlist_formats = []
1908 if split_discontinuity
1915 for line in m3u8_doc.splitlines():
1916 if not line.startswith('#'):
1917 format_info['files'].append(line)
1918 elif split_discontinuity and line.startswith('#EXT-X-DISCONTINUITY'):
1920 playlist_formats.append(format_info)
1926 playlist_formats.append(format_info)
1927 return playlist_formats
1929 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1931 playlist_formats = _extract_m3u8_playlist_formats(m3u8_doc, True)
1933 for format in playlist_formats:
1936 format_id.append(m3u8_id)
1937 format_index = format.get('index')
1939 format_id.append(str(format_index))
1941 'format_id': '-'.join(format_id),
1942 'format_index': format_index,
1945 'protocol': entry_protocol,
1946 'preference': preference,
1954 last_stream_inf = {}
1956 def extract_media(x_media_line):
1957 media = parse_m3u8_attributes(x_media_line)
1958 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1959 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1960 if not (media_type and group_id and name):
1962 groups.setdefault(group_id, []).append(media)
1963 if media_type not in ('VIDEO', 'AUDIO'):
1965 media_url = media.get('URI')
1967 manifest_url = format_url(media_url)
1969 playlist_formats = _extract_m3u8_playlist_formats(manifest_url)
1971 for format in playlist_formats:
1972 format_index = format.get('index')
1973 for v in (m3u8_id, group_id, name):
1977 format_id.append(str(format_index))
1979 'format_id': '-'.join(format_id),
1980 'format_index': format_index,
1981 'url': manifest_url,
1982 'manifest_url': m3u8_url,
1983 'language': media.get('LANGUAGE'),
1985 'protocol': entry_protocol,
1986 'preference': preference,
1989 if media_type == 'AUDIO':
1990 f['vcodec'] = 'none'
1993 def build_stream_name():
1994 # Despite specification does not mention NAME attribute for
1995 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1996 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1997 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1998 stream_name = last_stream_inf.get('NAME')
2001 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2002 # from corresponding rendition group
2003 stream_group_id = last_stream_inf.get('VIDEO')
2004 if not stream_group_id:
2006 stream_group = groups.get(stream_group_id)
2007 if not stream_group:
2008 return stream_group_id
2009 rendition = stream_group[0]
2010 return rendition.get('NAME') or stream_group_id
2012 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2013 # chance to detect video only formats when EXT-X-STREAM-INF tags
2014 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2015 for line in m3u8_doc.splitlines():
2016 if line.startswith('#EXT-X-MEDIA:'):
2019 for line in m3u8_doc.splitlines():
2020 if line.startswith('#EXT-X-STREAM-INF:'):
2021 last_stream_inf = parse_m3u8_attributes(line)
2022 elif line.startswith('#') or not line.strip():
2025 tbr = float_or_none(
2026 last_stream_inf.get('AVERAGE-BANDWIDTH')
2027 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2028 manifest_url = format_url(line.strip())
2030 playlist_formats = _extract_m3u8_playlist_formats(manifest_url)
2032 for format in playlist_formats:
2035 format_id.append(m3u8_id)
2036 format_index = format.get('index')
2037 stream_name = build_stream_name()
2038 # Bandwidth of live streams may differ over time thus making
2039 # format_id unpredictable. So it's better to keep provided
2042 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
2044 format_id.append(str(format_index))
2046 'format_id': '-'.join(format_id),
2047 'format_index': format_index,
2048 'url': manifest_url,
2049 'manifest_url': m3u8_url,
2052 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2053 'protocol': entry_protocol,
2054 'preference': preference,
2057 resolution = last_stream_inf.get('RESOLUTION')
2059 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2061 f['width'] = int(mobj.group('width'))
2062 f['height'] = int(mobj.group('height'))
2063 # Unified Streaming Platform
2065 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2067 abr, vbr = mobj.groups()
2068 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2073 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2075 audio_group_id = last_stream_inf.get('AUDIO')
2076 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2077 # references a rendition group MUST have a CODECS attribute.
2078 # However, this is not always respected, for example, [2]
2079 # contains EXT-X-STREAM-INF tag which references AUDIO
2080 # rendition group but does not have CODECS and despite
2081 # referencing an audio group it represents a complete
2082 # (with audio and video) format. So, for such cases we will
2083 # ignore references to rendition groups and treat them
2084 # as complete formats.
2085 if audio_group_id and codecs and f.get('vcodec') != 'none':
2086 audio_group = groups.get(audio_group_id)
2087 if audio_group and audio_group[0].get('URI'):
2088 # TODO: update acodec for audio only formats with
2090 f['acodec'] = 'none'
2094 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2097 del http_f['manifest_url']
2099 'format_id': f['format_id'].replace('hls-', 'http-'),
2101 'url': progressive_uri,
2103 formats.append(http_f)
2105 last_stream_inf = {}
2109 def _xpath_ns(path, namespace=None):
2113 for c in path.split('/'):
2114 if not c or c == '.':
2117 out.append('{%s}%s' % (namespace, c))
2118 return '/'.join(out)
2120 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2121 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2127 namespace = self._parse_smil_namespace(smil)
2129 return self._parse_smil_formats(
2130 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2132 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2133 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2136 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2138 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2139 return self._download_xml(
2140 smil_url, video_id, 'Downloading SMIL file',
2141 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2143 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2144 namespace = self._parse_smil_namespace(smil)
2146 formats = self._parse_smil_formats(
2147 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2148 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2150 video_id = os.path.splitext(url_basename(smil_url))[0]
2154 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2155 name = meta.attrib.get('name')
2156 content = meta.attrib.get('content')
2157 if not name or not content:
2159 if not title and name == 'title':
2161 elif not description and name in ('description', 'abstract'):
2162 description = content
2163 elif not upload_date and name == 'date':
2164 upload_date = unified_strdate(content)
2167 'id': image.get('type'),
2168 'url': image.get('src'),
2169 'width': int_or_none(image.get('width')),
2170 'height': int_or_none(image.get('height')),
2171 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2175 'title': title or video_id,
2176 'description': description,
2177 'upload_date': upload_date,
2178 'thumbnails': thumbnails,
2180 'subtitles': subtitles,
2183 def _parse_smil_namespace(self, smil):
2184 return self._search_regex(
2185 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2187 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2189 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2190 b = meta.get('base') or meta.get('httpBase')
2201 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2202 for medium in media:
2203 src = medium.get('src')
2204 if not src or src in srcs:
2208 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2209 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2210 width = int_or_none(medium.get('width'))
2211 height = int_or_none(medium.get('height'))
2212 proto = medium.get('proto')
2213 ext = medium.get('ext')
2214 src_ext = determine_ext(src)
2215 streamer = medium.get('streamer') or base
2217 if proto == 'rtmp' or streamer.startswith('rtmp'):
2223 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2225 'filesize': filesize,
2229 if transform_rtmp_url:
2230 streamer, src = transform_rtmp_url(streamer, src)
2231 formats[-1].update({
2237 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2238 src_url = src_url.strip()
2240 if proto == 'm3u8' or src_ext == 'm3u8':
2241 m3u8_formats = self._extract_m3u8_formats(
2242 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2243 if len(m3u8_formats) == 1:
2245 m3u8_formats[0].update({
2246 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2251 formats.extend(m3u8_formats)
2252 elif src_ext == 'f4m':
2257 'plugin': 'flowplayer-3.2.0.1',
2259 f4m_url += '&' if '?' in f4m_url else '?'
2260 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2261 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2262 elif src_ext == 'mpd':
2263 formats.extend(self._extract_mpd_formats(
2264 src_url, video_id, mpd_id='dash', fatal=False))
2265 elif re.search(r'\.ism/[Mm]anifest', src_url):
2266 formats.extend(self._extract_ism_formats(
2267 src_url, video_id, ism_id='mss', fatal=False))
2268 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2272 'ext': ext or src_ext or 'flv',
2273 'format_id': 'http-%d' % (bitrate or http_count),
2275 'filesize': filesize,
2282 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2285 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2286 src = textstream.get('src')
2287 if not src or src in urls:
2290 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2291 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2292 subtitles.setdefault(lang, []).append({
2298 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2299 xspf = self._download_xml(
2300 xspf_url, playlist_id, 'Downloading xpsf playlist',
2301 'Unable to download xspf manifest', fatal=fatal)
2304 return self._parse_xspf(
2305 xspf, playlist_id, xspf_url=xspf_url,
2306 xspf_base_url=base_url(xspf_url))
2308 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2310 'xspf': 'http://xspf.org/ns/0/',
2311 's1': 'http://static.streamone.nl/player/ns/0',
2315 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2317 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2318 description = xpath_text(
2319 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2320 thumbnail = xpath_text(
2321 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2322 duration = float_or_none(
2323 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2326 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2327 format_url = urljoin(xspf_base_url, location.text)
2332 'manifest_url': xspf_url,
2333 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2334 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2335 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2337 self._sort_formats(formats)
2342 'description': description,
2343 'thumbnail': thumbnail,
2344 'duration': duration,
2349 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2350 res = self._download_xml_handle(
2352 note=note or 'Downloading MPD manifest',
2353 errnote=errnote or 'Failed to download MPD manifest',
2354 fatal=fatal, data=data, headers=headers, query=query)
2360 mpd_base_url = base_url(urlh.geturl())
2362 return self._parse_mpd_formats(
2363 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2365 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2367 Parse formats
from MPD manifest
.
2369 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2370 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2371 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2373 if not self._downloader.params.get('dynamic_mpd'):
2374 if mpd_doc.get('type') == 'dynamic':
2377 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2380 return self._xpath_ns(path, namespace)
2382 def is_drm_protected(element):
2383 return element.find(_add_ns('ContentProtection')) is not None
2385 def extract_multisegment_info(element, ms_parent_info):
2386 ms_info = ms_parent_info.copy()
2388 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2389 # common attributes and elements. We will only extract relevant
2391 def extract_common(source):
2392 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2393 if segment_timeline is not None:
2394 s_e = segment_timeline.findall(_add_ns('S'))
2396 ms_info['total_number'] = 0
2399 r = int(s.get('r', 0))
2400 ms_info['total_number'] += 1 + r
2401 ms_info['s'].append({
2402 't': int(s.get('t', 0)),
2403 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2404 'd': int(s.attrib['d']),
2407 start_number = source.get('startNumber')
2409 ms_info['start_number'] = int(start_number)
2410 timescale = source.get('timescale')
2412 ms_info['timescale'] = int(timescale)
2413 segment_duration = source.get('duration')
2414 if segment_duration:
2415 ms_info['segment_duration'] = float(segment_duration)
2417 def extract_Initialization(source):
2418 initialization = source.find(_add_ns('Initialization'))
2419 if initialization is not None:
2420 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2422 segment_list = element.find(_add_ns('SegmentList'))
2423 if segment_list is not None:
2424 extract_common(segment_list)
2425 extract_Initialization(segment_list)
2426 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2428 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2430 segment_template = element.find(_add_ns('SegmentTemplate'))
2431 if segment_template is not None:
2432 extract_common(segment_template)
2433 media = segment_template.get('media')
2435 ms_info['media'] = media
2436 initialization = segment_template.get('initialization')
2438 ms_info['initialization'] = initialization
2440 extract_Initialization(segment_template)
2443 skip_unplayable = not self._downloader.params.get('allow_unplayable_formats')
2445 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2447 for period in mpd_doc.findall(_add_ns('Period')):
2448 period_duration = parse_duration(period.get('duration')) or mpd_duration
2449 period_ms_info = extract_multisegment_info(period, {
2453 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2454 if skip_unplayable and is_drm_protected(adaptation_set):
2456 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2457 for representation in adaptation_set.findall(_add_ns('Representation')):
2458 if skip_unplayable and is_drm_protected(representation):
2460 representation_attrib = adaptation_set.attrib.copy()
2461 representation_attrib.update(representation.attrib)
2462 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2463 mime_type = representation_attrib['mimeType']
2464 content_type = mime_type.split('/')[0]
2465 if content_type == 'text':
2466 # TODO implement WebVTT downloading
2468 elif content_type in ('video', 'audio'):
2470 for element in (representation, adaptation_set, period, mpd_doc):
2471 base_url_e = element.find(_add_ns('BaseURL'))
2472 if base_url_e is not None:
2473 base_url = base_url_e.text + base_url
2474 if re.match(r'^https?://', base_url):
2476 if mpd_base_url and not re.match(r'^https?://', base_url):
2477 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2479 base_url = mpd_base_url + base_url
2480 representation_id = representation_attrib.get('id')
2481 lang = representation_attrib.get('lang')
2482 url_el = representation.find(_add_ns('BaseURL'))
2483 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2484 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2486 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2487 'manifest_url': mpd_url,
2488 'ext': mimetype2ext(mime_type),
2489 'width': int_or_none(representation_attrib.get('width')),
2490 'height': int_or_none(representation_attrib.get('height')),
2491 'tbr': float_or_none(bandwidth, 1000),
2492 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2493 'fps': int_or_none(representation_attrib.get('frameRate')),
2494 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2495 'format_note': 'DASH %s' % content_type,
2496 'filesize': filesize,
2497 'container': mimetype2ext(mime_type) + '_dash',
2499 f.update(parse_codecs(representation_attrib.get('codecs')))
2500 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2502 def prepare_template(template_name, identifiers):
2503 tmpl = representation_ms_info[template_name]
2504 # First of, % characters outside $...$ templates
2505 # must be escaped by doubling for proper processing
2506 # by % operator string formatting used further (see
2507 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2513 in_template = not in_template
2514 elif c == '%' and not in_template:
2516 # Next, $...$ templates are translated to their
2517 # %(...) counterparts to be used with % operator
2518 t = t.replace('$RepresentationID$', representation_id)
2519 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2520 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2521 t.replace('$$', '$')
2524 # @initialization is a regular template like @media one
2525 # so it should be handled just the same way (see
2526 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2527 if 'initialization' in representation_ms_info:
2528 initialization_template = prepare_template(
2530 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2531 # $Time$ shall not be included for @initialization thus
2532 # only $Bandwidth$ remains
2534 representation_ms_info['initialization_url'] = initialization_template % {
2535 'Bandwidth': bandwidth,
2538 def location_key(location):
2539 return 'url' if re.match(r'^https?://', location) else 'path'
2541 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2543 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2544 media_location_key = location_key(media_template)
2546 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2547 # can't be used at the same time
2548 if '%(Number' in media_template and 's' not in representation_ms_info:
2549 segment_duration = None
2550 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2551 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2552 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2553 representation_ms_info['fragments'] = [{
2554 media_location_key: media_template % {
2555 'Number': segment_number,
2556 'Bandwidth': bandwidth,
2558 'duration': segment_duration,
2559 } for segment_number in range(
2560 representation_ms_info['start_number'],
2561 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2563 # $Number*$ or $Time$ in media template with S list available
2564 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2565 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2566 representation_ms_info['fragments'] = []
2569 segment_number = representation_ms_info['start_number']
2571 def add_segment_url():
2572 segment_url = media_template % {
2573 'Time': segment_time,
2574 'Bandwidth': bandwidth,
2575 'Number': segment_number,
2577 representation_ms_info['fragments'].append({
2578 media_location_key: segment_url,
2579 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2582 for num, s in enumerate(representation_ms_info['s']):
2583 segment_time = s.get('t') or segment_time
2587 for r in range(s.get('r', 0)):
2588 segment_time += segment_d
2591 segment_time += segment_d
2592 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2594 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2595 # or any YouTube dashsegments video
2598 timescale = representation_ms_info['timescale']
2599 for s in representation_ms_info['s']:
2600 duration = float_or_none(s['d'], timescale)
2601 for r in range(s.get('r', 0) + 1):
2602 segment_uri = representation_ms_info['segment_urls'][segment_index]
2604 location_key(segment_uri): segment_uri,
2605 'duration': duration,
2608 representation_ms_info['fragments'] = fragments
2609 elif 'segment_urls' in representation_ms_info:
2610 # Segment URLs with no SegmentTimeline
2611 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2612 # https://github.com/ytdl-org/youtube-dl/pull/14844
2614 segment_duration = float_or_none(
2615 representation_ms_info['segment_duration'],
2616 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2617 for segment_url in representation_ms_info['segment_urls']:
2619 location_key(segment_url): segment_url,
2621 if segment_duration:
2622 fragment['duration'] = segment_duration
2623 fragments.append(fragment)
2624 representation_ms_info['fragments'] = fragments
2625 # If there is a fragments key available then we correctly recognized fragmented media.
2626 # Otherwise we will assume unfragmented media with direct access. Technically, such
2627 # assumption is not necessarily correct since we may simply have no support for
2628 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2629 if 'fragments' in representation_ms_info:
2631 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2632 'url': mpd_url or base_url,
2633 'fragment_base_url': base_url,
2635 'protocol': 'http_dash_segments',
2637 if 'initialization_url' in representation_ms_info:
2638 initialization_url = representation_ms_info['initialization_url']
2639 if not f.get('url'):
2640 f['url'] = initialization_url
2641 f['fragments'].append({location_key(initialization_url): initialization_url})
2642 f['fragments'].extend(representation_ms_info['fragments'])
2644 # Assuming direct URL to unfragmented media.
2648 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2651 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2652 res = self._download_xml_handle(
2654 note=note or 'Downloading ISM manifest',
2655 errnote=errnote or 'Failed to download ISM manifest',
2656 fatal=fatal, data=data, headers=headers, query=query)
2663 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2665 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2667 Parse formats
from ISM manifest
.
2669 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2670 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2672 if ism_doc.get('IsLive') == 'TRUE':
2674 if (not self._downloader.params.get('allow_unplayable_formats')
2675 and ism_doc.find('Protection') is not None):
2678 duration = int(ism_doc.attrib['Duration'])
2679 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2682 for stream in ism_doc.findall('StreamIndex'):
2683 stream_type = stream.get('Type')
2684 if stream_type not in ('video', 'audio'):
2686 url_pattern = stream.attrib['Url']
2687 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2688 stream_name = stream.get('Name')
2689 for track in stream.findall('QualityLevel'):
2690 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2691 # TODO: add support for WVC1 and WMAP
2692 if fourcc not in ('H264', 'AVC1', 'AACL'):
2693 self.report_warning('%s is not a supported codec' % fourcc)
2695 tbr = int(track.attrib['Bitrate']) // 1000
2696 # [1] does not mention Width and Height attributes. However,
2697 # they're often present while MaxWidth and MaxHeight are
2698 # missing, so should be used as fallbacks
2699 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2700 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2701 sampling_rate = int_or_none(track.get('SamplingRate'))
2703 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2704 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2710 stream_fragments = stream.findall('c')
2711 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2712 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2713 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2714 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2715 if not fragment_ctx['duration']:
2717 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2719 next_fragment_time = duration
2720 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2721 for _ in range(fragment_repeat):
2723 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2724 'duration': fragment_ctx['duration'] / stream_timescale,
2726 fragment_ctx['time'] += fragment_ctx['duration']
2730 format_id.append(ism_id)
2732 format_id.append(stream_name)
2733 format_id.append(compat_str(tbr))
2736 'format_id': '-'.join(format_id),
2738 'manifest_url': ism_url,
2739 'ext': 'ismv' if stream_type == 'video' else 'isma',
2743 'asr': sampling_rate,
2744 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2745 'acodec': 'none' if stream_type == 'video' else fourcc,
2747 'fragments': fragments,
2748 '_download_params': {
2749 'duration': duration,
2750 'timescale': stream_timescale,
2751 'width': width or 0,
2752 'height': height or 0,
2754 'codec_private_data': track.get('CodecPrivateData'),
2755 'sampling_rate': sampling_rate,
2756 'channels': int_or_none(track.get('Channels', 2)),
2757 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2758 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2763 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None, quality=None):
2764 def absolute_url(item_url):
2765 return urljoin(base_url, item_url)
2767 def parse_content_type(content_type):
2768 if not content_type:
2770 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2772 mimetype, codecs = ctr.groups()
2773 f = parse_codecs(codecs)
2774 f['ext'] = mimetype2ext(mimetype)
2778 def _media_formats(src, cur_media_type, type_info={}):
2779 full_url = absolute_url(src)
2780 ext = type_info.get('ext') or determine_ext(full_url)
2782 is_plain_url = False
2783 formats = self._extract_m3u8_formats(
2784 full_url, video_id, ext='mp4',
2785 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2786 preference=preference, quality=quality, fatal=False)
2788 is_plain_url = False
2789 formats = self._extract_mpd_formats(
2790 full_url, video_id, mpd_id=mpd_id, fatal=False)
2795 'vcodec': 'none' if cur_media_type == 'audio' else None,
2797 return is_plain_url, formats
2800 # amp-video and amp-audio are very similar to their HTML5 counterparts
2801 # so we wll include them right here (see
2802 # https://www.ampproject.org/docs/reference/components/amp-video)
2803 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
2804 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
2805 media_tags = [(media_tag, media_tag_name, media_type, '')
2806 for media_tag, media_tag_name, media_type
2807 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
2808 media_tags.extend(re.findall(
2809 # We only allow video|audio followed by a whitespace or '>'.
2810 # Allowing more characters may end up in significant slow down (see
2811 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2812 # http://www.porntrex.com/maps/videositemap.xml).
2813 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
2814 for media_tag, _, media_type, media_content in media_tags:
2819 media_attributes = extract_attributes(media_tag)
2820 src = strip_or_none(media_attributes.get('src'))
2822 _, formats = _media_formats(src, media_type)
2823 media_info['formats'].extend(formats)
2824 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2826 for source_tag in re.findall(r'<source[^>]+>', media_content):
2827 s_attr = extract_attributes(source_tag)
2828 # data-video-src and data-src are non standard but seen
2829 # several times in the wild
2830 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2833 f = parse_content_type(s_attr.get('type'))
2834 is_plain_url, formats = _media_formats(src, media_type, f)
2836 # width, height, res, label and title attributes are
2837 # all not standard but seen several times in the wild
2840 for lbl in ('label', 'title')
2841 if str_or_none(s_attr.get(lbl))
2843 width = int_or_none(s_attr.get('width'))
2844 height = (int_or_none(s_attr.get('height'))
2845 or int_or_none(s_attr.get('res')))
2846 if not width or not height:
2848 resolution = parse_resolution(lbl)
2851 width = width or resolution.get('width')
2852 height = height or resolution.get('height')
2854 tbr = parse_bitrate(lbl)
2863 'format_id': s_attr.get('label') or s_attr.get('title'),
2865 f.update(formats[0])
2866 media_info['formats'].append(f)
2868 media_info['formats'].extend(formats)
2869 for track_tag in re.findall(r'<track[^>]+>', media_content):
2870 track_attributes = extract_attributes(track_tag)
2871 kind = track_attributes.get('kind')
2872 if not kind or kind in ('subtitles', 'captions'):
2873 src = strip_or_none(track_attributes.get('src'))
2876 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2877 media_info['subtitles'].setdefault(lang, []).append({
2878 'url': absolute_url(src),
2880 for f in media_info['formats']:
2881 f.setdefault('http_headers', {})['Referer'] = base_url
2882 if media_info['formats'] or media_info['subtitles']:
2883 entries.append(media_info)
2886 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2887 signed = 'hdnea=' in manifest_url
2889 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
2890 manifest_url = re.sub(
2891 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
2892 '', manifest_url).strip('?')
2896 hdcore_sign = 'hdcore=3.7.0'
2897 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2898 hds_host = hosts.get('hds')
2900 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2901 if 'hdcore=' not in f4m_url:
2902 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2903 f4m_formats = self._extract_f4m_formats(
2904 f4m_url, video_id, f4m_id='hds', fatal=False)
2905 for entry in f4m_formats:
2906 entry.update({'extra_param_to_segment_url': hdcore_sign})
2907 formats.extend(f4m_formats)
2909 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2910 hls_host = hosts.get('hls')
2912 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2913 m3u8_formats = self._extract_m3u8_formats(
2914 m3u8_url, video_id, 'mp4', 'm3u8_native',
2915 m3u8_id='hls', fatal=False)
2916 formats.extend(m3u8_formats)
2918 http_host = hosts.get('http')
2919 if http_host and m3u8_formats and not signed:
2920 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
2921 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
2922 qualities_length = len(qualities)
2923 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
2925 for f in m3u8_formats:
2926 if f['vcodec'] != 'none':
2927 for protocol in ('http', 'https'):
2929 del http_f['manifest_url']
2931 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
2933 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
2935 'protocol': protocol,
2937 formats.append(http_f)
2942 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2943 query = compat_urlparse.urlparse(url).query
2944 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2946 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2947 url_base = mobj.group('url')
2948 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2951 def manifest_url(manifest):
2952 m_url = '%s/%s' % (http_base_url, manifest)
2954 m_url += '?%s' % query
2957 if 'm3u8' not in skip_protocols:
2958 formats.extend(self._extract_m3u8_formats(
2959 manifest_url('playlist.m3u8'), video_id, 'mp4',
2960 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2961 if 'f4m' not in skip_protocols:
2962 formats.extend(self._extract_f4m_formats(
2963 manifest_url('manifest.f4m'),
2964 video_id, f4m_id='hds', fatal=False))
2965 if 'dash' not in skip_protocols:
2966 formats.extend(self._extract_mpd_formats(
2967 manifest_url('manifest.mpd'),
2968 video_id, mpd_id='dash', fatal=False))
2969 if re.search(r'(?:/smil:|\.smil)', url_base):
2970 if 'smil' not in skip_protocols:
2971 rtmp_formats = self._extract_smil_formats(
2972 manifest_url('jwplayer.smil'),
2973 video_id, fatal=False)
2974 for rtmp_format in rtmp_formats:
2975 rtsp_format = rtmp_format.copy()
2976 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2977 del rtsp_format['play_path']
2978 del rtsp_format['ext']
2979 rtsp_format.update({
2980 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2981 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2984 formats.extend([rtmp_format, rtsp_format])
2986 for protocol in ('rtmp', 'rtsp'):
2987 if protocol not in skip_protocols:
2989 'url': '%s:%s' % (protocol, url_base),
2990 'format_id': protocol,
2991 'protocol': protocol,
2995 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2997 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3001 jwplayer_data = self._parse_json(mobj.group('options'),
3003 transform_source=transform_source)
3004 except ExtractorError:
3007 if isinstance(jwplayer_data, dict):
3008 return jwplayer_data
3010 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3011 jwplayer_data = self._find_jwplayer_data(
3012 webpage, video_id, transform_source=js_to_json)
3013 return self._parse_jwplayer_data(
3014 jwplayer_data, video_id, *args, **kwargs)
3016 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3017 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3018 # JWPlayer backward compatibility: flattened playlists
3019 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3020 if 'playlist' not in jwplayer_data:
3021 jwplayer_data = {'playlist': [jwplayer_data]}
3025 # JWPlayer backward compatibility: single playlist item
3026 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3027 if not isinstance(jwplayer_data['playlist'], list):
3028 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3030 for video_data in jwplayer_data['playlist']:
3031 # JWPlayer backward compatibility: flattened sources
3032 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3033 if 'sources' not in video_data:
3034 video_data['sources'] = [video_data]
3036 this_video_id = video_id or video_data['mediaid']
3038 formats = self._parse_jwplayer_formats(
3039 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3040 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3043 tracks = video_data.get('tracks')
3044 if tracks and isinstance(tracks, list):
3045 for track in tracks:
3046 if not isinstance(track, dict):
3048 track_kind = track.get('kind')
3049 if not track_kind or not isinstance(track_kind, compat_str):
3051 if track_kind.lower() not in ('captions', 'subtitles'):
3053 track_url = urljoin(base_url, track.get('file'))
3056 subtitles.setdefault(track.get('label') or 'en', []).append({
3057 'url': self._proto_relative_url(track_url)
3061 'id': this_video_id,
3062 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3063 'description': clean_html(video_data.get('description')),
3064 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3065 'timestamp': int_or_none(video_data.get('pubdate')),
3066 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3067 'subtitles': subtitles,
3069 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3070 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3072 '_type': 'url_transparent',
3073 'url': formats[0]['url'],
3076 self._sort_formats(formats)
3077 entry['formats'] = formats
3078 entries.append(entry)
3079 if len(entries) == 1:
3082 return self.playlist_result(entries)
3084 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3085 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3088 for source in jwplayer_sources_data:
3089 if not isinstance(source, dict):
3091 source_url = urljoin(
3092 base_url, self._proto_relative_url(source.get('file')))
3093 if not source_url or source_url in urls:
3095 urls.append(source_url)
3096 source_type = source.get('type') or ''
3097 ext = mimetype2ext(source_type) or determine_ext(source_url)
3098 if source_type == 'hls' or ext == 'm3u8':
3099 formats.extend(self._extract_m3u8_formats(
3100 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3101 m3u8_id=m3u8_id, fatal=False))
3102 elif source_type == 'dash' or ext == 'mpd':
3103 formats.extend(self._extract_mpd_formats(
3104 source_url, video_id, mpd_id=mpd_id, fatal=False))
3106 formats.extend(self._extract_smil_formats(
3107 source_url, video_id, fatal=False))
3108 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3109 elif source_type.startswith('audio') or ext in (
3110 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3117 height = int_or_none(source.get('height'))
3119 # Often no height is provided but there is a label in
3120 # format like "1080p", "720p SD", or 1080.
3121 height = int_or_none(self._search_regex(
3122 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3123 'height', default=None))
3126 'width': int_or_none(source.get('width')),
3128 'tbr': int_or_none(source.get('bitrate')),
3131 if source_url.startswith('rtmp'):
3132 a_format['ext'] = 'flv'
3133 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3134 # of jwplayer.flash.swf
3135 rtmp_url_parts = re.split(
3136 r'((?:mp4|mp3|flv):)', source_url, 1)
3137 if len(rtmp_url_parts) == 3:
3138 rtmp_url, prefix, play_path = rtmp_url_parts
3141 'play_path': prefix + play_path,
3144 a_format.update(rtmp_params)
3145 formats.append(a_format)
3148 def _live_title(self, name):
3149 """ Generate the title
for a live video
"""
3150 now = datetime.datetime.now()
3151 now_str = now.strftime('%Y-%m-%d %H:%M')
3152 return name + ' ' + now_str
3154 def _int(self, v, name, fatal=False, **kwargs):
3155 res = int_or_none(v, **kwargs)
3156 if 'get_attr' in kwargs:
3157 print(getattr(v, kwargs['get_attr']))
3159 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3161 raise ExtractorError(msg)
3163 self._downloader.report_warning(msg)
3166 def _float(self, v, name, fatal=False, **kwargs):
3167 res = float_or_none(v, **kwargs)
3169 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3171 raise ExtractorError(msg)
3173 self._downloader.report_warning(msg)
3176 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3177 path='/', secure=False, discard=False, rest={}, **kwargs):
3178 cookie = compat_cookiejar_Cookie(
3179 0, name, value, port, port is not None, domain, True,
3180 domain.startswith('.'), path, True, secure, expire_time,
3181 discard, None, None, rest)
3182 self._downloader.cookiejar.set_cookie(cookie)
3184 def _get_cookies(self, url):
3185 """ Return a compat_cookies
.SimpleCookie
with the cookies
for the url
"""
3186 req = sanitized_Request(url)
3187 self._downloader.cookiejar.add_cookie_header(req)
3188 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
3190 def _apply_first_set_cookie_header(self, url_handle, cookie):
3192 Apply first Set
-Cookie header instead of the last
. Experimental
.
3194 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3195 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3196 than
second (new
). However
, as of RFC6265 the newer one cookie
3197 should be
set into cookie store what actually happens
.
3198 We will workaround this issue by resetting the cookie to
3199 the first one manually
.
3200 1. https
://new
.vk
.com
/
3201 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3202 3. https
://learning
.oreilly
.com
/
3204 for header, cookies in url_handle.headers.items():
3205 if header.lower() != 'set-cookie':
3207 if sys.version_info[0] >= 3:
3208 cookies = cookies.encode('iso-8859-1')
3209 cookies = cookies.decode('utf-8')
3210 cookie_value = re.search(
3211 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3213 value, domain = cookie_value.groups()
3214 self._set_cookie(domain, cookie, value)
3217 def get_testcases(self, include_onlymatching=False):
3218 t = getattr(self, '_TEST', None)
3220 assert not hasattr(self, '_TESTS'), \
3221 '%s has _TEST and _TESTS' % type(self).__name__
3224 tests = getattr(self, '_TESTS', [])
3226 if not include_onlymatching and t.get('only_matching', False):
3228 t['name'] = type(self).__name__[:-len('IE')]
3231 def is_suitable(self, age_limit):
3232 """ Test whether the extractor
is generally suitable
for the given
3233 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3235 any_restricted = False
3236 for tc in self.get_testcases(include_onlymatching=False):
3237 if tc.get('playlist', []):
3238 tc = tc['playlist'][0]
3239 is_restricted = age_restricted(
3240 tc.get('info_dict', {}).get('age_limit'), age_limit)
3241 if not is_restricted:
3243 any_restricted = any_restricted or is_restricted
3244 return not any_restricted
3246 def extract_subtitles(self, *args, **kwargs):
3247 if (self._downloader.params.get('writesubtitles', False)
3248 or self._downloader.params.get('listsubtitles')):
3249 return self._get_subtitles(*args, **kwargs)
3252 def _get_subtitles(self, *args, **kwargs):
3253 raise NotImplementedError('This method must be implemented by subclasses')
3256 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3257 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3258 will be dropped
. """
3259 list1_urls = set([item['url'] for item in subtitle_list1])
3260 ret = list(subtitle_list1)
3261 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3265 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
3266 """ Merge two subtitle dictionaries
, language by language
. """
3267 ret = dict(subtitle_dict1)
3268 for lang in subtitle_dict2:
3269 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
3272 def extract_automatic_captions(self, *args, **kwargs):
3273 if (self._downloader.params.get('writeautomaticsub', False)
3274 or self._downloader.params.get('listsubtitles')):
3275 return self._get_automatic_captions(*args, **kwargs)
3278 def _get_automatic_captions(self, *args, **kwargs):
3279 raise NotImplementedError('This method must be implemented by subclasses')
3281 def mark_watched(self, *args, **kwargs):
3282 if (self._downloader.params.get('mark_watched', False)
3283 and (self._get_login_info()[0] is not None
3284 or self._downloader.params.get('cookiefile') is not None)):
3285 self._mark_watched(*args, **kwargs)
3287 def _mark_watched(self, *args, **kwargs):
3288 raise NotImplementedError('This method must be implemented by subclasses')
3290 def geo_verification_headers(self):
3292 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
3293 if geo_verification_proxy:
3294 headers['Ytdl-request-proxy'] = geo_verification_proxy
3297 def _generic_id(self, url):
3298 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3300 def _generic_title(self, url):
3301 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3304 class SearchInfoExtractor(InfoExtractor):
3306 Base
class for paged search queries extractors
.
3307 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3308 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3312 def _make_valid_url(cls):
3313 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3316 def suitable(cls, url):
3317 return re.match(cls._make_valid_url(), url) is not None
3319 def _real_extract(self, query):
3320 mobj = re.match(self._make_valid_url(), query)
3322 raise ExtractorError('Invalid search query "%s"' % query)
3324 prefix = mobj.group('prefix')
3325 query = mobj.group('query')
3327 return self._get_n_results(query, 1)
3328 elif prefix == 'all':
3329 return self._get_n_results(query, self._MAX_RESULTS)
3333 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3334 elif n > self._MAX_RESULTS:
3335 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3336 n = self._MAX_RESULTS
3337 return self._get_n_results(query, n)
3339 def _get_n_results(self, query, n):
3340 """Get a specified number of results
for a query
"""
3341 raise NotImplementedError('This method must be implemented by subclasses')
3344 def SEARCH_KEY(self):
3345 return self._SEARCH_KEY