2 from __future__
import unicode_literals
18 from ..compat
import (
19 compat_cookiejar_Cookie
,
22 compat_etree_fromstring
,
29 compat_urllib_parse_unquote
,
30 compat_urllib_parse_urlencode
,
31 compat_urllib_request
,
33 compat_xml_parse_error
,
35 from ..downloader
import FileDownloader
36 from ..downloader
.f4m
import (
38 remove_encrypted_media
,
66 parse_m3u8_attributes
,
88 class InfoExtractor(object):
89 """Information Extractor class.
91 Information extractors are the classes that, given a URL, extract
92 information about the video (or videos) the URL refers to. This
93 information includes the real video URL, the video title, author and
94 others. The information is stored in a dictionary which is then
95 passed to the YoutubeDL. The YoutubeDL processes this
96 information possibly downloading the video to the file system, among
97 other possible outcomes.
99 The type field determines the type of the result.
100 By far the most common value (and the default if _type is missing) is
101 "video", which indicates a single video.
103 For a video, the dictionaries must include the following fields:
105 id: Video identifier.
106 title: Video title, unescaped.
108 Additionally, it must contain either a formats entry or a url one:
110 formats: A list of dictionaries for each format available, ordered
111 from worst to best quality.
114 * url The mandatory URL representing the media:
115 for plain file media - HTTP URL of this file,
117 for HLS - URL of the M3U8 media playlist,
118 for HDS - URL of the F4M manifest,
120 - HTTP URL to plain file media (in case of
122 - URL of the MPD manifest or base URL
123 representing the media if MPD manifest
124 is parsed from a string (in case of
126 for MSS - URL of the ISM manifest.
128 The URL of the manifest file in case of
130 for HLS - URL of the M3U8 master playlist,
131 for HDS - URL of the F4M manifest,
132 for DASH - URL of the MPD manifest,
133 for MSS - URL of the ISM manifest.
134 * ext Will be calculated from URL if missing
135 * format A human-readable description of the format
136 ("mp4 container with h264/opus").
137 Calculated from the format_id, width, height.
138 and format_note fields if missing.
139 * format_id A short description of the format
140 ("mp4_h264_opus" or "19").
141 Technically optional, but strongly recommended.
142 * format_note Additional info about the format
143 ("3D" or "DASH video")
144 * width Width of the video, if known
145 * height Height of the video, if known
146 * resolution Textual description of width and height
147 * tbr Average bitrate of audio and video in KBit/s
148 * abr Average audio bitrate in KBit/s
149 * acodec Name of the audio codec in use
150 * asr Audio sampling rate in Hertz
151 * vbr Average video bitrate in KBit/s
153 * vcodec Name of the video codec in use
154 * container Name of the container format
155 * filesize The number of bytes, if known in advance
156 * filesize_approx An estimate for the number of bytes
157 * player_url SWF Player URL (used for rtmpdump).
158 * protocol The protocol that will be used for the actual
159 download, lower-case.
160 "http", "https", "rtsp", "rtmp", "rtmpe",
161 "m3u8", "m3u8_native" or "http_dash_segments".
163 Base URL for fragments. Each fragment's path
164 value (if present) will be relative to
166 * fragments A list of fragments of a fragmented media.
167 Each fragment entry must contain either an url
168 or a path. If an url is present it should be
169 considered by a client. Otherwise both path and
170 fragment_base_url must be present. Here is
171 the list of all potential fields:
172 * "url" - fragment's URL
173 * "path" - fragment's path relative to
175 * "duration" (optional, int or float)
176 * "filesize" (optional, int)
177 * preference Order number of this format. If this field is
178 present and not None, the formats get sorted
179 by this field, regardless of all other values.
180 -1 for default (order by other properties),
181 -2 or smaller for less than default.
182 < -1000 to hide the format (if there is
183 another one which is strictly better)
184 * language Language code, e.g. "de" or "en-US".
185 * language_preference Is this in the language mentioned in
187 10 if it's what the URL is about,
188 -1 for default (don't know),
189 -10 otherwise, other values reserved for now.
190 * quality Order number of the video quality of this
191 format, irrespective of the file format.
192 -1 for default (order by other properties),
193 -2 or smaller for less than default.
194 * source_preference Order number for this video source
195 (quality takes higher priority)
196 -1 for default (order by other properties),
197 -2 or smaller for less than default.
198 * http_headers A dictionary of additional HTTP headers
199 to add to the request.
200 * stretched_ratio If given and not 1, indicates that the
201 video's pixels are not square.
202 width : height ratio as float.
203 * no_resume The server does not support resuming the
204 (HTTP or RTMP) download. Boolean.
205 * downloader_options A dictionary of downloader options as
206 described in FileDownloader
208 url: Final video URL.
209 ext: Video filename extension.
210 format: The video format, defaults to ext (used for --get-format)
211 player_url: SWF Player URL (used for rtmpdump).
213 The following fields are optional:
215 alt_title: A secondary title of the video.
216 display_id An alternative identifier for the video, not necessarily
217 unique, but available before title. Typically, id is
218 something like "4234987", title "Dancing naked mole rats",
219 and display_id "dancing-naked-mole-rats"
220 thumbnails: A list of dictionaries, with the following entries:
221 * "id" (optional, string) - Thumbnail format ID
223 * "preference" (optional, int) - quality of the image
224 * "width" (optional, int)
225 * "height" (optional, int)
226 * "resolution" (optional, string "{width}x{height}",
228 * "filesize" (optional, int)
229 thumbnail: Full URL to a video thumbnail image.
230 description: Full video description.
231 uploader: Full name of the video uploader.
232 license: License name the video is licensed under.
233 creator: The creator of the video.
234 release_date: The date (YYYYMMDD) when the video was released.
235 timestamp: UNIX timestamp of the moment the video became available.
236 upload_date: Video upload date (YYYYMMDD).
237 If not explicitly set, calculated from timestamp.
238 uploader_id: Nickname or id of the video uploader.
239 uploader_url: Full URL to a personal webpage of the video uploader.
240 channel: Full name of the channel the video is uploaded on.
241 Note that channel fields may or may not repeat uploader
242 fields. This depends on a particular extractor.
243 channel_id: Id of the channel.
244 channel_url: Full URL to a channel webpage.
245 location: Physical location where the video was filmed.
246 subtitles: The available subtitles as a dictionary in the format
247 {tag: subformats}. "tag" is usually a language code, and
248 "subformats" is a list sorted from lower to higher
249 preference, each element is a dictionary with the "ext"
251 * "data": The subtitles file contents
252 * "url": A URL pointing to the subtitles file
253 "ext" will be calculated from URL if missing
254 automatic_captions: Like 'subtitles', used by the YoutubeIE for
255 automatically generated captions
256 duration: Length of the video in seconds, as an integer or float.
257 view_count: How many users have watched the video on the platform.
258 like_count: Number of positive ratings of the video
259 dislike_count: Number of negative ratings of the video
260 repost_count: Number of reposts of the video
261 average_rating: Average rating give by users, the scale used depends on the webpage
262 comment_count: Number of comments on the video
263 comments: A list of comments, each with one or more of the following
264 properties (all but one of text or html optional):
265 * "author" - human-readable name of the comment author
266 * "author_id" - user ID of the comment author
268 * "html" - Comment as HTML
269 * "text" - Plain text of the comment
270 * "timestamp" - UNIX timestamp of comment
271 * "parent" - ID of the comment this one is replying to.
272 Set to "root" to indicate that this is a
273 comment to the original video.
274 age_limit: Age restriction for the video, as an integer (years)
275 webpage_url: The URL to the video webpage, if given to youtube-dlc it
276 should allow to get the same result again. (It will be set
277 by YoutubeDL if it's missing)
278 categories: A list of categories that the video falls in, for example
280 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
281 is_live: True, False, or None (=unknown). Whether this video is a
282 live stream that goes on instead of a fixed-length video.
283 start_time: Time in seconds where the reproduction should start, as
284 specified in the URL.
285 end_time: Time in seconds where the reproduction should end, as
286 specified in the URL.
287 chapters: A list of dictionaries, with the following entries:
288 * "start_time" - The start time of the chapter in seconds
289 * "end_time" - The end time of the chapter in seconds
290 * "title" (optional, string)
292 The following fields should only be used when the video belongs to some logical
295 chapter: Name or title of the chapter the video belongs to.
296 chapter_number: Number of the chapter the video belongs to, as an integer.
297 chapter_id: Id of the chapter the video belongs to, as a unicode string.
299 The following fields should only be used when the video is an episode of some
300 series, programme or podcast:
302 series: Title of the series or programme the video episode belongs to.
303 season: Title of the season the video episode belongs to.
304 season_number: Number of the season the video episode belongs to, as an integer.
305 season_id: Id of the season the video episode belongs to, as a unicode string.
306 episode: Title of the video episode. Unlike mandatory video title field,
307 this field should denote the exact title of the video episode
308 without any kind of decoration.
309 episode_number: Number of the video episode within a season, as an integer.
310 episode_id: Id of the video episode, as a unicode string.
312 The following fields should only be used when the media is a track or a part of
315 track: Title of the track.
316 track_number: Number of the track within an album or a disc, as an integer.
317 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
319 artist: Artist(s) of the track.
320 genre: Genre(s) of the track.
321 album: Title of the album the track belongs to.
322 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
323 album_artist: List of all artists appeared on the album (e.g.
324 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
326 disc_number: Number of the disc or other physical medium the track belongs to,
328 release_year: Year (YYYY) when the album was released.
330 Unless mentioned otherwise, the fields should be Unicode strings.
332 Unless mentioned otherwise, None is equivalent to absence of information.
335 _type "playlist" indicates multiple videos.
336 There must be a key "entries", which is a list, an iterable, or a PagedList
337 object, each element of which is a valid dictionary by this specification.
339 Additionally, playlists can have "id", "title", and any other relevent
340 attributes with the same semantics as videos (see above).
343 _type "multi_video" indicates that there are multiple videos that
344 form a single show, for examples multiple acts of an opera or TV episode.
345 It must have an entries key like a playlist and contain all the keys
346 required for a video at the same time.
349 _type "url" indicates that the video must be extracted from another
350 location, possibly by a different extractor. Its only required key is:
351 "url" - the next URL to extract.
352 The key "ie_key" can be set to the class name (minus the trailing "IE",
353 e.g. "Youtube") if the extractor class is known in advance.
354 Additionally, the dictionary may have any properties of the resolved entity
355 known in advance, for example "title" if the title of the referred video is
359 _type "url_transparent" entities have the same specification as "url", but
360 indicate that the given additional information is more precise than the one
361 associated with the resolved URL.
362 This is useful when a site employs a video service that hosts the video and
363 its technical metadata, but that video service does not embed a useful
364 title, description etc.
367 Subclasses of this one should re-define the _real_initialize() and
368 _real_extract() methods and define a _VALID_URL regexp.
369 Probably, they should also be added to the list of extractors.
371 _GEO_BYPASS attribute may be set to False in order to disable
372 geo restriction bypass mechanisms for a particular extractor.
373 Though it won't disable explicit geo restriction bypass based on
374 country code provided with geo_bypass_country.
376 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
377 countries for this extractor. One of these countries will be used by
378 geo restriction bypass mechanism right away in order to bypass
379 geo restriction, of course, if the mechanism is not disabled.
381 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
382 IP blocks in CIDR notation for this extractor. One of these IP blocks
383 will be used by geo restriction bypass mechanism similarly
386 Finally, the _WORKING attribute should be set to False for broken IEs
387 in order to warn the users and skip the tests.
392 _x_forwarded_for_ip
= None
394 _GEO_COUNTRIES
= None
395 _GEO_IP_BLOCKS
= None
398 def __init__(self
, downloader
=None):
399 """Constructor. Receives an optional downloader."""
401 self
._x
_forwarded
_for
_ip
= None
402 self
.set_downloader(downloader
)
405 def suitable(cls
, url
):
406 """Receives a URL and returns True if suitable for this IE."""
408 # This does not use has/getattr intentionally - we want to know whether
409 # we have cached the regexp for *this* class, whereas getattr would also
410 # match the superclass
411 if '_VALID_URL_RE' not in cls
.__dict
__:
412 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
413 return cls
._VALID
_URL
_RE
.match(url
) is not None
416 def _match_id(cls
, url
):
417 if '_VALID_URL_RE' not in cls
.__dict
__:
418 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
419 m
= cls
._VALID
_URL
_RE
.match(url
)
421 return compat_str(m
.group('id'))
425 """Getter method for _WORKING."""
428 def initialize(self
):
429 """Initializes an instance (authentication, etc)."""
430 self
._initialize
_geo
_bypass
({
431 'countries': self
._GEO
_COUNTRIES
,
432 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
435 self
._real
_initialize
()
438 def _initialize_geo_bypass(self
, geo_bypass_context
):
440 Initialize geo restriction bypass mechanism.
442 This method is used to initialize geo bypass mechanism based on faking
443 X-Forwarded-For HTTP header. A random country from provided country list
444 is selected and a random IP belonging to this country is generated. This
445 IP will be passed as X-Forwarded-For HTTP header in all subsequent
448 This method will be used for initial geo bypass mechanism initialization
449 during the instance initialization with _GEO_COUNTRIES and
452 You may also manually call it from extractor's code if geo bypass
453 information is not available beforehand (e.g. obtained during
454 extraction) or due to some other reason. In this case you should pass
455 this information in geo bypass context passed as first argument. It may
456 contain following fields:
458 countries: List of geo unrestricted countries (similar
460 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
461 (similar to _GEO_IP_BLOCKS)
464 if not self
._x
_forwarded
_for
_ip
:
466 # Geo bypass mechanism is explicitly disabled by user
467 if not self
._downloader
.params
.get('geo_bypass', True):
470 if not geo_bypass_context
:
471 geo_bypass_context
= {}
473 # Backward compatibility: previously _initialize_geo_bypass
474 # expected a list of countries, some 3rd party code may still use
476 if isinstance(geo_bypass_context
, (list, tuple)):
477 geo_bypass_context
= {
478 'countries': geo_bypass_context
,
481 # The whole point of geo bypass mechanism is to fake IP
482 # as X-Forwarded-For HTTP header based on some IP block or
485 # Path 1: bypassing based on IP block in CIDR notation
487 # Explicit IP block specified by user, use it right away
488 # regardless of whether extractor is geo bypassable or not
489 ip_block
= self
._downloader
.params
.get('geo_bypass_ip_block', None)
491 # Otherwise use random IP block from geo bypass context but only
492 # if extractor is known as geo bypassable
494 ip_blocks
= geo_bypass_context
.get('ip_blocks')
495 if self
._GEO
_BYPASS
and ip_blocks
:
496 ip_block
= random
.choice(ip_blocks
)
499 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
500 if self
._downloader
.params
.get('verbose', False):
501 self
._downloader
.to_screen(
502 '[debug] Using fake IP %s as X-Forwarded-For.'
503 % self
._x
_forwarded
_for
_ip
)
506 # Path 2: bypassing based on country code
508 # Explicit country code specified by user, use it right away
509 # regardless of whether extractor is geo bypassable or not
510 country
= self
._downloader
.params
.get('geo_bypass_country', None)
512 # Otherwise use random country code from geo bypass context but
513 # only if extractor is known as geo bypassable
515 countries
= geo_bypass_context
.get('countries')
516 if self
._GEO
_BYPASS
and countries
:
517 country
= random
.choice(countries
)
520 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
521 if self
._downloader
.params
.get('verbose', False):
522 self
._downloader
.to_screen(
523 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
524 % (self
._x
_forwarded
_for
_ip
, country
.upper()))
526 def extract(self
, url
):
527 """Extracts URL information and returns it in list of dicts."""
532 ie_result
= self
._real
_extract
(url
)
533 if self
._x
_forwarded
_for
_ip
:
534 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
536 except GeoRestrictedError
as e
:
537 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
540 except ExtractorError
:
542 except compat_http_client
.IncompleteRead
as e
:
543 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
544 except (KeyError, StopIteration) as e
:
545 raise ExtractorError('An extractor error has occurred.', cause
=e
)
547 def __maybe_fake_ip_and_retry(self
, countries
):
548 if (not self
._downloader
.params
.get('geo_bypass_country', None)
550 and self
._downloader
.params
.get('geo_bypass', True)
551 and not self
._x
_forwarded
_for
_ip
553 country_code
= random
.choice(countries
)
554 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
555 if self
._x
_forwarded
_for
_ip
:
557 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
558 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
562 def set_downloader(self
, downloader
):
563 """Sets the downloader for this IE."""
564 self
._downloader
= downloader
566 def _real_initialize(self
):
567 """Real initialization process. Redefine in subclasses."""
570 def _real_extract(self
, url
):
571 """Real extraction process. Redefine in subclasses."""
576 """A string for getting the InfoExtractor with get_info_extractor"""
577 return compat_str(cls
.__name
__[:-2])
581 return compat_str(type(self
).__name
__[:-2])
584 def __can_accept_status_code(err
, expected_status
):
585 assert isinstance(err
, compat_urllib_error
.HTTPError
)
586 if expected_status
is None:
588 if isinstance(expected_status
, compat_integer_types
):
589 return err
.code
== expected_status
590 elif isinstance(expected_status
, (list, tuple)):
591 return err
.code
in expected_status
592 elif callable(expected_status
):
593 return expected_status(err
.code
) is True
597 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
599 Return the response handle.
601 See _download_webpage docstring for arguments specification.
604 self
.report_download_webpage(video_id
)
605 elif note
is not False:
607 self
.to_screen('%s' % (note
,))
609 self
.to_screen('%s: %s' % (video_id
, note
))
611 # Some sites check X-Forwarded-For HTTP header in order to figure out
612 # the origin of the client behind proxy. This allows bypassing geo
613 # restriction by faking this header's value to IP that belongs to some
614 # geo unrestricted country. We will do so once we encounter any
615 # geo restriction error.
616 if self
._x
_forwarded
_for
_ip
:
617 if 'X-Forwarded-For' not in headers
:
618 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
620 if isinstance(url_or_request
, compat_urllib_request
.Request
):
621 url_or_request
= update_Request(
622 url_or_request
, data
=data
, headers
=headers
, query
=query
)
625 url_or_request
= update_url_query(url_or_request
, query
)
626 if data
is not None or headers
:
627 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
628 exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
629 if hasattr(ssl
, 'CertificateError'):
630 exceptions
.append(ssl
.CertificateError
)
632 return self
._downloader
.urlopen(url_or_request
)
633 except tuple(exceptions
) as err
:
634 if isinstance(err
, compat_urllib_error
.HTTPError
):
635 if self
.__can
_accept
_status
_code
(err
, expected_status
):
636 # Retain reference to error to prevent file object from
637 # being closed before it can be read. Works around the
638 # effects of <https://bugs.python.org/issue15002>
639 # introduced in Python 3.4.1.
646 errnote
= 'Unable to download webpage'
648 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
650 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
652 self
._downloader
.report_warning(errmsg
)
655 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
657 Return a tuple (page content as string, URL handle).
659 See _download_webpage docstring for arguments specification.
661 # Strip hashes from the URL (#1038)
662 if isinstance(url_or_request
, (compat_str
, str)):
663 url_or_request
= url_or_request
.partition('#')[0]
665 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
669 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
670 return (content
, urlh
)
673 def _guess_encoding_from_content(content_type
, webpage_bytes
):
674 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
676 encoding
= m
.group(1)
678 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
679 webpage_bytes[:1024])
681 encoding = m.group(1).decode('ascii')
682 elif webpage_bytes.startswith(b'\xff\xfe'):
689 def __check_blocked(self, content):
690 first_block = content[:512]
691 if ('<title>Access to this site is blocked</title>' in content
692 and 'Websense' in first_block):
693 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
694 blocked_iframe = self._html_search_regex(
695 r'<iframe src="([^
"]+)"', content,
696 'Websense information URL
', default=None)
698 msg += ' Visit
%s for more details
' % blocked_iframe
699 raise ExtractorError(msg, expected=True)
700 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
702 'Access to this webpage has been blocked by Indian censorship
. '
703 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
704 block_msg = self._html_search_regex(
705 r'</h1
><p
>(.*?
)</p
>',
706 content, 'block message
', default=None)
708 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
709 raise ExtractorError(msg, expected=True)
710 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
711 and 'blocklist
.rkn
.gov
.ru
' in content):
712 raise ExtractorError(
713 'Access to this webpage has been blocked by decision of the Russian government
. '
714 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
717 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
718 content_type = urlh.headers.get('Content
-Type
', '')
719 webpage_bytes = urlh.read()
720 if prefix is not None:
721 webpage_bytes = prefix + webpage_bytes
723 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
724 if self._downloader.params.get('dump_intermediate_pages
', False):
725 self.to_screen('Dumping request to
' + urlh.geturl())
726 dump = base64.b64encode(webpage_bytes).decode('ascii
')
727 self._downloader.to_screen(dump)
728 if self._downloader.params.get('write_pages
', False):
729 basen = '%s_%s' % (video_id, urlh.geturl())
731 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
732 basen = basen[:240 - len(h)] + h
733 raw_filename = basen + '.dump
'
734 filename = sanitize_filename(raw_filename, restricted=True)
735 self.to_screen('Saving request to
' + filename)
736 # Working around MAX_PATH limitation on Windows (see
737 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
738 if compat_os_name == 'nt
':
739 absfilepath = os.path.abspath(filename)
740 if len(absfilepath) > 259:
741 filename = '\\\\?
\\' + absfilepath
742 with open(filename, 'wb
') as outf:
743 outf.write(webpage_bytes)
746 content = webpage_bytes.decode(encoding, 'replace
')
748 content = webpage_bytes.decode('utf
-8', 'replace
')
750 self.__check_blocked(content)
754 def _download_webpage(
755 self, url_or_request, video_id, note=None, errnote=None,
756 fatal=True, tries=1, timeout=5, encoding=None, data=None,
757 headers={}, query={}, expected_status=None):
759 Return the data of the page as a string.
762 url_or_request -- plain text URL as a string or
763 a compat_urllib_request.Requestobject
764 video_id -- Video/playlist/item identifier (string)
767 note -- note printed before downloading (string)
768 errnote -- note printed in case of an error (string)
769 fatal -- flag denoting whether error should be considered fatal,
770 i.e. whether it should cause ExtractionError to be raised,
771 otherwise a warning will be reported and extraction continued
772 tries -- number of tries
773 timeout -- sleep interval between tries
774 encoding -- encoding for a page content decoding, guessed automatically
775 when not explicitly specified
776 data -- POST data (bytes)
777 headers -- HTTP headers (dict)
778 query -- URL query (dict)
779 expected_status -- allows to accept failed HTTP requests (non 2xx
780 status code) by explicitly specifying a set of accepted status
781 codes. Can be any of the following entities:
782 - an integer type specifying an exact failed status code to
784 - a list or a tuple of integer types specifying a list of
785 failed status codes to accept
786 - a callable accepting an actual failed status code and
787 returning True if it should be accepted
788 Note that this argument does not affect success status codes (2xx)
789 which are always accepted.
794 while success is False:
796 res = self._download_webpage_handle(
797 url_or_request, video_id, note, errnote, fatal,
798 encoding=encoding, data=data, headers=headers, query=query,
799 expected_status=expected_status)
801 except compat_http_client.IncompleteRead as e:
803 if try_count >= tries:
805 self._sleep(timeout, video_id)
812 def _download_xml_handle(
813 self, url_or_request, video_id, note='Downloading XML
',
814 errnote='Unable to download XML
', transform_source=None,
815 fatal=True, encoding=None, data=None, headers={}, query={},
816 expected_status=None):
818 Return a tuple (xml as an compat_etree_Element, URL handle).
820 See _download_webpage docstring for arguments specification.
822 res = self._download_webpage_handle(
823 url_or_request, video_id, note, errnote, fatal=fatal,
824 encoding=encoding, data=data, headers=headers, query=query,
825 expected_status=expected_status)
828 xml_string, urlh = res
829 return self._parse_xml(
830 xml_string, video_id, transform_source=transform_source,
834 self, url_or_request, video_id,
835 note='Downloading XML
', errnote='Unable to download XML
',
836 transform_source=None, fatal=True, encoding=None,
837 data=None, headers={}, query={}, expected_status=None):
839 Return the xml as an compat_etree_Element.
841 See _download_webpage docstring for arguments specification.
843 res = self._download_xml_handle(
844 url_or_request, video_id, note=note, errnote=errnote,
845 transform_source=transform_source, fatal=fatal, encoding=encoding,
846 data=data, headers=headers, query=query,
847 expected_status=expected_status)
848 return res if res is False else res[0]
850 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
852 xml_string = transform_source(xml_string)
854 return compat_etree_fromstring(xml_string.encode('utf
-8'))
855 except compat_xml_parse_error as ve:
856 errmsg = '%s: Failed to parse XML
' % video_id
858 raise ExtractorError(errmsg, cause=ve)
860 self.report_warning(errmsg + str(ve))
862 def _download_json_handle(
863 self, url_or_request, video_id, note='Downloading JSON metadata
',
864 errnote='Unable to download JSON metadata
', transform_source=None,
865 fatal=True, encoding=None, data=None, headers={}, query={},
866 expected_status=None):
868 Return a tuple (JSON object, URL handle).
870 See _download_webpage docstring for arguments specification.
872 res = self._download_webpage_handle(
873 url_or_request, video_id, note, errnote, fatal=fatal,
874 encoding=encoding, data=data, headers=headers, query=query,
875 expected_status=expected_status)
878 json_string, urlh = res
879 return self._parse_json(
880 json_string, video_id, transform_source=transform_source,
884 self, url_or_request, video_id, note='Downloading JSON metadata
',
885 errnote='Unable to download JSON metadata
', transform_source=None,
886 fatal=True, encoding=None, data=None, headers={}, query={},
887 expected_status=None):
889 Return the JSON object as a dict.
891 See _download_webpage docstring for arguments specification.
893 res = self._download_json_handle(
894 url_or_request, video_id, note=note, errnote=errnote,
895 transform_source=transform_source, fatal=fatal, encoding=encoding,
896 data=data, headers=headers, query=query,
897 expected_status=expected_status)
898 return res if res is False else res[0]
900 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
902 json_string = transform_source(json_string)
904 return json.loads(json_string)
905 except ValueError as ve:
906 errmsg = '%s: Failed to parse JSON
' % video_id
908 raise ExtractorError(errmsg, cause=ve)
910 self.report_warning(errmsg + str(ve))
912 def report_warning(self, msg, video_id=None):
913 idstr = '' if video_id is None else '%s: ' % video_id
914 self._downloader.report_warning(
915 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
917 def to_screen(self, msg):
918 """Print msg to screen, prefixing it with '[ie_name
]'"""
919 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
921 def report_extraction(self, id_or_name):
922 """Report information extraction."""
923 self.to_screen('%s: Extracting information
' % id_or_name)
925 def report_download_webpage(self, video_id):
926 """Report webpage download."""
927 self.to_screen('%s: Downloading webpage
' % video_id)
929 def report_age_confirmation(self):
930 """Report attempt to confirm age."""
931 self.to_screen('Confirming age
')
933 def report_login(self):
934 """Report attempt to log in."""
935 self.to_screen('Logging
in')
938 def raise_login_required(msg='This video
is only available
for registered users
'):
939 raise ExtractorError(
940 '%s. Use
--username
and --password
or --netrc to provide account credentials
.' % msg,
944 def raise_geo_restricted(msg='This video
is not available
from your location due to geo restriction
', countries=None):
945 raise GeoRestrictedError(msg, countries=countries)
947 # Methods for following #608
949 def url_result(url, ie=None, video_id=None, video_title=None):
950 """Returns a URL that points to a page that should be processed"""
951 # TODO: ie should be the class used for getting the info
952 video_info = {'_type
': 'url
',
955 if video_id is not None:
956 video_info['id'] = video_id
957 if video_title is not None:
958 video_info['title
'] = video_title
961 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
963 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
965 return self.playlist_result(
966 urls, playlist_id=playlist_id, playlist_title=playlist_title)
969 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
970 """Returns a playlist"""
971 video_info = {'_type
': 'playlist
',
973 video_info.update(kwargs)
975 video_info['id'] = playlist_id
977 video_info['title
'] = playlist_title
978 if playlist_description is not None:
979 video_info['description
'] = playlist_description
982 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
984 Perform a regex search on the given string, using a single or a list of
985 patterns returning the first matching group.
986 In case of failure return a default value or raise a WARNING or a
987 RegexNotFoundError, depending on fatal, specifying the field name.
989 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
990 mobj = re.search(pattern, string, flags)
993 mobj = re.search(p, string, flags)
997 if not self._downloader.params.get('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
998 _name = '\033[0;34m
%s\033[0m
' % name
1004 # return the first matching group
1005 return next(g for g in mobj.groups() if g is not None)
1007 return mobj.group(group)
1008 elif default is not NO_DEFAULT:
1011 raise RegexNotFoundError('Unable to extract
%s' % _name)
1013 self._downloader.report_warning('unable to extract
%s' % _name + bug_reports_message())
1016 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1018 Like _search_regex, but strips HTML tags and unescapes entities.
1020 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1022 return clean_html(res).strip()
1026 def _get_netrc_login_info(self, netrc_machine=None):
1029 netrc_machine = netrc_machine or self._NETRC_MACHINE
1031 if self._downloader.params.get('usenetrc
', False):
1033 info = netrc.netrc().authenticators(netrc_machine)
1034 if info is not None:
1038 raise netrc.NetrcParseError(
1039 'No authenticators
for %s' % netrc_machine)
1040 except (IOError, netrc.NetrcParseError) as err:
1041 self._downloader.report_warning(
1042 'parsing
.netrc
: %s' % error_to_compat_str(err))
1044 return username, password
1046 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1048 Get the login info as (username, password)
1049 First look for the manually specified credentials using username_option
1050 and password_option as keys in params dictionary. If no such credentials
1051 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1053 If there's no info available
, return (None, None)
1055 if self._downloader is None:
1058 downloader_params = self._downloader.params
1060 # Attempt to use provided username and password or .netrc data
1061 if downloader_params.get(username_option) is not None:
1062 username = downloader_params[username_option]
1063 password = downloader_params[password_option]
1065 username, password = self._get_netrc_login_info(netrc_machine)
1067 return username, password
1069 def _get_tfa_info(self, note='two-factor verification code'):
1071 Get the two
-factor authentication info
1072 TODO
- asking the user will be required
for sms
/phone verify
1073 currently just uses the command line option
1074 If there
's no info available, return None
1076 if self._downloader is None:
1078 downloader_params = self._downloader.params
1080 if downloader_params.get('twofactor
') is not None:
1081 return downloader_params['twofactor
']
1083 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1085 # Helper functions for extracting OpenGraph info
1087 def _og_regexes(prop):
1088 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1089 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1090 % {'prop': re.escape(prop)})
1091 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1093 template % (property_re, content_re),
1094 template % (content_re, property_re),
1098 def _meta_regex(prop):
1099 return r'''(?isx)<meta
1100 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1101 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1103 def _og_search_property(self, prop, html, name=None, **kargs):
1104 if not isinstance(prop, (list, tuple)):
1107 name = 'OpenGraph
%s' % prop[0]
1110 og_regexes.extend(self._og_regexes(p))
1111 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1114 return unescapeHTML(escaped)
1116 def _og_search_thumbnail(self, html, **kargs):
1117 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1119 def _og_search_description(self, html, **kargs):
1120 return self._og_search_property('description
', html, fatal=False, **kargs)
1122 def _og_search_title(self, html, **kargs):
1123 return self._og_search_property('title
', html, **kargs)
1125 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1126 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1128 regexes = self._og_regexes('video
:secure_url
') + regexes
1129 return self._html_search_regex(regexes, html, name, **kargs)
1131 def _og_search_url(self, html, **kargs):
1132 return self._og_search_property('url
', html, **kargs)
1134 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1135 if not isinstance(name, (list, tuple)):
1137 if display_name is None:
1138 display_name = name[0]
1139 return self._html_search_regex(
1140 [self._meta_regex(n) for n in name],
1141 html, display_name, fatal=fatal, group='content
', **kwargs)
1143 def _dc_search_uploader(self, html):
1144 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1146 def _rta_search(self, html):
1147 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1148 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1149 r' content
="RTA-5042-1996-1400-1577-RTA"',
1154 def _media_rating_search(self, html):
1155 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1156 rating = self._html_search_meta('rating
', html)
1168 return RATING_TABLE.get(rating.lower())
1170 def _family_friendly_search(self, html):
1171 # See http://schema.org/VideoObject
1172 family_friendly = self._html_search_meta(
1173 'isFamilyFriendly
', html, default=None)
1175 if not family_friendly:
1184 return RATING_TABLE.get(family_friendly.lower())
1186 def _twitter_search_player(self, html):
1187 return self._html_search_meta('twitter
:player
', html,
1188 'twitter card player
')
1190 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1191 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1192 default = kwargs.get('default
', NO_DEFAULT)
1193 # JSON-LD may be malformed and thus `fatal` should be respected.
1194 # At the same time `default` may be passed that assumes `fatal=False`
1195 # for _search_regex. Let's simulate the same behavior here
as well
.
1196 fatal
= kwargs
.get('fatal', True) if default
== NO_DEFAULT
else False
1198 for mobj
in json_ld_list
:
1199 json_ld_item
= self
._parse
_json
(
1200 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1201 if not json_ld_item
:
1203 if isinstance(json_ld_item
, dict):
1204 json_ld
.append(json_ld_item
)
1205 elif isinstance(json_ld_item
, (list, tuple)):
1206 json_ld
.extend(json_ld_item
)
1208 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1211 if default
is not NO_DEFAULT
:
1214 raise RegexNotFoundError('Unable to extract JSON-LD')
1216 self
._downloader
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1219 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1220 if isinstance(json_ld
, compat_str
):
1221 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1225 if not isinstance(json_ld
, (list, tuple, dict)):
1227 if isinstance(json_ld
, dict):
1230 INTERACTION_TYPE_MAP
= {
1231 'CommentAction': 'comment',
1232 'AgreeAction': 'like',
1233 'DisagreeAction': 'dislike',
1234 'LikeAction': 'like',
1235 'DislikeAction': 'dislike',
1236 'ListenAction': 'view',
1237 'WatchAction': 'view',
1238 'ViewAction': 'view',
1241 def extract_interaction_type(e
):
1242 interaction_type
= e
.get('interactionType')
1243 if isinstance(interaction_type
, dict):
1244 interaction_type
= interaction_type
.get('@type')
1245 return str_or_none(interaction_type
)
1247 def extract_interaction_statistic(e
):
1248 interaction_statistic
= e
.get('interactionStatistic')
1249 if isinstance(interaction_statistic
, dict):
1250 interaction_statistic
= [interaction_statistic
]
1251 if not isinstance(interaction_statistic
, list):
1253 for is_e
in interaction_statistic
:
1254 if not isinstance(is_e
, dict):
1256 if is_e
.get('@type') != 'InteractionCounter':
1258 interaction_type
= extract_interaction_type(is_e
)
1259 if not interaction_type
:
1261 # For interaction count some sites provide string instead of
1262 # an integer (as per spec) with non digit characters (e.g. ",")
1263 # so extracting count with more relaxed str_to_int
1264 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1265 if interaction_count
is None:
1267 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1270 count_key
= '%s_count' % count_kind
1271 if info
.get(count_key
) is not None:
1273 info
[count_key
] = interaction_count
1275 def extract_video_object(e
):
1276 assert e
['@type'] == 'VideoObject'
1278 'url': url_or_none(e
.get('contentUrl')),
1279 'title': unescapeHTML(e
.get('name')),
1280 'description': unescapeHTML(e
.get('description')),
1281 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1282 'duration': parse_duration(e
.get('duration')),
1283 'timestamp': unified_timestamp(e
.get('uploadDate')),
1284 'uploader': str_or_none(e
.get('author')),
1285 'filesize': float_or_none(e
.get('contentSize')),
1286 'tbr': int_or_none(e
.get('bitrate')),
1287 'width': int_or_none(e
.get('width')),
1288 'height': int_or_none(e
.get('height')),
1289 'view_count': int_or_none(e
.get('interactionCount')),
1291 extract_interaction_statistic(e
)
1295 item_type
= e
.get('@type')
1296 if expected_type
is not None and expected_type
!= item_type
:
1298 if item_type
in ('TVEpisode', 'Episode'):
1299 episode_name
= unescapeHTML(e
.get('name'))
1301 'episode': episode_name
,
1302 'episode_number': int_or_none(e
.get('episodeNumber')),
1303 'description': unescapeHTML(e
.get('description')),
1305 if not info
.get('title') and episode_name
:
1306 info
['title'] = episode_name
1307 part_of_season
= e
.get('partOfSeason')
1308 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1310 'season': unescapeHTML(part_of_season
.get('name')),
1311 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1313 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1314 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1315 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1316 elif item_type
== 'Movie':
1318 'title': unescapeHTML(e
.get('name')),
1319 'description': unescapeHTML(e
.get('description')),
1320 'duration': parse_duration(e
.get('duration')),
1321 'timestamp': unified_timestamp(e
.get('dateCreated')),
1323 elif item_type
in ('Article', 'NewsArticle'):
1325 'timestamp': parse_iso8601(e
.get('datePublished')),
1326 'title': unescapeHTML(e
.get('headline')),
1327 'description': unescapeHTML(e
.get('articleBody')),
1329 elif item_type
== 'VideoObject':
1330 extract_video_object(e
)
1331 if expected_type
is None:
1335 video
= e
.get('video')
1336 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1337 extract_video_object(video
)
1338 if expected_type
is None:
1342 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1345 def _hidden_inputs(html
):
1346 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1348 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1349 attrs
= extract_attributes(input)
1352 if attrs
.get('type') not in ('hidden', 'submit'):
1354 name
= attrs
.get('name') or attrs
.get('id')
1355 value
= attrs
.get('value')
1356 if name
and value
is not None:
1357 hidden_inputs
[name
] = value
1358 return hidden_inputs
1360 def _form_hidden_inputs(self
, form_id
, html
):
1361 form
= self
._search
_regex
(
1362 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1363 html, '%s form
' % form_id, group='form
')
1364 return self._hidden_inputs(form)
1367 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<seperator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1369 default = ('hidden
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1370 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1371 'proto
', 'ext
', 'has_audio
', 'source
', 'format_id
') # These must not be aliases
1374 'vcodec
': {'type': 'ordered
', 'regex
': True,
1375 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1376 'acodec
': {'type': 'ordered
', 'regex
': True,
1377 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1378 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1379 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', 'm3u8
', '.*dash
', '', 'mms|rtsp
', 'none
', 'f4
']},
1380 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1381 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1382 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1383 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1384 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1385 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1386 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1387 'ie_pref
': {'priority': True, 'type': 'extractor', 'field': 'extractor_preference'},
1388 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1389 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1390 'lang
': {'priority': True, 'convert': 'ignore', 'field': 'language_preference'},
1391 'quality
': {'convert': 'float_none'},
1392 'filesize
': {'convert': 'bytes'},
1393 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1394 'id': {'convert': 'string', 'field': 'format_id'},
1395 'height
': {'convert': 'float_none'},
1396 'width
': {'convert': 'float_none'},
1397 'fps
': {'convert': 'float_none'},
1398 'tbr
': {'convert': 'float_none'},
1399 'vbr
': {'convert': 'float_none'},
1400 'abr
': {'convert': 'float_none'},
1401 'asr
': {'convert': 'float_none'},
1402 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1404 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1405 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1406 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1407 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1408 'res
': {'type': 'multiple', 'field': ('height', 'width'), 'function': min},
1410 # Most of these exist only for compatibility reasons
1411 'dimension
': {'type': 'alias', 'field': 'res'},
1412 'resolution
': {'type': 'alias', 'field': 'res'},
1413 'extension
': {'type': 'alias', 'field': 'ext'},
1414 'bitrate
': {'type': 'alias', 'field': 'br'},
1415 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1416 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1417 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1418 'framerate
': {'type': 'alias', 'field': 'fps'},
1419 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1420 'protocol
': {'type': 'alias', 'field': 'proto'},
1421 'source_preference
': {'type': 'alias', 'field': 'source'},
1422 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1423 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1424 'samplerate
': {'type': 'alias', 'field': 'asr'},
1425 'video_ext
': {'type': 'alias', 'field': 'vext'},
1426 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1427 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1428 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1429 'video
': {'type': 'alias', 'field': 'hasvid'},
1430 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1431 'audio
': {'type': 'alias', 'field': 'hasaud'},
1432 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1433 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1434 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1435 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1436 'format_id
': {'type': 'alias', 'field': 'id'},
1441 def _get_field_setting(self, field, key):
1442 if field not in self.settings:
1443 self.settings[field] = {}
1444 propObj = self.settings[field]
1445 if key not in propObj:
1446 type = propObj.get('type')
1448 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1449 elif key == 'convert
':
1450 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1452 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,), 'function': max}.get(key, None)
1453 propObj[key] = default
1456 def _resolve_field_value(self, field, value, convertNone=False):
1461 value = value.lower()
1462 conversion = self._get_field_setting(field, 'convert
')
1463 if conversion == 'ignore
':
1465 if conversion == 'string
':
1467 elif conversion == 'float_none
':
1468 return float_or_none(value)
1469 elif conversion == 'bytes':
1470 return FileDownloader.parse_bytes(value)
1471 elif conversion == 'order
':
1472 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1473 use_regex = self._get_field_setting(field, 'regex
')
1474 list_length = len(order_list)
1475 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1476 if use_regex and value is not None:
1477 for i, regex in enumerate(order_list):
1478 if regex and re.match(regex, value):
1479 return list_length - i
1480 return list_length - empty_pos # not in list
1481 else: # not regex or value = None
1482 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1484 if value.isnumeric():
1487 self.settings[field]['convert
'] = 'string
'
1490 def evaluate_params(self, params, sort_extractor):
1491 self._use_free_order = params.get('prefer_free_formats
', False)
1492 self._sort_user = params.get('format_sort
', [])
1493 self._sort_extractor = sort_extractor
1495 def add_item(field, reverse, closest, limit_text):
1496 field = field.lower()
1497 if field in self._order:
1499 self._order.append(field)
1500 limit = self._resolve_field_value(field, limit_text)
1503 'closest
': False if limit is None else closest,
1504 'limit_text
': limit_text,
1506 if field in self.settings:
1507 self.settings[field].update(data)
1509 self.settings[field] = data
1512 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1513 + (tuple() if params.get('format_sort_force
', False)
1514 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1515 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1517 for item in sort_list:
1518 match = re.match(self.regex, item)
1520 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1521 field = match.group('field
')
1524 if self._get_field_setting(field, 'type') == 'alias
':
1525 field = self._get_field_setting(field, 'field
')
1526 reverse = match.group('reverse
') is not None
1527 closest = match.group('seperator
') == '~
'
1528 limit_text = match.group('limit
')
1530 has_limit = limit_text is not None
1531 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1532 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1534 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1535 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1536 limit_count = len(limits)
1537 for (i, f) in enumerate(fields):
1538 add_item(f, reverse, closest,
1539 limits[i] if i < limit_count
1540 else limits[0] if has_limit and not has_multiple_limits
1543 def print_verbose_info(self, to_screen):
1544 to_screen('[debug
] Sort order given by user
: %s' % ','.join(self._sort_user))
1545 if self._sort_extractor:
1546 to_screen('[debug
] Sort order given by extractor
: %s' % ','.join(self._sort_extractor))
1547 to_screen('[debug
] Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1548 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1549 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1550 self._get_field_setting(field, 'limit_text
'),
1551 self._get_field_setting(field, 'limit
'))
1552 if self._get_field_setting(field, 'limit_text
') is not None else '')
1553 for field in self._order if self._get_field_setting(field, 'visible
')]))
1555 def _calculate_field_preference_from_value(self, format, field, type, value):
1556 reverse = self._get_field_setting(field, 'reverse
')
1557 closest = self._get_field_setting(field, 'closest
')
1558 limit = self._get_field_setting(field, 'limit
')
1560 if type == 'extractor
':
1561 maximum = self._get_field_setting(field, 'max')
1562 if value is None or (maximum is not None and value >= maximum):
1564 elif type == 'boolean
':
1565 in_list = self._get_field_setting(field, 'in_list
')
1566 not_in_list = self._get_field_setting(field, 'not_in_list
')
1567 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
1568 elif type == 'ordered
':
1569 value = self._resolve_field_value(field, value, True)
1571 # try to convert to number
1572 val_num = float_or_none(value)
1573 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1577 return ((-10, 0) if value is None
1578 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1579 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1580 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1581 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1582 else (-1, value, 0))
1584 def _calculate_field_preference(self, format, field):
1585 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1586 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1587 if type == 'multiple
':
1588 type = 'field
' # Only 'field
' is allowed in multiple for now
1589 actual_fields = self._get_field_setting(field, 'field
')
1591 def wrapped_function(values):
1592 values = tuple(filter(lambda x: x is not None, values))
1593 return (self._get_field_setting(field, 'function
')(*values) if len(values) > 1
1594 else values[0] if values
1597 value = wrapped_function((get_value(f) for f in actual_fields))
1599 value = get_value(field)
1600 return self._calculate_field_preference_from_value(format, field, type, value)
1602 def calculate_preference(self, format):
1603 # Determine missing protocol
1604 if not format.get('protocol
'):
1605 format['protocol
'] = determine_protocol(format)
1607 # Determine missing ext
1608 if not format.get('ext
') and 'url
' in format:
1609 format['ext
'] = determine_ext(format['url
'])
1610 if format.get('vcodec
') == 'none
':
1611 format['audio_ext
'] = format['ext
']
1612 format['video_ext
'] = 'none
'
1614 format['video_ext
'] = format['ext
']
1615 format['audio_ext
'] = 'none
'
1616 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1617 # format['preference
'] = -1000
1619 # Determine missing bitrates
1620 if format.get('tbr
') is None:
1621 if format.get('vbr
') is not None and format.get('abr
') is not None:
1622 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1624 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1625 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1626 if format.get('acodec
') != "none" and format.get('abr
') is None:
1627 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1629 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1631 def _sort_formats(self, formats, field_preference=[]):
1633 raise ExtractorError('No video formats found
')
1634 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1635 format_sort.evaluate_params(self._downloader.params, field_preference)
1636 if self._downloader.params.get('verbose
', False):
1637 format_sort.print_verbose_info(self._downloader.to_screen)
1638 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1640 def _check_formats(self, formats, video_id):
1642 formats[:] = filter(
1643 lambda f: self._is_valid_url(
1645 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1649 def _remove_duplicate_formats(formats):
1653 if f['url
'] not in format_urls:
1654 format_urls.add(f['url
'])
1655 unique_formats.append(f)
1656 formats[:] = unique_formats
1658 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1659 url = self._proto_relative_url(url, scheme='http
:')
1660 # For now assume non HTTP(S) URLs always valid
1661 if not (url.startswith('http
://') or url.startswith('https
://')):
1664 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1666 except ExtractorError as e:
1668 '%s: %s URL
is invalid
, skipping
: %s'
1669 % (video_id, item, error_to_compat_str(e.cause)))
1672 def http_scheme(self):
1673 """ Either "http:" or "https:", depending on the user's preferences
"""
1676 if self._downloader.params.get('prefer_insecure', False)
1679 def _proto_relative_url(self, url, scheme=None):
1682 if url.startswith('//'):
1684 scheme = self.http_scheme()
1689 def _sleep(self, timeout, video_id, msg_template=None):
1690 if msg_template is None:
1691 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1692 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1696 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1697 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1698 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1699 manifest = self._download_xml(
1700 manifest_url, video_id, 'Downloading f4m manifest',
1701 'Unable to download f4m manifest',
1702 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1703 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1704 transform_source=transform_source,
1705 fatal=fatal, data=data, headers=headers, query=query)
1707 if manifest is False:
1710 return self._parse_f4m_formats(
1711 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1712 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1714 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1715 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1716 fatal=True, m3u8_id=None):
1717 if not isinstance(manifest, compat_etree_Element) and not fatal:
1720 # currently youtube-dlc cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1721 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1722 if akamai_pv is not None and ';' in akamai_pv.text:
1723 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1724 if playerVerificationChallenge.strip() != '':
1728 manifest_version = '1.0'
1729 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1731 manifest_version = '2.0'
1732 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1733 # Remove unsupported DRM protected media from final formats
1734 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1735 media_nodes = remove_encrypted_media(media_nodes)
1739 manifest_base_url = get_base_url(manifest)
1741 bootstrap_info = xpath_element(
1742 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1743 'bootstrap info', default=None)
1746 mime_type = xpath_text(
1747 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1748 'base URL', default=None)
1749 if mime_type and mime_type.startswith('audio/'):
1752 for i, media_el in enumerate(media_nodes):
1753 tbr = int_or_none(media_el.attrib.get('bitrate'))
1754 width = int_or_none(media_el.attrib.get('width'))
1755 height = int_or_none(media_el.attrib.get('height'))
1756 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1757 # If <bootstrapInfo> is present, the specified f4m is a
1758 # stream-level manifest, and only set-level manifests may refer to
1759 # external resources. See section 11.4 and section 4 of F4M spec
1760 if bootstrap_info is None:
1762 # @href is introduced in 2.0, see section 11.6 of F4M spec
1763 if manifest_version == '2.0':
1764 media_url = media_el.attrib.get('href')
1765 if media_url is None:
1766 media_url = media_el.attrib.get('url')
1770 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1771 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1772 # If media_url is itself a f4m manifest do the recursive extraction
1773 # since bitrates in parent manifest (this one) and media_url manifest
1774 # may differ leading to inability to resolve the format by requested
1775 # bitrate in f4m downloader
1776 ext = determine_ext(manifest_url)
1778 f4m_formats = self._extract_f4m_formats(
1779 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1780 transform_source=transform_source, fatal=fatal)
1781 # Sometimes stream-level manifest contains single media entry that
1782 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1783 # At the same time parent's media entry in set-level manifest may
1784 # contain it. We will copy it from parent in such cases.
1785 if len(f4m_formats) == 1:
1788 'tbr': f.get('tbr') or tbr,
1789 'width': f.get('width') or width,
1790 'height': f.get('height') or height,
1791 'format_id': f.get('format_id') if not tbr else format_id,
1794 formats.extend(f4m_formats)
1797 formats.extend(self._extract_m3u8_formats(
1798 manifest_url, video_id, 'mp4', preference=preference,
1799 m3u8_id=m3u8_id, fatal=fatal))
1802 'format_id': format_id,
1803 'url': manifest_url,
1804 'manifest_url': manifest_url,
1805 'ext': 'flv' if bootstrap_info is not None else None,
1811 'preference': preference,
1815 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1817 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1821 'preference': preference - 100 if preference else -100,
1822 'resolution': 'multiple',
1823 'format_note': 'Quality selection URL',
1826 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1827 entry_protocol='m3u8', preference=None,
1828 m3u8_id=None, note=None, errnote=None,
1829 fatal=True, live=False, data=None, headers={},
1831 res = self._download_webpage_handle(
1833 note=note or 'Downloading m3u8 information',
1834 errnote=errnote or 'Failed to download m3u8 information',
1835 fatal=fatal, data=data, headers=headers, query=query)
1840 m3u8_doc, urlh = res
1841 m3u8_url = urlh.geturl()
1843 return self._parse_m3u8_formats(
1844 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1845 preference=preference, m3u8_id=m3u8_id, live=live)
1847 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1848 entry_protocol='m3u8', preference=None,
1849 m3u8_id=None, live=False):
1850 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1853 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1858 format_url = lambda u: (
1860 if re.match(r'^https?://', u)
1861 else compat_urlparse.urljoin(m3u8_url, u))
1864 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1865 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1866 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1868 # We should try extracting formats only from master playlists [1, 4.3.4],
1869 # i.e. playlists that describe available qualities. On the other hand
1870 # media playlists [1, 4.3.3] should be returned as is since they contain
1871 # just the media without qualities renditions.
1872 # Fortunately, master playlist can be easily distinguished from media
1873 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1874 # master playlist tags MUST NOT appear in a media playlist and vice versa.
1875 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1876 # media playlist and MUST NOT appear in master playlist thus we can
1877 # clearly detect media playlist with this criterion.
1879 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1882 'format_id': m3u8_id,
1884 'protocol': entry_protocol,
1885 'preference': preference,
1889 last_stream_inf = {}
1891 def extract_media(x_media_line):
1892 media = parse_m3u8_attributes(x_media_line)
1893 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1894 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1895 if not (media_type and group_id and name):
1897 groups.setdefault(group_id, []).append(media)
1898 if media_type not in ('VIDEO', 'AUDIO'):
1900 media_url = media.get('URI')
1903 for v in (m3u8_id, group_id, name):
1907 'format_id': '-'.join(format_id),
1908 'url': format_url(media_url),
1909 'manifest_url': m3u8_url,
1910 'language': media.get('LANGUAGE'),
1912 'protocol': entry_protocol,
1913 'preference': preference,
1915 if media_type == 'AUDIO':
1916 f['vcodec'] = 'none'
1919 def build_stream_name():
1920 # Despite specification does not mention NAME attribute for
1921 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1922 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1923 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1924 stream_name = last_stream_inf.get('NAME')
1927 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1928 # from corresponding rendition group
1929 stream_group_id = last_stream_inf.get('VIDEO')
1930 if not stream_group_id:
1932 stream_group = groups.get(stream_group_id)
1933 if not stream_group:
1934 return stream_group_id
1935 rendition = stream_group[0]
1936 return rendition.get('NAME') or stream_group_id
1938 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1939 # chance to detect video only formats when EXT-X-STREAM-INF tags
1940 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1941 for line in m3u8_doc.splitlines():
1942 if line.startswith('#EXT-X-MEDIA:'):
1945 for line in m3u8_doc.splitlines():
1946 if line.startswith('#EXT-X-STREAM-INF:'):
1947 last_stream_inf = parse_m3u8_attributes(line)
1948 elif line.startswith('#') or not line.strip():
1951 tbr = float_or_none(
1952 last_stream_inf.get('AVERAGE-BANDWIDTH')
1953 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1956 format_id.append(m3u8_id)
1957 stream_name = build_stream_name()
1958 # Bandwidth of live streams may differ over time thus making
1959 # format_id unpredictable. So it's better to keep provided
1962 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1963 manifest_url = format_url(line.strip())
1965 'format_id': '-'.join(format_id),
1966 'url': manifest_url,
1967 'manifest_url': m3u8_url,
1970 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1971 'protocol': entry_protocol,
1972 'preference': preference,
1974 resolution = last_stream_inf.get('RESOLUTION')
1976 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1978 f['width'] = int(mobj.group('width'))
1979 f['height'] = int(mobj.group('height'))
1980 # Unified Streaming Platform
1982 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1984 abr, vbr = mobj.groups()
1985 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1990 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1992 audio_group_id = last_stream_inf.get('AUDIO')
1993 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1994 # references a rendition group MUST have a CODECS attribute.
1995 # However, this is not always respected, for example, [2]
1996 # contains EXT-X-STREAM-INF tag which references AUDIO
1997 # rendition group but does not have CODECS and despite
1998 # referencing an audio group it represents a complete
1999 # (with audio and video) format. So, for such cases we will
2000 # ignore references to rendition groups and treat them
2001 # as complete formats.
2002 if audio_group_id and codecs and f.get('vcodec') != 'none':
2003 audio_group = groups.get(audio_group_id)
2004 if audio_group and audio_group[0].get('URI'):
2005 # TODO: update acodec for audio only formats with
2007 f['acodec'] = 'none'
2011 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2014 del http_f['manifest_url']
2016 'format_id': f['format_id'].replace('hls-', 'http-'),
2018 'url': progressive_uri,
2020 formats.append(http_f)
2022 last_stream_inf = {}
2026 def _xpath_ns(path, namespace=None):
2030 for c in path.split('/'):
2031 if not c or c == '.':
2034 out.append('{%s}%s' % (namespace, c))
2035 return '/'.join(out)
2037 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2038 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2044 namespace = self._parse_smil_namespace(smil)
2046 return self._parse_smil_formats(
2047 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2049 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2050 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2053 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2055 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2056 return self._download_xml(
2057 smil_url, video_id, 'Downloading SMIL file',
2058 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2060 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2061 namespace = self._parse_smil_namespace(smil)
2063 formats = self._parse_smil_formats(
2064 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2065 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2067 video_id = os.path.splitext(url_basename(smil_url))[0]
2071 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2072 name = meta.attrib.get('name')
2073 content = meta.attrib.get('content')
2074 if not name or not content:
2076 if not title and name == 'title':
2078 elif not description and name in ('description', 'abstract'):
2079 description = content
2080 elif not upload_date and name == 'date':
2081 upload_date = unified_strdate(content)
2084 'id': image.get('type'),
2085 'url': image.get('src'),
2086 'width': int_or_none(image.get('width')),
2087 'height': int_or_none(image.get('height')),
2088 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2092 'title': title or video_id,
2093 'description': description,
2094 'upload_date': upload_date,
2095 'thumbnails': thumbnails,
2097 'subtitles': subtitles,
2100 def _parse_smil_namespace(self, smil):
2101 return self._search_regex(
2102 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2104 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2106 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2107 b = meta.get('base') or meta.get('httpBase')
2118 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2119 for medium in media:
2120 src = medium.get('src')
2121 if not src or src in srcs:
2125 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2126 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2127 width = int_or_none(medium.get('width'))
2128 height = int_or_none(medium.get('height'))
2129 proto = medium.get('proto')
2130 ext = medium.get('ext')
2131 src_ext = determine_ext(src)
2132 streamer = medium.get('streamer') or base
2134 if proto == 'rtmp' or streamer.startswith('rtmp'):
2140 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2142 'filesize': filesize,
2146 if transform_rtmp_url:
2147 streamer, src = transform_rtmp_url(streamer, src)
2148 formats[-1].update({
2154 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2155 src_url = src_url.strip()
2157 if proto == 'm3u8' or src_ext == 'm3u8':
2158 m3u8_formats = self._extract_m3u8_formats(
2159 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2160 if len(m3u8_formats) == 1:
2162 m3u8_formats[0].update({
2163 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2168 formats.extend(m3u8_formats)
2169 elif src_ext == 'f4m':
2174 'plugin': 'flowplayer-3.2.0.1',
2176 f4m_url += '&' if '?' in f4m_url else '?'
2177 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2178 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2179 elif src_ext == 'mpd':
2180 formats.extend(self._extract_mpd_formats(
2181 src_url, video_id, mpd_id='dash', fatal=False))
2182 elif re.search(r'\.ism/[Mm]anifest', src_url):
2183 formats.extend(self._extract_ism_formats(
2184 src_url, video_id, ism_id='mss', fatal=False))
2185 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2189 'ext': ext or src_ext or 'flv',
2190 'format_id': 'http-%d' % (bitrate or http_count),
2192 'filesize': filesize,
2199 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2202 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2203 src = textstream.get('src')
2204 if not src or src in urls:
2207 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2208 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2209 subtitles.setdefault(lang, []).append({
2215 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2216 xspf = self._download_xml(
2217 xspf_url, playlist_id, 'Downloading xpsf playlist',
2218 'Unable to download xspf manifest', fatal=fatal)
2221 return self._parse_xspf(
2222 xspf, playlist_id, xspf_url=xspf_url,
2223 xspf_base_url=base_url(xspf_url))
2225 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2227 'xspf': 'http://xspf.org/ns/0/',
2228 's1': 'http://static.streamone.nl/player/ns/0',
2232 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2234 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2235 description = xpath_text(
2236 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2237 thumbnail = xpath_text(
2238 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2239 duration = float_or_none(
2240 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2243 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2244 format_url = urljoin(xspf_base_url, location.text)
2249 'manifest_url': xspf_url,
2250 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2251 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2252 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2254 self._sort_formats(formats)
2259 'description': description,
2260 'thumbnail': thumbnail,
2261 'duration': duration,
2266 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2267 res = self._download_xml_handle(
2269 note=note or 'Downloading MPD manifest',
2270 errnote=errnote or 'Failed to download MPD manifest',
2271 fatal=fatal, data=data, headers=headers, query=query)
2277 mpd_base_url = base_url(urlh.geturl())
2279 return self._parse_mpd_formats(
2280 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2282 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2284 Parse formats
from MPD manifest
.
2286 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2287 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2288 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2290 if not self._downloader.params.get('dynamic_mpd'):
2291 if mpd_doc.get('type') == 'dynamic':
2294 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2297 return self._xpath_ns(path, namespace)
2299 def is_drm_protected(element):
2300 return element.find(_add_ns('ContentProtection')) is not None
2302 def extract_multisegment_info(element, ms_parent_info):
2303 ms_info = ms_parent_info.copy()
2305 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2306 # common attributes and elements. We will only extract relevant
2308 def extract_common(source):
2309 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2310 if segment_timeline is not None:
2311 s_e = segment_timeline.findall(_add_ns('S'))
2313 ms_info['total_number'] = 0
2316 r = int(s.get('r', 0))
2317 ms_info['total_number'] += 1 + r
2318 ms_info['s'].append({
2319 't': int(s.get('t', 0)),
2320 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2321 'd': int(s.attrib['d']),
2324 start_number = source.get('startNumber')
2326 ms_info['start_number'] = int(start_number)
2327 timescale = source.get('timescale')
2329 ms_info['timescale'] = int(timescale)
2330 segment_duration = source.get('duration')
2331 if segment_duration:
2332 ms_info['segment_duration'] = float(segment_duration)
2334 def extract_Initialization(source):
2335 initialization = source.find(_add_ns('Initialization'))
2336 if initialization is not None:
2337 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2339 segment_list = element.find(_add_ns('SegmentList'))
2340 if segment_list is not None:
2341 extract_common(segment_list)
2342 extract_Initialization(segment_list)
2343 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2345 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2347 segment_template = element.find(_add_ns('SegmentTemplate'))
2348 if segment_template is not None:
2349 extract_common(segment_template)
2350 media = segment_template.get('media')
2352 ms_info['media'] = media
2353 initialization = segment_template.get('initialization')
2355 ms_info['initialization'] = initialization
2357 extract_Initialization(segment_template)
2360 skip_unplayable = not self._downloader.params.get('allow_unplayable_formats')
2362 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2364 for period in mpd_doc.findall(_add_ns('Period')):
2365 period_duration = parse_duration(period.get('duration')) or mpd_duration
2366 period_ms_info = extract_multisegment_info(period, {
2370 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2371 if skip_unplayable and is_drm_protected(adaptation_set):
2373 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2374 for representation in adaptation_set.findall(_add_ns('Representation')):
2375 if skip_unplayable and is_drm_protected(representation):
2377 representation_attrib = adaptation_set.attrib.copy()
2378 representation_attrib.update(representation.attrib)
2379 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2380 mime_type = representation_attrib['mimeType']
2381 content_type = mime_type.split('/')[0]
2382 if content_type == 'text':
2383 # TODO implement WebVTT downloading
2385 elif content_type in ('video', 'audio'):
2387 for element in (representation, adaptation_set, period, mpd_doc):
2388 base_url_e = element.find(_add_ns('BaseURL'))
2389 if base_url_e is not None:
2390 base_url = base_url_e.text + base_url
2391 if re.match(r'^https?://', base_url):
2393 if mpd_base_url and not re.match(r'^https?://', base_url):
2394 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2396 base_url = mpd_base_url + base_url
2397 representation_id = representation_attrib.get('id')
2398 lang = representation_attrib.get('lang')
2399 url_el = representation.find(_add_ns('BaseURL'))
2400 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2401 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2403 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2404 'manifest_url': mpd_url,
2405 'ext': mimetype2ext(mime_type),
2406 'width': int_or_none(representation_attrib.get('width')),
2407 'height': int_or_none(representation_attrib.get('height')),
2408 'tbr': float_or_none(bandwidth, 1000),
2409 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2410 'fps': int_or_none(representation_attrib.get('frameRate')),
2411 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2412 'format_note': 'DASH %s' % content_type,
2413 'filesize': filesize,
2414 'container': mimetype2ext(mime_type) + '_dash',
2416 f.update(parse_codecs(representation_attrib.get('codecs')))
2417 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2419 def prepare_template(template_name, identifiers):
2420 tmpl = representation_ms_info[template_name]
2421 # First of, % characters outside $...$ templates
2422 # must be escaped by doubling for proper processing
2423 # by % operator string formatting used further (see
2424 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2430 in_template = not in_template
2431 elif c == '%' and not in_template:
2433 # Next, $...$ templates are translated to their
2434 # %(...) counterparts to be used with % operator
2435 t = t.replace('$RepresentationID$', representation_id)
2436 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2437 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2438 t.replace('$$', '$')
2441 # @initialization is a regular template like @media one
2442 # so it should be handled just the same way (see
2443 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2444 if 'initialization' in representation_ms_info:
2445 initialization_template = prepare_template(
2447 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2448 # $Time$ shall not be included for @initialization thus
2449 # only $Bandwidth$ remains
2451 representation_ms_info['initialization_url'] = initialization_template % {
2452 'Bandwidth': bandwidth,
2455 def location_key(location):
2456 return 'url' if re.match(r'^https?://', location) else 'path'
2458 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2460 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2461 media_location_key = location_key(media_template)
2463 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2464 # can't be used at the same time
2465 if '%(Number' in media_template and 's' not in representation_ms_info:
2466 segment_duration = None
2467 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2468 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2469 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2470 representation_ms_info['fragments'] = [{
2471 media_location_key: media_template % {
2472 'Number': segment_number,
2473 'Bandwidth': bandwidth,
2475 'duration': segment_duration,
2476 } for segment_number in range(
2477 representation_ms_info['start_number'],
2478 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2480 # $Number*$ or $Time$ in media template with S list available
2481 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2482 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2483 representation_ms_info['fragments'] = []
2486 segment_number = representation_ms_info['start_number']
2488 def add_segment_url():
2489 segment_url = media_template % {
2490 'Time': segment_time,
2491 'Bandwidth': bandwidth,
2492 'Number': segment_number,
2494 representation_ms_info['fragments'].append({
2495 media_location_key: segment_url,
2496 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2499 for num, s in enumerate(representation_ms_info['s']):
2500 segment_time = s.get('t') or segment_time
2504 for r in range(s.get('r', 0)):
2505 segment_time += segment_d
2508 segment_time += segment_d
2509 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2511 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2512 # or any YouTube dashsegments video
2515 timescale = representation_ms_info['timescale']
2516 for s in representation_ms_info['s']:
2517 duration = float_or_none(s['d'], timescale)
2518 for r in range(s.get('r', 0) + 1):
2519 segment_uri = representation_ms_info['segment_urls'][segment_index]
2521 location_key(segment_uri): segment_uri,
2522 'duration': duration,
2525 representation_ms_info['fragments'] = fragments
2526 elif 'segment_urls' in representation_ms_info:
2527 # Segment URLs with no SegmentTimeline
2528 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2529 # https://github.com/ytdl-org/youtube-dl/pull/14844
2531 segment_duration = float_or_none(
2532 representation_ms_info['segment_duration'],
2533 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2534 for segment_url in representation_ms_info['segment_urls']:
2536 location_key(segment_url): segment_url,
2538 if segment_duration:
2539 fragment['duration'] = segment_duration
2540 fragments.append(fragment)
2541 representation_ms_info['fragments'] = fragments
2542 # If there is a fragments key available then we correctly recognized fragmented media.
2543 # Otherwise we will assume unfragmented media with direct access. Technically, such
2544 # assumption is not necessarily correct since we may simply have no support for
2545 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2546 if 'fragments' in representation_ms_info:
2548 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2549 'url': mpd_url or base_url,
2550 'fragment_base_url': base_url,
2552 'protocol': 'http_dash_segments',
2554 if 'initialization_url' in representation_ms_info:
2555 initialization_url = representation_ms_info['initialization_url']
2556 if not f.get('url'):
2557 f['url'] = initialization_url
2558 f['fragments'].append({location_key(initialization_url): initialization_url})
2559 f['fragments'].extend(representation_ms_info['fragments'])
2561 # Assuming direct URL to unfragmented media.
2565 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2568 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2569 res = self._download_xml_handle(
2571 note=note or 'Downloading ISM manifest',
2572 errnote=errnote or 'Failed to download ISM manifest',
2573 fatal=fatal, data=data, headers=headers, query=query)
2580 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2582 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2584 Parse formats
from ISM manifest
.
2586 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2587 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2589 if ism_doc.get('IsLive') == 'TRUE':
2591 if (not self._downloader.params.get('allow_unplayable_formats')
2592 and ism_doc.find('Protection') is not None):
2595 duration = int(ism_doc.attrib['Duration'])
2596 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2599 for stream in ism_doc.findall('StreamIndex'):
2600 stream_type = stream.get('Type')
2601 if stream_type not in ('video', 'audio'):
2603 url_pattern = stream.attrib['Url']
2604 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2605 stream_name = stream.get('Name')
2606 for track in stream.findall('QualityLevel'):
2607 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2608 # TODO: add support for WVC1 and WMAP
2609 if fourcc not in ('H264', 'AVC1', 'AACL'):
2610 self.report_warning('%s is not a supported codec' % fourcc)
2612 tbr = int(track.attrib['Bitrate']) // 1000
2613 # [1] does not mention Width and Height attributes. However,
2614 # they're often present while MaxWidth and MaxHeight are
2615 # missing, so should be used as fallbacks
2616 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2617 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2618 sampling_rate = int_or_none(track.get('SamplingRate'))
2620 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2621 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2627 stream_fragments = stream.findall('c')
2628 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2629 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2630 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2631 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2632 if not fragment_ctx['duration']:
2634 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2636 next_fragment_time = duration
2637 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2638 for _ in range(fragment_repeat):
2640 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2641 'duration': fragment_ctx['duration'] / stream_timescale,
2643 fragment_ctx['time'] += fragment_ctx['duration']
2647 format_id.append(ism_id)
2649 format_id.append(stream_name)
2650 format_id.append(compat_str(tbr))
2653 'format_id': '-'.join(format_id),
2655 'manifest_url': ism_url,
2656 'ext': 'ismv' if stream_type == 'video' else 'isma',
2660 'asr': sampling_rate,
2661 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2662 'acodec': 'none' if stream_type == 'video' else fourcc,
2664 'fragments': fragments,
2665 '_download_params': {
2666 'duration': duration,
2667 'timescale': stream_timescale,
2668 'width': width or 0,
2669 'height': height or 0,
2671 'codec_private_data': track.get('CodecPrivateData'),
2672 'sampling_rate': sampling_rate,
2673 'channels': int_or_none(track.get('Channels', 2)),
2674 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2675 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2680 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2681 def absolute_url(item_url):
2682 return urljoin(base_url, item_url)
2684 def parse_content_type(content_type):
2685 if not content_type:
2687 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2689 mimetype, codecs = ctr.groups()
2690 f = parse_codecs(codecs)
2691 f['ext'] = mimetype2ext(mimetype)
2695 def _media_formats(src, cur_media_type, type_info={}):
2696 full_url = absolute_url(src)
2697 ext = type_info.get('ext') or determine_ext(full_url)
2699 is_plain_url = False
2700 formats = self._extract_m3u8_formats(
2701 full_url, video_id, ext='mp4',
2702 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2703 preference=preference, fatal=False)
2705 is_plain_url = False
2706 formats = self._extract_mpd_formats(
2707 full_url, video_id, mpd_id=mpd_id, fatal=False)
2712 'vcodec': 'none' if cur_media_type == 'audio' else None,
2714 return is_plain_url, formats
2717 # amp-video and amp-audio are very similar to their HTML5 counterparts
2718 # so we wll include them right here (see
2719 # https://www.ampproject.org/docs/reference/components/amp-video)
2720 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
2721 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
2722 media_tags = [(media_tag, media_tag_name, media_type, '')
2723 for media_tag, media_tag_name, media_type
2724 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
2725 media_tags.extend(re.findall(
2726 # We only allow video|audio followed by a whitespace or '>'.
2727 # Allowing more characters may end up in significant slow down (see
2728 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2729 # http://www.porntrex.com/maps/videositemap.xml).
2730 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
2731 for media_tag, _, media_type, media_content in media_tags:
2736 media_attributes = extract_attributes(media_tag)
2737 src = strip_or_none(media_attributes.get('src'))
2739 _, formats = _media_formats(src, media_type)
2740 media_info['formats'].extend(formats)
2741 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2743 for source_tag in re.findall(r'<source[^>]+>', media_content):
2744 s_attr = extract_attributes(source_tag)
2745 # data-video-src and data-src are non standard but seen
2746 # several times in the wild
2747 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2750 f = parse_content_type(s_attr.get('type'))
2751 is_plain_url, formats = _media_formats(src, media_type, f)
2753 # width, height, res, label and title attributes are
2754 # all not standard but seen several times in the wild
2757 for lbl in ('label', 'title')
2758 if str_or_none(s_attr.get(lbl))
2760 width = int_or_none(s_attr.get('width'))
2761 height = (int_or_none(s_attr.get('height'))
2762 or int_or_none(s_attr.get('res')))
2763 if not width or not height:
2765 resolution = parse_resolution(lbl)
2768 width = width or resolution.get('width')
2769 height = height or resolution.get('height')
2771 tbr = parse_bitrate(lbl)
2780 'format_id': s_attr.get('label') or s_attr.get('title'),
2782 f.update(formats[0])
2783 media_info['formats'].append(f)
2785 media_info['formats'].extend(formats)
2786 for track_tag in re.findall(r'<track[^>]+>', media_content):
2787 track_attributes = extract_attributes(track_tag)
2788 kind = track_attributes.get('kind')
2789 if not kind or kind in ('subtitles', 'captions'):
2790 src = strip_or_none(track_attributes.get('src'))
2793 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2794 media_info['subtitles'].setdefault(lang, []).append({
2795 'url': absolute_url(src),
2797 for f in media_info['formats']:
2798 f.setdefault('http_headers', {})['Referer'] = base_url
2799 if media_info['formats'] or media_info['subtitles']:
2800 entries.append(media_info)
2803 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2804 signed = 'hdnea=' in manifest_url
2806 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
2807 manifest_url = re.sub(
2808 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
2809 '', manifest_url).strip('?')
2813 hdcore_sign = 'hdcore=3.7.0'
2814 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2815 hds_host = hosts.get('hds')
2817 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2818 if 'hdcore=' not in f4m_url:
2819 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2820 f4m_formats = self._extract_f4m_formats(
2821 f4m_url, video_id, f4m_id='hds', fatal=False)
2822 for entry in f4m_formats:
2823 entry.update({'extra_param_to_segment_url': hdcore_sign})
2824 formats.extend(f4m_formats)
2826 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2827 hls_host = hosts.get('hls')
2829 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2830 m3u8_formats = self._extract_m3u8_formats(
2831 m3u8_url, video_id, 'mp4', 'm3u8_native',
2832 m3u8_id='hls', fatal=False)
2833 formats.extend(m3u8_formats)
2835 http_host = hosts.get('http')
2836 if http_host and m3u8_formats and not signed:
2837 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
2838 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
2839 qualities_length = len(qualities)
2840 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
2842 for f in m3u8_formats:
2843 if f['vcodec'] != 'none':
2844 for protocol in ('http', 'https'):
2846 del http_f['manifest_url']
2848 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
2850 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
2852 'protocol': protocol,
2854 formats.append(http_f)
2859 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2860 query = compat_urlparse.urlparse(url).query
2861 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2863 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2864 url_base = mobj.group('url')
2865 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2868 def manifest_url(manifest):
2869 m_url = '%s/%s' % (http_base_url, manifest)
2871 m_url += '?%s' % query
2874 if 'm3u8' not in skip_protocols:
2875 formats.extend(self._extract_m3u8_formats(
2876 manifest_url('playlist.m3u8'), video_id, 'mp4',
2877 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2878 if 'f4m' not in skip_protocols:
2879 formats.extend(self._extract_f4m_formats(
2880 manifest_url('manifest.f4m'),
2881 video_id, f4m_id='hds', fatal=False))
2882 if 'dash' not in skip_protocols:
2883 formats.extend(self._extract_mpd_formats(
2884 manifest_url('manifest.mpd'),
2885 video_id, mpd_id='dash', fatal=False))
2886 if re.search(r'(?:/smil:|\.smil)', url_base):
2887 if 'smil' not in skip_protocols:
2888 rtmp_formats = self._extract_smil_formats(
2889 manifest_url('jwplayer.smil'),
2890 video_id, fatal=False)
2891 for rtmp_format in rtmp_formats:
2892 rtsp_format = rtmp_format.copy()
2893 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2894 del rtsp_format['play_path']
2895 del rtsp_format['ext']
2896 rtsp_format.update({
2897 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2898 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2901 formats.extend([rtmp_format, rtsp_format])
2903 for protocol in ('rtmp', 'rtsp'):
2904 if protocol not in skip_protocols:
2906 'url': '%s:%s' % (protocol, url_base),
2907 'format_id': protocol,
2908 'protocol': protocol,
2912 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2914 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2918 jwplayer_data = self._parse_json(mobj.group('options'),
2920 transform_source=transform_source)
2921 except ExtractorError:
2924 if isinstance(jwplayer_data, dict):
2925 return jwplayer_data
2927 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2928 jwplayer_data = self._find_jwplayer_data(
2929 webpage, video_id, transform_source=js_to_json)
2930 return self._parse_jwplayer_data(
2931 jwplayer_data, video_id, *args, **kwargs)
2933 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2934 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2935 # JWPlayer backward compatibility: flattened playlists
2936 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2937 if 'playlist' not in jwplayer_data:
2938 jwplayer_data = {'playlist': [jwplayer_data]}
2942 # JWPlayer backward compatibility: single playlist item
2943 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2944 if not isinstance(jwplayer_data['playlist'], list):
2945 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2947 for video_data in jwplayer_data['playlist']:
2948 # JWPlayer backward compatibility: flattened sources
2949 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2950 if 'sources' not in video_data:
2951 video_data['sources'] = [video_data]
2953 this_video_id = video_id or video_data['mediaid']
2955 formats = self._parse_jwplayer_formats(
2956 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2957 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2960 tracks = video_data.get('tracks')
2961 if tracks and isinstance(tracks, list):
2962 for track in tracks:
2963 if not isinstance(track, dict):
2965 track_kind = track.get('kind')
2966 if not track_kind or not isinstance(track_kind, compat_str):
2968 if track_kind.lower() not in ('captions', 'subtitles'):
2970 track_url = urljoin(base_url, track.get('file'))
2973 subtitles.setdefault(track.get('label') or 'en', []).append({
2974 'url': self._proto_relative_url(track_url)
2978 'id': this_video_id,
2979 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2980 'description': clean_html(video_data.get('description')),
2981 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2982 'timestamp': int_or_none(video_data.get('pubdate')),
2983 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2984 'subtitles': subtitles,
2986 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2987 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2989 '_type': 'url_transparent',
2990 'url': formats[0]['url'],
2993 self._sort_formats(formats)
2994 entry['formats'] = formats
2995 entries.append(entry)
2996 if len(entries) == 1:
2999 return self.playlist_result(entries)
3001 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3002 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3005 for source in jwplayer_sources_data:
3006 if not isinstance(source, dict):
3008 source_url = urljoin(
3009 base_url, self._proto_relative_url(source.get('file')))
3010 if not source_url or source_url in urls:
3012 urls.append(source_url)
3013 source_type = source.get('type') or ''
3014 ext = mimetype2ext(source_type) or determine_ext(source_url)
3015 if source_type == 'hls' or ext == 'm3u8':
3016 formats.extend(self._extract_m3u8_formats(
3017 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3018 m3u8_id=m3u8_id, fatal=False))
3019 elif source_type == 'dash' or ext == 'mpd':
3020 formats.extend(self._extract_mpd_formats(
3021 source_url, video_id, mpd_id=mpd_id, fatal=False))
3023 formats.extend(self._extract_smil_formats(
3024 source_url, video_id, fatal=False))
3025 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3026 elif source_type.startswith('audio') or ext in (
3027 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3034 height = int_or_none(source.get('height'))
3036 # Often no height is provided but there is a label in
3037 # format like "1080p", "720p SD", or 1080.
3038 height = int_or_none(self._search_regex(
3039 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3040 'height', default=None))
3043 'width': int_or_none(source.get('width')),
3045 'tbr': int_or_none(source.get('bitrate')),
3048 if source_url.startswith('rtmp'):
3049 a_format['ext'] = 'flv'
3050 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3051 # of jwplayer.flash.swf
3052 rtmp_url_parts = re.split(
3053 r'((?:mp4|mp3|flv):)', source_url, 1)
3054 if len(rtmp_url_parts) == 3:
3055 rtmp_url, prefix, play_path = rtmp_url_parts
3058 'play_path': prefix + play_path,
3061 a_format.update(rtmp_params)
3062 formats.append(a_format)
3065 def _live_title(self, name):
3066 """ Generate the title
for a live video
"""
3067 now = datetime.datetime.now()
3068 now_str = now.strftime('%Y-%m-%d %H:%M')
3069 return name + ' ' + now_str
3071 def _int(self, v, name, fatal=False, **kwargs):
3072 res = int_or_none(v, **kwargs)
3073 if 'get_attr' in kwargs:
3074 print(getattr(v, kwargs['get_attr']))
3076 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3078 raise ExtractorError(msg)
3080 self._downloader.report_warning(msg)
3083 def _float(self, v, name, fatal=False, **kwargs):
3084 res = float_or_none(v, **kwargs)
3086 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3088 raise ExtractorError(msg)
3090 self._downloader.report_warning(msg)
3093 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3094 path='/', secure=False, discard=False, rest={}, **kwargs):
3095 cookie = compat_cookiejar_Cookie(
3096 0, name, value, port, port is not None, domain, True,
3097 domain.startswith('.'), path, True, secure, expire_time,
3098 discard, None, None, rest)
3099 self._downloader.cookiejar.set_cookie(cookie)
3101 def _get_cookies(self, url):
3102 """ Return a compat_cookies
.SimpleCookie
with the cookies
for the url
"""
3103 req = sanitized_Request(url)
3104 self._downloader.cookiejar.add_cookie_header(req)
3105 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
3107 def _apply_first_set_cookie_header(self, url_handle, cookie):
3109 Apply first Set
-Cookie header instead of the last
. Experimental
.
3111 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3112 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3113 than
second (new
). However
, as of RFC6265 the newer one cookie
3114 should be
set into cookie store what actually happens
.
3115 We will workaround this issue by resetting the cookie to
3116 the first one manually
.
3117 1. https
://new
.vk
.com
/
3118 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3119 3. https
://learning
.oreilly
.com
/
3121 for header, cookies in url_handle.headers.items():
3122 if header.lower() != 'set-cookie':
3124 if sys.version_info[0] >= 3:
3125 cookies = cookies.encode('iso-8859-1')
3126 cookies = cookies.decode('utf-8')
3127 cookie_value = re.search(
3128 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3130 value, domain = cookie_value.groups()
3131 self._set_cookie(domain, cookie, value)
3134 def get_testcases(self, include_onlymatching=False):
3135 t = getattr(self, '_TEST', None)
3137 assert not hasattr(self, '_TESTS'), \
3138 '%s has _TEST and _TESTS' % type(self).__name__
3141 tests = getattr(self, '_TESTS', [])
3143 if not include_onlymatching and t.get('only_matching', False):
3145 t['name'] = type(self).__name__[:-len('IE')]
3148 def is_suitable(self, age_limit):
3149 """ Test whether the extractor
is generally suitable
for the given
3150 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3152 any_restricted = False
3153 for tc in self.get_testcases(include_onlymatching=False):
3154 if tc.get('playlist', []):
3155 tc = tc['playlist'][0]
3156 is_restricted = age_restricted(
3157 tc.get('info_dict', {}).get('age_limit'), age_limit)
3158 if not is_restricted:
3160 any_restricted = any_restricted or is_restricted
3161 return not any_restricted
3163 def extract_subtitles(self, *args, **kwargs):
3164 if (self._downloader.params.get('writesubtitles', False)
3165 or self._downloader.params.get('listsubtitles')):
3166 return self._get_subtitles(*args, **kwargs)
3169 def _get_subtitles(self, *args, **kwargs):
3170 raise NotImplementedError('This method must be implemented by subclasses')
3173 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3174 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3175 will be dropped
. """
3176 list1_urls = set([item['url'] for item in subtitle_list1])
3177 ret = list(subtitle_list1)
3178 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3182 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
3183 """ Merge two subtitle dictionaries
, language by language
. """
3184 ret = dict(subtitle_dict1)
3185 for lang in subtitle_dict2:
3186 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
3189 def extract_automatic_captions(self, *args, **kwargs):
3190 if (self._downloader.params.get('writeautomaticsub', False)
3191 or self._downloader.params.get('listsubtitles')):
3192 return self._get_automatic_captions(*args, **kwargs)
3195 def _get_automatic_captions(self, *args, **kwargs):
3196 raise NotImplementedError('This method must be implemented by subclasses')
3198 def mark_watched(self, *args, **kwargs):
3199 if (self._downloader.params.get('mark_watched', False)
3200 and (self._get_login_info()[0] is not None
3201 or self._downloader.params.get('cookiefile') is not None)):
3202 self._mark_watched(*args, **kwargs)
3204 def _mark_watched(self, *args, **kwargs):
3205 raise NotImplementedError('This method must be implemented by subclasses')
3207 def geo_verification_headers(self):
3209 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
3210 if geo_verification_proxy:
3211 headers['Ytdl-request-proxy'] = geo_verification_proxy
3214 def _generic_id(self, url):
3215 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3217 def _generic_title(self, url):
3218 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3221 class SearchInfoExtractor(InfoExtractor):
3223 Base
class for paged search queries extractors
.
3224 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3225 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3229 def _make_valid_url(cls):
3230 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3233 def suitable(cls, url):
3234 return re.match(cls._make_valid_url(), url) is not None
3236 def _real_extract(self, query):
3237 mobj = re.match(self._make_valid_url(), query)
3239 raise ExtractorError('Invalid search query "%s"' % query)
3241 prefix = mobj.group('prefix')
3242 query = mobj.group('query')
3244 return self._get_n_results(query, 1)
3245 elif prefix == 'all':
3246 return self._get_n_results(query, self._MAX_RESULTS)
3250 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3251 elif n > self._MAX_RESULTS:
3252 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3253 n = self._MAX_RESULTS
3254 return self._get_n_results(query, n)
3256 def _get_n_results(self, query, n):
3257 """Get a specified number of results
for a query
"""
3258 raise NotImplementedError('This method must be implemented by subclasses')
3261 def SEARCH_KEY(self):
3262 return self._SEARCH_KEY