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", "description", "uploader",
340 "uploader_id", "uploader_url", "duration" attributes with the same semantics
341 as videos (see above).
344 _type "multi_video" indicates that there are multiple videos that
345 form a single show, for examples multiple acts of an opera or TV episode.
346 It must have an entries key like a playlist and contain all the keys
347 required for a video at the same time.
350 _type "url" indicates that the video must be extracted from another
351 location, possibly by a different extractor. Its only required key is:
352 "url" - the next URL to extract.
353 The key "ie_key" can be set to the class name (minus the trailing "IE",
354 e.g. "Youtube") if the extractor class is known in advance.
355 Additionally, the dictionary may have any properties of the resolved entity
356 known in advance, for example "title" if the title of the referred video is
360 _type "url_transparent" entities have the same specification as "url", but
361 indicate that the given additional information is more precise than the one
362 associated with the resolved URL.
363 This is useful when a site employs a video service that hosts the video and
364 its technical metadata, but that video service does not embed a useful
365 title, description etc.
368 Subclasses of this one should re-define the _real_initialize() and
369 _real_extract() methods and define a _VALID_URL regexp.
370 Probably, they should also be added to the list of extractors.
372 _GEO_BYPASS attribute may be set to False in order to disable
373 geo restriction bypass mechanisms for a particular extractor.
374 Though it won't disable explicit geo restriction bypass based on
375 country code provided with geo_bypass_country.
377 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
378 countries for this extractor. One of these countries will be used by
379 geo restriction bypass mechanism right away in order to bypass
380 geo restriction, of course, if the mechanism is not disabled.
382 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
383 IP blocks in CIDR notation for this extractor. One of these IP blocks
384 will be used by geo restriction bypass mechanism similarly
387 Finally, the _WORKING attribute should be set to False for broken IEs
388 in order to warn the users and skip the tests.
393 _x_forwarded_for_ip
= None
395 _GEO_COUNTRIES
= None
396 _GEO_IP_BLOCKS
= None
399 def __init__(self
, downloader
=None):
400 """Constructor. Receives an optional downloader."""
402 self
._x
_forwarded
_for
_ip
= None
403 self
.set_downloader(downloader
)
406 def suitable(cls
, url
):
407 """Receives a URL and returns True if suitable for this IE."""
409 # This does not use has/getattr intentionally - we want to know whether
410 # we have cached the regexp for *this* class, whereas getattr would also
411 # match the superclass
412 if '_VALID_URL_RE' not in cls
.__dict
__:
413 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
414 return cls
._VALID
_URL
_RE
.match(url
) is not None
417 def _match_id(cls
, url
):
418 if '_VALID_URL_RE' not in cls
.__dict
__:
419 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
420 m
= cls
._VALID
_URL
_RE
.match(url
)
422 return compat_str(m
.group('id'))
426 """Getter method for _WORKING."""
429 def initialize(self
):
430 """Initializes an instance (authentication, etc)."""
431 self
._initialize
_geo
_bypass
({
432 'countries': self
._GEO
_COUNTRIES
,
433 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
436 self
._real
_initialize
()
439 def _initialize_geo_bypass(self
, geo_bypass_context
):
441 Initialize geo restriction bypass mechanism.
443 This method is used to initialize geo bypass mechanism based on faking
444 X-Forwarded-For HTTP header. A random country from provided country list
445 is selected and a random IP belonging to this country is generated. This
446 IP will be passed as X-Forwarded-For HTTP header in all subsequent
449 This method will be used for initial geo bypass mechanism initialization
450 during the instance initialization with _GEO_COUNTRIES and
453 You may also manually call it from extractor's code if geo bypass
454 information is not available beforehand (e.g. obtained during
455 extraction) or due to some other reason. In this case you should pass
456 this information in geo bypass context passed as first argument. It may
457 contain following fields:
459 countries: List of geo unrestricted countries (similar
461 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
462 (similar to _GEO_IP_BLOCKS)
465 if not self
._x
_forwarded
_for
_ip
:
467 # Geo bypass mechanism is explicitly disabled by user
468 if not self
._downloader
.params
.get('geo_bypass', True):
471 if not geo_bypass_context
:
472 geo_bypass_context
= {}
474 # Backward compatibility: previously _initialize_geo_bypass
475 # expected a list of countries, some 3rd party code may still use
477 if isinstance(geo_bypass_context
, (list, tuple)):
478 geo_bypass_context
= {
479 'countries': geo_bypass_context
,
482 # The whole point of geo bypass mechanism is to fake IP
483 # as X-Forwarded-For HTTP header based on some IP block or
486 # Path 1: bypassing based on IP block in CIDR notation
488 # Explicit IP block specified by user, use it right away
489 # regardless of whether extractor is geo bypassable or not
490 ip_block
= self
._downloader
.params
.get('geo_bypass_ip_block', None)
492 # Otherwise use random IP block from geo bypass context but only
493 # if extractor is known as geo bypassable
495 ip_blocks
= geo_bypass_context
.get('ip_blocks')
496 if self
._GEO
_BYPASS
and ip_blocks
:
497 ip_block
= random
.choice(ip_blocks
)
500 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
501 if self
._downloader
.params
.get('verbose', False):
502 self
._downloader
.to_screen(
503 '[debug] Using fake IP %s as X-Forwarded-For.'
504 % self
._x
_forwarded
_for
_ip
)
507 # Path 2: bypassing based on country code
509 # Explicit country code specified by user, use it right away
510 # regardless of whether extractor is geo bypassable or not
511 country
= self
._downloader
.params
.get('geo_bypass_country', None)
513 # Otherwise use random country code from geo bypass context but
514 # only if extractor is known as geo bypassable
516 countries
= geo_bypass_context
.get('countries')
517 if self
._GEO
_BYPASS
and countries
:
518 country
= random
.choice(countries
)
521 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
522 if self
._downloader
.params
.get('verbose', False):
523 self
._downloader
.to_screen(
524 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
525 % (self
._x
_forwarded
_for
_ip
, country
.upper()))
527 def extract(self
, url
):
528 """Extracts URL information and returns it in list of dicts."""
533 ie_result
= self
._real
_extract
(url
)
534 if self
._x
_forwarded
_for
_ip
:
535 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
537 except GeoRestrictedError
as e
:
538 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
541 except ExtractorError
:
543 except compat_http_client
.IncompleteRead
as e
:
544 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
545 except (KeyError, StopIteration) as e
:
546 raise ExtractorError('An extractor error has occurred.', cause
=e
)
548 def __maybe_fake_ip_and_retry(self
, countries
):
549 if (not self
._downloader
.params
.get('geo_bypass_country', None)
551 and self
._downloader
.params
.get('geo_bypass', True)
552 and not self
._x
_forwarded
_for
_ip
554 country_code
= random
.choice(countries
)
555 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
556 if self
._x
_forwarded
_for
_ip
:
558 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
559 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
563 def set_downloader(self
, downloader
):
564 """Sets the downloader for this IE."""
565 self
._downloader
= downloader
567 def _real_initialize(self
):
568 """Real initialization process. Redefine in subclasses."""
571 def _real_extract(self
, url
):
572 """Real extraction process. Redefine in subclasses."""
577 """A string for getting the InfoExtractor with get_info_extractor"""
578 return compat_str(cls
.__name
__[:-2])
582 return compat_str(type(self
).__name
__[:-2])
585 def __can_accept_status_code(err
, expected_status
):
586 assert isinstance(err
, compat_urllib_error
.HTTPError
)
587 if expected_status
is None:
589 if isinstance(expected_status
, compat_integer_types
):
590 return err
.code
== expected_status
591 elif isinstance(expected_status
, (list, tuple)):
592 return err
.code
in expected_status
593 elif callable(expected_status
):
594 return expected_status(err
.code
) is True
598 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
600 Return the response handle.
602 See _download_webpage docstring for arguments specification.
605 self
.report_download_webpage(video_id
)
606 elif note
is not False:
608 self
.to_screen('%s' % (note
,))
610 self
.to_screen('%s: %s' % (video_id
, note
))
612 # Some sites check X-Forwarded-For HTTP header in order to figure out
613 # the origin of the client behind proxy. This allows bypassing geo
614 # restriction by faking this header's value to IP that belongs to some
615 # geo unrestricted country. We will do so once we encounter any
616 # geo restriction error.
617 if self
._x
_forwarded
_for
_ip
:
618 if 'X-Forwarded-For' not in headers
:
619 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
621 if isinstance(url_or_request
, compat_urllib_request
.Request
):
622 url_or_request
= update_Request(
623 url_or_request
, data
=data
, headers
=headers
, query
=query
)
626 url_or_request
= update_url_query(url_or_request
, query
)
627 if data
is not None or headers
:
628 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
629 exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
630 if hasattr(ssl
, 'CertificateError'):
631 exceptions
.append(ssl
.CertificateError
)
633 return self
._downloader
.urlopen(url_or_request
)
634 except tuple(exceptions
) as err
:
635 if isinstance(err
, compat_urllib_error
.HTTPError
):
636 if self
.__can
_accept
_status
_code
(err
, expected_status
):
637 # Retain reference to error to prevent file object from
638 # being closed before it can be read. Works around the
639 # effects of <https://bugs.python.org/issue15002>
640 # introduced in Python 3.4.1.
647 errnote
= 'Unable to download webpage'
649 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
651 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
653 self
._downloader
.report_warning(errmsg
)
656 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):
658 Return a tuple (page content as string, URL handle).
660 See _download_webpage docstring for arguments specification.
662 # Strip hashes from the URL (#1038)
663 if isinstance(url_or_request
, (compat_str
, str)):
664 url_or_request
= url_or_request
.partition('#')[0]
666 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
670 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
671 return (content
, urlh
)
674 def _guess_encoding_from_content(content_type
, webpage_bytes
):
675 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
677 encoding
= m
.group(1)
679 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
680 webpage_bytes[:1024])
682 encoding = m.group(1).decode('ascii')
683 elif webpage_bytes.startswith(b'\xff\xfe'):
690 def __check_blocked(self, content):
691 first_block = content[:512]
692 if ('<title>Access to this site is blocked</title>' in content
693 and 'Websense' in first_block):
694 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
695 blocked_iframe = self._html_search_regex(
696 r'<iframe src="([^
"]+)"', content,
697 'Websense information URL
', default=None)
699 msg += ' Visit
%s for more details
' % blocked_iframe
700 raise ExtractorError(msg, expected=True)
701 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
703 'Access to this webpage has been blocked by Indian censorship
. '
704 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
705 block_msg = self._html_search_regex(
706 r'</h1
><p
>(.*?
)</p
>',
707 content, 'block message
', default=None)
709 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
710 raise ExtractorError(msg, expected=True)
711 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
712 and 'blocklist
.rkn
.gov
.ru
' in content):
713 raise ExtractorError(
714 'Access to this webpage has been blocked by decision of the Russian government
. '
715 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
718 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
719 content_type = urlh.headers.get('Content
-Type
', '')
720 webpage_bytes = urlh.read()
721 if prefix is not None:
722 webpage_bytes = prefix + webpage_bytes
724 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
725 if self._downloader.params.get('dump_intermediate_pages
', False):
726 self.to_screen('Dumping request to
' + urlh.geturl())
727 dump = base64.b64encode(webpage_bytes).decode('ascii
')
728 self._downloader.to_screen(dump)
729 if self._downloader.params.get('write_pages
', False):
730 basen = '%s_%s' % (video_id, urlh.geturl())
732 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
733 basen = basen[:240 - len(h)] + h
734 raw_filename = basen + '.dump
'
735 filename = sanitize_filename(raw_filename, restricted=True)
736 self.to_screen('Saving request to
' + filename)
737 # Working around MAX_PATH limitation on Windows (see
738 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
739 if compat_os_name == 'nt
':
740 absfilepath = os.path.abspath(filename)
741 if len(absfilepath) > 259:
742 filename = '\\\\?
\\' + absfilepath
743 with open(filename, 'wb
') as outf:
744 outf.write(webpage_bytes)
747 content = webpage_bytes.decode(encoding, 'replace
')
749 content = webpage_bytes.decode('utf
-8', 'replace
')
751 self.__check_blocked(content)
755 def _download_webpage(
756 self, url_or_request, video_id, note=None, errnote=None,
757 fatal=True, tries=1, timeout=5, encoding=None, data=None,
758 headers={}, query={}, expected_status=None):
760 Return the data of the page as a string.
763 url_or_request -- plain text URL as a string or
764 a compat_urllib_request.Requestobject
765 video_id -- Video/playlist/item identifier (string)
768 note -- note printed before downloading (string)
769 errnote -- note printed in case of an error (string)
770 fatal -- flag denoting whether error should be considered fatal,
771 i.e. whether it should cause ExtractionError to be raised,
772 otherwise a warning will be reported and extraction continued
773 tries -- number of tries
774 timeout -- sleep interval between tries
775 encoding -- encoding for a page content decoding, guessed automatically
776 when not explicitly specified
777 data -- POST data (bytes)
778 headers -- HTTP headers (dict)
779 query -- URL query (dict)
780 expected_status -- allows to accept failed HTTP requests (non 2xx
781 status code) by explicitly specifying a set of accepted status
782 codes. Can be any of the following entities:
783 - an integer type specifying an exact failed status code to
785 - a list or a tuple of integer types specifying a list of
786 failed status codes to accept
787 - a callable accepting an actual failed status code and
788 returning True if it should be accepted
789 Note that this argument does not affect success status codes (2xx)
790 which are always accepted.
795 while success is False:
797 res = self._download_webpage_handle(
798 url_or_request, video_id, note, errnote, fatal,
799 encoding=encoding, data=data, headers=headers, query=query,
800 expected_status=expected_status)
802 except compat_http_client.IncompleteRead as e:
804 if try_count >= tries:
806 self._sleep(timeout, video_id)
813 def _download_xml_handle(
814 self, url_or_request, video_id, note='Downloading XML
',
815 errnote='Unable to download XML
', transform_source=None,
816 fatal=True, encoding=None, data=None, headers={}, query={},
817 expected_status=None):
819 Return a tuple (xml as an compat_etree_Element, URL handle).
821 See _download_webpage docstring for arguments specification.
823 res = self._download_webpage_handle(
824 url_or_request, video_id, note, errnote, fatal=fatal,
825 encoding=encoding, data=data, headers=headers, query=query,
826 expected_status=expected_status)
829 xml_string, urlh = res
830 return self._parse_xml(
831 xml_string, video_id, transform_source=transform_source,
835 self, url_or_request, video_id,
836 note='Downloading XML
', errnote='Unable to download XML
',
837 transform_source=None, fatal=True, encoding=None,
838 data=None, headers={}, query={}, expected_status=None):
840 Return the xml as an compat_etree_Element.
842 See _download_webpage docstring for arguments specification.
844 res = self._download_xml_handle(
845 url_or_request, video_id, note=note, errnote=errnote,
846 transform_source=transform_source, fatal=fatal, encoding=encoding,
847 data=data, headers=headers, query=query,
848 expected_status=expected_status)
849 return res if res is False else res[0]
851 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
853 xml_string = transform_source(xml_string)
855 return compat_etree_fromstring(xml_string.encode('utf
-8'))
856 except compat_xml_parse_error as ve:
857 errmsg = '%s: Failed to parse XML
' % video_id
859 raise ExtractorError(errmsg, cause=ve)
861 self.report_warning(errmsg + str(ve))
863 def _download_json_handle(
864 self, url_or_request, video_id, note='Downloading JSON metadata
',
865 errnote='Unable to download JSON metadata
', transform_source=None,
866 fatal=True, encoding=None, data=None, headers={}, query={},
867 expected_status=None):
869 Return a tuple (JSON object, URL handle).
871 See _download_webpage docstring for arguments specification.
873 res = self._download_webpage_handle(
874 url_or_request, video_id, note, errnote, fatal=fatal,
875 encoding=encoding, data=data, headers=headers, query=query,
876 expected_status=expected_status)
879 json_string, urlh = res
880 return self._parse_json(
881 json_string, video_id, transform_source=transform_source,
885 self, url_or_request, video_id, note='Downloading JSON metadata
',
886 errnote='Unable to download JSON metadata
', transform_source=None,
887 fatal=True, encoding=None, data=None, headers={}, query={},
888 expected_status=None):
890 Return the JSON object as a dict.
892 See _download_webpage docstring for arguments specification.
894 res = self._download_json_handle(
895 url_or_request, video_id, note=note, errnote=errnote,
896 transform_source=transform_source, fatal=fatal, encoding=encoding,
897 data=data, headers=headers, query=query,
898 expected_status=expected_status)
899 return res if res is False else res[0]
901 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
903 json_string = transform_source(json_string)
905 return json.loads(json_string)
906 except ValueError as ve:
907 errmsg = '%s: Failed to parse JSON
' % video_id
909 raise ExtractorError(errmsg, cause=ve)
911 self.report_warning(errmsg + str(ve))
913 def report_warning(self, msg, video_id=None):
914 idstr = '' if video_id is None else '%s: ' % video_id
915 self._downloader.report_warning(
916 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
918 def to_screen(self, msg):
919 """Print msg to screen, prefixing it with '[ie_name
]'"""
920 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
922 def report_extraction(self, id_or_name):
923 """Report information extraction."""
924 self.to_screen('%s: Extracting information
' % id_or_name)
926 def report_download_webpage(self, video_id):
927 """Report webpage download."""
928 self.to_screen('%s: Downloading webpage
' % video_id)
930 def report_age_confirmation(self):
931 """Report attempt to confirm age."""
932 self.to_screen('Confirming age
')
934 def report_login(self):
935 """Report attempt to log in."""
936 self.to_screen('Logging
in')
939 def raise_login_required(msg='This video
is only available
for registered users
'):
940 raise ExtractorError(
941 '%s. Use
--username
and --password
or --netrc to provide account credentials
.' % msg,
945 def raise_geo_restricted(msg='This video
is not available
from your location due to geo restriction
', countries=None):
946 raise GeoRestrictedError(msg, countries=countries)
948 # Methods for following #608
950 def url_result(url, ie=None, video_id=None, video_title=None):
951 """Returns a URL that points to a page that should be processed"""
952 # TODO: ie should be the class used for getting the info
953 video_info = {'_type
': 'url
',
956 if video_id is not None:
957 video_info['id'] = video_id
958 if video_title is not None:
959 video_info['title
'] = video_title
962 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
964 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
966 return self.playlist_result(
967 urls, playlist_id=playlist_id, playlist_title=playlist_title)
970 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
971 """Returns a playlist"""
972 video_info = {'_type
': 'playlist
',
975 video_info['id'] = playlist_id
977 video_info['title
'] = playlist_title
978 if playlist_description:
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
', 'has_video
', 'extractor
', 'lang
', 'quality
',
1370 'res
', 'fps
', 'codec
', 'size
', 'br
', 'asr
',
1371 'proto
', 'ext
', 'has_audio
', 'source
', 'format_id
')
1374 'vcodec
': {'type': 'ordered
', 'regex
': True,
1375 'order
': ['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 'protocol
': {'type': 'ordered
', 'regex
': True,
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 'extractor_preference
': {'priority': True, 'type': 'extractor'},
1388 'has_video
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1389 'has_audio
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1390 'language_preference
': {'priority': True, 'convert': 'ignore'},
1391 'quality
': {'priority': True, 'convert': 'float_none'},
1392 'filesize
': {'convert': 'bytes'},
1393 'filesize_approx
': {'convert': 'bytes'},
1394 'format_id
': {'convert': 'string'},
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_preference
': {'convert': 'ignore'},
1403 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1404 'bitrate
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1405 'filesize_estimate
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'filesize_approx')},
1406 'extension
': {'type': 'combined', 'field': ('vext', 'aext')},
1407 'dimension
': {'type': 'multiple', 'field': ('height', 'width'), 'function': min}, # not named as 'resolution
' because such a field exists
1408 'res
': {'type': 'alias', 'field': 'dimension'},
1409 'ext
': {'type': 'alias', 'field': 'extension'},
1410 'br
': {'type': 'alias', 'field': 'bitrate'},
1411 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1412 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1413 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1414 'framerate
': {'type': 'alias', 'field': 'fps'},
1415 'lang
': {'type': 'alias', 'field': 'language_preference'}, # not named as 'language
' because such a field exists
1416 'proto
': {'type': 'alias', 'field': 'protocol'},
1417 'source
': {'type': 'alias', 'field': 'source_preference'},
1418 'size
': {'type': 'alias', 'field': 'filesize_estimate'},
1419 'samplerate
': {'type': 'alias', 'field': 'asr'},
1420 'video_ext
': {'type': 'alias', 'field': 'vext'},
1421 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1422 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1423 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1424 'video
': {'type': 'alias', 'field': 'has_video'},
1425 'audio
': {'type': 'alias', 'field': 'has_audio'},
1426 'extractor
': {'type': 'alias', 'field': 'extractor_preference'},
1427 'preference
': {'type': 'alias', 'field': 'extractor_preference'}}
1431 def _get_field_setting(self, field, key):
1432 if field not in self.settings:
1433 self.settings[field] = {}
1434 propObj = self.settings[field]
1435 if key not in propObj:
1436 type = propObj.get('type')
1438 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1439 elif key == 'convert
':
1440 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1442 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,), 'function': max}.get(key, None)
1443 propObj[key] = default
1446 def _resolve_field_value(self, field, value, convertNone=False):
1451 value = value.lower()
1452 conversion = self._get_field_setting(field, 'convert
')
1453 if conversion == 'ignore
':
1455 if conversion == 'string
':
1457 elif conversion == 'float_none
':
1458 return float_or_none(value)
1459 elif conversion == 'bytes':
1460 return FileDownloader.parse_bytes(value)
1461 elif conversion == 'order
':
1462 order_free = self._get_field_setting(field, 'order_free
')
1463 order_list = order_free if order_free and self._use_free_order else self._get_field_setting(field, 'order
')
1464 use_regex = self._get_field_setting(field, 'regex
')
1465 list_length = len(order_list)
1466 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1467 if use_regex and value is not None:
1468 for (i, regex) in enumerate(order_list):
1469 if regex and re.match(regex, value):
1470 return list_length - i
1471 return list_length - empty_pos # not in list
1472 else: # not regex or value = None
1473 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1475 if value.isnumeric():
1478 self.settings[field]['convert
'] = 'string
'
1481 def evaluate_params(self, params, sort_extractor):
1482 self._use_free_order = params.get('prefer_free_formats
', False)
1483 self._sort_user = params.get('format_sort
', [])
1484 self._sort_extractor = sort_extractor
1486 def add_item(field, reverse, closest, limit_text):
1487 field = field.lower()
1488 if field in self._order:
1490 self._order.append(field)
1491 limit = self._resolve_field_value(field, limit_text)
1494 'closest
': False if limit is None else closest,
1495 'limit_text
': limit_text,
1497 if field in self.settings:
1498 self.settings[field].update(data)
1500 self.settings[field] = data
1503 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1504 + (tuple() if params.get('format_sort_force
', False)
1505 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1506 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1508 for item in sort_list:
1509 match = re.match(self.regex, item)
1511 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1512 field = match.group('field
')
1515 if self._get_field_setting(field, 'type') == 'alias
':
1516 field = self._get_field_setting(field, 'field
')
1517 reverse = match.group('reverse
') is not None
1518 closest = match.group('seperator
') == '~
'
1519 limit_text = match.group('limit
')
1521 has_limit = limit_text is not None
1522 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1523 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1525 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1526 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1527 limit_count = len(limits)
1528 for (i, f) in enumerate(fields):
1529 add_item(f, reverse, closest,
1530 limits[i] if i < limit_count
1531 else limits[0] if has_limit and not has_multiple_limits
1534 def print_verbose_info(self, to_screen):
1535 to_screen('[debug
] Sort order given by user
: %s' % ','.join(self._sort_user))
1536 if self._sort_extractor:
1537 to_screen('[debug
] Sort order given by extractor
: %s' % ','.join(self._sort_extractor))
1538 to_screen('[debug
] Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1539 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1540 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1541 self._get_field_setting(field, 'limit_text
'),
1542 self._get_field_setting(field, 'limit
'))
1543 if self._get_field_setting(field, 'limit_text
') is not None else '')
1544 for field in self._order if self._get_field_setting(field, 'visible
')]))
1546 def _calculate_field_preference_from_value(self, format, field, type, value):
1547 reverse = self._get_field_setting(field, 'reverse
')
1548 closest = self._get_field_setting(field, 'closest
')
1549 limit = self._get_field_setting(field, 'limit
')
1551 if type == 'extractor
':
1552 maximum = self._get_field_setting(field, 'max')
1553 if value is None or (maximum is not None and value >= maximum):
1555 elif type == 'boolean
':
1556 in_list = self._get_field_setting(field, 'in_list
')
1557 not_in_list = self._get_field_setting(field, 'not_in_list
')
1558 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
1559 elif type == 'ordered
':
1560 value = self._resolve_field_value(field, value, True)
1562 # try to convert to number
1563 val_num = float_or_none(value)
1564 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1568 return ((-10, 0) if value is None
1569 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1570 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1571 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1572 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1573 else (-1, value, 0))
1575 def _calculate_field_preference(self, format, field):
1576 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1577 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1578 if type == 'multiple
':
1579 type = 'field
' # Only 'field
' is allowed in multiple for now
1580 actual_fields = self._get_field_setting(field, 'field
')
1582 def wrapped_function(values):
1583 values = tuple(filter(lambda x: x is not None, values))
1584 return (self._get_field_setting(field, 'function
')(*values) if len(values) > 1
1585 else values[0] if values
1588 value = wrapped_function((get_value(f) for f in actual_fields))
1590 value = get_value(field)
1591 return self._calculate_field_preference_from_value(format, field, type, value)
1593 def calculate_preference(self, format):
1594 # Determine missing protocol
1595 if not format.get('protocol
'):
1596 format['protocol
'] = determine_protocol(format)
1598 # Determine missing ext
1599 if not format.get('ext
') and 'url
' in format:
1600 format['ext
'] = determine_ext(format['url
'])
1601 if format.get('vcodec
') == 'none
':
1602 format['audio_ext
'] = format['ext
']
1603 format['video_ext
'] = 'none
'
1605 format['video_ext
'] = format['ext
']
1606 format['audio_ext
'] = 'none
'
1607 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1608 # format['preference
'] = -1000
1610 # Determine missing bitrates
1611 if format.get('tbr
') is None:
1612 if format.get('vbr
') is not None and format.get('abr
') is not None:
1613 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1615 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1616 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1617 if format.get('acodec
') != "none" and format.get('abr
') is None:
1618 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1620 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1622 def _sort_formats(self, formats, field_preference=[]):
1624 raise ExtractorError('No video formats found
')
1625 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1626 format_sort.evaluate_params(self._downloader.params, field_preference)
1627 if self._downloader.params.get('verbose
', False):
1628 format_sort.print_verbose_info(self._downloader.to_screen)
1629 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1631 def _check_formats(self, formats, video_id):
1633 formats[:] = filter(
1634 lambda f: self._is_valid_url(
1636 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1640 def _remove_duplicate_formats(formats):
1644 if f['url
'] not in format_urls:
1645 format_urls.add(f['url
'])
1646 unique_formats.append(f)
1647 formats[:] = unique_formats
1649 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1650 url = self._proto_relative_url(url, scheme='http
:')
1651 # For now assume non HTTP(S) URLs always valid
1652 if not (url.startswith('http
://') or url.startswith('https
://')):
1655 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1657 except ExtractorError as e:
1659 '%s: %s URL
is invalid
, skipping
: %s'
1660 % (video_id, item, error_to_compat_str(e.cause)))
1663 def http_scheme(self):
1664 """ Either "http:" or "https:", depending on the user's preferences
"""
1667 if self._downloader.params.get('prefer_insecure', False)
1670 def _proto_relative_url(self, url, scheme=None):
1673 if url.startswith('//'):
1675 scheme = self.http_scheme()
1680 def _sleep(self, timeout, video_id, msg_template=None):
1681 if msg_template is None:
1682 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1683 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1687 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1688 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1689 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1690 manifest = self._download_xml(
1691 manifest_url, video_id, 'Downloading f4m manifest',
1692 'Unable to download f4m manifest',
1693 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1694 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1695 transform_source=transform_source,
1696 fatal=fatal, data=data, headers=headers, query=query)
1698 if manifest is False:
1701 return self._parse_f4m_formats(
1702 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1703 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1705 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1706 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1707 fatal=True, m3u8_id=None):
1708 if not isinstance(manifest, compat_etree_Element) and not fatal:
1711 # currently youtube-dlc cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1712 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1713 if akamai_pv is not None and ';' in akamai_pv.text:
1714 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1715 if playerVerificationChallenge.strip() != '':
1719 manifest_version = '1.0'
1720 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1722 manifest_version = '2.0'
1723 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1724 # Remove unsupported DRM protected media from final formats
1725 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1726 media_nodes = remove_encrypted_media(media_nodes)
1730 manifest_base_url = get_base_url(manifest)
1732 bootstrap_info = xpath_element(
1733 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1734 'bootstrap info', default=None)
1737 mime_type = xpath_text(
1738 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1739 'base URL', default=None)
1740 if mime_type and mime_type.startswith('audio/'):
1743 for i, media_el in enumerate(media_nodes):
1744 tbr = int_or_none(media_el.attrib.get('bitrate'))
1745 width = int_or_none(media_el.attrib.get('width'))
1746 height = int_or_none(media_el.attrib.get('height'))
1747 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1748 # If <bootstrapInfo> is present, the specified f4m is a
1749 # stream-level manifest, and only set-level manifests may refer to
1750 # external resources. See section 11.4 and section 4 of F4M spec
1751 if bootstrap_info is None:
1753 # @href is introduced in 2.0, see section 11.6 of F4M spec
1754 if manifest_version == '2.0':
1755 media_url = media_el.attrib.get('href')
1756 if media_url is None:
1757 media_url = media_el.attrib.get('url')
1761 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1762 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1763 # If media_url is itself a f4m manifest do the recursive extraction
1764 # since bitrates in parent manifest (this one) and media_url manifest
1765 # may differ leading to inability to resolve the format by requested
1766 # bitrate in f4m downloader
1767 ext = determine_ext(manifest_url)
1769 f4m_formats = self._extract_f4m_formats(
1770 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1771 transform_source=transform_source, fatal=fatal)
1772 # Sometimes stream-level manifest contains single media entry that
1773 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1774 # At the same time parent's media entry in set-level manifest may
1775 # contain it. We will copy it from parent in such cases.
1776 if len(f4m_formats) == 1:
1779 'tbr': f.get('tbr') or tbr,
1780 'width': f.get('width') or width,
1781 'height': f.get('height') or height,
1782 'format_id': f.get('format_id') if not tbr else format_id,
1785 formats.extend(f4m_formats)
1788 formats.extend(self._extract_m3u8_formats(
1789 manifest_url, video_id, 'mp4', preference=preference,
1790 m3u8_id=m3u8_id, fatal=fatal))
1793 'format_id': format_id,
1794 'url': manifest_url,
1795 'manifest_url': manifest_url,
1796 'ext': 'flv' if bootstrap_info is not None else None,
1802 'preference': preference,
1806 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1808 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1812 'preference': preference - 100 if preference else -100,
1813 'resolution': 'multiple',
1814 'format_note': 'Quality selection URL',
1817 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1818 entry_protocol='m3u8', preference=None,
1819 m3u8_id=None, note=None, errnote=None,
1820 fatal=True, live=False, data=None, headers={},
1822 res = self._download_webpage_handle(
1824 note=note or 'Downloading m3u8 information',
1825 errnote=errnote or 'Failed to download m3u8 information',
1826 fatal=fatal, data=data, headers=headers, query=query)
1831 m3u8_doc, urlh = res
1832 m3u8_url = urlh.geturl()
1834 return self._parse_m3u8_formats(
1835 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1836 preference=preference, m3u8_id=m3u8_id, live=live)
1838 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1839 entry_protocol='m3u8', preference=None,
1840 m3u8_id=None, live=False):
1841 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1844 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1849 format_url = lambda u: (
1851 if re.match(r'^https?://', u)
1852 else compat_urlparse.urljoin(m3u8_url, u))
1855 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1856 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1857 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1859 # We should try extracting formats only from master playlists [1, 4.3.4],
1860 # i.e. playlists that describe available qualities. On the other hand
1861 # media playlists [1, 4.3.3] should be returned as is since they contain
1862 # just the media without qualities renditions.
1863 # Fortunately, master playlist can be easily distinguished from media
1864 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1865 # master playlist tags MUST NOT appear in a media playlist and vice versa.
1866 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1867 # media playlist and MUST NOT appear in master playlist thus we can
1868 # clearly detect media playlist with this criterion.
1870 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1873 'format_id': m3u8_id,
1875 'protocol': entry_protocol,
1876 'preference': preference,
1880 last_stream_inf = {}
1882 def extract_media(x_media_line):
1883 media = parse_m3u8_attributes(x_media_line)
1884 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1885 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1886 if not (media_type and group_id and name):
1888 groups.setdefault(group_id, []).append(media)
1889 if media_type not in ('VIDEO', 'AUDIO'):
1891 media_url = media.get('URI')
1894 for v in (m3u8_id, group_id, name):
1898 'format_id': '-'.join(format_id),
1899 'url': format_url(media_url),
1900 'manifest_url': m3u8_url,
1901 'language': media.get('LANGUAGE'),
1903 'protocol': entry_protocol,
1904 'preference': preference,
1906 if media_type == 'AUDIO':
1907 f['vcodec'] = 'none'
1910 def build_stream_name():
1911 # Despite specification does not mention NAME attribute for
1912 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1913 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1914 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1915 stream_name = last_stream_inf.get('NAME')
1918 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1919 # from corresponding rendition group
1920 stream_group_id = last_stream_inf.get('VIDEO')
1921 if not stream_group_id:
1923 stream_group = groups.get(stream_group_id)
1924 if not stream_group:
1925 return stream_group_id
1926 rendition = stream_group[0]
1927 return rendition.get('NAME') or stream_group_id
1929 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1930 # chance to detect video only formats when EXT-X-STREAM-INF tags
1931 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1932 for line in m3u8_doc.splitlines():
1933 if line.startswith('#EXT-X-MEDIA:'):
1936 for line in m3u8_doc.splitlines():
1937 if line.startswith('#EXT-X-STREAM-INF:'):
1938 last_stream_inf = parse_m3u8_attributes(line)
1939 elif line.startswith('#') or not line.strip():
1942 tbr = float_or_none(
1943 last_stream_inf.get('AVERAGE-BANDWIDTH')
1944 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1947 format_id.append(m3u8_id)
1948 stream_name = build_stream_name()
1949 # Bandwidth of live streams may differ over time thus making
1950 # format_id unpredictable. So it's better to keep provided
1953 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1954 manifest_url = format_url(line.strip())
1956 'format_id': '-'.join(format_id),
1957 'url': manifest_url,
1958 'manifest_url': m3u8_url,
1961 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1962 'protocol': entry_protocol,
1963 'preference': preference,
1965 resolution = last_stream_inf.get('RESOLUTION')
1967 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1969 f['width'] = int(mobj.group('width'))
1970 f['height'] = int(mobj.group('height'))
1971 # Unified Streaming Platform
1973 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1975 abr, vbr = mobj.groups()
1976 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1981 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1983 audio_group_id = last_stream_inf.get('AUDIO')
1984 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1985 # references a rendition group MUST have a CODECS attribute.
1986 # However, this is not always respected, for example, [2]
1987 # contains EXT-X-STREAM-INF tag which references AUDIO
1988 # rendition group but does not have CODECS and despite
1989 # referencing an audio group it represents a complete
1990 # (with audio and video) format. So, for such cases we will
1991 # ignore references to rendition groups and treat them
1992 # as complete formats.
1993 if audio_group_id and codecs and f.get('vcodec') != 'none':
1994 audio_group = groups.get(audio_group_id)
1995 if audio_group and audio_group[0].get('URI'):
1996 # TODO: update acodec for audio only formats with
1998 f['acodec'] = 'none'
2002 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2005 del http_f['manifest_url']
2007 'format_id': f['format_id'].replace('hls-', 'http-'),
2009 'url': progressive_uri,
2011 formats.append(http_f)
2013 last_stream_inf = {}
2017 def _xpath_ns(path, namespace=None):
2021 for c in path.split('/'):
2022 if not c or c == '.':
2025 out.append('{%s}%s' % (namespace, c))
2026 return '/'.join(out)
2028 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2029 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2035 namespace = self._parse_smil_namespace(smil)
2037 return self._parse_smil_formats(
2038 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2040 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2041 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2044 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2046 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2047 return self._download_xml(
2048 smil_url, video_id, 'Downloading SMIL file',
2049 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2051 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2052 namespace = self._parse_smil_namespace(smil)
2054 formats = self._parse_smil_formats(
2055 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2056 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2058 video_id = os.path.splitext(url_basename(smil_url))[0]
2062 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2063 name = meta.attrib.get('name')
2064 content = meta.attrib.get('content')
2065 if not name or not content:
2067 if not title and name == 'title':
2069 elif not description and name in ('description', 'abstract'):
2070 description = content
2071 elif not upload_date and name == 'date':
2072 upload_date = unified_strdate(content)
2075 'id': image.get('type'),
2076 'url': image.get('src'),
2077 'width': int_or_none(image.get('width')),
2078 'height': int_or_none(image.get('height')),
2079 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2083 'title': title or video_id,
2084 'description': description,
2085 'upload_date': upload_date,
2086 'thumbnails': thumbnails,
2088 'subtitles': subtitles,
2091 def _parse_smil_namespace(self, smil):
2092 return self._search_regex(
2093 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2095 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2097 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2098 b = meta.get('base') or meta.get('httpBase')
2109 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2110 for medium in media:
2111 src = medium.get('src')
2112 if not src or src in srcs:
2116 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2117 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2118 width = int_or_none(medium.get('width'))
2119 height = int_or_none(medium.get('height'))
2120 proto = medium.get('proto')
2121 ext = medium.get('ext')
2122 src_ext = determine_ext(src)
2123 streamer = medium.get('streamer') or base
2125 if proto == 'rtmp' or streamer.startswith('rtmp'):
2131 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2133 'filesize': filesize,
2137 if transform_rtmp_url:
2138 streamer, src = transform_rtmp_url(streamer, src)
2139 formats[-1].update({
2145 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2146 src_url = src_url.strip()
2148 if proto == 'm3u8' or src_ext == 'm3u8':
2149 m3u8_formats = self._extract_m3u8_formats(
2150 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2151 if len(m3u8_formats) == 1:
2153 m3u8_formats[0].update({
2154 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2159 formats.extend(m3u8_formats)
2160 elif src_ext == 'f4m':
2165 'plugin': 'flowplayer-3.2.0.1',
2167 f4m_url += '&' if '?' in f4m_url else '?'
2168 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2169 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2170 elif src_ext == 'mpd':
2171 formats.extend(self._extract_mpd_formats(
2172 src_url, video_id, mpd_id='dash', fatal=False))
2173 elif re.search(r'\.ism/[Mm]anifest', src_url):
2174 formats.extend(self._extract_ism_formats(
2175 src_url, video_id, ism_id='mss', fatal=False))
2176 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2180 'ext': ext or src_ext or 'flv',
2181 'format_id': 'http-%d' % (bitrate or http_count),
2183 'filesize': filesize,
2190 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2193 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2194 src = textstream.get('src')
2195 if not src or src in urls:
2198 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2199 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2200 subtitles.setdefault(lang, []).append({
2206 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2207 xspf = self._download_xml(
2208 xspf_url, playlist_id, 'Downloading xpsf playlist',
2209 'Unable to download xspf manifest', fatal=fatal)
2212 return self._parse_xspf(
2213 xspf, playlist_id, xspf_url=xspf_url,
2214 xspf_base_url=base_url(xspf_url))
2216 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2218 'xspf': 'http://xspf.org/ns/0/',
2219 's1': 'http://static.streamone.nl/player/ns/0',
2223 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2225 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2226 description = xpath_text(
2227 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2228 thumbnail = xpath_text(
2229 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2230 duration = float_or_none(
2231 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2234 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2235 format_url = urljoin(xspf_base_url, location.text)
2240 'manifest_url': xspf_url,
2241 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2242 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2243 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2245 self._sort_formats(formats)
2250 'description': description,
2251 'thumbnail': thumbnail,
2252 'duration': duration,
2257 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
2258 res = self._download_xml_handle(
2260 note=note or 'Downloading MPD manifest',
2261 errnote=errnote or 'Failed to download MPD manifest',
2262 fatal=fatal, data=data, headers=headers, query=query)
2268 mpd_base_url = base_url(urlh.geturl())
2270 return self._parse_mpd_formats(
2271 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2272 formats_dict=formats_dict, mpd_url=mpd_url)
2274 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2276 Parse formats
from MPD manifest
.
2278 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2279 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2280 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2282 if not self._downloader.params.get('dynamic_mpd'):
2283 if mpd_doc.get('type') == 'dynamic':
2286 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2289 return self._xpath_ns(path, namespace)
2291 def is_drm_protected(element):
2292 return element.find(_add_ns('ContentProtection')) is not None
2294 def extract_multisegment_info(element, ms_parent_info):
2295 ms_info = ms_parent_info.copy()
2297 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2298 # common attributes and elements. We will only extract relevant
2300 def extract_common(source):
2301 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2302 if segment_timeline is not None:
2303 s_e = segment_timeline.findall(_add_ns('S'))
2305 ms_info['total_number'] = 0
2308 r = int(s.get('r', 0))
2309 ms_info['total_number'] += 1 + r
2310 ms_info['s'].append({
2311 't': int(s.get('t', 0)),
2312 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2313 'd': int(s.attrib['d']),
2316 start_number = source.get('startNumber')
2318 ms_info['start_number'] = int(start_number)
2319 timescale = source.get('timescale')
2321 ms_info['timescale'] = int(timescale)
2322 segment_duration = source.get('duration')
2323 if segment_duration:
2324 ms_info['segment_duration'] = float(segment_duration)
2326 def extract_Initialization(source):
2327 initialization = source.find(_add_ns('Initialization'))
2328 if initialization is not None:
2329 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2331 segment_list = element.find(_add_ns('SegmentList'))
2332 if segment_list is not None:
2333 extract_common(segment_list)
2334 extract_Initialization(segment_list)
2335 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2337 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2339 segment_template = element.find(_add_ns('SegmentTemplate'))
2340 if segment_template is not None:
2341 extract_common(segment_template)
2342 media = segment_template.get('media')
2344 ms_info['media'] = media
2345 initialization = segment_template.get('initialization')
2347 ms_info['initialization'] = initialization
2349 extract_Initialization(segment_template)
2352 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2354 for period in mpd_doc.findall(_add_ns('Period')):
2355 period_duration = parse_duration(period.get('duration')) or mpd_duration
2356 period_ms_info = extract_multisegment_info(period, {
2360 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2361 if is_drm_protected(adaptation_set):
2363 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2364 for representation in adaptation_set.findall(_add_ns('Representation')):
2365 if is_drm_protected(representation):
2367 representation_attrib = adaptation_set.attrib.copy()
2368 representation_attrib.update(representation.attrib)
2369 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2370 mime_type = representation_attrib['mimeType']
2371 content_type = mime_type.split('/')[0]
2372 if content_type == 'text':
2373 # TODO implement WebVTT downloading
2375 elif content_type in ('video', 'audio'):
2377 for element in (representation, adaptation_set, period, mpd_doc):
2378 base_url_e = element.find(_add_ns('BaseURL'))
2379 if base_url_e is not None:
2380 base_url = base_url_e.text + base_url
2381 if re.match(r'^https?://', base_url):
2383 if mpd_base_url and not re.match(r'^https?://', base_url):
2384 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2386 base_url = mpd_base_url + base_url
2387 representation_id = representation_attrib.get('id')
2388 lang = representation_attrib.get('lang')
2389 url_el = representation.find(_add_ns('BaseURL'))
2390 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2391 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2393 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2394 'manifest_url': mpd_url,
2395 'ext': mimetype2ext(mime_type),
2396 'width': int_or_none(representation_attrib.get('width')),
2397 'height': int_or_none(representation_attrib.get('height')),
2398 'tbr': float_or_none(bandwidth, 1000),
2399 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2400 'fps': int_or_none(representation_attrib.get('frameRate')),
2401 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2402 'format_note': 'DASH %s' % content_type,
2403 'filesize': filesize,
2404 'container': mimetype2ext(mime_type) + '_dash',
2406 f.update(parse_codecs(representation_attrib.get('codecs')))
2407 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2409 def prepare_template(template_name, identifiers):
2410 tmpl = representation_ms_info[template_name]
2411 # First of, % characters outside $...$ templates
2412 # must be escaped by doubling for proper processing
2413 # by % operator string formatting used further (see
2414 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2420 in_template = not in_template
2421 elif c == '%' and not in_template:
2423 # Next, $...$ templates are translated to their
2424 # %(...) counterparts to be used with % operator
2425 t = t.replace('$RepresentationID$', representation_id)
2426 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2427 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2428 t.replace('$$', '$')
2431 # @initialization is a regular template like @media one
2432 # so it should be handled just the same way (see
2433 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2434 if 'initialization' in representation_ms_info:
2435 initialization_template = prepare_template(
2437 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2438 # $Time$ shall not be included for @initialization thus
2439 # only $Bandwidth$ remains
2441 representation_ms_info['initialization_url'] = initialization_template % {
2442 'Bandwidth': bandwidth,
2445 def location_key(location):
2446 return 'url' if re.match(r'^https?://', location) else 'path'
2448 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2450 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2451 media_location_key = location_key(media_template)
2453 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2454 # can't be used at the same time
2455 if '%(Number' in media_template and 's' not in representation_ms_info:
2456 segment_duration = None
2457 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2458 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2459 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2460 representation_ms_info['fragments'] = [{
2461 media_location_key: media_template % {
2462 'Number': segment_number,
2463 'Bandwidth': bandwidth,
2465 'duration': segment_duration,
2466 } for segment_number in range(
2467 representation_ms_info['start_number'],
2468 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2470 # $Number*$ or $Time$ in media template with S list available
2471 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2472 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2473 representation_ms_info['fragments'] = []
2476 segment_number = representation_ms_info['start_number']
2478 def add_segment_url():
2479 segment_url = media_template % {
2480 'Time': segment_time,
2481 'Bandwidth': bandwidth,
2482 'Number': segment_number,
2484 representation_ms_info['fragments'].append({
2485 media_location_key: segment_url,
2486 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2489 for num, s in enumerate(representation_ms_info['s']):
2490 segment_time = s.get('t') or segment_time
2494 for r in range(s.get('r', 0)):
2495 segment_time += segment_d
2498 segment_time += segment_d
2499 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2501 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2502 # or any YouTube dashsegments video
2505 timescale = representation_ms_info['timescale']
2506 for s in representation_ms_info['s']:
2507 duration = float_or_none(s['d'], timescale)
2508 for r in range(s.get('r', 0) + 1):
2509 segment_uri = representation_ms_info['segment_urls'][segment_index]
2511 location_key(segment_uri): segment_uri,
2512 'duration': duration,
2515 representation_ms_info['fragments'] = fragments
2516 elif 'segment_urls' in representation_ms_info:
2517 # Segment URLs with no SegmentTimeline
2518 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2519 # https://github.com/ytdl-org/youtube-dl/pull/14844
2521 segment_duration = float_or_none(
2522 representation_ms_info['segment_duration'],
2523 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2524 for segment_url in representation_ms_info['segment_urls']:
2526 location_key(segment_url): segment_url,
2528 if segment_duration:
2529 fragment['duration'] = segment_duration
2530 fragments.append(fragment)
2531 representation_ms_info['fragments'] = fragments
2532 # If there is a fragments key available then we correctly recognized fragmented media.
2533 # Otherwise we will assume unfragmented media with direct access. Technically, such
2534 # assumption is not necessarily correct since we may simply have no support for
2535 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2536 if 'fragments' in representation_ms_info:
2538 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2539 'url': mpd_url or base_url,
2540 'fragment_base_url': base_url,
2542 'protocol': 'http_dash_segments',
2544 if 'initialization_url' in representation_ms_info:
2545 initialization_url = representation_ms_info['initialization_url']
2546 if not f.get('url'):
2547 f['url'] = initialization_url
2548 f['fragments'].append({location_key(initialization_url): initialization_url})
2549 f['fragments'].extend(representation_ms_info['fragments'])
2551 # Assuming direct URL to unfragmented media.
2554 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2555 # is not necessarily unique within a Period thus formats with
2556 # the same `format_id` are quite possible. There are numerous examples
2557 # of such manifests (see https://github.com/ytdl-org/youtube-dl/issues/15111,
2558 # https://github.com/ytdl-org/youtube-dl/issues/13919)
2559 full_info = formats_dict.get(representation_id, {}).copy()
2561 formats.append(full_info)
2563 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2566 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2567 res = self._download_xml_handle(
2569 note=note or 'Downloading ISM manifest',
2570 errnote=errnote or 'Failed to download ISM manifest',
2571 fatal=fatal, data=data, headers=headers, query=query)
2578 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2580 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2582 Parse formats
from ISM manifest
.
2584 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2585 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2587 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2590 duration = int(ism_doc.attrib['Duration'])
2591 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2594 for stream in ism_doc.findall('StreamIndex'):
2595 stream_type = stream.get('Type')
2596 if stream_type not in ('video', 'audio'):
2598 url_pattern = stream.attrib['Url']
2599 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2600 stream_name = stream.get('Name')
2601 for track in stream.findall('QualityLevel'):
2602 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2603 # TODO: add support for WVC1 and WMAP
2604 if fourcc not in ('H264', 'AVC1', 'AACL'):
2605 self.report_warning('%s is not a supported codec' % fourcc)
2607 tbr = int(track.attrib['Bitrate']) // 1000
2608 # [1] does not mention Width and Height attributes. However,
2609 # they're often present while MaxWidth and MaxHeight are
2610 # missing, so should be used as fallbacks
2611 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2612 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2613 sampling_rate = int_or_none(track.get('SamplingRate'))
2615 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2616 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2622 stream_fragments = stream.findall('c')
2623 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2624 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2625 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2626 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2627 if not fragment_ctx['duration']:
2629 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2631 next_fragment_time = duration
2632 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2633 for _ in range(fragment_repeat):
2635 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2636 'duration': fragment_ctx['duration'] / stream_timescale,
2638 fragment_ctx['time'] += fragment_ctx['duration']
2642 format_id.append(ism_id)
2644 format_id.append(stream_name)
2645 format_id.append(compat_str(tbr))
2648 'format_id': '-'.join(format_id),
2650 'manifest_url': ism_url,
2651 'ext': 'ismv' if stream_type == 'video' else 'isma',
2655 'asr': sampling_rate,
2656 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2657 'acodec': 'none' if stream_type == 'video' else fourcc,
2659 'fragments': fragments,
2660 '_download_params': {
2661 'duration': duration,
2662 'timescale': stream_timescale,
2663 'width': width or 0,
2664 'height': height or 0,
2666 'codec_private_data': track.get('CodecPrivateData'),
2667 'sampling_rate': sampling_rate,
2668 'channels': int_or_none(track.get('Channels', 2)),
2669 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2670 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2675 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2676 def absolute_url(item_url):
2677 return urljoin(base_url, item_url)
2679 def parse_content_type(content_type):
2680 if not content_type:
2682 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2684 mimetype, codecs = ctr.groups()
2685 f = parse_codecs(codecs)
2686 f['ext'] = mimetype2ext(mimetype)
2690 def _media_formats(src, cur_media_type, type_info={}):
2691 full_url = absolute_url(src)
2692 ext = type_info.get('ext') or determine_ext(full_url)
2694 is_plain_url = False
2695 formats = self._extract_m3u8_formats(
2696 full_url, video_id, ext='mp4',
2697 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2698 preference=preference, fatal=False)
2700 is_plain_url = False
2701 formats = self._extract_mpd_formats(
2702 full_url, video_id, mpd_id=mpd_id, fatal=False)
2707 'vcodec': 'none' if cur_media_type == 'audio' else None,
2709 return is_plain_url, formats
2712 # amp-video and amp-audio are very similar to their HTML5 counterparts
2713 # so we wll include them right here (see
2714 # https://www.ampproject.org/docs/reference/components/amp-video)
2715 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
2716 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
2717 media_tags = [(media_tag, media_tag_name, media_type, '')
2718 for media_tag, media_tag_name, media_type
2719 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
2720 media_tags.extend(re.findall(
2721 # We only allow video|audio followed by a whitespace or '>'.
2722 # Allowing more characters may end up in significant slow down (see
2723 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2724 # http://www.porntrex.com/maps/videositemap.xml).
2725 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
2726 for media_tag, _, media_type, media_content in media_tags:
2731 media_attributes = extract_attributes(media_tag)
2732 src = strip_or_none(media_attributes.get('src'))
2734 _, formats = _media_formats(src, media_type)
2735 media_info['formats'].extend(formats)
2736 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2738 for source_tag in re.findall(r'<source[^>]+>', media_content):
2739 s_attr = extract_attributes(source_tag)
2740 # data-video-src and data-src are non standard but seen
2741 # several times in the wild
2742 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2745 f = parse_content_type(s_attr.get('type'))
2746 is_plain_url, formats = _media_formats(src, media_type, f)
2748 # width, height, res, label and title attributes are
2749 # all not standard but seen several times in the wild
2752 for lbl in ('label', 'title')
2753 if str_or_none(s_attr.get(lbl))
2755 width = int_or_none(s_attr.get('width'))
2756 height = (int_or_none(s_attr.get('height'))
2757 or int_or_none(s_attr.get('res')))
2758 if not width or not height:
2760 resolution = parse_resolution(lbl)
2763 width = width or resolution.get('width')
2764 height = height or resolution.get('height')
2766 tbr = parse_bitrate(lbl)
2775 'format_id': s_attr.get('label') or s_attr.get('title'),
2777 f.update(formats[0])
2778 media_info['formats'].append(f)
2780 media_info['formats'].extend(formats)
2781 for track_tag in re.findall(r'<track[^>]+>', media_content):
2782 track_attributes = extract_attributes(track_tag)
2783 kind = track_attributes.get('kind')
2784 if not kind or kind in ('subtitles', 'captions'):
2785 src = strip_or_none(track_attributes.get('src'))
2788 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2789 media_info['subtitles'].setdefault(lang, []).append({
2790 'url': absolute_url(src),
2792 for f in media_info['formats']:
2793 f.setdefault('http_headers', {})['Referer'] = base_url
2794 if media_info['formats'] or media_info['subtitles']:
2795 entries.append(media_info)
2798 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2799 signed = 'hdnea=' in manifest_url
2801 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
2802 manifest_url = re.sub(
2803 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
2804 '', manifest_url).strip('?')
2808 hdcore_sign = 'hdcore=3.7.0'
2809 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2810 hds_host = hosts.get('hds')
2812 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2813 if 'hdcore=' not in f4m_url:
2814 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2815 f4m_formats = self._extract_f4m_formats(
2816 f4m_url, video_id, f4m_id='hds', fatal=False)
2817 for entry in f4m_formats:
2818 entry.update({'extra_param_to_segment_url': hdcore_sign})
2819 formats.extend(f4m_formats)
2821 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2822 hls_host = hosts.get('hls')
2824 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2825 m3u8_formats = self._extract_m3u8_formats(
2826 m3u8_url, video_id, 'mp4', 'm3u8_native',
2827 m3u8_id='hls', fatal=False)
2828 formats.extend(m3u8_formats)
2830 http_host = hosts.get('http')
2831 if http_host and m3u8_formats and not signed:
2832 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
2833 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
2834 qualities_length = len(qualities)
2835 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
2837 for f in m3u8_formats:
2838 if f['vcodec'] != 'none':
2839 for protocol in ('http', 'https'):
2841 del http_f['manifest_url']
2843 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
2845 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
2847 'protocol': protocol,
2849 formats.append(http_f)
2854 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2855 query = compat_urlparse.urlparse(url).query
2856 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2858 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2859 url_base = mobj.group('url')
2860 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2863 def manifest_url(manifest):
2864 m_url = '%s/%s' % (http_base_url, manifest)
2866 m_url += '?%s' % query
2869 if 'm3u8' not in skip_protocols:
2870 formats.extend(self._extract_m3u8_formats(
2871 manifest_url('playlist.m3u8'), video_id, 'mp4',
2872 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2873 if 'f4m' not in skip_protocols:
2874 formats.extend(self._extract_f4m_formats(
2875 manifest_url('manifest.f4m'),
2876 video_id, f4m_id='hds', fatal=False))
2877 if 'dash' not in skip_protocols:
2878 formats.extend(self._extract_mpd_formats(
2879 manifest_url('manifest.mpd'),
2880 video_id, mpd_id='dash', fatal=False))
2881 if re.search(r'(?:/smil:|\.smil)', url_base):
2882 if 'smil' not in skip_protocols:
2883 rtmp_formats = self._extract_smil_formats(
2884 manifest_url('jwplayer.smil'),
2885 video_id, fatal=False)
2886 for rtmp_format in rtmp_formats:
2887 rtsp_format = rtmp_format.copy()
2888 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2889 del rtsp_format['play_path']
2890 del rtsp_format['ext']
2891 rtsp_format.update({
2892 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2893 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2896 formats.extend([rtmp_format, rtsp_format])
2898 for protocol in ('rtmp', 'rtsp'):
2899 if protocol not in skip_protocols:
2901 'url': '%s:%s' % (protocol, url_base),
2902 'format_id': protocol,
2903 'protocol': protocol,
2907 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2909 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2913 jwplayer_data = self._parse_json(mobj.group('options'),
2915 transform_source=transform_source)
2916 except ExtractorError:
2919 if isinstance(jwplayer_data, dict):
2920 return jwplayer_data
2922 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2923 jwplayer_data = self._find_jwplayer_data(
2924 webpage, video_id, transform_source=js_to_json)
2925 return self._parse_jwplayer_data(
2926 jwplayer_data, video_id, *args, **kwargs)
2928 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2929 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2930 # JWPlayer backward compatibility: flattened playlists
2931 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2932 if 'playlist' not in jwplayer_data:
2933 jwplayer_data = {'playlist': [jwplayer_data]}
2937 # JWPlayer backward compatibility: single playlist item
2938 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2939 if not isinstance(jwplayer_data['playlist'], list):
2940 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2942 for video_data in jwplayer_data['playlist']:
2943 # JWPlayer backward compatibility: flattened sources
2944 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2945 if 'sources' not in video_data:
2946 video_data['sources'] = [video_data]
2948 this_video_id = video_id or video_data['mediaid']
2950 formats = self._parse_jwplayer_formats(
2951 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2952 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2955 tracks = video_data.get('tracks')
2956 if tracks and isinstance(tracks, list):
2957 for track in tracks:
2958 if not isinstance(track, dict):
2960 track_kind = track.get('kind')
2961 if not track_kind or not isinstance(track_kind, compat_str):
2963 if track_kind.lower() not in ('captions', 'subtitles'):
2965 track_url = urljoin(base_url, track.get('file'))
2968 subtitles.setdefault(track.get('label') or 'en', []).append({
2969 'url': self._proto_relative_url(track_url)
2973 'id': this_video_id,
2974 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2975 'description': clean_html(video_data.get('description')),
2976 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2977 'timestamp': int_or_none(video_data.get('pubdate')),
2978 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2979 'subtitles': subtitles,
2981 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2982 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2984 '_type': 'url_transparent',
2985 'url': formats[0]['url'],
2988 self._sort_formats(formats)
2989 entry['formats'] = formats
2990 entries.append(entry)
2991 if len(entries) == 1:
2994 return self.playlist_result(entries)
2996 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2997 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3000 for source in jwplayer_sources_data:
3001 if not isinstance(source, dict):
3003 source_url = urljoin(
3004 base_url, self._proto_relative_url(source.get('file')))
3005 if not source_url or source_url in urls:
3007 urls.append(source_url)
3008 source_type = source.get('type') or ''
3009 ext = mimetype2ext(source_type) or determine_ext(source_url)
3010 if source_type == 'hls' or ext == 'm3u8':
3011 formats.extend(self._extract_m3u8_formats(
3012 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3013 m3u8_id=m3u8_id, fatal=False))
3014 elif source_type == 'dash' or ext == 'mpd':
3015 formats.extend(self._extract_mpd_formats(
3016 source_url, video_id, mpd_id=mpd_id, fatal=False))
3018 formats.extend(self._extract_smil_formats(
3019 source_url, video_id, fatal=False))
3020 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3021 elif source_type.startswith('audio') or ext in (
3022 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3029 height = int_or_none(source.get('height'))
3031 # Often no height is provided but there is a label in
3032 # format like "1080p", "720p SD", or 1080.
3033 height = int_or_none(self._search_regex(
3034 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3035 'height', default=None))
3038 'width': int_or_none(source.get('width')),
3040 'tbr': int_or_none(source.get('bitrate')),
3043 if source_url.startswith('rtmp'):
3044 a_format['ext'] = 'flv'
3045 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3046 # of jwplayer.flash.swf
3047 rtmp_url_parts = re.split(
3048 r'((?:mp4|mp3|flv):)', source_url, 1)
3049 if len(rtmp_url_parts) == 3:
3050 rtmp_url, prefix, play_path = rtmp_url_parts
3053 'play_path': prefix + play_path,
3056 a_format.update(rtmp_params)
3057 formats.append(a_format)
3060 def _live_title(self, name):
3061 """ Generate the title
for a live video
"""
3062 now = datetime.datetime.now()
3063 now_str = now.strftime('%Y-%m-%d %H:%M')
3064 return name + ' ' + now_str
3066 def _int(self, v, name, fatal=False, **kwargs):
3067 res = int_or_none(v, **kwargs)
3068 if 'get_attr' in kwargs:
3069 print(getattr(v, kwargs['get_attr']))
3071 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3073 raise ExtractorError(msg)
3075 self._downloader.report_warning(msg)
3078 def _float(self, v, name, fatal=False, **kwargs):
3079 res = float_or_none(v, **kwargs)
3081 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3083 raise ExtractorError(msg)
3085 self._downloader.report_warning(msg)
3088 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3089 path='/', secure=False, discard=False, rest={}, **kwargs):
3090 cookie = compat_cookiejar_Cookie(
3091 0, name, value, port, port is not None, domain, True,
3092 domain.startswith('.'), path, True, secure, expire_time,
3093 discard, None, None, rest)
3094 self._downloader.cookiejar.set_cookie(cookie)
3096 def _get_cookies(self, url):
3097 """ Return a compat_cookies
.SimpleCookie
with the cookies
for the url
"""
3098 req = sanitized_Request(url)
3099 self._downloader.cookiejar.add_cookie_header(req)
3100 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
3102 def _apply_first_set_cookie_header(self, url_handle, cookie):
3104 Apply first Set
-Cookie header instead of the last
. Experimental
.
3106 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3107 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3108 than
second (new
). However
, as of RFC6265 the newer one cookie
3109 should be
set into cookie store what actually happens
.
3110 We will workaround this issue by resetting the cookie to
3111 the first one manually
.
3112 1. https
://new
.vk
.com
/
3113 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3114 3. https
://learning
.oreilly
.com
/
3116 for header, cookies in url_handle.headers.items():
3117 if header.lower() != 'set-cookie':
3119 if sys.version_info[0] >= 3:
3120 cookies = cookies.encode('iso-8859-1')
3121 cookies = cookies.decode('utf-8')
3122 cookie_value = re.search(
3123 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3125 value, domain = cookie_value.groups()
3126 self._set_cookie(domain, cookie, value)
3129 def get_testcases(self, include_onlymatching=False):
3130 t = getattr(self, '_TEST', None)
3132 assert not hasattr(self, '_TESTS'), \
3133 '%s has _TEST and _TESTS' % type(self).__name__
3136 tests = getattr(self, '_TESTS', [])
3138 if not include_onlymatching and t.get('only_matching', False):
3140 t['name'] = type(self).__name__[:-len('IE')]
3143 def is_suitable(self, age_limit):
3144 """ Test whether the extractor
is generally suitable
for the given
3145 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3147 any_restricted = False
3148 for tc in self.get_testcases(include_onlymatching=False):
3149 if tc.get('playlist', []):
3150 tc = tc['playlist'][0]
3151 is_restricted = age_restricted(
3152 tc.get('info_dict', {}).get('age_limit'), age_limit)
3153 if not is_restricted:
3155 any_restricted = any_restricted or is_restricted
3156 return not any_restricted
3158 def extract_subtitles(self, *args, **kwargs):
3159 if (self._downloader.params.get('writesubtitles', False)
3160 or self._downloader.params.get('listsubtitles')):
3161 return self._get_subtitles(*args, **kwargs)
3164 def _get_subtitles(self, *args, **kwargs):
3165 raise NotImplementedError('This method must be implemented by subclasses')
3168 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3169 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3170 will be dropped
. """
3171 list1_urls = set([item['url'] for item in subtitle_list1])
3172 ret = list(subtitle_list1)
3173 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3177 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
3178 """ Merge two subtitle dictionaries
, language by language
. """
3179 ret = dict(subtitle_dict1)
3180 for lang in subtitle_dict2:
3181 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
3184 def extract_automatic_captions(self, *args, **kwargs):
3185 if (self._downloader.params.get('writeautomaticsub', False)
3186 or self._downloader.params.get('listsubtitles')):
3187 return self._get_automatic_captions(*args, **kwargs)
3190 def _get_automatic_captions(self, *args, **kwargs):
3191 raise NotImplementedError('This method must be implemented by subclasses')
3193 def mark_watched(self, *args, **kwargs):
3194 if (self._downloader.params.get('mark_watched', False)
3195 and (self._get_login_info()[0] is not None
3196 or self._downloader.params.get('cookiefile') is not None)):
3197 self._mark_watched(*args, **kwargs)
3199 def _mark_watched(self, *args, **kwargs):
3200 raise NotImplementedError('This method must be implemented by subclasses')
3202 def geo_verification_headers(self):
3204 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
3205 if geo_verification_proxy:
3206 headers['Ytdl-request-proxy'] = geo_verification_proxy
3209 def _generic_id(self, url):
3210 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3212 def _generic_title(self, url):
3213 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3216 class SearchInfoExtractor(InfoExtractor):
3218 Base
class for paged search queries extractors
.
3219 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3220 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3224 def _make_valid_url(cls):
3225 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3228 def suitable(cls, url):
3229 return re.match(cls._make_valid_url(), url) is not None
3231 def _real_extract(self, query):
3232 mobj = re.match(self._make_valid_url(), query)
3234 raise ExtractorError('Invalid search query "%s"' % query)
3236 prefix = mobj.group('prefix')
3237 query = mobj.group('query')
3239 return self._get_n_results(query, 1)
3240 elif prefix == 'all':
3241 return self._get_n_results(query, self._MAX_RESULTS)
3245 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3246 elif n > self._MAX_RESULTS:
3247 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3248 n = self._MAX_RESULTS
3249 return self._get_n_results(query, n)
3251 def _get_n_results(self, query, n):
3252 """Get a specified number of results
for a query
"""
3253 raise NotImplementedError('This method must be implemented by subclasses')
3256 def SEARCH_KEY(self):
3257 return self._SEARCH_KEY