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 youtube-dlc 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 start_time: Time in seconds where the reproduction should start, as
284 specified in the URL.
285 end_time: Time in seconds where the reproduction should end, as
286 specified in the URL.
287 chapters: A list of dictionaries, with the following entries:
288 * "start_time" - The start time of the chapter in seconds
289 * "end_time" - The end time of the chapter in seconds
290 * "title" (optional, string)
292 The following fields should only be used when the video belongs to some logical
295 chapter: Name or title of the chapter the video belongs to.
296 chapter_number: Number of the chapter the video belongs to, as an integer.
297 chapter_id: Id of the chapter the video belongs to, as a unicode string.
299 The following fields should only be used when the video is an episode of some
300 series, programme or podcast:
302 series: Title of the series or programme the video episode belongs to.
303 season: Title of the season the video episode belongs to.
304 season_number: Number of the season the video episode belongs to, as an integer.
305 season_id: Id of the season the video episode belongs to, as a unicode string.
306 episode: Title of the video episode. Unlike mandatory video title field,
307 this field should denote the exact title of the video episode
308 without any kind of decoration.
309 episode_number: Number of the video episode within a season, as an integer.
310 episode_id: Id of the video episode, as a unicode string.
312 The following fields should only be used when the media is a track or a part of
315 track: Title of the track.
316 track_number: Number of the track within an album or a disc, as an integer.
317 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
319 artist: Artist(s) of the track.
320 genre: Genre(s) of the track.
321 album: Title of the album the track belongs to.
322 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
323 album_artist: List of all artists appeared on the album (e.g.
324 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
326 disc_number: Number of the disc or other physical medium the track belongs to,
328 release_year: Year (YYYY) when the album was released.
330 Unless mentioned otherwise, the fields should be Unicode strings.
332 Unless mentioned otherwise, None is equivalent to absence of information.
335 _type "playlist" indicates multiple videos.
336 There must be a key "entries", which is a list, an iterable, or a PagedList
337 object, each element of which is a valid dictionary by this specification.
339 Additionally, playlists can have "id", "title", and any other relevent
340 attributes with the same semantics as videos (see above).
343 _type "multi_video" indicates that there are multiple videos that
344 form a single show, for examples multiple acts of an opera or TV episode.
345 It must have an entries key like a playlist and contain all the keys
346 required for a video at the same time.
349 _type "url" indicates that the video must be extracted from another
350 location, possibly by a different extractor. Its only required key is:
351 "url" - the next URL to extract.
352 The key "ie_key" can be set to the class name (minus the trailing "IE",
353 e.g. "Youtube") if the extractor class is known in advance.
354 Additionally, the dictionary may have any properties of the resolved entity
355 known in advance, for example "title" if the title of the referred video is
359 _type "url_transparent" entities have the same specification as "url", but
360 indicate that the given additional information is more precise than the one
361 associated with the resolved URL.
362 This is useful when a site employs a video service that hosts the video and
363 its technical metadata, but that video service does not embed a useful
364 title, description etc.
367 Subclasses of this one should re-define the _real_initialize() and
368 _real_extract() methods and define a _VALID_URL regexp.
369 Probably, they should also be added to the list of extractors.
371 _GEO_BYPASS attribute may be set to False in order to disable
372 geo restriction bypass mechanisms for a particular extractor.
373 Though it won't disable explicit geo restriction bypass based on
374 country code provided with geo_bypass_country.
376 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
377 countries for this extractor. One of these countries will be used by
378 geo restriction bypass mechanism right away in order to bypass
379 geo restriction, of course, if the mechanism is not disabled.
381 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
382 IP blocks in CIDR notation for this extractor. One of these IP blocks
383 will be used by geo restriction bypass mechanism similarly
386 Finally, the _WORKING attribute should be set to False for broken IEs
387 in order to warn the users and skip the tests.
392 _x_forwarded_for_ip
= None
394 _GEO_COUNTRIES
= None
395 _GEO_IP_BLOCKS
= None
398 def __init__(self
, downloader
=None):
399 """Constructor. Receives an optional downloader."""
401 self
._x
_forwarded
_for
_ip
= None
402 self
.set_downloader(downloader
)
405 def suitable(cls
, url
):
406 """Receives a URL and returns True if suitable for this IE."""
408 # This does not use has/getattr intentionally - we want to know whether
409 # we have cached the regexp for *this* class, whereas getattr would also
410 # match the superclass
411 if '_VALID_URL_RE' not in cls
.__dict
__:
412 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
413 return cls
._VALID
_URL
_RE
.match(url
) is not None
416 def _match_id(cls
, url
):
417 if '_VALID_URL_RE' not in cls
.__dict
__:
418 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
419 m
= cls
._VALID
_URL
_RE
.match(url
)
421 return compat_str(m
.group('id'))
425 """Getter method for _WORKING."""
428 def initialize(self
):
429 """Initializes an instance (authentication, etc)."""
430 self
._initialize
_geo
_bypass
({
431 'countries': self
._GEO
_COUNTRIES
,
432 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
435 self
._real
_initialize
()
438 def _initialize_geo_bypass(self
, geo_bypass_context
):
440 Initialize geo restriction bypass mechanism.
442 This method is used to initialize geo bypass mechanism based on faking
443 X-Forwarded-For HTTP header. A random country from provided country list
444 is selected and a random IP belonging to this country is generated. This
445 IP will be passed as X-Forwarded-For HTTP header in all subsequent
448 This method will be used for initial geo bypass mechanism initialization
449 during the instance initialization with _GEO_COUNTRIES and
452 You may also manually call it from extractor's code if geo bypass
453 information is not available beforehand (e.g. obtained during
454 extraction) or due to some other reason. In this case you should pass
455 this information in geo bypass context passed as first argument. It may
456 contain following fields:
458 countries: List of geo unrestricted countries (similar
460 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
461 (similar to _GEO_IP_BLOCKS)
464 if not self
._x
_forwarded
_for
_ip
:
466 # Geo bypass mechanism is explicitly disabled by user
467 if not self
._downloader
.params
.get('geo_bypass', True):
470 if not geo_bypass_context
:
471 geo_bypass_context
= {}
473 # Backward compatibility: previously _initialize_geo_bypass
474 # expected a list of countries, some 3rd party code may still use
476 if isinstance(geo_bypass_context
, (list, tuple)):
477 geo_bypass_context
= {
478 'countries': geo_bypass_context
,
481 # The whole point of geo bypass mechanism is to fake IP
482 # as X-Forwarded-For HTTP header based on some IP block or
485 # Path 1: bypassing based on IP block in CIDR notation
487 # Explicit IP block specified by user, use it right away
488 # regardless of whether extractor is geo bypassable or not
489 ip_block
= self
._downloader
.params
.get('geo_bypass_ip_block', None)
491 # Otherwise use random IP block from geo bypass context but only
492 # if extractor is known as geo bypassable
494 ip_blocks
= geo_bypass_context
.get('ip_blocks')
495 if self
._GEO
_BYPASS
and ip_blocks
:
496 ip_block
= random
.choice(ip_blocks
)
499 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
500 if self
._downloader
.params
.get('verbose', False):
501 self
._downloader
.to_screen(
502 '[debug] Using fake IP %s as X-Forwarded-For.'
503 % self
._x
_forwarded
_for
_ip
)
506 # Path 2: bypassing based on country code
508 # Explicit country code specified by user, use it right away
509 # regardless of whether extractor is geo bypassable or not
510 country
= self
._downloader
.params
.get('geo_bypass_country', None)
512 # Otherwise use random country code from geo bypass context but
513 # only if extractor is known as geo bypassable
515 countries
= geo_bypass_context
.get('countries')
516 if self
._GEO
_BYPASS
and countries
:
517 country
= random
.choice(countries
)
520 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
521 if self
._downloader
.params
.get('verbose', False):
522 self
._downloader
.to_screen(
523 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
524 % (self
._x
_forwarded
_for
_ip
, country
.upper()))
526 def extract(self
, url
):
527 """Extracts URL information and returns it in list of dicts."""
532 ie_result
= self
._real
_extract
(url
)
533 if self
._x
_forwarded
_for
_ip
:
534 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
536 except GeoRestrictedError
as e
:
537 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
540 except ExtractorError
:
542 except compat_http_client
.IncompleteRead
as e
:
543 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
544 except (KeyError, StopIteration) as e
:
545 raise ExtractorError('An extractor error has occurred.', cause
=e
)
547 def __maybe_fake_ip_and_retry(self
, countries
):
548 if (not self
._downloader
.params
.get('geo_bypass_country', None)
550 and self
._downloader
.params
.get('geo_bypass', True)
551 and not self
._x
_forwarded
_for
_ip
553 country_code
= random
.choice(countries
)
554 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
555 if self
._x
_forwarded
_for
_ip
:
557 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
558 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
562 def set_downloader(self
, downloader
):
563 """Sets the downloader for this IE."""
564 self
._downloader
= downloader
566 def _real_initialize(self
):
567 """Real initialization process. Redefine in subclasses."""
570 def _real_extract(self
, url
):
571 """Real extraction process. Redefine in subclasses."""
576 """A string for getting the InfoExtractor with get_info_extractor"""
577 return compat_str(cls
.__name
__[:-2])
581 return compat_str(type(self
).__name
__[:-2])
584 def __can_accept_status_code(err
, expected_status
):
585 assert isinstance(err
, compat_urllib_error
.HTTPError
)
586 if expected_status
is None:
588 if isinstance(expected_status
, compat_integer_types
):
589 return err
.code
== expected_status
590 elif isinstance(expected_status
, (list, tuple)):
591 return err
.code
in expected_status
592 elif callable(expected_status
):
593 return expected_status(err
.code
) is True
597 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
599 Return the response handle.
601 See _download_webpage docstring for arguments specification.
604 self
.report_download_webpage(video_id
)
605 elif note
is not False:
607 self
.to_screen('%s' % (note
,))
609 self
.to_screen('%s: %s' % (video_id
, note
))
611 # Some sites check X-Forwarded-For HTTP header in order to figure out
612 # the origin of the client behind proxy. This allows bypassing geo
613 # restriction by faking this header's value to IP that belongs to some
614 # geo unrestricted country. We will do so once we encounter any
615 # geo restriction error.
616 if self
._x
_forwarded
_for
_ip
:
617 if 'X-Forwarded-For' not in headers
:
618 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
620 if isinstance(url_or_request
, compat_urllib_request
.Request
):
621 url_or_request
= update_Request(
622 url_or_request
, data
=data
, headers
=headers
, query
=query
)
625 url_or_request
= update_url_query(url_or_request
, query
)
626 if data
is not None or headers
:
627 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
628 exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
629 if hasattr(ssl
, 'CertificateError'):
630 exceptions
.append(ssl
.CertificateError
)
632 return self
._downloader
.urlopen(url_or_request
)
633 except tuple(exceptions
) as err
:
634 if isinstance(err
, compat_urllib_error
.HTTPError
):
635 if self
.__can
_accept
_status
_code
(err
, expected_status
):
636 # Retain reference to error to prevent file object from
637 # being closed before it can be read. Works around the
638 # effects of <https://bugs.python.org/issue15002>
639 # introduced in Python 3.4.1.
646 errnote
= 'Unable to download webpage'
648 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
650 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
652 self
._downloader
.report_warning(errmsg
)
655 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):
657 Return a tuple (page content as string, URL handle).
659 See _download_webpage docstring for arguments specification.
661 # Strip hashes from the URL (#1038)
662 if isinstance(url_or_request
, (compat_str
, str)):
663 url_or_request
= url_or_request
.partition('#')[0]
665 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
669 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
670 return (content
, urlh
)
673 def _guess_encoding_from_content(content_type
, webpage_bytes
):
674 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
676 encoding
= m
.group(1)
678 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
679 webpage_bytes[:1024])
681 encoding = m.group(1).decode('ascii')
682 elif webpage_bytes.startswith(b'\xff\xfe'):
689 def __check_blocked(self, content):
690 first_block = content[:512]
691 if ('<title>Access to this site is blocked</title>' in content
692 and 'Websense' in first_block):
693 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
694 blocked_iframe = self._html_search_regex(
695 r'<iframe src="([^
"]+)"', content,
696 'Websense information URL
', default=None)
698 msg += ' Visit
%s for more details
' % blocked_iframe
699 raise ExtractorError(msg, expected=True)
700 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
702 'Access to this webpage has been blocked by Indian censorship
. '
703 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
704 block_msg = self._html_search_regex(
705 r'</h1
><p
>(.*?
)</p
>',
706 content, 'block message
', default=None)
708 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
709 raise ExtractorError(msg, expected=True)
710 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
711 and 'blocklist
.rkn
.gov
.ru
' in content):
712 raise ExtractorError(
713 'Access to this webpage has been blocked by decision of the Russian government
. '
714 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
717 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
718 content_type = urlh.headers.get('Content
-Type
', '')
719 webpage_bytes = urlh.read()
720 if prefix is not None:
721 webpage_bytes = prefix + webpage_bytes
723 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
724 if self._downloader.params.get('dump_intermediate_pages
', False):
725 self.to_screen('Dumping request to
' + urlh.geturl())
726 dump = base64.b64encode(webpage_bytes).decode('ascii
')
727 self._downloader.to_screen(dump)
728 if self._downloader.params.get('write_pages
', False):
729 basen = '%s_%s' % (video_id, urlh.geturl())
731 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
732 basen = basen[:240 - len(h)] + h
733 raw_filename = basen + '.dump
'
734 filename = sanitize_filename(raw_filename, restricted=True)
735 self.to_screen('Saving request to
' + filename)
736 # Working around MAX_PATH limitation on Windows (see
737 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
738 if compat_os_name == 'nt
':
739 absfilepath = os.path.abspath(filename)
740 if len(absfilepath) > 259:
741 filename = '\\\\?
\\' + absfilepath
742 with open(filename, 'wb
') as outf:
743 outf.write(webpage_bytes)
746 content = webpage_bytes.decode(encoding, 'replace
')
748 content = webpage_bytes.decode('utf
-8', 'replace
')
750 self.__check_blocked(content)
754 def _download_webpage(
755 self, url_or_request, video_id, note=None, errnote=None,
756 fatal=True, tries=1, timeout=5, encoding=None, data=None,
757 headers={}, query={}, expected_status=None):
759 Return the data of the page as a string.
762 url_or_request -- plain text URL as a string or
763 a compat_urllib_request.Requestobject
764 video_id -- Video/playlist/item identifier (string)
767 note -- note printed before downloading (string)
768 errnote -- note printed in case of an error (string)
769 fatal -- flag denoting whether error should be considered fatal,
770 i.e. whether it should cause ExtractionError to be raised,
771 otherwise a warning will be reported and extraction continued
772 tries -- number of tries
773 timeout -- sleep interval between tries
774 encoding -- encoding for a page content decoding, guessed automatically
775 when not explicitly specified
776 data -- POST data (bytes)
777 headers -- HTTP headers (dict)
778 query -- URL query (dict)
779 expected_status -- allows to accept failed HTTP requests (non 2xx
780 status code) by explicitly specifying a set of accepted status
781 codes. Can be any of the following entities:
782 - an integer type specifying an exact failed status code to
784 - a list or a tuple of integer types specifying a list of
785 failed status codes to accept
786 - a callable accepting an actual failed status code and
787 returning True if it should be accepted
788 Note that this argument does not affect success status codes (2xx)
789 which are always accepted.
794 while success is False:
796 res = self._download_webpage_handle(
797 url_or_request, video_id, note, errnote, fatal,
798 encoding=encoding, data=data, headers=headers, query=query,
799 expected_status=expected_status)
801 except compat_http_client.IncompleteRead as e:
803 if try_count >= tries:
805 self._sleep(timeout, video_id)
812 def _download_xml_handle(
813 self, url_or_request, video_id, note='Downloading XML
',
814 errnote='Unable to download XML
', transform_source=None,
815 fatal=True, encoding=None, data=None, headers={}, query={},
816 expected_status=None):
818 Return a tuple (xml as an compat_etree_Element, URL handle).
820 See _download_webpage docstring for arguments specification.
822 res = self._download_webpage_handle(
823 url_or_request, video_id, note, errnote, fatal=fatal,
824 encoding=encoding, data=data, headers=headers, query=query,
825 expected_status=expected_status)
828 xml_string, urlh = res
829 return self._parse_xml(
830 xml_string, video_id, transform_source=transform_source,
834 self, url_or_request, video_id,
835 note='Downloading XML
', errnote='Unable to download XML
',
836 transform_source=None, fatal=True, encoding=None,
837 data=None, headers={}, query={}, expected_status=None):
839 Return the xml as an compat_etree_Element.
841 See _download_webpage docstring for arguments specification.
843 res = self._download_xml_handle(
844 url_or_request, video_id, note=note, errnote=errnote,
845 transform_source=transform_source, fatal=fatal, encoding=encoding,
846 data=data, headers=headers, query=query,
847 expected_status=expected_status)
848 return res if res is False else res[0]
850 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
852 xml_string = transform_source(xml_string)
854 return compat_etree_fromstring(xml_string.encode('utf
-8'))
855 except compat_xml_parse_error as ve:
856 errmsg = '%s: Failed to parse XML
' % video_id
858 raise ExtractorError(errmsg, cause=ve)
860 self.report_warning(errmsg + str(ve))
862 def _download_json_handle(
863 self, url_or_request, video_id, note='Downloading JSON metadata
',
864 errnote='Unable to download JSON metadata
', transform_source=None,
865 fatal=True, encoding=None, data=None, headers={}, query={},
866 expected_status=None):
868 Return a tuple (JSON object, URL handle).
870 See _download_webpage docstring for arguments specification.
872 res = self._download_webpage_handle(
873 url_or_request, video_id, note, errnote, fatal=fatal,
874 encoding=encoding, data=data, headers=headers, query=query,
875 expected_status=expected_status)
878 json_string, urlh = res
879 return self._parse_json(
880 json_string, video_id, transform_source=transform_source,
884 self, url_or_request, video_id, note='Downloading JSON metadata
',
885 errnote='Unable to download JSON metadata
', transform_source=None,
886 fatal=True, encoding=None, data=None, headers={}, query={},
887 expected_status=None):
889 Return the JSON object as a dict.
891 See _download_webpage docstring for arguments specification.
893 res = self._download_json_handle(
894 url_or_request, video_id, note=note, errnote=errnote,
895 transform_source=transform_source, fatal=fatal, encoding=encoding,
896 data=data, headers=headers, query=query,
897 expected_status=expected_status)
898 return res if res is False else res[0]
900 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
902 json_string = transform_source(json_string)
904 return json.loads(json_string)
905 except ValueError as ve:
906 errmsg = '%s: Failed to parse JSON
' % video_id
908 raise ExtractorError(errmsg, cause=ve)
910 self.report_warning(errmsg + str(ve))
912 def report_warning(self, msg, video_id=None):
913 idstr = '' if video_id is None else '%s: ' % video_id
914 self._downloader.report_warning(
915 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
917 def to_screen(self, msg):
918 """Print msg to screen, prefixing it with '[ie_name
]'"""
919 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
921 def report_extraction(self, id_or_name):
922 """Report information extraction."""
923 self.to_screen('%s: Extracting information
' % id_or_name)
925 def report_download_webpage(self, video_id):
926 """Report webpage download."""
927 self.to_screen('%s: Downloading webpage
' % video_id)
929 def report_age_confirmation(self):
930 """Report attempt to confirm age."""
931 self.to_screen('Confirming age
')
933 def report_login(self):
934 """Report attempt to log in."""
935 self.to_screen('Logging
in')
938 def raise_login_required(msg='This video
is only available
for registered users
'):
939 raise ExtractorError(
940 '%s. Use
--username
and --password
or --netrc to provide account credentials
.' % msg,
944 def raise_geo_restricted(msg='This video
is not available
from your location due to geo restriction
', countries=None):
945 raise GeoRestrictedError(msg, countries=countries)
947 # Methods for following #608
949 def url_result(url, ie=None, video_id=None, video_title=None):
950 """Returns a URL that points to a page that should be processed"""
951 # TODO: ie should be the class used for getting the info
952 video_info = {'_type
': 'url
',
955 if video_id is not None:
956 video_info['id'] = video_id
957 if video_title is not None:
958 video_info['title
'] = video_title
961 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
963 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
965 return self.playlist_result(
966 urls, playlist_id=playlist_id, playlist_title=playlist_title)
969 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
970 """Returns a playlist"""
971 video_info = {'_type
': 'playlist
',
973 video_info.update(kwargs)
975 video_info['id'] = playlist_id
977 video_info['title
'] = playlist_title
978 if playlist_description is not None:
979 video_info['description
'] = playlist_description
982 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
984 Perform a regex search on the given string, using a single or a list of
985 patterns returning the first matching group.
986 In case of failure return a default value or raise a WARNING or a
987 RegexNotFoundError, depending on fatal, specifying the field name.
989 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
990 mobj = re.search(pattern, string, flags)
993 mobj = re.search(p, string, flags)
997 if not self._downloader.params.get('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
998 _name = '\033[0;34m
%s\033[0m
' % name
1004 # return the first matching group
1005 return next(g for g in mobj.groups() if g is not None)
1007 return mobj.group(group)
1008 elif default is not NO_DEFAULT:
1011 raise RegexNotFoundError('Unable to extract
%s' % _name)
1013 self._downloader.report_warning('unable to extract
%s' % _name + bug_reports_message())
1016 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1018 Like _search_regex, but strips HTML tags and unescapes entities.
1020 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1022 return clean_html(res).strip()
1026 def _get_netrc_login_info(self, netrc_machine=None):
1029 netrc_machine = netrc_machine or self._NETRC_MACHINE
1031 if self._downloader.params.get('usenetrc
', False):
1033 info = netrc.netrc().authenticators(netrc_machine)
1034 if info is not None:
1038 raise netrc.NetrcParseError(
1039 'No authenticators
for %s' % netrc_machine)
1040 except (IOError, netrc.NetrcParseError) as err:
1041 self._downloader.report_warning(
1042 'parsing
.netrc
: %s' % error_to_compat_str(err))
1044 return username, password
1046 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1048 Get the login info as (username, password)
1049 First look for the manually specified credentials using username_option
1050 and password_option as keys in params dictionary. If no such credentials
1051 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1053 If there's no info available
, return (None, None)
1055 if self._downloader is None:
1058 downloader_params = self._downloader.params
1060 # Attempt to use provided username and password or .netrc data
1061 if downloader_params.get(username_option) is not None:
1062 username = downloader_params[username_option]
1063 password = downloader_params[password_option]
1065 username, password = self._get_netrc_login_info(netrc_machine)
1067 return username, password
1069 def _get_tfa_info(self, note='two-factor verification code'):
1071 Get the two
-factor authentication info
1072 TODO
- asking the user will be required
for sms
/phone verify
1073 currently just uses the command line option
1074 If there
's no info available, return None
1076 if self._downloader is None:
1078 downloader_params = self._downloader.params
1080 if downloader_params.get('twofactor
') is not None:
1081 return downloader_params['twofactor
']
1083 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1085 # Helper functions for extracting OpenGraph info
1087 def _og_regexes(prop):
1088 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1089 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1090 % {'prop': re.escape(prop)})
1091 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1093 template % (property_re, content_re),
1094 template % (content_re, property_re),
1098 def _meta_regex(prop):
1099 return r'''(?isx)<meta
1100 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1101 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1103 def _og_search_property(self, prop, html, name=None, **kargs):
1104 if not isinstance(prop, (list, tuple)):
1107 name = 'OpenGraph
%s' % prop[0]
1110 og_regexes.extend(self._og_regexes(p))
1111 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1114 return unescapeHTML(escaped)
1116 def _og_search_thumbnail(self, html, **kargs):
1117 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1119 def _og_search_description(self, html, **kargs):
1120 return self._og_search_property('description
', html, fatal=False, **kargs)
1122 def _og_search_title(self, html, **kargs):
1123 return self._og_search_property('title
', html, **kargs)
1125 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1126 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1128 regexes = self._og_regexes('video
:secure_url
') + regexes
1129 return self._html_search_regex(regexes, html, name, **kargs)
1131 def _og_search_url(self, html, **kargs):
1132 return self._og_search_property('url
', html, **kargs)
1134 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1135 if not isinstance(name, (list, tuple)):
1137 if display_name is None:
1138 display_name = name[0]
1139 return self._html_search_regex(
1140 [self._meta_regex(n) for n in name],
1141 html, display_name, fatal=fatal, group='content
', **kwargs)
1143 def _dc_search_uploader(self, html):
1144 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1146 def _rta_search(self, html):
1147 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1148 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1149 r' content
="RTA-5042-1996-1400-1577-RTA"',
1154 def _media_rating_search(self, html):
1155 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1156 rating = self._html_search_meta('rating
', html)
1168 return RATING_TABLE.get(rating.lower())
1170 def _family_friendly_search(self, html):
1171 # See http://schema.org/VideoObject
1172 family_friendly = self._html_search_meta(
1173 'isFamilyFriendly
', html, default=None)
1175 if not family_friendly:
1184 return RATING_TABLE.get(family_friendly.lower())
1186 def _twitter_search_player(self, html):
1187 return self._html_search_meta('twitter
:player
', html,
1188 'twitter card player
')
1190 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1191 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1192 default = kwargs.get('default
', NO_DEFAULT)
1193 # JSON-LD may be malformed and thus `fatal` should be respected.
1194 # At the same time `default` may be passed that assumes `fatal=False`
1195 # for _search_regex. Let's simulate the same behavior here
as well
.
1196 fatal
= kwargs
.get('fatal', True) if default
== NO_DEFAULT
else False
1198 for mobj
in json_ld_list
:
1199 json_ld_item
= self
._parse
_json
(
1200 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1201 if not json_ld_item
:
1203 if isinstance(json_ld_item
, dict):
1204 json_ld
.append(json_ld_item
)
1205 elif isinstance(json_ld_item
, (list, tuple)):
1206 json_ld
.extend(json_ld_item
)
1208 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1211 if default
is not NO_DEFAULT
:
1214 raise RegexNotFoundError('Unable to extract JSON-LD')
1216 self
._downloader
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1219 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1220 if isinstance(json_ld
, compat_str
):
1221 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1225 if not isinstance(json_ld
, (list, tuple, dict)):
1227 if isinstance(json_ld
, dict):
1230 INTERACTION_TYPE_MAP
= {
1231 'CommentAction': 'comment',
1232 'AgreeAction': 'like',
1233 'DisagreeAction': 'dislike',
1234 'LikeAction': 'like',
1235 'DislikeAction': 'dislike',
1236 'ListenAction': 'view',
1237 'WatchAction': 'view',
1238 'ViewAction': 'view',
1241 def extract_interaction_type(e
):
1242 interaction_type
= e
.get('interactionType')
1243 if isinstance(interaction_type
, dict):
1244 interaction_type
= interaction_type
.get('@type')
1245 return str_or_none(interaction_type
)
1247 def extract_interaction_statistic(e
):
1248 interaction_statistic
= e
.get('interactionStatistic')
1249 if isinstance(interaction_statistic
, dict):
1250 interaction_statistic
= [interaction_statistic
]
1251 if not isinstance(interaction_statistic
, list):
1253 for is_e
in interaction_statistic
:
1254 if not isinstance(is_e
, dict):
1256 if is_e
.get('@type') != 'InteractionCounter':
1258 interaction_type
= extract_interaction_type(is_e
)
1259 if not interaction_type
:
1261 # For interaction count some sites provide string instead of
1262 # an integer (as per spec) with non digit characters (e.g. ",")
1263 # so extracting count with more relaxed str_to_int
1264 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1265 if interaction_count
is None:
1267 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1270 count_key
= '%s_count' % count_kind
1271 if info
.get(count_key
) is not None:
1273 info
[count_key
] = interaction_count
1275 def extract_video_object(e
):
1276 assert e
['@type'] == 'VideoObject'
1278 'url': url_or_none(e
.get('contentUrl')),
1279 'title': unescapeHTML(e
.get('name')),
1280 'description': unescapeHTML(e
.get('description')),
1281 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1282 'duration': parse_duration(e
.get('duration')),
1283 'timestamp': unified_timestamp(e
.get('uploadDate')),
1284 'uploader': str_or_none(e
.get('author')),
1285 'filesize': float_or_none(e
.get('contentSize')),
1286 'tbr': int_or_none(e
.get('bitrate')),
1287 'width': int_or_none(e
.get('width')),
1288 'height': int_or_none(e
.get('height')),
1289 'view_count': int_or_none(e
.get('interactionCount')),
1291 extract_interaction_statistic(e
)
1295 item_type
= e
.get('@type')
1296 if expected_type
is not None and expected_type
!= item_type
:
1298 if item_type
in ('TVEpisode', 'Episode'):
1299 episode_name
= unescapeHTML(e
.get('name'))
1301 'episode': episode_name
,
1302 'episode_number': int_or_none(e
.get('episodeNumber')),
1303 'description': unescapeHTML(e
.get('description')),
1305 if not info
.get('title') and episode_name
:
1306 info
['title'] = episode_name
1307 part_of_season
= e
.get('partOfSeason')
1308 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1310 'season': unescapeHTML(part_of_season
.get('name')),
1311 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1313 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1314 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1315 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1316 elif item_type
== 'Movie':
1318 'title': unescapeHTML(e
.get('name')),
1319 'description': unescapeHTML(e
.get('description')),
1320 'duration': parse_duration(e
.get('duration')),
1321 'timestamp': unified_timestamp(e
.get('dateCreated')),
1323 elif item_type
in ('Article', 'NewsArticle'):
1325 'timestamp': parse_iso8601(e
.get('datePublished')),
1326 'title': unescapeHTML(e
.get('headline')),
1327 'description': unescapeHTML(e
.get('articleBody')),
1329 elif item_type
== 'VideoObject':
1330 extract_video_object(e
)
1331 if expected_type
is None:
1335 video
= e
.get('video')
1336 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1337 extract_video_object(video
)
1338 if expected_type
is None:
1342 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1345 def _hidden_inputs(html
):
1346 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1348 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1349 attrs
= extract_attributes(input)
1352 if attrs
.get('type') not in ('hidden', 'submit'):
1354 name
= attrs
.get('name') or attrs
.get('id')
1355 value
= attrs
.get('value')
1356 if name
and value
is not None:
1357 hidden_inputs
[name
] = value
1358 return hidden_inputs
1360 def _form_hidden_inputs(self
, form_id
, html
):
1361 form
= self
._search
_regex
(
1362 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1363 html, '%s form
' % form_id, group='form
')
1364 return self._hidden_inputs(form)
1367 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<seperator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1369 default = ('hidden
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1370 'res
', 'fps
', 'codec
:vp9
', 'size
', 'br
', 'asr
',
1371 'proto
', 'ext
', 'has_audio
', 'source
', 'format_id
') # These must not be aliases
1374 'vcodec
': {'type': 'ordered
', 'regex
': True,
1375 'order
': ['av0?
1', 'vp9
', '(h265|he?vc?
)', '(h264|avc
)', 'vp8
', '(mp4v|h263
)', 'theora
', '', None, 'none
']},
1376 'acodec
': {'type': 'ordered
', 'regex
': True,
1377 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1378 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1379 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', 'm3u8
', '.*dash
', '', 'mms|rtsp
', 'none
', 'f4
']},
1380 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1381 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1382 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1383 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1384 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1385 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1386 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1387 'ie_pref
': {'priority': True, 'type': 'extractor', 'field': 'extractor_preference'},
1388 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1389 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1390 'lang
': {'priority': True, 'convert': 'ignore', 'field': 'language_preference'},
1391 'quality
': {'priority': True, 'convert': 'float_none'},
1392 'filesize
': {'convert': 'bytes'},
1393 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1394 'id': {'convert': 'string', 'field': 'format_id'},
1395 'height
': {'convert': 'float_none'},
1396 'width
': {'convert': 'float_none'},
1397 'fps
': {'convert': 'float_none'},
1398 'tbr
': {'convert': 'float_none'},
1399 'vbr
': {'convert': 'float_none'},
1400 'abr
': {'convert': 'float_none'},
1401 'asr
': {'convert': 'float_none'},
1402 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1404 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1405 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1406 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1407 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1408 'res
': {'type': 'multiple', 'field': ('height', 'width'), 'function': min},
1410 # Most of these exist only for compatibility reasons
1411 'dimension
': {'type': 'alias', 'field': 'res'},
1412 'resolution
': {'type': 'alias', 'field': 'res'},
1413 'extension
': {'type': 'alias', 'field': 'ext'},
1414 'bitrate
': {'type': 'alias', 'field': 'br'},
1415 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1416 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1417 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1418 'framerate
': {'type': 'alias', 'field': 'fps'},
1419 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1420 'protocol
': {'type': 'alias', 'field': 'proto'},
1421 'source_preference
': {'type': 'alias', 'field': 'source'},
1422 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1423 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1424 'samplerate
': {'type': 'alias', 'field': 'asr'},
1425 'video_ext
': {'type': 'alias', 'field': 'vext'},
1426 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1427 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1428 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1429 'video
': {'type': 'alias', 'field': 'hasvid'},
1430 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1431 'audio
': {'type': 'alias', 'field': 'hasaud'},
1432 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1433 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1434 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1435 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1436 'format_id
': {'type': 'alias', 'field': 'id'},
1441 def _get_field_setting(self, field, key):
1442 if field not in self.settings:
1443 self.settings[field] = {}
1444 propObj = self.settings[field]
1445 if key not in propObj:
1446 type = propObj.get('type')
1448 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1449 elif key == 'convert
':
1450 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1452 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,), 'function': max}.get(key, None)
1453 propObj[key] = default
1456 def _resolve_field_value(self, field, value, convertNone=False):
1461 value = value.lower()
1462 conversion = self._get_field_setting(field, 'convert
')
1463 if conversion == 'ignore
':
1465 if conversion == 'string
':
1467 elif conversion == 'float_none
':
1468 return float_or_none(value)
1469 elif conversion == 'bytes':
1470 return FileDownloader.parse_bytes(value)
1471 elif conversion == 'order
':
1472 order_free = self._get_field_setting(field, 'order_free
')
1473 order_list = order_free if order_free and self._use_free_order else self._get_field_setting(field, 'order
')
1474 use_regex = self._get_field_setting(field, 'regex
')
1475 list_length = len(order_list)
1476 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1477 if use_regex and value is not None:
1478 for (i, regex) in enumerate(order_list):
1479 if regex and re.match(regex, value):
1480 return list_length - i
1481 return list_length - empty_pos # not in list
1482 else: # not regex or value = None
1483 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1485 if value.isnumeric():
1488 self.settings[field]['convert
'] = 'string
'
1491 def evaluate_params(self, params, sort_extractor):
1492 self._use_free_order = params.get('prefer_free_formats
', False)
1493 self._sort_user = params.get('format_sort
', [])
1494 self._sort_extractor = sort_extractor
1496 def add_item(field, reverse, closest, limit_text):
1497 field = field.lower()
1498 if field in self._order:
1500 self._order.append(field)
1501 limit = self._resolve_field_value(field, limit_text)
1504 'closest
': False if limit is None else closest,
1505 'limit_text
': limit_text,
1507 if field in self.settings:
1508 self.settings[field].update(data)
1510 self.settings[field] = data
1513 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1514 + (tuple() if params.get('format_sort_force
', False)
1515 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1516 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1518 for item in sort_list:
1519 match = re.match(self.regex, item)
1521 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1522 field = match.group('field
')
1525 if self._get_field_setting(field, 'type') == 'alias
':
1526 field = self._get_field_setting(field, 'field
')
1527 reverse = match.group('reverse
') is not None
1528 closest = match.group('seperator
') == '~
'
1529 limit_text = match.group('limit
')
1531 has_limit = limit_text is not None
1532 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1533 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1535 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1536 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1537 limit_count = len(limits)
1538 for (i, f) in enumerate(fields):
1539 add_item(f, reverse, closest,
1540 limits[i] if i < limit_count
1541 else limits[0] if has_limit and not has_multiple_limits
1544 def print_verbose_info(self, to_screen):
1545 to_screen('[debug
] Sort order given by user
: %s' % ','.join(self._sort_user))
1546 if self._sort_extractor:
1547 to_screen('[debug
] Sort order given by extractor
: %s' % ','.join(self._sort_extractor))
1548 to_screen('[debug
] Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1549 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1550 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1551 self._get_field_setting(field, 'limit_text
'),
1552 self._get_field_setting(field, 'limit
'))
1553 if self._get_field_setting(field, 'limit_text
') is not None else '')
1554 for field in self._order if self._get_field_setting(field, 'visible
')]))
1556 def _calculate_field_preference_from_value(self, format, field, type, value):
1557 reverse = self._get_field_setting(field, 'reverse
')
1558 closest = self._get_field_setting(field, 'closest
')
1559 limit = self._get_field_setting(field, 'limit
')
1561 if type == 'extractor
':
1562 maximum = self._get_field_setting(field, 'max')
1563 if value is None or (maximum is not None and value >= maximum):
1565 elif type == 'boolean
':
1566 in_list = self._get_field_setting(field, 'in_list
')
1567 not_in_list = self._get_field_setting(field, 'not_in_list
')
1568 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
1569 elif type == 'ordered
':
1570 value = self._resolve_field_value(field, value, True)
1572 # try to convert to number
1573 val_num = float_or_none(value)
1574 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1578 return ((-10, 0) if value is None
1579 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1580 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1581 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1582 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1583 else (-1, value, 0))
1585 def _calculate_field_preference(self, format, field):
1586 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1587 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1588 if type == 'multiple
':
1589 type = 'field
' # Only 'field
' is allowed in multiple for now
1590 actual_fields = self._get_field_setting(field, 'field
')
1592 def wrapped_function(values):
1593 values = tuple(filter(lambda x: x is not None, values))
1594 return (self._get_field_setting(field, 'function
')(*values) if len(values) > 1
1595 else values[0] if values
1598 value = wrapped_function((get_value(f) for f in actual_fields))
1600 value = get_value(field)
1601 return self._calculate_field_preference_from_value(format, field, type, value)
1603 def calculate_preference(self, format):
1604 # Determine missing protocol
1605 if not format.get('protocol
'):
1606 format['protocol
'] = determine_protocol(format)
1608 # Determine missing ext
1609 if not format.get('ext
') and 'url
' in format:
1610 format['ext
'] = determine_ext(format['url
'])
1611 if format.get('vcodec
') == 'none
':
1612 format['audio_ext
'] = format['ext
']
1613 format['video_ext
'] = 'none
'
1615 format['video_ext
'] = format['ext
']
1616 format['audio_ext
'] = 'none
'
1617 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1618 # format['preference
'] = -1000
1620 # Determine missing bitrates
1621 if format.get('tbr
') is None:
1622 if format.get('vbr
') is not None and format.get('abr
') is not None:
1623 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1625 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1626 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1627 if format.get('acodec
') != "none" and format.get('abr
') is None:
1628 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1630 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1632 def _sort_formats(self, formats, field_preference=[]):
1634 raise ExtractorError('No video formats found
')
1635 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1636 format_sort.evaluate_params(self._downloader.params, field_preference)
1637 if self._downloader.params.get('verbose
', False):
1638 format_sort.print_verbose_info(self._downloader.to_screen)
1639 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1641 def _check_formats(self, formats, video_id):
1643 formats[:] = filter(
1644 lambda f: self._is_valid_url(
1646 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1650 def _remove_duplicate_formats(formats):
1654 if f['url
'] not in format_urls:
1655 format_urls.add(f['url
'])
1656 unique_formats.append(f)
1657 formats[:] = unique_formats
1659 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1660 url = self._proto_relative_url(url, scheme='http
:')
1661 # For now assume non HTTP(S) URLs always valid
1662 if not (url.startswith('http
://') or url.startswith('https
://')):
1665 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1667 except ExtractorError as e:
1669 '%s: %s URL
is invalid
, skipping
: %s'
1670 % (video_id, item, error_to_compat_str(e.cause)))
1673 def http_scheme(self):
1674 """ Either "http:" or "https:", depending on the user's preferences
"""
1677 if self._downloader.params.get('prefer_insecure', False)
1680 def _proto_relative_url(self, url, scheme=None):
1683 if url.startswith('//'):
1685 scheme = self.http_scheme()
1690 def _sleep(self, timeout, video_id, msg_template=None):
1691 if msg_template is None:
1692 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1693 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1697 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1698 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1699 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1700 manifest = self._download_xml(
1701 manifest_url, video_id, 'Downloading f4m manifest',
1702 'Unable to download f4m manifest',
1703 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1704 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1705 transform_source=transform_source,
1706 fatal=fatal, data=data, headers=headers, query=query)
1708 if manifest is False:
1711 return self._parse_f4m_formats(
1712 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1713 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1715 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1716 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1717 fatal=True, m3u8_id=None):
1718 if not isinstance(manifest, compat_etree_Element) and not fatal:
1721 # currently youtube-dlc cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1722 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1723 if akamai_pv is not None and ';' in akamai_pv.text:
1724 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1725 if playerVerificationChallenge.strip() != '':
1729 manifest_version = '1.0'
1730 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1732 manifest_version = '2.0'
1733 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1734 # Remove unsupported DRM protected media from final formats
1735 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1736 media_nodes = remove_encrypted_media(media_nodes)
1740 manifest_base_url = get_base_url(manifest)
1742 bootstrap_info = xpath_element(
1743 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1744 'bootstrap info', default=None)
1747 mime_type = xpath_text(
1748 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1749 'base URL', default=None)
1750 if mime_type and mime_type.startswith('audio/'):
1753 for i, media_el in enumerate(media_nodes):
1754 tbr = int_or_none(media_el.attrib.get('bitrate'))
1755 width = int_or_none(media_el.attrib.get('width'))
1756 height = int_or_none(media_el.attrib.get('height'))
1757 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1758 # If <bootstrapInfo> is present, the specified f4m is a
1759 # stream-level manifest, and only set-level manifests may refer to
1760 # external resources. See section 11.4 and section 4 of F4M spec
1761 if bootstrap_info is None:
1763 # @href is introduced in 2.0, see section 11.6 of F4M spec
1764 if manifest_version == '2.0':
1765 media_url = media_el.attrib.get('href')
1766 if media_url is None:
1767 media_url = media_el.attrib.get('url')
1771 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1772 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1773 # If media_url is itself a f4m manifest do the recursive extraction
1774 # since bitrates in parent manifest (this one) and media_url manifest
1775 # may differ leading to inability to resolve the format by requested
1776 # bitrate in f4m downloader
1777 ext = determine_ext(manifest_url)
1779 f4m_formats = self._extract_f4m_formats(
1780 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1781 transform_source=transform_source, fatal=fatal)
1782 # Sometimes stream-level manifest contains single media entry that
1783 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1784 # At the same time parent's media entry in set-level manifest may
1785 # contain it. We will copy it from parent in such cases.
1786 if len(f4m_formats) == 1:
1789 'tbr': f.get('tbr') or tbr,
1790 'width': f.get('width') or width,
1791 'height': f.get('height') or height,
1792 'format_id': f.get('format_id') if not tbr else format_id,
1795 formats.extend(f4m_formats)
1798 formats.extend(self._extract_m3u8_formats(
1799 manifest_url, video_id, 'mp4', preference=preference,
1800 m3u8_id=m3u8_id, fatal=fatal))
1803 'format_id': format_id,
1804 'url': manifest_url,
1805 'manifest_url': manifest_url,
1806 'ext': 'flv' if bootstrap_info is not None else None,
1812 'preference': preference,
1816 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1818 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1822 'preference': preference - 100 if preference else -100,
1823 'resolution': 'multiple',
1824 'format_note': 'Quality selection URL',
1827 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1828 entry_protocol='m3u8', preference=None,
1829 m3u8_id=None, note=None, errnote=None,
1830 fatal=True, live=False, data=None, headers={},
1832 res = self._download_webpage_handle(
1834 note=note or 'Downloading m3u8 information',
1835 errnote=errnote or 'Failed to download m3u8 information',
1836 fatal=fatal, data=data, headers=headers, query=query)
1841 m3u8_doc, urlh = res
1842 m3u8_url = urlh.geturl()
1844 return self._parse_m3u8_formats(
1845 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1846 preference=preference, m3u8_id=m3u8_id, live=live)
1848 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1849 entry_protocol='m3u8', preference=None,
1850 m3u8_id=None, live=False):
1851 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1854 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1859 format_url = lambda u: (
1861 if re.match(r'^https?://', u)
1862 else compat_urlparse.urljoin(m3u8_url, u))
1865 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1866 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1867 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1869 # We should try extracting formats only from master playlists [1, 4.3.4],
1870 # i.e. playlists that describe available qualities. On the other hand
1871 # media playlists [1, 4.3.3] should be returned as is since they contain
1872 # just the media without qualities renditions.
1873 # Fortunately, master playlist can be easily distinguished from media
1874 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1875 # master playlist tags MUST NOT appear in a media playlist and vice versa.
1876 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1877 # media playlist and MUST NOT appear in master playlist thus we can
1878 # clearly detect media playlist with this criterion.
1880 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1883 'format_id': m3u8_id,
1885 'protocol': entry_protocol,
1886 'preference': preference,
1890 last_stream_inf = {}
1892 def extract_media(x_media_line):
1893 media = parse_m3u8_attributes(x_media_line)
1894 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1895 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1896 if not (media_type and group_id and name):
1898 groups.setdefault(group_id, []).append(media)
1899 if media_type not in ('VIDEO', 'AUDIO'):
1901 media_url = media.get('URI')
1904 for v in (m3u8_id, group_id, name):
1908 'format_id': '-'.join(format_id),
1909 'url': format_url(media_url),
1910 'manifest_url': m3u8_url,
1911 'language': media.get('LANGUAGE'),
1913 'protocol': entry_protocol,
1914 'preference': preference,
1916 if media_type == 'AUDIO':
1917 f['vcodec'] = 'none'
1920 def build_stream_name():
1921 # Despite specification does not mention NAME attribute for
1922 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1923 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1924 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1925 stream_name = last_stream_inf.get('NAME')
1928 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1929 # from corresponding rendition group
1930 stream_group_id = last_stream_inf.get('VIDEO')
1931 if not stream_group_id:
1933 stream_group = groups.get(stream_group_id)
1934 if not stream_group:
1935 return stream_group_id
1936 rendition = stream_group[0]
1937 return rendition.get('NAME') or stream_group_id
1939 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1940 # chance to detect video only formats when EXT-X-STREAM-INF tags
1941 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1942 for line in m3u8_doc.splitlines():
1943 if line.startswith('#EXT-X-MEDIA:'):
1946 for line in m3u8_doc.splitlines():
1947 if line.startswith('#EXT-X-STREAM-INF:'):
1948 last_stream_inf = parse_m3u8_attributes(line)
1949 elif line.startswith('#') or not line.strip():
1952 tbr = float_or_none(
1953 last_stream_inf.get('AVERAGE-BANDWIDTH')
1954 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1957 format_id.append(m3u8_id)
1958 stream_name = build_stream_name()
1959 # Bandwidth of live streams may differ over time thus making
1960 # format_id unpredictable. So it's better to keep provided
1963 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1964 manifest_url = format_url(line.strip())
1966 'format_id': '-'.join(format_id),
1967 'url': manifest_url,
1968 'manifest_url': m3u8_url,
1971 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1972 'protocol': entry_protocol,
1973 'preference': preference,
1975 resolution = last_stream_inf.get('RESOLUTION')
1977 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1979 f['width'] = int(mobj.group('width'))
1980 f['height'] = int(mobj.group('height'))
1981 # Unified Streaming Platform
1983 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1985 abr, vbr = mobj.groups()
1986 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1991 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1993 audio_group_id = last_stream_inf.get('AUDIO')
1994 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1995 # references a rendition group MUST have a CODECS attribute.
1996 # However, this is not always respected, for example, [2]
1997 # contains EXT-X-STREAM-INF tag which references AUDIO
1998 # rendition group but does not have CODECS and despite
1999 # referencing an audio group it represents a complete
2000 # (with audio and video) format. So, for such cases we will
2001 # ignore references to rendition groups and treat them
2002 # as complete formats.
2003 if audio_group_id and codecs and f.get('vcodec') != 'none':
2004 audio_group = groups.get(audio_group_id)
2005 if audio_group and audio_group[0].get('URI'):
2006 # TODO: update acodec for audio only formats with
2008 f['acodec'] = 'none'
2012 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2015 del http_f['manifest_url']
2017 'format_id': f['format_id'].replace('hls-', 'http-'),
2019 'url': progressive_uri,
2021 formats.append(http_f)
2023 last_stream_inf = {}
2027 def _xpath_ns(path, namespace=None):
2031 for c in path.split('/'):
2032 if not c or c == '.':
2035 out.append('{%s}%s' % (namespace, c))
2036 return '/'.join(out)
2038 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2039 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2045 namespace = self._parse_smil_namespace(smil)
2047 return self._parse_smil_formats(
2048 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2050 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2051 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2054 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2056 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2057 return self._download_xml(
2058 smil_url, video_id, 'Downloading SMIL file',
2059 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2061 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2062 namespace = self._parse_smil_namespace(smil)
2064 formats = self._parse_smil_formats(
2065 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2066 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2068 video_id = os.path.splitext(url_basename(smil_url))[0]
2072 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2073 name = meta.attrib.get('name')
2074 content = meta.attrib.get('content')
2075 if not name or not content:
2077 if not title and name == 'title':
2079 elif not description and name in ('description', 'abstract'):
2080 description = content
2081 elif not upload_date and name == 'date':
2082 upload_date = unified_strdate(content)
2085 'id': image.get('type'),
2086 'url': image.get('src'),
2087 'width': int_or_none(image.get('width')),
2088 'height': int_or_none(image.get('height')),
2089 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2093 'title': title or video_id,
2094 'description': description,
2095 'upload_date': upload_date,
2096 'thumbnails': thumbnails,
2098 'subtitles': subtitles,
2101 def _parse_smil_namespace(self, smil):
2102 return self._search_regex(
2103 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2105 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2107 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2108 b = meta.get('base') or meta.get('httpBase')
2119 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2120 for medium in media:
2121 src = medium.get('src')
2122 if not src or src in srcs:
2126 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2127 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2128 width = int_or_none(medium.get('width'))
2129 height = int_or_none(medium.get('height'))
2130 proto = medium.get('proto')
2131 ext = medium.get('ext')
2132 src_ext = determine_ext(src)
2133 streamer = medium.get('streamer') or base
2135 if proto == 'rtmp' or streamer.startswith('rtmp'):
2141 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2143 'filesize': filesize,
2147 if transform_rtmp_url:
2148 streamer, src = transform_rtmp_url(streamer, src)
2149 formats[-1].update({
2155 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2156 src_url = src_url.strip()
2158 if proto == 'm3u8' or src_ext == 'm3u8':
2159 m3u8_formats = self._extract_m3u8_formats(
2160 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2161 if len(m3u8_formats) == 1:
2163 m3u8_formats[0].update({
2164 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2169 formats.extend(m3u8_formats)
2170 elif src_ext == 'f4m':
2175 'plugin': 'flowplayer-3.2.0.1',
2177 f4m_url += '&' if '?' in f4m_url else '?'
2178 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2179 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2180 elif src_ext == 'mpd':
2181 formats.extend(self._extract_mpd_formats(
2182 src_url, video_id, mpd_id='dash', fatal=False))
2183 elif re.search(r'\.ism/[Mm]anifest', src_url):
2184 formats.extend(self._extract_ism_formats(
2185 src_url, video_id, ism_id='mss', fatal=False))
2186 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2190 'ext': ext or src_ext or 'flv',
2191 'format_id': 'http-%d' % (bitrate or http_count),
2193 'filesize': filesize,
2200 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2203 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2204 src = textstream.get('src')
2205 if not src or src in urls:
2208 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2209 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2210 subtitles.setdefault(lang, []).append({
2216 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2217 xspf = self._download_xml(
2218 xspf_url, playlist_id, 'Downloading xpsf playlist',
2219 'Unable to download xspf manifest', fatal=fatal)
2222 return self._parse_xspf(
2223 xspf, playlist_id, xspf_url=xspf_url,
2224 xspf_base_url=base_url(xspf_url))
2226 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2228 'xspf': 'http://xspf.org/ns/0/',
2229 's1': 'http://static.streamone.nl/player/ns/0',
2233 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2235 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2236 description = xpath_text(
2237 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2238 thumbnail = xpath_text(
2239 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2240 duration = float_or_none(
2241 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2244 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2245 format_url = urljoin(xspf_base_url, location.text)
2250 'manifest_url': xspf_url,
2251 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2252 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2253 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2255 self._sort_formats(formats)
2260 'description': description,
2261 'thumbnail': thumbnail,
2262 'duration': duration,
2267 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2268 res = self._download_xml_handle(
2270 note=note or 'Downloading MPD manifest',
2271 errnote=errnote or 'Failed to download MPD manifest',
2272 fatal=fatal, data=data, headers=headers, query=query)
2278 mpd_base_url = base_url(urlh.geturl())
2280 return self._parse_mpd_formats(
2281 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2283 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2285 Parse formats
from MPD manifest
.
2287 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2288 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2289 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2291 if not self._downloader.params.get('dynamic_mpd'):
2292 if mpd_doc.get('type') == 'dynamic':
2295 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2298 return self._xpath_ns(path, namespace)
2300 def is_drm_protected(element):
2301 return element.find(_add_ns('ContentProtection')) is not None
2303 def extract_multisegment_info(element, ms_parent_info):
2304 ms_info = ms_parent_info.copy()
2306 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2307 # common attributes and elements. We will only extract relevant
2309 def extract_common(source):
2310 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2311 if segment_timeline is not None:
2312 s_e = segment_timeline.findall(_add_ns('S'))
2314 ms_info['total_number'] = 0
2317 r = int(s.get('r', 0))
2318 ms_info['total_number'] += 1 + r
2319 ms_info['s'].append({
2320 't': int(s.get('t', 0)),
2321 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2322 'd': int(s.attrib['d']),
2325 start_number = source.get('startNumber')
2327 ms_info['start_number'] = int(start_number)
2328 timescale = source.get('timescale')
2330 ms_info['timescale'] = int(timescale)
2331 segment_duration = source.get('duration')
2332 if segment_duration:
2333 ms_info['segment_duration'] = float(segment_duration)
2335 def extract_Initialization(source):
2336 initialization = source.find(_add_ns('Initialization'))
2337 if initialization is not None:
2338 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2340 segment_list = element.find(_add_ns('SegmentList'))
2341 if segment_list is not None:
2342 extract_common(segment_list)
2343 extract_Initialization(segment_list)
2344 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2346 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2348 segment_template = element.find(_add_ns('SegmentTemplate'))
2349 if segment_template is not None:
2350 extract_common(segment_template)
2351 media = segment_template.get('media')
2353 ms_info['media'] = media
2354 initialization = segment_template.get('initialization')
2356 ms_info['initialization'] = initialization
2358 extract_Initialization(segment_template)
2361 skip_unplayable = not self._downloader.params.get('allow_unplayable_formats')
2363 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2365 for period in mpd_doc.findall(_add_ns('Period')):
2366 period_duration = parse_duration(period.get('duration')) or mpd_duration
2367 period_ms_info = extract_multisegment_info(period, {
2371 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2372 if skip_unplayable and is_drm_protected(adaptation_set):
2374 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2375 for representation in adaptation_set.findall(_add_ns('Representation')):
2376 if skip_unplayable and is_drm_protected(representation):
2378 representation_attrib = adaptation_set.attrib.copy()
2379 representation_attrib.update(representation.attrib)
2380 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2381 mime_type = representation_attrib['mimeType']
2382 content_type = mime_type.split('/')[0]
2383 if content_type == 'text':
2384 # TODO implement WebVTT downloading
2386 elif content_type in ('video', 'audio'):
2388 for element in (representation, adaptation_set, period, mpd_doc):
2389 base_url_e = element.find(_add_ns('BaseURL'))
2390 if base_url_e is not None:
2391 base_url = base_url_e.text + base_url
2392 if re.match(r'^https?://', base_url):
2394 if mpd_base_url and not re.match(r'^https?://', base_url):
2395 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2397 base_url = mpd_base_url + base_url
2398 representation_id = representation_attrib.get('id')
2399 lang = representation_attrib.get('lang')
2400 url_el = representation.find(_add_ns('BaseURL'))
2401 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2402 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2404 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2405 'manifest_url': mpd_url,
2406 'ext': mimetype2ext(mime_type),
2407 'width': int_or_none(representation_attrib.get('width')),
2408 'height': int_or_none(representation_attrib.get('height')),
2409 'tbr': float_or_none(bandwidth, 1000),
2410 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2411 'fps': int_or_none(representation_attrib.get('frameRate')),
2412 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2413 'format_note': 'DASH %s' % content_type,
2414 'filesize': filesize,
2415 'container': mimetype2ext(mime_type) + '_dash',
2417 f.update(parse_codecs(representation_attrib.get('codecs')))
2418 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2420 def prepare_template(template_name, identifiers):
2421 tmpl = representation_ms_info[template_name]
2422 # First of, % characters outside $...$ templates
2423 # must be escaped by doubling for proper processing
2424 # by % operator string formatting used further (see
2425 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2431 in_template = not in_template
2432 elif c == '%' and not in_template:
2434 # Next, $...$ templates are translated to their
2435 # %(...) counterparts to be used with % operator
2436 t = t.replace('$RepresentationID$', representation_id)
2437 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2438 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2439 t.replace('$$', '$')
2442 # @initialization is a regular template like @media one
2443 # so it should be handled just the same way (see
2444 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2445 if 'initialization' in representation_ms_info:
2446 initialization_template = prepare_template(
2448 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2449 # $Time$ shall not be included for @initialization thus
2450 # only $Bandwidth$ remains
2452 representation_ms_info['initialization_url'] = initialization_template % {
2453 'Bandwidth': bandwidth,
2456 def location_key(location):
2457 return 'url' if re.match(r'^https?://', location) else 'path'
2459 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2461 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2462 media_location_key = location_key(media_template)
2464 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2465 # can't be used at the same time
2466 if '%(Number' in media_template and 's' not in representation_ms_info:
2467 segment_duration = None
2468 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2469 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2470 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2471 representation_ms_info['fragments'] = [{
2472 media_location_key: media_template % {
2473 'Number': segment_number,
2474 'Bandwidth': bandwidth,
2476 'duration': segment_duration,
2477 } for segment_number in range(
2478 representation_ms_info['start_number'],
2479 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2481 # $Number*$ or $Time$ in media template with S list available
2482 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2483 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2484 representation_ms_info['fragments'] = []
2487 segment_number = representation_ms_info['start_number']
2489 def add_segment_url():
2490 segment_url = media_template % {
2491 'Time': segment_time,
2492 'Bandwidth': bandwidth,
2493 'Number': segment_number,
2495 representation_ms_info['fragments'].append({
2496 media_location_key: segment_url,
2497 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2500 for num, s in enumerate(representation_ms_info['s']):
2501 segment_time = s.get('t') or segment_time
2505 for r in range(s.get('r', 0)):
2506 segment_time += segment_d
2509 segment_time += segment_d
2510 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2512 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2513 # or any YouTube dashsegments video
2516 timescale = representation_ms_info['timescale']
2517 for s in representation_ms_info['s']:
2518 duration = float_or_none(s['d'], timescale)
2519 for r in range(s.get('r', 0) + 1):
2520 segment_uri = representation_ms_info['segment_urls'][segment_index]
2522 location_key(segment_uri): segment_uri,
2523 'duration': duration,
2526 representation_ms_info['fragments'] = fragments
2527 elif 'segment_urls' in representation_ms_info:
2528 # Segment URLs with no SegmentTimeline
2529 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2530 # https://github.com/ytdl-org/youtube-dl/pull/14844
2532 segment_duration = float_or_none(
2533 representation_ms_info['segment_duration'],
2534 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2535 for segment_url in representation_ms_info['segment_urls']:
2537 location_key(segment_url): segment_url,
2539 if segment_duration:
2540 fragment['duration'] = segment_duration
2541 fragments.append(fragment)
2542 representation_ms_info['fragments'] = fragments
2543 # If there is a fragments key available then we correctly recognized fragmented media.
2544 # Otherwise we will assume unfragmented media with direct access. Technically, such
2545 # assumption is not necessarily correct since we may simply have no support for
2546 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2547 if 'fragments' in representation_ms_info:
2549 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2550 'url': mpd_url or base_url,
2551 'fragment_base_url': base_url,
2553 'protocol': 'http_dash_segments',
2555 if 'initialization_url' in representation_ms_info:
2556 initialization_url = representation_ms_info['initialization_url']
2557 if not f.get('url'):
2558 f['url'] = initialization_url
2559 f['fragments'].append({location_key(initialization_url): initialization_url})
2560 f['fragments'].extend(representation_ms_info['fragments'])
2562 # Assuming direct URL to unfragmented media.
2566 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2569 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2570 res = self._download_xml_handle(
2572 note=note or 'Downloading ISM manifest',
2573 errnote=errnote or 'Failed to download ISM manifest',
2574 fatal=fatal, data=data, headers=headers, query=query)
2581 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2583 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2585 Parse formats
from ISM manifest
.
2587 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2588 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2590 if ism_doc.get('IsLive') == 'TRUE':
2592 if (not self._downloader.params.get('allow_unplayable_formats')
2593 and ism_doc.find('Protection') is not None):
2596 duration = int(ism_doc.attrib['Duration'])
2597 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2600 for stream in ism_doc.findall('StreamIndex'):
2601 stream_type = stream.get('Type')
2602 if stream_type not in ('video', 'audio'):
2604 url_pattern = stream.attrib['Url']
2605 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2606 stream_name = stream.get('Name')
2607 for track in stream.findall('QualityLevel'):
2608 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2609 # TODO: add support for WVC1 and WMAP
2610 if fourcc not in ('H264', 'AVC1', 'AACL'):
2611 self.report_warning('%s is not a supported codec' % fourcc)
2613 tbr = int(track.attrib['Bitrate']) // 1000
2614 # [1] does not mention Width and Height attributes. However,
2615 # they're often present while MaxWidth and MaxHeight are
2616 # missing, so should be used as fallbacks
2617 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2618 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2619 sampling_rate = int_or_none(track.get('SamplingRate'))
2621 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2622 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2628 stream_fragments = stream.findall('c')
2629 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2630 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2631 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2632 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2633 if not fragment_ctx['duration']:
2635 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2637 next_fragment_time = duration
2638 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2639 for _ in range(fragment_repeat):
2641 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2642 'duration': fragment_ctx['duration'] / stream_timescale,
2644 fragment_ctx['time'] += fragment_ctx['duration']
2648 format_id.append(ism_id)
2650 format_id.append(stream_name)
2651 format_id.append(compat_str(tbr))
2654 'format_id': '-'.join(format_id),
2656 'manifest_url': ism_url,
2657 'ext': 'ismv' if stream_type == 'video' else 'isma',
2661 'asr': sampling_rate,
2662 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2663 'acodec': 'none' if stream_type == 'video' else fourcc,
2665 'fragments': fragments,
2666 '_download_params': {
2667 'duration': duration,
2668 'timescale': stream_timescale,
2669 'width': width or 0,
2670 'height': height or 0,
2672 'codec_private_data': track.get('CodecPrivateData'),
2673 'sampling_rate': sampling_rate,
2674 'channels': int_or_none(track.get('Channels', 2)),
2675 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2676 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2681 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2682 def absolute_url(item_url):
2683 return urljoin(base_url, item_url)
2685 def parse_content_type(content_type):
2686 if not content_type:
2688 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2690 mimetype, codecs = ctr.groups()
2691 f = parse_codecs(codecs)
2692 f['ext'] = mimetype2ext(mimetype)
2696 def _media_formats(src, cur_media_type, type_info={}):
2697 full_url = absolute_url(src)
2698 ext = type_info.get('ext') or determine_ext(full_url)
2700 is_plain_url = False
2701 formats = self._extract_m3u8_formats(
2702 full_url, video_id, ext='mp4',
2703 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2704 preference=preference, fatal=False)
2706 is_plain_url = False
2707 formats = self._extract_mpd_formats(
2708 full_url, video_id, mpd_id=mpd_id, fatal=False)
2713 'vcodec': 'none' if cur_media_type == 'audio' else None,
2715 return is_plain_url, formats
2718 # amp-video and amp-audio are very similar to their HTML5 counterparts
2719 # so we wll include them right here (see
2720 # https://www.ampproject.org/docs/reference/components/amp-video)
2721 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
2722 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
2723 media_tags = [(media_tag, media_tag_name, media_type, '')
2724 for media_tag, media_tag_name, media_type
2725 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
2726 media_tags.extend(re.findall(
2727 # We only allow video|audio followed by a whitespace or '>'.
2728 # Allowing more characters may end up in significant slow down (see
2729 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2730 # http://www.porntrex.com/maps/videositemap.xml).
2731 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
2732 for media_tag, _, media_type, media_content in media_tags:
2737 media_attributes = extract_attributes(media_tag)
2738 src = strip_or_none(media_attributes.get('src'))
2740 _, formats = _media_formats(src, media_type)
2741 media_info['formats'].extend(formats)
2742 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2744 for source_tag in re.findall(r'<source[^>]+>', media_content):
2745 s_attr = extract_attributes(source_tag)
2746 # data-video-src and data-src are non standard but seen
2747 # several times in the wild
2748 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2751 f = parse_content_type(s_attr.get('type'))
2752 is_plain_url, formats = _media_formats(src, media_type, f)
2754 # width, height, res, label and title attributes are
2755 # all not standard but seen several times in the wild
2758 for lbl in ('label', 'title')
2759 if str_or_none(s_attr.get(lbl))
2761 width = int_or_none(s_attr.get('width'))
2762 height = (int_or_none(s_attr.get('height'))
2763 or int_or_none(s_attr.get('res')))
2764 if not width or not height:
2766 resolution = parse_resolution(lbl)
2769 width = width or resolution.get('width')
2770 height = height or resolution.get('height')
2772 tbr = parse_bitrate(lbl)
2781 'format_id': s_attr.get('label') or s_attr.get('title'),
2783 f.update(formats[0])
2784 media_info['formats'].append(f)
2786 media_info['formats'].extend(formats)
2787 for track_tag in re.findall(r'<track[^>]+>', media_content):
2788 track_attributes = extract_attributes(track_tag)
2789 kind = track_attributes.get('kind')
2790 if not kind or kind in ('subtitles', 'captions'):
2791 src = strip_or_none(track_attributes.get('src'))
2794 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2795 media_info['subtitles'].setdefault(lang, []).append({
2796 'url': absolute_url(src),
2798 for f in media_info['formats']:
2799 f.setdefault('http_headers', {})['Referer'] = base_url
2800 if media_info['formats'] or media_info['subtitles']:
2801 entries.append(media_info)
2804 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2805 signed = 'hdnea=' in manifest_url
2807 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
2808 manifest_url = re.sub(
2809 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
2810 '', manifest_url).strip('?')
2814 hdcore_sign = 'hdcore=3.7.0'
2815 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2816 hds_host = hosts.get('hds')
2818 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2819 if 'hdcore=' not in f4m_url:
2820 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2821 f4m_formats = self._extract_f4m_formats(
2822 f4m_url, video_id, f4m_id='hds', fatal=False)
2823 for entry in f4m_formats:
2824 entry.update({'extra_param_to_segment_url': hdcore_sign})
2825 formats.extend(f4m_formats)
2827 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2828 hls_host = hosts.get('hls')
2830 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2831 m3u8_formats = self._extract_m3u8_formats(
2832 m3u8_url, video_id, 'mp4', 'm3u8_native',
2833 m3u8_id='hls', fatal=False)
2834 formats.extend(m3u8_formats)
2836 http_host = hosts.get('http')
2837 if http_host and m3u8_formats and not signed:
2838 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
2839 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
2840 qualities_length = len(qualities)
2841 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
2843 for f in m3u8_formats:
2844 if f['vcodec'] != 'none':
2845 for protocol in ('http', 'https'):
2847 del http_f['manifest_url']
2849 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
2851 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
2853 'protocol': protocol,
2855 formats.append(http_f)
2860 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2861 query = compat_urlparse.urlparse(url).query
2862 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2864 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2865 url_base = mobj.group('url')
2866 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2869 def manifest_url(manifest):
2870 m_url = '%s/%s' % (http_base_url, manifest)
2872 m_url += '?%s' % query
2875 if 'm3u8' not in skip_protocols:
2876 formats.extend(self._extract_m3u8_formats(
2877 manifest_url('playlist.m3u8'), video_id, 'mp4',
2878 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2879 if 'f4m' not in skip_protocols:
2880 formats.extend(self._extract_f4m_formats(
2881 manifest_url('manifest.f4m'),
2882 video_id, f4m_id='hds', fatal=False))
2883 if 'dash' not in skip_protocols:
2884 formats.extend(self._extract_mpd_formats(
2885 manifest_url('manifest.mpd'),
2886 video_id, mpd_id='dash', fatal=False))
2887 if re.search(r'(?:/smil:|\.smil)', url_base):
2888 if 'smil' not in skip_protocols:
2889 rtmp_formats = self._extract_smil_formats(
2890 manifest_url('jwplayer.smil'),
2891 video_id, fatal=False)
2892 for rtmp_format in rtmp_formats:
2893 rtsp_format = rtmp_format.copy()
2894 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2895 del rtsp_format['play_path']
2896 del rtsp_format['ext']
2897 rtsp_format.update({
2898 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2899 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2902 formats.extend([rtmp_format, rtsp_format])
2904 for protocol in ('rtmp', 'rtsp'):
2905 if protocol not in skip_protocols:
2907 'url': '%s:%s' % (protocol, url_base),
2908 'format_id': protocol,
2909 'protocol': protocol,
2913 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2915 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2919 jwplayer_data = self._parse_json(mobj.group('options'),
2921 transform_source=transform_source)
2922 except ExtractorError:
2925 if isinstance(jwplayer_data, dict):
2926 return jwplayer_data
2928 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2929 jwplayer_data = self._find_jwplayer_data(
2930 webpage, video_id, transform_source=js_to_json)
2931 return self._parse_jwplayer_data(
2932 jwplayer_data, video_id, *args, **kwargs)
2934 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2935 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2936 # JWPlayer backward compatibility: flattened playlists
2937 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2938 if 'playlist' not in jwplayer_data:
2939 jwplayer_data = {'playlist': [jwplayer_data]}
2943 # JWPlayer backward compatibility: single playlist item
2944 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2945 if not isinstance(jwplayer_data['playlist'], list):
2946 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2948 for video_data in jwplayer_data['playlist']:
2949 # JWPlayer backward compatibility: flattened sources
2950 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2951 if 'sources' not in video_data:
2952 video_data['sources'] = [video_data]
2954 this_video_id = video_id or video_data['mediaid']
2956 formats = self._parse_jwplayer_formats(
2957 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2958 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2961 tracks = video_data.get('tracks')
2962 if tracks and isinstance(tracks, list):
2963 for track in tracks:
2964 if not isinstance(track, dict):
2966 track_kind = track.get('kind')
2967 if not track_kind or not isinstance(track_kind, compat_str):
2969 if track_kind.lower() not in ('captions', 'subtitles'):
2971 track_url = urljoin(base_url, track.get('file'))
2974 subtitles.setdefault(track.get('label') or 'en', []).append({
2975 'url': self._proto_relative_url(track_url)
2979 'id': this_video_id,
2980 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2981 'description': clean_html(video_data.get('description')),
2982 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2983 'timestamp': int_or_none(video_data.get('pubdate')),
2984 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2985 'subtitles': subtitles,
2987 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2988 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2990 '_type': 'url_transparent',
2991 'url': formats[0]['url'],
2994 self._sort_formats(formats)
2995 entry['formats'] = formats
2996 entries.append(entry)
2997 if len(entries) == 1:
3000 return self.playlist_result(entries)
3002 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3003 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3006 for source in jwplayer_sources_data:
3007 if not isinstance(source, dict):
3009 source_url = urljoin(
3010 base_url, self._proto_relative_url(source.get('file')))
3011 if not source_url or source_url in urls:
3013 urls.append(source_url)
3014 source_type = source.get('type') or ''
3015 ext = mimetype2ext(source_type) or determine_ext(source_url)
3016 if source_type == 'hls' or ext == 'm3u8':
3017 formats.extend(self._extract_m3u8_formats(
3018 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3019 m3u8_id=m3u8_id, fatal=False))
3020 elif source_type == 'dash' or ext == 'mpd':
3021 formats.extend(self._extract_mpd_formats(
3022 source_url, video_id, mpd_id=mpd_id, fatal=False))
3024 formats.extend(self._extract_smil_formats(
3025 source_url, video_id, fatal=False))
3026 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3027 elif source_type.startswith('audio') or ext in (
3028 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3035 height = int_or_none(source.get('height'))
3037 # Often no height is provided but there is a label in
3038 # format like "1080p", "720p SD", or 1080.
3039 height = int_or_none(self._search_regex(
3040 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3041 'height', default=None))
3044 'width': int_or_none(source.get('width')),
3046 'tbr': int_or_none(source.get('bitrate')),
3049 if source_url.startswith('rtmp'):
3050 a_format['ext'] = 'flv'
3051 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3052 # of jwplayer.flash.swf
3053 rtmp_url_parts = re.split(
3054 r'((?:mp4|mp3|flv):)', source_url, 1)
3055 if len(rtmp_url_parts) == 3:
3056 rtmp_url, prefix, play_path = rtmp_url_parts
3059 'play_path': prefix + play_path,
3062 a_format.update(rtmp_params)
3063 formats.append(a_format)
3066 def _live_title(self, name):
3067 """ Generate the title
for a live video
"""
3068 now = datetime.datetime.now()
3069 now_str = now.strftime('%Y-%m-%d %H:%M')
3070 return name + ' ' + now_str
3072 def _int(self, v, name, fatal=False, **kwargs):
3073 res = int_or_none(v, **kwargs)
3074 if 'get_attr' in kwargs:
3075 print(getattr(v, kwargs['get_attr']))
3077 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3079 raise ExtractorError(msg)
3081 self._downloader.report_warning(msg)
3084 def _float(self, v, name, fatal=False, **kwargs):
3085 res = float_or_none(v, **kwargs)
3087 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3089 raise ExtractorError(msg)
3091 self._downloader.report_warning(msg)
3094 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3095 path='/', secure=False, discard=False, rest={}, **kwargs):
3096 cookie = compat_cookiejar_Cookie(
3097 0, name, value, port, port is not None, domain, True,
3098 domain.startswith('.'), path, True, secure, expire_time,
3099 discard, None, None, rest)
3100 self._downloader.cookiejar.set_cookie(cookie)
3102 def _get_cookies(self, url):
3103 """ Return a compat_cookies
.SimpleCookie
with the cookies
for the url
"""
3104 req = sanitized_Request(url)
3105 self._downloader.cookiejar.add_cookie_header(req)
3106 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
3108 def _apply_first_set_cookie_header(self, url_handle, cookie):
3110 Apply first Set
-Cookie header instead of the last
. Experimental
.
3112 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3113 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3114 than
second (new
). However
, as of RFC6265 the newer one cookie
3115 should be
set into cookie store what actually happens
.
3116 We will workaround this issue by resetting the cookie to
3117 the first one manually
.
3118 1. https
://new
.vk
.com
/
3119 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3120 3. https
://learning
.oreilly
.com
/
3122 for header, cookies in url_handle.headers.items():
3123 if header.lower() != 'set-cookie':
3125 if sys.version_info[0] >= 3:
3126 cookies = cookies.encode('iso-8859-1')
3127 cookies = cookies.decode('utf-8')
3128 cookie_value = re.search(
3129 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3131 value, domain = cookie_value.groups()
3132 self._set_cookie(domain, cookie, value)
3135 def get_testcases(self, include_onlymatching=False):
3136 t = getattr(self, '_TEST', None)
3138 assert not hasattr(self, '_TESTS'), \
3139 '%s has _TEST and _TESTS' % type(self).__name__
3142 tests = getattr(self, '_TESTS', [])
3144 if not include_onlymatching and t.get('only_matching', False):
3146 t['name'] = type(self).__name__[:-len('IE')]
3149 def is_suitable(self, age_limit):
3150 """ Test whether the extractor
is generally suitable
for the given
3151 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3153 any_restricted = False
3154 for tc in self.get_testcases(include_onlymatching=False):
3155 if tc.get('playlist', []):
3156 tc = tc['playlist'][0]
3157 is_restricted = age_restricted(
3158 tc.get('info_dict', {}).get('age_limit'), age_limit)
3159 if not is_restricted:
3161 any_restricted = any_restricted or is_restricted
3162 return not any_restricted
3164 def extract_subtitles(self, *args, **kwargs):
3165 if (self._downloader.params.get('writesubtitles', False)
3166 or self._downloader.params.get('listsubtitles')):
3167 return self._get_subtitles(*args, **kwargs)
3170 def _get_subtitles(self, *args, **kwargs):
3171 raise NotImplementedError('This method must be implemented by subclasses')
3174 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3175 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3176 will be dropped
. """
3177 list1_urls = set([item['url'] for item in subtitle_list1])
3178 ret = list(subtitle_list1)
3179 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3183 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
3184 """ Merge two subtitle dictionaries
, language by language
. """
3185 ret = dict(subtitle_dict1)
3186 for lang in subtitle_dict2:
3187 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
3190 def extract_automatic_captions(self, *args, **kwargs):
3191 if (self._downloader.params.get('writeautomaticsub', False)
3192 or self._downloader.params.get('listsubtitles')):
3193 return self._get_automatic_captions(*args, **kwargs)
3196 def _get_automatic_captions(self, *args, **kwargs):
3197 raise NotImplementedError('This method must be implemented by subclasses')
3199 def mark_watched(self, *args, **kwargs):
3200 if (self._downloader.params.get('mark_watched', False)
3201 and (self._get_login_info()[0] is not None
3202 or self._downloader.params.get('cookiefile') is not None)):
3203 self._mark_watched(*args, **kwargs)
3205 def _mark_watched(self, *args, **kwargs):
3206 raise NotImplementedError('This method must be implemented by subclasses')
3208 def geo_verification_headers(self):
3210 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
3211 if geo_verification_proxy:
3212 headers['Ytdl-request-proxy'] = geo_verification_proxy
3215 def _generic_id(self, url):
3216 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3218 def _generic_title(self, url):
3219 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3222 class SearchInfoExtractor(InfoExtractor):
3224 Base
class for paged search queries extractors
.
3225 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3226 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3230 def _make_valid_url(cls):
3231 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3234 def suitable(cls, url):
3235 return re.match(cls._make_valid_url(), url) is not None
3237 def _real_extract(self, query):
3238 mobj = re.match(self._make_valid_url(), query)
3240 raise ExtractorError('Invalid search query "%s"' % query)
3242 prefix = mobj.group('prefix')
3243 query = mobj.group('query')
3245 return self._get_n_results(query, 1)
3246 elif prefix == 'all':
3247 return self._get_n_results(query, self._MAX_RESULTS)
3251 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3252 elif n > self._MAX_RESULTS:
3253 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3254 n = self._MAX_RESULTS
3255 return self._get_n_results(query, n)
3257 def _get_n_results(self, query, n):
3258 """Get a specified number of results
for a query
"""
3259 raise NotImplementedError('This method must be implemented by subclasses')
3262 def SEARCH_KEY(self):
3263 return self._SEARCH_KEY