2 from __future__
import unicode_literals
18 from ..compat
import (
19 compat_cookiejar_Cookie
,
22 compat_etree_fromstring
,
29 compat_urllib_parse_unquote
,
30 compat_urllib_parse_urlencode
,
31 compat_urllib_request
,
33 compat_xml_parse_error
,
35 from ..downloader
import FileDownloader
36 from ..downloader
.f4m
import (
38 remove_encrypted_media
,
66 parse_m3u8_attributes
,
88 class InfoExtractor(object):
89 """Information Extractor class.
91 Information extractors are the classes that, given a URL, extract
92 information about the video (or videos) the URL refers to. This
93 information includes the real video URL, the video title, author and
94 others. The information is stored in a dictionary which is then
95 passed to the YoutubeDL. The YoutubeDL processes this
96 information possibly downloading the video to the file system, among
97 other possible outcomes.
99 The type field determines the type of the result.
100 By far the most common value (and the default if _type is missing) is
101 "video", which indicates a single video.
103 For a video, the dictionaries must include the following fields:
105 id: Video identifier.
106 title: Video title, unescaped.
108 Additionally, it must contain either a formats entry or a url one:
110 formats: A list of dictionaries for each format available, ordered
111 from worst to best quality.
114 * url The mandatory URL representing the media:
115 for plain file media - HTTP URL of this file,
117 for HLS - URL of the M3U8 media playlist,
118 for HDS - URL of the F4M manifest,
120 - HTTP URL to plain file media (in case of
122 - URL of the MPD manifest or base URL
123 representing the media if MPD manifest
124 is parsed from a string (in case of
126 for MSS - URL of the ISM manifest.
128 The URL of the manifest file in case of
130 for HLS - URL of the M3U8 master playlist,
131 for HDS - URL of the F4M manifest,
132 for DASH - URL of the MPD manifest,
133 for MSS - URL of the ISM manifest.
134 * ext Will be calculated from URL if missing
135 * format A human-readable description of the format
136 ("mp4 container with h264/opus").
137 Calculated from the format_id, width, height.
138 and format_note fields if missing.
139 * format_id A short description of the format
140 ("mp4_h264_opus" or "19").
141 Technically optional, but strongly recommended.
142 * format_note Additional info about the format
143 ("3D" or "DASH video")
144 * width Width of the video, if known
145 * height Height of the video, if known
146 * resolution Textual description of width and height
147 * tbr Average bitrate of audio and video in KBit/s
148 * abr Average audio bitrate in KBit/s
149 * acodec Name of the audio codec in use
150 * asr Audio sampling rate in Hertz
151 * vbr Average video bitrate in KBit/s
153 * vcodec Name of the video codec in use
154 * container Name of the container format
155 * filesize The number of bytes, if known in advance
156 * filesize_approx An estimate for the number of bytes
157 * player_url SWF Player URL (used for rtmpdump).
158 * protocol The protocol that will be used for the actual
159 download, lower-case.
160 "http", "https", "rtsp", "rtmp", "rtmpe",
161 "m3u8", "m3u8_native" or "http_dash_segments".
163 Base URL for fragments. Each fragment's path
164 value (if present) will be relative to
166 * fragments A list of fragments of a fragmented media.
167 Each fragment entry must contain either an url
168 or a path. If an url is present it should be
169 considered by a client. Otherwise both path and
170 fragment_base_url must be present. Here is
171 the list of all potential fields:
172 * "url" - fragment's URL
173 * "path" - fragment's path relative to
175 * "duration" (optional, int or float)
176 * "filesize" (optional, int)
177 * preference Order number of this format. If this field is
178 present and not None, the formats get sorted
179 by this field, regardless of all other values.
180 -1 for default (order by other properties),
181 -2 or smaller for less than default.
182 < -1000 to hide the format (if there is
183 another one which is strictly better)
184 * language Language code, e.g. "de" or "en-US".
185 * language_preference Is this in the language mentioned in
187 10 if it's what the URL is about,
188 -1 for default (don't know),
189 -10 otherwise, other values reserved for now.
190 * quality Order number of the video quality of this
191 format, irrespective of the file format.
192 -1 for default (order by other properties),
193 -2 or smaller for less than default.
194 * source_preference Order number for this video source
195 (quality takes higher priority)
196 -1 for default (order by other properties),
197 -2 or smaller for less than default.
198 * http_headers A dictionary of additional HTTP headers
199 to add to the request.
200 * stretched_ratio If given and not 1, indicates that the
201 video's pixels are not square.
202 width : height ratio as float.
203 * no_resume The server does not support resuming the
204 (HTTP or RTMP) download. Boolean.
205 * downloader_options A dictionary of downloader options as
206 described in FileDownloader
208 url: Final video URL.
209 ext: Video filename extension.
210 format: The video format, defaults to ext (used for --get-format)
211 player_url: SWF Player URL (used for rtmpdump).
213 The following fields are optional:
215 alt_title: A secondary title of the video.
216 display_id An alternative identifier for the video, not necessarily
217 unique, but available before title. Typically, id is
218 something like "4234987", title "Dancing naked mole rats",
219 and display_id "dancing-naked-mole-rats"
220 thumbnails: A list of dictionaries, with the following entries:
221 * "id" (optional, string) - Thumbnail format ID
223 * "preference" (optional, int) - quality of the image
224 * "width" (optional, int)
225 * "height" (optional, int)
226 * "resolution" (optional, string "{width}x{height}",
228 * "filesize" (optional, int)
229 thumbnail: Full URL to a video thumbnail image.
230 description: Full video description.
231 uploader: Full name of the video uploader.
232 license: License name the video is licensed under.
233 creator: The creator of the video.
234 release_date: The date (YYYYMMDD) when the video was released.
235 timestamp: UNIX timestamp of the moment the video became available.
236 upload_date: Video upload date (YYYYMMDD).
237 If not explicitly set, calculated from timestamp.
238 uploader_id: Nickname or id of the video uploader.
239 uploader_url: Full URL to a personal webpage of the video uploader.
240 channel: Full name of the channel the video is uploaded on.
241 Note that channel fields may or may not repeat uploader
242 fields. This depends on a particular extractor.
243 channel_id: Id of the channel.
244 channel_url: Full URL to a channel webpage.
245 location: Physical location where the video was filmed.
246 subtitles: The available subtitles as a dictionary in the format
247 {tag: subformats}. "tag" is usually a language code, and
248 "subformats" is a list sorted from lower to higher
249 preference, each element is a dictionary with the "ext"
251 * "data": The subtitles file contents
252 * "url": A URL pointing to the subtitles file
253 "ext" will be calculated from URL if missing
254 automatic_captions: Like 'subtitles', used by the YoutubeIE for
255 automatically generated captions
256 duration: Length of the video in seconds, as an integer or float.
257 view_count: How many users have watched the video on the platform.
258 like_count: Number of positive ratings of the video
259 dislike_count: Number of negative ratings of the video
260 repost_count: Number of reposts of the video
261 average_rating: Average rating give by users, the scale used depends on the webpage
262 comment_count: Number of comments on the video
263 comments: A list of comments, each with one or more of the following
264 properties (all but one of text or html optional):
265 * "author" - human-readable name of the comment author
266 * "author_id" - user ID of the comment author
268 * "html" - Comment as HTML
269 * "text" - Plain text of the comment
270 * "timestamp" - UNIX timestamp of comment
271 * "parent" - ID of the comment this one is replying to.
272 Set to "root" to indicate that this is a
273 comment to the original video.
274 age_limit: Age restriction for the video, as an integer (years)
275 webpage_url: The URL to the video webpage, if given to youtube-dlc it
276 should allow to get the same result again. (It will be set
277 by YoutubeDL if it's missing)
278 categories: A list of categories that the video falls in, for example
280 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
281 is_live: True, False, or None (=unknown). Whether this video is a
282 live stream that goes on instead of a fixed-length video.
283 start_time: Time in seconds where the reproduction should start, as
284 specified in the URL.
285 end_time: Time in seconds where the reproduction should end, as
286 specified in the URL.
287 chapters: A list of dictionaries, with the following entries:
288 * "start_time" - The start time of the chapter in seconds
289 * "end_time" - The end time of the chapter in seconds
290 * "title" (optional, string)
292 The following fields should only be used when the video belongs to some logical
295 chapter: Name or title of the chapter the video belongs to.
296 chapter_number: Number of the chapter the video belongs to, as an integer.
297 chapter_id: Id of the chapter the video belongs to, as a unicode string.
299 The following fields should only be used when the video is an episode of some
300 series, programme or podcast:
302 series: Title of the series or programme the video episode belongs to.
303 season: Title of the season the video episode belongs to.
304 season_number: Number of the season the video episode belongs to, as an integer.
305 season_id: Id of the season the video episode belongs to, as a unicode string.
306 episode: Title of the video episode. Unlike mandatory video title field,
307 this field should denote the exact title of the video episode
308 without any kind of decoration.
309 episode_number: Number of the video episode within a season, as an integer.
310 episode_id: Id of the video episode, as a unicode string.
312 The following fields should only be used when the media is a track or a part of
315 track: Title of the track.
316 track_number: Number of the track within an album or a disc, as an integer.
317 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
319 artist: Artist(s) of the track.
320 genre: Genre(s) of the track.
321 album: Title of the album the track belongs to.
322 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
323 album_artist: List of all artists appeared on the album (e.g.
324 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
326 disc_number: Number of the disc or other physical medium the track belongs to,
328 release_year: Year (YYYY) when the album was released.
330 Unless mentioned otherwise, the fields should be Unicode strings.
332 Unless mentioned otherwise, None is equivalent to absence of information.
335 _type "playlist" indicates multiple videos.
336 There must be a key "entries", which is a list, an iterable, or a PagedList
337 object, each element of which is a valid dictionary by this specification.
339 Additionally, playlists can have "id", "title", "description", "uploader",
340 "uploader_id", "uploader_url" attributes with the same semantics as videos
344 _type "multi_video" indicates that there are multiple videos that
345 form a single show, for examples multiple acts of an opera or TV episode.
346 It must have an entries key like a playlist and contain all the keys
347 required for a video at the same time.
350 _type "url" indicates that the video must be extracted from another
351 location, possibly by a different extractor. Its only required key is:
352 "url" - the next URL to extract.
353 The key "ie_key" can be set to the class name (minus the trailing "IE",
354 e.g. "Youtube") if the extractor class is known in advance.
355 Additionally, the dictionary may have any properties of the resolved entity
356 known in advance, for example "title" if the title of the referred video is
360 _type "url_transparent" entities have the same specification as "url", but
361 indicate that the given additional information is more precise than the one
362 associated with the resolved URL.
363 This is useful when a site employs a video service that hosts the video and
364 its technical metadata, but that video service does not embed a useful
365 title, description etc.
368 Subclasses of this one should re-define the _real_initialize() and
369 _real_extract() methods and define a _VALID_URL regexp.
370 Probably, they should also be added to the list of extractors.
372 _GEO_BYPASS attribute may be set to False in order to disable
373 geo restriction bypass mechanisms for a particular extractor.
374 Though it won't disable explicit geo restriction bypass based on
375 country code provided with geo_bypass_country.
377 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
378 countries for this extractor. One of these countries will be used by
379 geo restriction bypass mechanism right away in order to bypass
380 geo restriction, of course, if the mechanism is not disabled.
382 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
383 IP blocks in CIDR notation for this extractor. One of these IP blocks
384 will be used by geo restriction bypass mechanism similarly
387 Finally, the _WORKING attribute should be set to False for broken IEs
388 in order to warn the users and skip the tests.
393 _x_forwarded_for_ip
= None
395 _GEO_COUNTRIES
= None
396 _GEO_IP_BLOCKS
= None
399 def __init__(self
, downloader
=None):
400 """Constructor. Receives an optional downloader."""
402 self
._x
_forwarded
_for
_ip
= None
403 self
.set_downloader(downloader
)
406 def suitable(cls
, url
):
407 """Receives a URL and returns True if suitable for this IE."""
409 # This does not use has/getattr intentionally - we want to know whether
410 # we have cached the regexp for *this* class, whereas getattr would also
411 # match the superclass
412 if '_VALID_URL_RE' not in cls
.__dict
__:
413 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
414 return cls
._VALID
_URL
_RE
.match(url
) is not None
417 def _match_id(cls
, url
):
418 if '_VALID_URL_RE' not in cls
.__dict
__:
419 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
420 m
= cls
._VALID
_URL
_RE
.match(url
)
422 return compat_str(m
.group('id'))
426 """Getter method for _WORKING."""
429 def initialize(self
):
430 """Initializes an instance (authentication, etc)."""
431 self
._initialize
_geo
_bypass
({
432 'countries': self
._GEO
_COUNTRIES
,
433 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
436 self
._real
_initialize
()
439 def _initialize_geo_bypass(self
, geo_bypass_context
):
441 Initialize geo restriction bypass mechanism.
443 This method is used to initialize geo bypass mechanism based on faking
444 X-Forwarded-For HTTP header. A random country from provided country list
445 is selected and a random IP belonging to this country is generated. This
446 IP will be passed as X-Forwarded-For HTTP header in all subsequent
449 This method will be used for initial geo bypass mechanism initialization
450 during the instance initialization with _GEO_COUNTRIES and
453 You may also manually call it from extractor's code if geo bypass
454 information is not available beforehand (e.g. obtained during
455 extraction) or due to some other reason. In this case you should pass
456 this information in geo bypass context passed as first argument. It may
457 contain following fields:
459 countries: List of geo unrestricted countries (similar
461 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
462 (similar to _GEO_IP_BLOCKS)
465 if not self
._x
_forwarded
_for
_ip
:
467 # Geo bypass mechanism is explicitly disabled by user
468 if not self
._downloader
.params
.get('geo_bypass', True):
471 if not geo_bypass_context
:
472 geo_bypass_context
= {}
474 # Backward compatibility: previously _initialize_geo_bypass
475 # expected a list of countries, some 3rd party code may still use
477 if isinstance(geo_bypass_context
, (list, tuple)):
478 geo_bypass_context
= {
479 'countries': geo_bypass_context
,
482 # The whole point of geo bypass mechanism is to fake IP
483 # as X-Forwarded-For HTTP header based on some IP block or
486 # Path 1: bypassing based on IP block in CIDR notation
488 # Explicit IP block specified by user, use it right away
489 # regardless of whether extractor is geo bypassable or not
490 ip_block
= self
._downloader
.params
.get('geo_bypass_ip_block', None)
492 # Otherwise use random IP block from geo bypass context but only
493 # if extractor is known as geo bypassable
495 ip_blocks
= geo_bypass_context
.get('ip_blocks')
496 if self
._GEO
_BYPASS
and ip_blocks
:
497 ip_block
= random
.choice(ip_blocks
)
500 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
501 if self
._downloader
.params
.get('verbose', False):
502 self
._downloader
.to_screen(
503 '[debug] Using fake IP %s as X-Forwarded-For.'
504 % self
._x
_forwarded
_for
_ip
)
507 # Path 2: bypassing based on country code
509 # Explicit country code specified by user, use it right away
510 # regardless of whether extractor is geo bypassable or not
511 country
= self
._downloader
.params
.get('geo_bypass_country', None)
513 # Otherwise use random country code from geo bypass context but
514 # only if extractor is known as geo bypassable
516 countries
= geo_bypass_context
.get('countries')
517 if self
._GEO
_BYPASS
and countries
:
518 country
= random
.choice(countries
)
521 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
522 if self
._downloader
.params
.get('verbose', False):
523 self
._downloader
.to_screen(
524 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
525 % (self
._x
_forwarded
_for
_ip
, country
.upper()))
527 def extract(self
, url
):
528 """Extracts URL information and returns it in list of dicts."""
533 ie_result
= self
._real
_extract
(url
)
534 if self
._x
_forwarded
_for
_ip
:
535 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
537 except GeoRestrictedError
as e
:
538 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
541 except ExtractorError
:
543 except compat_http_client
.IncompleteRead
as e
:
544 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
545 except (KeyError, StopIteration) as e
:
546 raise ExtractorError('An extractor error has occurred.', cause
=e
)
548 def __maybe_fake_ip_and_retry(self
, countries
):
549 if (not self
._downloader
.params
.get('geo_bypass_country', None)
551 and self
._downloader
.params
.get('geo_bypass', True)
552 and not self
._x
_forwarded
_for
_ip
554 country_code
= random
.choice(countries
)
555 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
556 if self
._x
_forwarded
_for
_ip
:
558 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
559 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
563 def set_downloader(self
, downloader
):
564 """Sets the downloader for this IE."""
565 self
._downloader
= downloader
567 def _real_initialize(self
):
568 """Real initialization process. Redefine in subclasses."""
571 def _real_extract(self
, url
):
572 """Real extraction process. Redefine in subclasses."""
577 """A string for getting the InfoExtractor with get_info_extractor"""
578 return compat_str(cls
.__name
__[:-2])
582 return compat_str(type(self
).__name
__[:-2])
585 def __can_accept_status_code(err
, expected_status
):
586 assert isinstance(err
, compat_urllib_error
.HTTPError
)
587 if expected_status
is None:
589 if isinstance(expected_status
, compat_integer_types
):
590 return err
.code
== expected_status
591 elif isinstance(expected_status
, (list, tuple)):
592 return err
.code
in expected_status
593 elif callable(expected_status
):
594 return expected_status(err
.code
) is True
598 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
600 Return the response handle.
602 See _download_webpage docstring for arguments specification.
605 self
.report_download_webpage(video_id
)
606 elif note
is not False:
608 self
.to_screen('%s' % (note
,))
610 self
.to_screen('%s: %s' % (video_id
, note
))
612 # Some sites check X-Forwarded-For HTTP header in order to figure out
613 # the origin of the client behind proxy. This allows bypassing geo
614 # restriction by faking this header's value to IP that belongs to some
615 # geo unrestricted country. We will do so once we encounter any
616 # geo restriction error.
617 if self
._x
_forwarded
_for
_ip
:
618 if 'X-Forwarded-For' not in headers
:
619 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
621 if isinstance(url_or_request
, compat_urllib_request
.Request
):
622 url_or_request
= update_Request(
623 url_or_request
, data
=data
, headers
=headers
, query
=query
)
626 url_or_request
= update_url_query(url_or_request
, query
)
627 if data
is not None or headers
:
628 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
629 exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
630 if hasattr(ssl
, 'CertificateError'):
631 exceptions
.append(ssl
.CertificateError
)
633 return self
._downloader
.urlopen(url_or_request
)
634 except tuple(exceptions
) as err
:
635 if isinstance(err
, compat_urllib_error
.HTTPError
):
636 if self
.__can
_accept
_status
_code
(err
, expected_status
):
637 # Retain reference to error to prevent file object from
638 # being closed before it can be read. Works around the
639 # effects of <https://bugs.python.org/issue15002>
640 # introduced in Python 3.4.1.
647 errnote
= 'Unable to download webpage'
649 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
651 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
653 self
._downloader
.report_warning(errmsg
)
656 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
658 Return a tuple (page content as string, URL handle).
660 See _download_webpage docstring for arguments specification.
662 # Strip hashes from the URL (#1038)
663 if isinstance(url_or_request
, (compat_str
, str)):
664 url_or_request
= url_or_request
.partition('#')[0]
666 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
670 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
671 return (content
, urlh
)
674 def _guess_encoding_from_content(content_type
, webpage_bytes
):
675 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
677 encoding
= m
.group(1)
679 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
680 webpage_bytes[:1024])
682 encoding = m.group(1).decode('ascii')
683 elif webpage_bytes.startswith(b'\xff\xfe'):
690 def __check_blocked(self, content):
691 first_block = content[:512]
692 if ('<title>Access to this site is blocked</title>' in content
693 and 'Websense' in first_block):
694 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
695 blocked_iframe = self._html_search_regex(
696 r'<iframe src="([^
"]+)"', content,
697 'Websense information URL
', default=None)
699 msg += ' Visit
%s for more details
' % blocked_iframe
700 raise ExtractorError(msg, expected=True)
701 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
703 'Access to this webpage has been blocked by Indian censorship
. '
704 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
705 block_msg = self._html_search_regex(
706 r'</h1
><p
>(.*?
)</p
>',
707 content, 'block message
', default=None)
709 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
710 raise ExtractorError(msg, expected=True)
711 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
712 and 'blocklist
.rkn
.gov
.ru
' in content):
713 raise ExtractorError(
714 'Access to this webpage has been blocked by decision of the Russian government
. '
715 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
718 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
719 content_type = urlh.headers.get('Content
-Type
', '')
720 webpage_bytes = urlh.read()
721 if prefix is not None:
722 webpage_bytes = prefix + webpage_bytes
724 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
725 if self._downloader.params.get('dump_intermediate_pages
', False):
726 self.to_screen('Dumping request to
' + urlh.geturl())
727 dump = base64.b64encode(webpage_bytes).decode('ascii
')
728 self._downloader.to_screen(dump)
729 if self._downloader.params.get('write_pages
', False):
730 basen = '%s_%s' % (video_id, urlh.geturl())
732 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
733 basen = basen[:240 - len(h)] + h
734 raw_filename = basen + '.dump
'
735 filename = sanitize_filename(raw_filename, restricted=True)
736 self.to_screen('Saving request to
' + filename)
737 # Working around MAX_PATH limitation on Windows (see
738 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
739 if compat_os_name == 'nt
':
740 absfilepath = os.path.abspath(filename)
741 if len(absfilepath) > 259:
742 filename = '\\\\?
\\' + absfilepath
743 with open(filename, 'wb
') as outf:
744 outf.write(webpage_bytes)
747 content = webpage_bytes.decode(encoding, 'replace
')
749 content = webpage_bytes.decode('utf
-8', 'replace
')
751 self.__check_blocked(content)
755 def _download_webpage(
756 self, url_or_request, video_id, note=None, errnote=None,
757 fatal=True, tries=1, timeout=5, encoding=None, data=None,
758 headers={}, query={}, expected_status=None):
760 Return the data of the page as a string.
763 url_or_request -- plain text URL as a string or
764 a compat_urllib_request.Requestobject
765 video_id -- Video/playlist/item identifier (string)
768 note -- note printed before downloading (string)
769 errnote -- note printed in case of an error (string)
770 fatal -- flag denoting whether error should be considered fatal,
771 i.e. whether it should cause ExtractionError to be raised,
772 otherwise a warning will be reported and extraction continued
773 tries -- number of tries
774 timeout -- sleep interval between tries
775 encoding -- encoding for a page content decoding, guessed automatically
776 when not explicitly specified
777 data -- POST data (bytes)
778 headers -- HTTP headers (dict)
779 query -- URL query (dict)
780 expected_status -- allows to accept failed HTTP requests (non 2xx
781 status code) by explicitly specifying a set of accepted status
782 codes. Can be any of the following entities:
783 - an integer type specifying an exact failed status code to
785 - a list or a tuple of integer types specifying a list of
786 failed status codes to accept
787 - a callable accepting an actual failed status code and
788 returning True if it should be accepted
789 Note that this argument does not affect success status codes (2xx)
790 which are always accepted.
795 while success is False:
797 res = self._download_webpage_handle(
798 url_or_request, video_id, note, errnote, fatal,
799 encoding=encoding, data=data, headers=headers, query=query,
800 expected_status=expected_status)
802 except compat_http_client.IncompleteRead as e:
804 if try_count >= tries:
806 self._sleep(timeout, video_id)
813 def _download_xml_handle(
814 self, url_or_request, video_id, note='Downloading XML
',
815 errnote='Unable to download XML
', transform_source=None,
816 fatal=True, encoding=None, data=None, headers={}, query={},
817 expected_status=None):
819 Return a tuple (xml as an compat_etree_Element, URL handle).
821 See _download_webpage docstring for arguments specification.
823 res = self._download_webpage_handle(
824 url_or_request, video_id, note, errnote, fatal=fatal,
825 encoding=encoding, data=data, headers=headers, query=query,
826 expected_status=expected_status)
829 xml_string, urlh = res
830 return self._parse_xml(
831 xml_string, video_id, transform_source=transform_source,
835 self, url_or_request, video_id,
836 note='Downloading XML
', errnote='Unable to download XML
',
837 transform_source=None, fatal=True, encoding=None,
838 data=None, headers={}, query={}, expected_status=None):
840 Return the xml as an compat_etree_Element.
842 See _download_webpage docstring for arguments specification.
844 res = self._download_xml_handle(
845 url_or_request, video_id, note=note, errnote=errnote,
846 transform_source=transform_source, fatal=fatal, encoding=encoding,
847 data=data, headers=headers, query=query,
848 expected_status=expected_status)
849 return res if res is False else res[0]
851 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
853 xml_string = transform_source(xml_string)
855 return compat_etree_fromstring(xml_string.encode('utf
-8'))
856 except compat_xml_parse_error as ve:
857 errmsg = '%s: Failed to parse XML
' % video_id
859 raise ExtractorError(errmsg, cause=ve)
861 self.report_warning(errmsg + str(ve))
863 def _download_json_handle(
864 self, url_or_request, video_id, note='Downloading JSON metadata
',
865 errnote='Unable to download JSON metadata
', transform_source=None,
866 fatal=True, encoding=None, data=None, headers={}, query={},
867 expected_status=None):
869 Return a tuple (JSON object, URL handle).
871 See _download_webpage docstring for arguments specification.
873 res = self._download_webpage_handle(
874 url_or_request, video_id, note, errnote, fatal=fatal,
875 encoding=encoding, data=data, headers=headers, query=query,
876 expected_status=expected_status)
879 json_string, urlh = res
880 return self._parse_json(
881 json_string, video_id, transform_source=transform_source,
885 self, url_or_request, video_id, note='Downloading JSON metadata
',
886 errnote='Unable to download JSON metadata
', transform_source=None,
887 fatal=True, encoding=None, data=None, headers={}, query={},
888 expected_status=None):
890 Return the JSON object as a dict.
892 See _download_webpage docstring for arguments specification.
894 res = self._download_json_handle(
895 url_or_request, video_id, note=note, errnote=errnote,
896 transform_source=transform_source, fatal=fatal, encoding=encoding,
897 data=data, headers=headers, query=query,
898 expected_status=expected_status)
899 return res if res is False else res[0]
901 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
903 json_string = transform_source(json_string)
905 return json.loads(json_string)
906 except ValueError as ve:
907 errmsg = '%s: Failed to parse JSON
' % video_id
909 raise ExtractorError(errmsg, cause=ve)
911 self.report_warning(errmsg + str(ve))
913 def report_warning(self, msg, video_id=None):
914 idstr = '' if video_id is None else '%s: ' % video_id
915 self._downloader.report_warning(
916 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
918 def to_screen(self, msg):
919 """Print msg to screen, prefixing it with '[ie_name
]'"""
920 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
922 def report_extraction(self, id_or_name):
923 """Report information extraction."""
924 self.to_screen('%s: Extracting information
' % id_or_name)
926 def report_download_webpage(self, video_id):
927 """Report webpage download."""
928 self.to_screen('%s: Downloading webpage
' % video_id)
930 def report_age_confirmation(self):
931 """Report attempt to confirm age."""
932 self.to_screen('Confirming age
')
934 def report_login(self):
935 """Report attempt to log in."""
936 self.to_screen('Logging
in')
939 def raise_login_required(msg='This video
is only available
for registered users
'):
940 raise ExtractorError(
941 '%s. Use
--username
and --password
or --netrc to provide account credentials
.' % msg,
945 def raise_geo_restricted(msg='This video
is not available
from your location due to geo restriction
', countries=None):
946 raise GeoRestrictedError(msg, countries=countries)
948 # Methods for following #608
950 def url_result(url, ie=None, video_id=None, video_title=None):
951 """Returns a URL that points to a page that should be processed"""
952 # TODO: ie should be the class used for getting the info
953 video_info = {'_type
': 'url
',
956 if video_id is not None:
957 video_info['id'] = video_id
958 if video_title is not None:
959 video_info['title
'] = video_title
962 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
964 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
966 return self.playlist_result(
967 urls, playlist_id=playlist_id, playlist_title=playlist_title)
970 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
971 """Returns a playlist"""
972 video_info = {'_type
': 'playlist
',
975 video_info['id'] = playlist_id
977 video_info['title
'] = playlist_title
978 if playlist_description:
979 video_info['description
'] = playlist_description
982 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
984 Perform a regex search on the given string, using a single or a list of
985 patterns returning the first matching group.
986 In case of failure return a default value or raise a WARNING or a
987 RegexNotFoundError, depending on fatal, specifying the field name.
989 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
990 mobj = re.search(pattern, string, flags)
993 mobj = re.search(p, string, flags)
997 if not self._downloader.params.get('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
998 _name = '\033[0;34m
%s\033[0m
' % name
1004 # return the first matching group
1005 return next(g for g in mobj.groups() if g is not None)
1007 return mobj.group(group)
1008 elif default is not NO_DEFAULT:
1011 raise RegexNotFoundError('Unable to extract
%s' % _name)
1013 self._downloader.report_warning('unable to extract
%s' % _name + bug_reports_message())
1016 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1018 Like _search_regex, but strips HTML tags and unescapes entities.
1020 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1022 return clean_html(res).strip()
1026 def _get_netrc_login_info(self, netrc_machine=None):
1029 netrc_machine = netrc_machine or self._NETRC_MACHINE
1031 if self._downloader.params.get('usenetrc
', False):
1033 info = netrc.netrc().authenticators(netrc_machine)
1034 if info is not None:
1038 raise netrc.NetrcParseError(
1039 'No authenticators
for %s' % netrc_machine)
1040 except (IOError, netrc.NetrcParseError) as err:
1041 self._downloader.report_warning(
1042 'parsing
.netrc
: %s' % error_to_compat_str(err))
1044 return username, password
1046 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1048 Get the login info as (username, password)
1049 First look for the manually specified credentials using username_option
1050 and password_option as keys in params dictionary. If no such credentials
1051 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1053 If there's no info available
, return (None, None)
1055 if self._downloader is None:
1058 downloader_params = self._downloader.params
1060 # Attempt to use provided username and password or .netrc data
1061 if downloader_params.get(username_option) is not None:
1062 username = downloader_params[username_option]
1063 password = downloader_params[password_option]
1065 username, password = self._get_netrc_login_info(netrc_machine)
1067 return username, password
1069 def _get_tfa_info(self, note='two-factor verification code'):
1071 Get the two
-factor authentication info
1072 TODO
- asking the user will be required
for sms
/phone verify
1073 currently just uses the command line option
1074 If there
's no info available, return None
1076 if self._downloader is None:
1078 downloader_params = self._downloader.params
1080 if downloader_params.get('twofactor
') is not None:
1081 return downloader_params['twofactor
']
1083 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1085 # Helper functions for extracting OpenGraph info
1087 def _og_regexes(prop):
1088 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1089 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1090 % {'prop': re.escape(prop)})
1091 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1093 template % (property_re, content_re),
1094 template % (content_re, property_re),
1098 def _meta_regex(prop):
1099 return r'''(?isx)<meta
1100 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1101 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1103 def _og_search_property(self, prop, html, name=None, **kargs):
1104 if not isinstance(prop, (list, tuple)):
1107 name = 'OpenGraph
%s' % prop[0]
1110 og_regexes.extend(self._og_regexes(p))
1111 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1114 return unescapeHTML(escaped)
1116 def _og_search_thumbnail(self, html, **kargs):
1117 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1119 def _og_search_description(self, html, **kargs):
1120 return self._og_search_property('description
', html, fatal=False, **kargs)
1122 def _og_search_title(self, html, **kargs):
1123 return self._og_search_property('title
', html, **kargs)
1125 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1126 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1128 regexes = self._og_regexes('video
:secure_url
') + regexes
1129 return self._html_search_regex(regexes, html, name, **kargs)
1131 def _og_search_url(self, html, **kargs):
1132 return self._og_search_property('url
', html, **kargs)
1134 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1135 if not isinstance(name, (list, tuple)):
1137 if display_name is None:
1138 display_name = name[0]
1139 return self._html_search_regex(
1140 [self._meta_regex(n) for n in name],
1141 html, display_name, fatal=fatal, group='content
', **kwargs)
1143 def _dc_search_uploader(self, html):
1144 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1146 def _rta_search(self, html):
1147 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1148 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1149 r' content
="RTA-5042-1996-1400-1577-RTA"',
1154 def _media_rating_search(self, html):
1155 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1156 rating = self._html_search_meta('rating
', html)
1168 return RATING_TABLE.get(rating.lower())
1170 def _family_friendly_search(self, html):
1171 # See http://schema.org/VideoObject
1172 family_friendly = self._html_search_meta(
1173 'isFamilyFriendly
', html, default=None)
1175 if not family_friendly:
1184 return RATING_TABLE.get(family_friendly.lower())
1186 def _twitter_search_player(self, html):
1187 return self._html_search_meta('twitter
:player
', html,
1188 'twitter card player
')
1190 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1191 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1192 default = kwargs.get('default
', NO_DEFAULT)
1193 # JSON-LD may be malformed and thus `fatal` should be respected.
1194 # At the same time `default` may be passed that assumes `fatal=False`
1195 # for _search_regex. Let's simulate the same behavior here
as well
.
1196 fatal
= kwargs
.get('fatal', True) if default
== NO_DEFAULT
else False
1198 for mobj
in json_ld_list
:
1199 json_ld_item
= self
._parse
_json
(
1200 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1201 if not json_ld_item
:
1203 if isinstance(json_ld_item
, dict):
1204 json_ld
.append(json_ld_item
)
1205 elif isinstance(json_ld_item
, (list, tuple)):
1206 json_ld
.extend(json_ld_item
)
1208 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1211 if default
is not NO_DEFAULT
:
1214 raise RegexNotFoundError('Unable to extract JSON-LD')
1216 self
._downloader
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1219 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1220 if isinstance(json_ld
, compat_str
):
1221 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1225 if not isinstance(json_ld
, (list, tuple, dict)):
1227 if isinstance(json_ld
, dict):
1230 INTERACTION_TYPE_MAP
= {
1231 'CommentAction': 'comment',
1232 'AgreeAction': 'like',
1233 'DisagreeAction': 'dislike',
1234 'LikeAction': 'like',
1235 'DislikeAction': 'dislike',
1236 'ListenAction': 'view',
1237 'WatchAction': 'view',
1238 'ViewAction': 'view',
1241 def extract_interaction_statistic(e
):
1242 interaction_statistic
= e
.get('interactionStatistic')
1243 if not isinstance(interaction_statistic
, list):
1245 for is_e
in interaction_statistic
:
1246 if not isinstance(is_e
, dict):
1248 if is_e
.get('@type') != 'InteractionCounter':
1250 interaction_type
= is_e
.get('interactionType')
1251 if not isinstance(interaction_type
, compat_str
):
1253 # For interaction count some sites provide string instead of
1254 # an integer (as per spec) with non digit characters (e.g. ",")
1255 # so extracting count with more relaxed str_to_int
1256 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1257 if interaction_count
is None:
1259 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1262 count_key
= '%s_count' % count_kind
1263 if info
.get(count_key
) is not None:
1265 info
[count_key
] = interaction_count
1267 def extract_video_object(e
):
1268 assert e
['@type'] == 'VideoObject'
1270 'url': url_or_none(e
.get('contentUrl')),
1271 'title': unescapeHTML(e
.get('name')),
1272 'description': unescapeHTML(e
.get('description')),
1273 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1274 'duration': parse_duration(e
.get('duration')),
1275 'timestamp': unified_timestamp(e
.get('uploadDate')),
1276 'uploader': str_or_none(e
.get('author')),
1277 'filesize': float_or_none(e
.get('contentSize')),
1278 'tbr': int_or_none(e
.get('bitrate')),
1279 'width': int_or_none(e
.get('width')),
1280 'height': int_or_none(e
.get('height')),
1281 'view_count': int_or_none(e
.get('interactionCount')),
1283 extract_interaction_statistic(e
)
1287 item_type
= e
.get('@type')
1288 if expected_type
is not None and expected_type
!= item_type
:
1290 if item_type
in ('TVEpisode', 'Episode'):
1291 episode_name
= unescapeHTML(e
.get('name'))
1293 'episode': episode_name
,
1294 'episode_number': int_or_none(e
.get('episodeNumber')),
1295 'description': unescapeHTML(e
.get('description')),
1297 if not info
.get('title') and episode_name
:
1298 info
['title'] = episode_name
1299 part_of_season
= e
.get('partOfSeason')
1300 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1302 'season': unescapeHTML(part_of_season
.get('name')),
1303 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1305 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1306 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1307 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1308 elif item_type
== 'Movie':
1310 'title': unescapeHTML(e
.get('name')),
1311 'description': unescapeHTML(e
.get('description')),
1312 'duration': parse_duration(e
.get('duration')),
1313 'timestamp': unified_timestamp(e
.get('dateCreated')),
1315 elif item_type
in ('Article', 'NewsArticle'):
1317 'timestamp': parse_iso8601(e
.get('datePublished')),
1318 'title': unescapeHTML(e
.get('headline')),
1319 'description': unescapeHTML(e
.get('articleBody')),
1321 elif item_type
== 'VideoObject':
1322 extract_video_object(e
)
1323 if expected_type
is None:
1327 video
= e
.get('video')
1328 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1329 extract_video_object(video
)
1330 if expected_type
is None:
1334 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1337 def _hidden_inputs(html
):
1338 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1340 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1341 attrs
= extract_attributes(input)
1344 if attrs
.get('type') not in ('hidden', 'submit'):
1346 name
= attrs
.get('name') or attrs
.get('id')
1347 value
= attrs
.get('value')
1348 if name
and value
is not None:
1349 hidden_inputs
[name
] = value
1350 return hidden_inputs
1352 def _form_hidden_inputs(self
, form_id
, html
):
1353 form
= self
._search
_regex
(
1354 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1355 html, '%s form
' % form_id, group='form
')
1356 return self._hidden_inputs(form)
1359 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<seperator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1361 default = ('hidden
', 'has_video
', 'extractor
', 'lang
', 'quality
',
1362 'tbr
', 'filesize
', 'vbr
', 'height
', 'width
', 'protocol
', 'vext
',
1363 'abr
', 'aext
', 'fps
', 'filesize_approx
', 'source_preference
', 'format_id
')
1366 'vcodec
': {'type': 'ordered
', 'regex
': True,
1367 'order
': ['av01
', 'vp9
', '(h265|he?vc?
)', '(h264|avc
)', 'vp8
', '(mp4v|h263
)', 'theora
', '', None, 'none
']},
1368 'acodec
': {'type': 'ordered
', 'regex
': True,
1369 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1370 'protocol
': {'type': 'ordered
', 'regex
': True,
1371 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', 'm3u8
', '.*dash
', '', 'mms|rtsp
', 'none
', 'f4
']},
1372 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1373 'order
': ('mp4
', 'flv
', 'webm
', '', 'none
'), # Why is flv prefered over webm???
1374 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1375 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1376 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1377 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1378 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1379 'extractor_preference
': {'priority': True, 'type': 'extractor'},
1380 'has_video
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1381 'has_audio
': {'priority': False, 'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1382 'language_preference
': {'priority': True, 'convert': 'ignore'},
1383 'quality
': {'priority': True, 'convert': 'float_none'},
1384 'filesize
': {'convert': 'bytes'},
1385 'filesize_approx
': {'convert': 'bytes'},
1386 'format_id
': {'convert': 'string'},
1387 'height
': {'convert': 'float_none'},
1388 'width
': {'convert': 'float_none'},
1389 'fps
': {'convert': 'float_none'},
1390 'tbr
': {'convert': 'float_none'},
1391 'vbr
': {'convert': 'float_none'},
1392 'abr
': {'convert': 'float_none'},
1393 'asr
': {'convert': 'float_none'},
1394 'source_preference
': {'convert': 'ignore'},
1395 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1396 'bitrate
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1397 'filesize_estimate
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'filesize_approx')},
1398 'extension
': {'type': 'combined', 'field': ('vext', 'aext')},
1399 'dimension
': {'type': 'multiple', 'field': ('height', 'width'), 'function': min}, # not named as 'resolution
' because such a field exists
1400 'res
': {'type': 'alias', 'field': 'dimension'},
1401 'ext
': {'type': 'alias', 'field': 'extension'},
1402 'br
': {'type': 'alias', 'field': 'bitrate'},
1403 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1404 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1405 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1406 'framerate
': {'type': 'alias', 'field': 'fps'},
1407 'lang
': {'type': 'alias', 'field': 'language_preference'}, # not named as 'language
' because such a field exists
1408 'proto
': {'type': 'alias', 'field': 'protocol'},
1409 'source
': {'type': 'alias', 'field': 'source_preference'},
1410 'size
': {'type': 'alias', 'field': 'filesize_estimate'},
1411 'samplerate
': {'type': 'alias', 'field': 'asr'},
1412 'video_ext
': {'type': 'alias', 'field': 'vext'},
1413 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1414 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1415 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1416 'video
': {'type': 'alias', 'field': 'has_video'},
1417 'audio
': {'type': 'alias', 'field': 'has_audio'},
1418 'extractor
': {'type': 'alias', 'field': 'extractor_preference'},
1419 'preference
': {'type': 'alias', 'field': 'extractor_preference'}}
1423 def _get_field_setting(self, field, key):
1424 if field not in self.settings:
1425 self.settings[field] = {}
1426 propObj = self.settings[field]
1427 if key not in propObj:
1428 type = propObj.get('type')
1430 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1431 elif key == 'convert
':
1432 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1434 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,), 'function': max}.get(key, None)
1435 propObj[key] = default
1438 def _resolve_field_value(self, field, value, convertNone=False):
1443 value = value.lower()
1444 conversion = self._get_field_setting(field, 'convert
')
1445 if conversion == 'ignore
':
1447 if conversion == 'string
':
1449 elif conversion == 'float_none
':
1450 return float_or_none(value)
1451 elif conversion == 'bytes':
1452 return FileDownloader.parse_bytes(value)
1453 elif conversion == 'order
':
1454 order_free = self._get_field_setting(field, 'order_free
')
1455 order_list = order_free if order_free and self._use_free_order else self._get_field_setting(field, 'order
')
1456 use_regex = self._get_field_setting(field, 'regex
')
1457 list_length = len(order_list)
1458 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1459 if use_regex and value is not None:
1460 for (i, regex) in enumerate(order_list):
1461 if regex and re.match(regex, value):
1462 return list_length - i
1463 return list_length - empty_pos # not in list
1464 else: # not regex or value = None
1465 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1467 if value.isnumeric():
1470 self.settings[field]['convert
'] = 'string
'
1473 def evaluate_params(self, params, sort_extractor):
1474 self._use_free_order = params.get('prefer_free_formats
', False)
1475 self._sort_user = params.get('format_sort
', [])
1476 self._sort_extractor = sort_extractor
1478 def add_item(field, reverse, closest, limit_text):
1479 field = field.lower()
1480 if field in self._order:
1482 self._order.append(field)
1483 limit = self._resolve_field_value(field, limit_text)
1486 'closest
': False if limit is None else closest,
1487 'limit_text
': limit_text,
1489 if field in self.settings:
1490 self.settings[field].update(data)
1492 self.settings[field] = data
1495 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1496 + (tuple() if params.get('format_sort_force
', False)
1497 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1498 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1500 for item in sort_list:
1501 match = re.match(self.regex, item)
1503 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1504 field = match.group('field
')
1507 if self._get_field_setting(field, 'type') == 'alias
':
1508 field = self._get_field_setting(field, 'field
')
1509 reverse = match.group('reverse
') is not None
1510 closest = match.group('seperator
') == '~
'
1511 limit_text = match.group('limit
')
1513 has_limit = limit_text is not None
1514 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1515 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1517 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1518 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1519 limit_count = len(limits)
1520 for (i, f) in enumerate(fields):
1521 add_item(f, reverse, closest,
1522 limits[i] if i < limit_count
1523 else limits[0] if has_limit and not has_multiple_limits
1526 def print_verbose_info(self, to_screen):
1527 to_screen('[debug
] Sort order given by user
: %s' % ','.join(self._sort_user))
1528 if self._sort_extractor:
1529 to_screen('[debug
] Sort order given by extractor
: %s' % ','.join(self._sort_extractor))
1530 to_screen('[debug
] Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1531 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1532 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1533 self._get_field_setting(field, 'limit_text
'),
1534 self._get_field_setting(field, 'limit
'))
1535 if self._get_field_setting(field, 'limit_text
') is not None else '')
1536 for field in self._order if self._get_field_setting(field, 'visible
')]))
1538 def _calculate_field_preference_from_value(self, format, field, type, value):
1539 reverse = self._get_field_setting(field, 'reverse
')
1540 closest = self._get_field_setting(field, 'closest
')
1541 limit = self._get_field_setting(field, 'limit
')
1543 if type == 'extractor
':
1544 maximum = self._get_field_setting(field, 'max')
1545 if value is None or (maximum is not None and value >= maximum):
1547 elif type == 'boolean
':
1548 in_list = self._get_field_setting(field, 'in_list
')
1549 not_in_list = self._get_field_setting(field, 'not_in_list
')
1550 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
1551 elif type == 'ordered
':
1552 value = self._resolve_field_value(field, value, True)
1554 # try to convert to number
1555 val_num = float_or_none(value)
1556 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1560 return ((-10, 0) if value is None
1561 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1562 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1563 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1564 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1565 else (-1, value, 0))
1567 def _calculate_field_preference(self, format, field):
1568 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1569 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1570 if type == 'multiple
':
1571 type = 'field
' # Only 'field
' is allowed in multiple for now
1572 actual_fields = self._get_field_setting(field, 'field
')
1574 def wrapped_function(values):
1575 values = tuple(filter(lambda x: x is not None, values))
1576 return (self._get_field_setting(field, 'function
')(*values) if len(values) > 1
1577 else values[0] if values
1580 value = wrapped_function((get_value(f) for f in actual_fields))
1582 value = get_value(field)
1583 return self._calculate_field_preference_from_value(format, field, type, value)
1585 def calculate_preference(self, format):
1586 # Determine missing protocol
1587 if not format.get('protocol
'):
1588 format['protocol
'] = determine_protocol(format)
1590 # Determine missing ext
1591 if not format.get('ext
') and 'url
' in format:
1592 format['ext
'] = determine_ext(format['url
'])
1593 if format.get('vcodec
') == 'none
':
1594 format['audio_ext
'] = format['ext
']
1595 format['video_ext
'] = 'none
'
1597 format['video_ext
'] = format['ext
']
1598 format['audio_ext
'] = 'none
'
1599 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1600 # format['preference
'] = -1000
1602 # Determine missing bitrates
1603 if format.get('tbr
') is None:
1604 if format.get('vbr
') is not None and format.get('abr
') is not None:
1605 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1607 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1608 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1609 if format.get('acodec
') != "none" and format.get('abr
') is None:
1610 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1612 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1614 def _sort_formats(self, formats, field_preference=[]):
1616 raise ExtractorError('No video formats found
')
1617 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1618 format_sort.evaluate_params(self._downloader.params, field_preference)
1619 if self._downloader.params.get('verbose
', False):
1620 format_sort.print_verbose_info(self._downloader.to_screen)
1621 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1623 def _check_formats(self, formats, video_id):
1625 formats[:] = filter(
1626 lambda f: self._is_valid_url(
1628 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1632 def _remove_duplicate_formats(formats):
1636 if f['url
'] not in format_urls:
1637 format_urls.add(f['url
'])
1638 unique_formats.append(f)
1639 formats[:] = unique_formats
1641 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1642 url = self._proto_relative_url(url, scheme='http
:')
1643 # For now assume non HTTP(S) URLs always valid
1644 if not (url.startswith('http
://') or url.startswith('https
://')):
1647 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1649 except ExtractorError as e:
1651 '%s: %s URL
is invalid
, skipping
: %s'
1652 % (video_id, item, error_to_compat_str(e.cause)))
1655 def http_scheme(self):
1656 """ Either "http:" or "https:", depending on the user's preferences
"""
1659 if self._downloader.params.get('prefer_insecure', False)
1662 def _proto_relative_url(self, url, scheme=None):
1665 if url.startswith('//'):
1667 scheme = self.http_scheme()
1672 def _sleep(self, timeout, video_id, msg_template=None):
1673 if msg_template is None:
1674 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1675 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1679 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1680 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1681 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1682 manifest = self._download_xml(
1683 manifest_url, video_id, 'Downloading f4m manifest',
1684 'Unable to download f4m manifest',
1685 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1686 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1687 transform_source=transform_source,
1688 fatal=fatal, data=data, headers=headers, query=query)
1690 if manifest is False:
1693 return self._parse_f4m_formats(
1694 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1695 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1697 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1698 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1699 fatal=True, m3u8_id=None):
1700 if not isinstance(manifest, compat_etree_Element) and not fatal:
1703 # currently youtube-dlc cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1704 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1705 if akamai_pv is not None and ';' in akamai_pv.text:
1706 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1707 if playerVerificationChallenge.strip() != '':
1711 manifest_version = '1.0'
1712 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1714 manifest_version = '2.0'
1715 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1716 # Remove unsupported DRM protected media from final formats
1717 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1718 media_nodes = remove_encrypted_media(media_nodes)
1722 manifest_base_url = get_base_url(manifest)
1724 bootstrap_info = xpath_element(
1725 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1726 'bootstrap info', default=None)
1729 mime_type = xpath_text(
1730 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1731 'base URL', default=None)
1732 if mime_type and mime_type.startswith('audio/'):
1735 for i, media_el in enumerate(media_nodes):
1736 tbr = int_or_none(media_el.attrib.get('bitrate'))
1737 width = int_or_none(media_el.attrib.get('width'))
1738 height = int_or_none(media_el.attrib.get('height'))
1739 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1740 # If <bootstrapInfo> is present, the specified f4m is a
1741 # stream-level manifest, and only set-level manifests may refer to
1742 # external resources. See section 11.4 and section 4 of F4M spec
1743 if bootstrap_info is None:
1745 # @href is introduced in 2.0, see section 11.6 of F4M spec
1746 if manifest_version == '2.0':
1747 media_url = media_el.attrib.get('href')
1748 if media_url is None:
1749 media_url = media_el.attrib.get('url')
1753 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1754 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1755 # If media_url is itself a f4m manifest do the recursive extraction
1756 # since bitrates in parent manifest (this one) and media_url manifest
1757 # may differ leading to inability to resolve the format by requested
1758 # bitrate in f4m downloader
1759 ext = determine_ext(manifest_url)
1761 f4m_formats = self._extract_f4m_formats(
1762 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1763 transform_source=transform_source, fatal=fatal)
1764 # Sometimes stream-level manifest contains single media entry that
1765 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1766 # At the same time parent's media entry in set-level manifest may
1767 # contain it. We will copy it from parent in such cases.
1768 if len(f4m_formats) == 1:
1771 'tbr': f.get('tbr') or tbr,
1772 'width': f.get('width') or width,
1773 'height': f.get('height') or height,
1774 'format_id': f.get('format_id') if not tbr else format_id,
1777 formats.extend(f4m_formats)
1780 formats.extend(self._extract_m3u8_formats(
1781 manifest_url, video_id, 'mp4', preference=preference,
1782 m3u8_id=m3u8_id, fatal=fatal))
1785 'format_id': format_id,
1786 'url': manifest_url,
1787 'manifest_url': manifest_url,
1788 'ext': 'flv' if bootstrap_info is not None else None,
1794 'preference': preference,
1798 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1800 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1804 'preference': preference - 100 if preference else -100,
1805 'resolution': 'multiple',
1806 'format_note': 'Quality selection URL',
1809 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1810 entry_protocol='m3u8', preference=None,
1811 m3u8_id=None, note=None, errnote=None,
1812 fatal=True, live=False, data=None, headers={},
1814 res = self._download_webpage_handle(
1816 note=note or 'Downloading m3u8 information',
1817 errnote=errnote or 'Failed to download m3u8 information',
1818 fatal=fatal, data=data, headers=headers, query=query)
1823 m3u8_doc, urlh = res
1824 m3u8_url = urlh.geturl()
1826 return self._parse_m3u8_formats(
1827 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1828 preference=preference, m3u8_id=m3u8_id, live=live)
1830 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1831 entry_protocol='m3u8', preference=None,
1832 m3u8_id=None, live=False):
1833 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1836 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1841 format_url = lambda u: (
1843 if re.match(r'^https?://', u)
1844 else compat_urlparse.urljoin(m3u8_url, u))
1847 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1848 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1849 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1851 # We should try extracting formats only from master playlists [1, 4.3.4],
1852 # i.e. playlists that describe available qualities. On the other hand
1853 # media playlists [1, 4.3.3] should be returned as is since they contain
1854 # just the media without qualities renditions.
1855 # Fortunately, master playlist can be easily distinguished from media
1856 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1857 # master playlist tags MUST NOT appear in a media playlist and vice versa.
1858 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1859 # media playlist and MUST NOT appear in master playlist thus we can
1860 # clearly detect media playlist with this criterion.
1862 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1865 'format_id': m3u8_id,
1867 'protocol': entry_protocol,
1868 'preference': preference,
1872 last_stream_inf = {}
1874 def extract_media(x_media_line):
1875 media = parse_m3u8_attributes(x_media_line)
1876 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1877 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1878 if not (media_type and group_id and name):
1880 groups.setdefault(group_id, []).append(media)
1881 if media_type not in ('VIDEO', 'AUDIO'):
1883 media_url = media.get('URI')
1886 for v in (m3u8_id, group_id, name):
1890 'format_id': '-'.join(format_id),
1891 'url': format_url(media_url),
1892 'manifest_url': m3u8_url,
1893 'language': media.get('LANGUAGE'),
1895 'protocol': entry_protocol,
1896 'preference': preference,
1898 if media_type == 'AUDIO':
1899 f['vcodec'] = 'none'
1902 def build_stream_name():
1903 # Despite specification does not mention NAME attribute for
1904 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1905 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1906 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1907 stream_name = last_stream_inf.get('NAME')
1910 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1911 # from corresponding rendition group
1912 stream_group_id = last_stream_inf.get('VIDEO')
1913 if not stream_group_id:
1915 stream_group = groups.get(stream_group_id)
1916 if not stream_group:
1917 return stream_group_id
1918 rendition = stream_group[0]
1919 return rendition.get('NAME') or stream_group_id
1921 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1922 # chance to detect video only formats when EXT-X-STREAM-INF tags
1923 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1924 for line in m3u8_doc.splitlines():
1925 if line.startswith('#EXT-X-MEDIA:'):
1928 for line in m3u8_doc.splitlines():
1929 if line.startswith('#EXT-X-STREAM-INF:'):
1930 last_stream_inf = parse_m3u8_attributes(line)
1931 elif line.startswith('#') or not line.strip():
1934 tbr = float_or_none(
1935 last_stream_inf.get('AVERAGE-BANDWIDTH')
1936 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1939 format_id.append(m3u8_id)
1940 stream_name = build_stream_name()
1941 # Bandwidth of live streams may differ over time thus making
1942 # format_id unpredictable. So it's better to keep provided
1945 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1946 manifest_url = format_url(line.strip())
1948 'format_id': '-'.join(format_id),
1949 'url': manifest_url,
1950 'manifest_url': m3u8_url,
1953 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1954 'protocol': entry_protocol,
1955 'preference': preference,
1957 resolution = last_stream_inf.get('RESOLUTION')
1959 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1961 f['width'] = int(mobj.group('width'))
1962 f['height'] = int(mobj.group('height'))
1963 # Unified Streaming Platform
1965 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1967 abr, vbr = mobj.groups()
1968 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1973 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1975 audio_group_id = last_stream_inf.get('AUDIO')
1976 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1977 # references a rendition group MUST have a CODECS attribute.
1978 # However, this is not always respected, for example, [2]
1979 # contains EXT-X-STREAM-INF tag which references AUDIO
1980 # rendition group but does not have CODECS and despite
1981 # referencing an audio group it represents a complete
1982 # (with audio and video) format. So, for such cases we will
1983 # ignore references to rendition groups and treat them
1984 # as complete formats.
1985 if audio_group_id and codecs and f.get('vcodec') != 'none':
1986 audio_group = groups.get(audio_group_id)
1987 if audio_group and audio_group[0].get('URI'):
1988 # TODO: update acodec for audio only formats with
1990 f['acodec'] = 'none'
1994 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
1997 del http_f['manifest_url']
1999 'format_id': f['format_id'].replace('hls-', 'http-'),
2001 'url': progressive_uri,
2003 formats.append(http_f)
2005 last_stream_inf = {}
2009 def _xpath_ns(path, namespace=None):
2013 for c in path.split('/'):
2014 if not c or c == '.':
2017 out.append('{%s}%s' % (namespace, c))
2018 return '/'.join(out)
2020 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2021 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2027 namespace = self._parse_smil_namespace(smil)
2029 return self._parse_smil_formats(
2030 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2032 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2033 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2036 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2038 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2039 return self._download_xml(
2040 smil_url, video_id, 'Downloading SMIL file',
2041 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2043 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2044 namespace = self._parse_smil_namespace(smil)
2046 formats = self._parse_smil_formats(
2047 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2048 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2050 video_id = os.path.splitext(url_basename(smil_url))[0]
2054 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2055 name = meta.attrib.get('name')
2056 content = meta.attrib.get('content')
2057 if not name or not content:
2059 if not title and name == 'title':
2061 elif not description and name in ('description', 'abstract'):
2062 description = content
2063 elif not upload_date and name == 'date':
2064 upload_date = unified_strdate(content)
2067 'id': image.get('type'),
2068 'url': image.get('src'),
2069 'width': int_or_none(image.get('width')),
2070 'height': int_or_none(image.get('height')),
2071 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2075 'title': title or video_id,
2076 'description': description,
2077 'upload_date': upload_date,
2078 'thumbnails': thumbnails,
2080 'subtitles': subtitles,
2083 def _parse_smil_namespace(self, smil):
2084 return self._search_regex(
2085 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2087 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2089 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2090 b = meta.get('base') or meta.get('httpBase')
2101 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2102 for medium in media:
2103 src = medium.get('src')
2104 if not src or src in srcs:
2108 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2109 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2110 width = int_or_none(medium.get('width'))
2111 height = int_or_none(medium.get('height'))
2112 proto = medium.get('proto')
2113 ext = medium.get('ext')
2114 src_ext = determine_ext(src)
2115 streamer = medium.get('streamer') or base
2117 if proto == 'rtmp' or streamer.startswith('rtmp'):
2123 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2125 'filesize': filesize,
2129 if transform_rtmp_url:
2130 streamer, src = transform_rtmp_url(streamer, src)
2131 formats[-1].update({
2137 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2138 src_url = src_url.strip()
2140 if proto == 'm3u8' or src_ext == 'm3u8':
2141 m3u8_formats = self._extract_m3u8_formats(
2142 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2143 if len(m3u8_formats) == 1:
2145 m3u8_formats[0].update({
2146 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2151 formats.extend(m3u8_formats)
2152 elif src_ext == 'f4m':
2157 'plugin': 'flowplayer-3.2.0.1',
2159 f4m_url += '&' if '?' in f4m_url else '?'
2160 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2161 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2162 elif src_ext == 'mpd':
2163 formats.extend(self._extract_mpd_formats(
2164 src_url, video_id, mpd_id='dash', fatal=False))
2165 elif re.search(r'\.ism/[Mm]anifest', src_url):
2166 formats.extend(self._extract_ism_formats(
2167 src_url, video_id, ism_id='mss', fatal=False))
2168 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2172 'ext': ext or src_ext or 'flv',
2173 'format_id': 'http-%d' % (bitrate or http_count),
2175 'filesize': filesize,
2182 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2185 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2186 src = textstream.get('src')
2187 if not src or src in urls:
2190 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2191 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2192 subtitles.setdefault(lang, []).append({
2198 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2199 xspf = self._download_xml(
2200 xspf_url, playlist_id, 'Downloading xpsf playlist',
2201 'Unable to download xspf manifest', fatal=fatal)
2204 return self._parse_xspf(
2205 xspf, playlist_id, xspf_url=xspf_url,
2206 xspf_base_url=base_url(xspf_url))
2208 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2210 'xspf': 'http://xspf.org/ns/0/',
2211 's1': 'http://static.streamone.nl/player/ns/0',
2215 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2217 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2218 description = xpath_text(
2219 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2220 thumbnail = xpath_text(
2221 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2222 duration = float_or_none(
2223 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2226 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2227 format_url = urljoin(xspf_base_url, location.text)
2232 'manifest_url': xspf_url,
2233 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2234 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2235 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2237 self._sort_formats(formats)
2242 'description': description,
2243 'thumbnail': thumbnail,
2244 'duration': duration,
2249 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
2250 res = self._download_xml_handle(
2252 note=note or 'Downloading MPD manifest',
2253 errnote=errnote or 'Failed to download MPD manifest',
2254 fatal=fatal, data=data, headers=headers, query=query)
2260 mpd_base_url = base_url(urlh.geturl())
2262 return self._parse_mpd_formats(
2263 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2264 formats_dict=formats_dict, mpd_url=mpd_url)
2266 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2268 Parse formats
from MPD manifest
.
2270 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2271 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2272 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2274 if not self._downloader.params.get('dynamic_mpd'):
2275 if mpd_doc.get('type') == 'dynamic':
2278 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2281 return self._xpath_ns(path, namespace)
2283 def is_drm_protected(element):
2284 return element.find(_add_ns('ContentProtection')) is not None
2286 def extract_multisegment_info(element, ms_parent_info):
2287 ms_info = ms_parent_info.copy()
2289 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2290 # common attributes and elements. We will only extract relevant
2292 def extract_common(source):
2293 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2294 if segment_timeline is not None:
2295 s_e = segment_timeline.findall(_add_ns('S'))
2297 ms_info['total_number'] = 0
2300 r = int(s.get('r', 0))
2301 ms_info['total_number'] += 1 + r
2302 ms_info['s'].append({
2303 't': int(s.get('t', 0)),
2304 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2305 'd': int(s.attrib['d']),
2308 start_number = source.get('startNumber')
2310 ms_info['start_number'] = int(start_number)
2311 timescale = source.get('timescale')
2313 ms_info['timescale'] = int(timescale)
2314 segment_duration = source.get('duration')
2315 if segment_duration:
2316 ms_info['segment_duration'] = float(segment_duration)
2318 def extract_Initialization(source):
2319 initialization = source.find(_add_ns('Initialization'))
2320 if initialization is not None:
2321 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2323 segment_list = element.find(_add_ns('SegmentList'))
2324 if segment_list is not None:
2325 extract_common(segment_list)
2326 extract_Initialization(segment_list)
2327 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2329 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2331 segment_template = element.find(_add_ns('SegmentTemplate'))
2332 if segment_template is not None:
2333 extract_common(segment_template)
2334 media = segment_template.get('media')
2336 ms_info['media'] = media
2337 initialization = segment_template.get('initialization')
2339 ms_info['initialization'] = initialization
2341 extract_Initialization(segment_template)
2344 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2346 for period in mpd_doc.findall(_add_ns('Period')):
2347 period_duration = parse_duration(period.get('duration')) or mpd_duration
2348 period_ms_info = extract_multisegment_info(period, {
2352 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2353 if is_drm_protected(adaptation_set):
2355 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2356 for representation in adaptation_set.findall(_add_ns('Representation')):
2357 if is_drm_protected(representation):
2359 representation_attrib = adaptation_set.attrib.copy()
2360 representation_attrib.update(representation.attrib)
2361 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2362 mime_type = representation_attrib['mimeType']
2363 content_type = mime_type.split('/')[0]
2364 if content_type == 'text':
2365 # TODO implement WebVTT downloading
2367 elif content_type in ('video', 'audio'):
2369 for element in (representation, adaptation_set, period, mpd_doc):
2370 base_url_e = element.find(_add_ns('BaseURL'))
2371 if base_url_e is not None:
2372 base_url = base_url_e.text + base_url
2373 if re.match(r'^https?://', base_url):
2375 if mpd_base_url and not re.match(r'^https?://', base_url):
2376 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2378 base_url = mpd_base_url + base_url
2379 representation_id = representation_attrib.get('id')
2380 lang = representation_attrib.get('lang')
2381 url_el = representation.find(_add_ns('BaseURL'))
2382 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2383 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2385 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2386 'manifest_url': mpd_url,
2387 'ext': mimetype2ext(mime_type),
2388 'width': int_or_none(representation_attrib.get('width')),
2389 'height': int_or_none(representation_attrib.get('height')),
2390 'tbr': float_or_none(bandwidth, 1000),
2391 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2392 'fps': int_or_none(representation_attrib.get('frameRate')),
2393 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2394 'format_note': 'DASH %s' % content_type,
2395 'filesize': filesize,
2396 'container': mimetype2ext(mime_type) + '_dash',
2398 f.update(parse_codecs(representation_attrib.get('codecs')))
2399 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2401 def prepare_template(template_name, identifiers):
2402 tmpl = representation_ms_info[template_name]
2403 # First of, % characters outside $...$ templates
2404 # must be escaped by doubling for proper processing
2405 # by % operator string formatting used further (see
2406 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2412 in_template = not in_template
2413 elif c == '%' and not in_template:
2415 # Next, $...$ templates are translated to their
2416 # %(...) counterparts to be used with % operator
2417 t = t.replace('$RepresentationID$', representation_id)
2418 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2419 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2420 t.replace('$$', '$')
2423 # @initialization is a regular template like @media one
2424 # so it should be handled just the same way (see
2425 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2426 if 'initialization' in representation_ms_info:
2427 initialization_template = prepare_template(
2429 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2430 # $Time$ shall not be included for @initialization thus
2431 # only $Bandwidth$ remains
2433 representation_ms_info['initialization_url'] = initialization_template % {
2434 'Bandwidth': bandwidth,
2437 def location_key(location):
2438 return 'url' if re.match(r'^https?://', location) else 'path'
2440 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2442 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2443 media_location_key = location_key(media_template)
2445 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2446 # can't be used at the same time
2447 if '%(Number' in media_template and 's' not in representation_ms_info:
2448 segment_duration = None
2449 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2450 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2451 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2452 representation_ms_info['fragments'] = [{
2453 media_location_key: media_template % {
2454 'Number': segment_number,
2455 'Bandwidth': bandwidth,
2457 'duration': segment_duration,
2458 } for segment_number in range(
2459 representation_ms_info['start_number'],
2460 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2462 # $Number*$ or $Time$ in media template with S list available
2463 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2464 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2465 representation_ms_info['fragments'] = []
2468 segment_number = representation_ms_info['start_number']
2470 def add_segment_url():
2471 segment_url = media_template % {
2472 'Time': segment_time,
2473 'Bandwidth': bandwidth,
2474 'Number': segment_number,
2476 representation_ms_info['fragments'].append({
2477 media_location_key: segment_url,
2478 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2481 for num, s in enumerate(representation_ms_info['s']):
2482 segment_time = s.get('t') or segment_time
2486 for r in range(s.get('r', 0)):
2487 segment_time += segment_d
2490 segment_time += segment_d
2491 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2493 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2494 # or any YouTube dashsegments video
2497 timescale = representation_ms_info['timescale']
2498 for s in representation_ms_info['s']:
2499 duration = float_or_none(s['d'], timescale)
2500 for r in range(s.get('r', 0) + 1):
2501 segment_uri = representation_ms_info['segment_urls'][segment_index]
2503 location_key(segment_uri): segment_uri,
2504 'duration': duration,
2507 representation_ms_info['fragments'] = fragments
2508 elif 'segment_urls' in representation_ms_info:
2509 # Segment URLs with no SegmentTimeline
2510 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2511 # https://github.com/ytdl-org/youtube-dl/pull/14844
2513 segment_duration = float_or_none(
2514 representation_ms_info['segment_duration'],
2515 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2516 for segment_url in representation_ms_info['segment_urls']:
2518 location_key(segment_url): segment_url,
2520 if segment_duration:
2521 fragment['duration'] = segment_duration
2522 fragments.append(fragment)
2523 representation_ms_info['fragments'] = fragments
2524 # If there is a fragments key available then we correctly recognized fragmented media.
2525 # Otherwise we will assume unfragmented media with direct access. Technically, such
2526 # assumption is not necessarily correct since we may simply have no support for
2527 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2528 if 'fragments' in representation_ms_info:
2530 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2531 'url': mpd_url or base_url,
2532 'fragment_base_url': base_url,
2534 'protocol': 'http_dash_segments',
2536 if 'initialization_url' in representation_ms_info:
2537 initialization_url = representation_ms_info['initialization_url']
2538 if not f.get('url'):
2539 f['url'] = initialization_url
2540 f['fragments'].append({location_key(initialization_url): initialization_url})
2541 f['fragments'].extend(representation_ms_info['fragments'])
2543 # Assuming direct URL to unfragmented media.
2546 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2547 # is not necessarily unique within a Period thus formats with
2548 # the same `format_id` are quite possible. There are numerous examples
2549 # of such manifests (see https://github.com/ytdl-org/youtube-dl/issues/15111,
2550 # https://github.com/ytdl-org/youtube-dl/issues/13919)
2551 full_info = formats_dict.get(representation_id, {}).copy()
2553 formats.append(full_info)
2555 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2558 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2559 res = self._download_xml_handle(
2561 note=note or 'Downloading ISM manifest',
2562 errnote=errnote or 'Failed to download ISM manifest',
2563 fatal=fatal, data=data, headers=headers, query=query)
2570 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2572 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2574 Parse formats
from ISM manifest
.
2576 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2577 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2579 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2582 duration = int(ism_doc.attrib['Duration'])
2583 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2586 for stream in ism_doc.findall('StreamIndex'):
2587 stream_type = stream.get('Type')
2588 if stream_type not in ('video', 'audio'):
2590 url_pattern = stream.attrib['Url']
2591 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2592 stream_name = stream.get('Name')
2593 for track in stream.findall('QualityLevel'):
2594 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2595 # TODO: add support for WVC1 and WMAP
2596 if fourcc not in ('H264', 'AVC1', 'AACL'):
2597 self.report_warning('%s is not a supported codec' % fourcc)
2599 tbr = int(track.attrib['Bitrate']) // 1000
2600 # [1] does not mention Width and Height attributes. However,
2601 # they're often present while MaxWidth and MaxHeight are
2602 # missing, so should be used as fallbacks
2603 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2604 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2605 sampling_rate = int_or_none(track.get('SamplingRate'))
2607 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2608 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2614 stream_fragments = stream.findall('c')
2615 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2616 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2617 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2618 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2619 if not fragment_ctx['duration']:
2621 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2623 next_fragment_time = duration
2624 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2625 for _ in range(fragment_repeat):
2627 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2628 'duration': fragment_ctx['duration'] / stream_timescale,
2630 fragment_ctx['time'] += fragment_ctx['duration']
2634 format_id.append(ism_id)
2636 format_id.append(stream_name)
2637 format_id.append(compat_str(tbr))
2640 'format_id': '-'.join(format_id),
2642 'manifest_url': ism_url,
2643 'ext': 'ismv' if stream_type == 'video' else 'isma',
2647 'asr': sampling_rate,
2648 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2649 'acodec': 'none' if stream_type == 'video' else fourcc,
2651 'fragments': fragments,
2652 '_download_params': {
2653 'duration': duration,
2654 'timescale': stream_timescale,
2655 'width': width or 0,
2656 'height': height or 0,
2658 'codec_private_data': track.get('CodecPrivateData'),
2659 'sampling_rate': sampling_rate,
2660 'channels': int_or_none(track.get('Channels', 2)),
2661 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2662 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2667 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2668 def absolute_url(item_url):
2669 return urljoin(base_url, item_url)
2671 def parse_content_type(content_type):
2672 if not content_type:
2674 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2676 mimetype, codecs = ctr.groups()
2677 f = parse_codecs(codecs)
2678 f['ext'] = mimetype2ext(mimetype)
2682 def _media_formats(src, cur_media_type, type_info={}):
2683 full_url = absolute_url(src)
2684 ext = type_info.get('ext') or determine_ext(full_url)
2686 is_plain_url = False
2687 formats = self._extract_m3u8_formats(
2688 full_url, video_id, ext='mp4',
2689 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2690 preference=preference, fatal=False)
2692 is_plain_url = False
2693 formats = self._extract_mpd_formats(
2694 full_url, video_id, mpd_id=mpd_id, fatal=False)
2699 'vcodec': 'none' if cur_media_type == 'audio' else None,
2701 return is_plain_url, formats
2704 # amp-video and amp-audio are very similar to their HTML5 counterparts
2705 # so we wll include them right here (see
2706 # https://www.ampproject.org/docs/reference/components/amp-video)
2707 media_tags = [(media_tag, media_type, '')
2708 for media_tag, media_type
2709 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2710 media_tags.extend(re.findall(
2711 # We only allow video|audio followed by a whitespace or '>'.
2712 # Allowing more characters may end up in significant slow down (see
2713 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2714 # http://www.porntrex.com/maps/videositemap.xml).
2715 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2716 for media_tag, media_type, media_content in media_tags:
2721 media_attributes = extract_attributes(media_tag)
2722 src = strip_or_none(media_attributes.get('src'))
2724 _, formats = _media_formats(src, media_type)
2725 media_info['formats'].extend(formats)
2726 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2728 for source_tag in re.findall(r'<source[^>]+>', media_content):
2729 s_attr = extract_attributes(source_tag)
2730 # data-video-src and data-src are non standard but seen
2731 # several times in the wild
2732 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2735 f = parse_content_type(s_attr.get('type'))
2736 is_plain_url, formats = _media_formats(src, media_type, f)
2738 # width, height, res, label and title attributes are
2739 # all not standard but seen several times in the wild
2742 for lbl in ('label', 'title')
2743 if str_or_none(s_attr.get(lbl))
2745 width = int_or_none(s_attr.get('width'))
2746 height = (int_or_none(s_attr.get('height'))
2747 or int_or_none(s_attr.get('res')))
2748 if not width or not height:
2750 resolution = parse_resolution(lbl)
2753 width = width or resolution.get('width')
2754 height = height or resolution.get('height')
2756 tbr = parse_bitrate(lbl)
2765 'format_id': s_attr.get('label') or s_attr.get('title'),
2767 f.update(formats[0])
2768 media_info['formats'].append(f)
2770 media_info['formats'].extend(formats)
2771 for track_tag in re.findall(r'<track[^>]+>', media_content):
2772 track_attributes = extract_attributes(track_tag)
2773 kind = track_attributes.get('kind')
2774 if not kind or kind in ('subtitles', 'captions'):
2775 src = strip_or_none(track_attributes.get('src'))
2778 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2779 media_info['subtitles'].setdefault(lang, []).append({
2780 'url': absolute_url(src),
2782 for f in media_info['formats']:
2783 f.setdefault('http_headers', {})['Referer'] = base_url
2784 if media_info['formats'] or media_info['subtitles']:
2785 entries.append(media_info)
2788 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2791 hdcore_sign = 'hdcore=3.7.0'
2792 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2793 hds_host = hosts.get('hds')
2795 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2796 if 'hdcore=' not in f4m_url:
2797 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2798 f4m_formats = self._extract_f4m_formats(
2799 f4m_url, video_id, f4m_id='hds', fatal=False)
2800 for entry in f4m_formats:
2801 entry.update({'extra_param_to_segment_url': hdcore_sign})
2802 formats.extend(f4m_formats)
2804 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2805 hls_host = hosts.get('hls')
2807 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2808 formats.extend(self._extract_m3u8_formats(
2809 m3u8_url, video_id, 'mp4', 'm3u8_native',
2810 m3u8_id='hls', fatal=False))
2812 http_host = hosts.get('http')
2813 if http_host and 'hdnea=' not in manifest_url:
2814 REPL_REGEX = r'https://[^/]+/i/([^,]+),([^/]+),([^/]+).csmil/.+'
2815 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
2816 qualities_length = len(qualities)
2817 if len(formats) in (qualities_length + 1, qualities_length * 2 + 1):
2821 if f['protocol'] == 'm3u8_native' and f['vcodec'] != 'none':
2822 for protocol in ('http', 'https'):
2824 del http_f['manifest_url']
2826 REPL_REGEX, protocol + r'://%s/\1%s\3' % (http_host, qualities[i]), f['url'])
2828 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
2830 'protocol': protocol,
2832 http_formats.append(http_f)
2834 formats.extend(http_formats)
2838 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2839 query = compat_urlparse.urlparse(url).query
2840 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2842 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2843 url_base = mobj.group('url')
2844 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2847 def manifest_url(manifest):
2848 m_url = '%s/%s' % (http_base_url, manifest)
2850 m_url += '?%s' % query
2853 if 'm3u8' not in skip_protocols:
2854 formats.extend(self._extract_m3u8_formats(
2855 manifest_url('playlist.m3u8'), video_id, 'mp4',
2856 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2857 if 'f4m' not in skip_protocols:
2858 formats.extend(self._extract_f4m_formats(
2859 manifest_url('manifest.f4m'),
2860 video_id, f4m_id='hds', fatal=False))
2861 if 'dash' not in skip_protocols:
2862 formats.extend(self._extract_mpd_formats(
2863 manifest_url('manifest.mpd'),
2864 video_id, mpd_id='dash', fatal=False))
2865 if re.search(r'(?:/smil:|\.smil)', url_base):
2866 if 'smil' not in skip_protocols:
2867 rtmp_formats = self._extract_smil_formats(
2868 manifest_url('jwplayer.smil'),
2869 video_id, fatal=False)
2870 for rtmp_format in rtmp_formats:
2871 rtsp_format = rtmp_format.copy()
2872 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2873 del rtsp_format['play_path']
2874 del rtsp_format['ext']
2875 rtsp_format.update({
2876 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2877 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2880 formats.extend([rtmp_format, rtsp_format])
2882 for protocol in ('rtmp', 'rtsp'):
2883 if protocol not in skip_protocols:
2885 'url': '%s:%s' % (protocol, url_base),
2886 'format_id': protocol,
2887 'protocol': protocol,
2891 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2893 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2897 jwplayer_data = self._parse_json(mobj.group('options'),
2899 transform_source=transform_source)
2900 except ExtractorError:
2903 if isinstance(jwplayer_data, dict):
2904 return jwplayer_data
2906 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2907 jwplayer_data = self._find_jwplayer_data(
2908 webpage, video_id, transform_source=js_to_json)
2909 return self._parse_jwplayer_data(
2910 jwplayer_data, video_id, *args, **kwargs)
2912 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2913 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2914 # JWPlayer backward compatibility: flattened playlists
2915 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2916 if 'playlist' not in jwplayer_data:
2917 jwplayer_data = {'playlist': [jwplayer_data]}
2921 # JWPlayer backward compatibility: single playlist item
2922 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2923 if not isinstance(jwplayer_data['playlist'], list):
2924 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2926 for video_data in jwplayer_data['playlist']:
2927 # JWPlayer backward compatibility: flattened sources
2928 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2929 if 'sources' not in video_data:
2930 video_data['sources'] = [video_data]
2932 this_video_id = video_id or video_data['mediaid']
2934 formats = self._parse_jwplayer_formats(
2935 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2936 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2939 tracks = video_data.get('tracks')
2940 if tracks and isinstance(tracks, list):
2941 for track in tracks:
2942 if not isinstance(track, dict):
2944 track_kind = track.get('kind')
2945 if not track_kind or not isinstance(track_kind, compat_str):
2947 if track_kind.lower() not in ('captions', 'subtitles'):
2949 track_url = urljoin(base_url, track.get('file'))
2952 subtitles.setdefault(track.get('label') or 'en', []).append({
2953 'url': self._proto_relative_url(track_url)
2957 'id': this_video_id,
2958 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2959 'description': clean_html(video_data.get('description')),
2960 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2961 'timestamp': int_or_none(video_data.get('pubdate')),
2962 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2963 'subtitles': subtitles,
2965 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2966 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2968 '_type': 'url_transparent',
2969 'url': formats[0]['url'],
2972 self._sort_formats(formats)
2973 entry['formats'] = formats
2974 entries.append(entry)
2975 if len(entries) == 1:
2978 return self.playlist_result(entries)
2980 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2981 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2984 for source in jwplayer_sources_data:
2985 if not isinstance(source, dict):
2987 source_url = urljoin(
2988 base_url, self._proto_relative_url(source.get('file')))
2989 if not source_url or source_url in urls:
2991 urls.append(source_url)
2992 source_type = source.get('type') or ''
2993 ext = mimetype2ext(source_type) or determine_ext(source_url)
2994 if source_type == 'hls' or ext == 'm3u8':
2995 formats.extend(self._extract_m3u8_formats(
2996 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2997 m3u8_id=m3u8_id, fatal=False))
2998 elif source_type == 'dash' or ext == 'mpd':
2999 formats.extend(self._extract_mpd_formats(
3000 source_url, video_id, mpd_id=mpd_id, fatal=False))
3002 formats.extend(self._extract_smil_formats(
3003 source_url, video_id, fatal=False))
3004 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3005 elif source_type.startswith('audio') or ext in (
3006 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3013 height = int_or_none(source.get('height'))
3015 # Often no height is provided but there is a label in
3016 # format like "1080p", "720p SD", or 1080.
3017 height = int_or_none(self._search_regex(
3018 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3019 'height', default=None))
3022 'width': int_or_none(source.get('width')),
3024 'tbr': int_or_none(source.get('bitrate')),
3027 if source_url.startswith('rtmp'):
3028 a_format['ext'] = 'flv'
3029 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3030 # of jwplayer.flash.swf
3031 rtmp_url_parts = re.split(
3032 r'((?:mp4|mp3|flv):)', source_url, 1)
3033 if len(rtmp_url_parts) == 3:
3034 rtmp_url, prefix, play_path = rtmp_url_parts
3037 'play_path': prefix + play_path,
3040 a_format.update(rtmp_params)
3041 formats.append(a_format)
3044 def _live_title(self, name):
3045 """ Generate the title
for a live video
"""
3046 now = datetime.datetime.now()
3047 now_str = now.strftime('%Y-%m-%d %H:%M')
3048 return name + ' ' + now_str
3050 def _int(self, v, name, fatal=False, **kwargs):
3051 res = int_or_none(v, **kwargs)
3052 if 'get_attr' in kwargs:
3053 print(getattr(v, kwargs['get_attr']))
3055 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3057 raise ExtractorError(msg)
3059 self._downloader.report_warning(msg)
3062 def _float(self, v, name, fatal=False, **kwargs):
3063 res = float_or_none(v, **kwargs)
3065 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3067 raise ExtractorError(msg)
3069 self._downloader.report_warning(msg)
3072 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3073 path='/', secure=False, discard=False, rest={}, **kwargs):
3074 cookie = compat_cookiejar_Cookie(
3075 0, name, value, port, port is not None, domain, True,
3076 domain.startswith('.'), path, True, secure, expire_time,
3077 discard, None, None, rest)
3078 self._downloader.cookiejar.set_cookie(cookie)
3080 def _get_cookies(self, url):
3081 """ Return a compat_cookies
.SimpleCookie
with the cookies
for the url
"""
3082 req = sanitized_Request(url)
3083 self._downloader.cookiejar.add_cookie_header(req)
3084 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
3086 def _apply_first_set_cookie_header(self, url_handle, cookie):
3088 Apply first Set
-Cookie header instead of the last
. Experimental
.
3090 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3091 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3092 than
second (new
). However
, as of RFC6265 the newer one cookie
3093 should be
set into cookie store what actually happens
.
3094 We will workaround this issue by resetting the cookie to
3095 the first one manually
.
3096 1. https
://new
.vk
.com
/
3097 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3098 3. https
://learning
.oreilly
.com
/
3100 for header, cookies in url_handle.headers.items():
3101 if header.lower() != 'set-cookie':
3103 if sys.version_info[0] >= 3:
3104 cookies = cookies.encode('iso-8859-1')
3105 cookies = cookies.decode('utf-8')
3106 cookie_value = re.search(
3107 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3109 value, domain = cookie_value.groups()
3110 self._set_cookie(domain, cookie, value)
3113 def get_testcases(self, include_onlymatching=False):
3114 t = getattr(self, '_TEST', None)
3116 assert not hasattr(self, '_TESTS'), \
3117 '%s has _TEST and _TESTS' % type(self).__name__
3120 tests = getattr(self, '_TESTS', [])
3122 if not include_onlymatching and t.get('only_matching', False):
3124 t['name'] = type(self).__name__[:-len('IE')]
3127 def is_suitable(self, age_limit):
3128 """ Test whether the extractor
is generally suitable
for the given
3129 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3131 any_restricted = False
3132 for tc in self.get_testcases(include_onlymatching=False):
3133 if tc.get('playlist', []):
3134 tc = tc['playlist'][0]
3135 is_restricted = age_restricted(
3136 tc.get('info_dict', {}).get('age_limit'), age_limit)
3137 if not is_restricted:
3139 any_restricted = any_restricted or is_restricted
3140 return not any_restricted
3142 def extract_subtitles(self, *args, **kwargs):
3143 if (self._downloader.params.get('writesubtitles', False)
3144 or self._downloader.params.get('listsubtitles')):
3145 return self._get_subtitles(*args, **kwargs)
3148 def _get_subtitles(self, *args, **kwargs):
3149 raise NotImplementedError('This method must be implemented by subclasses')
3152 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3153 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3154 will be dropped
. """
3155 list1_urls = set([item['url'] for item in subtitle_list1])
3156 ret = list(subtitle_list1)
3157 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3161 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
3162 """ Merge two subtitle dictionaries
, language by language
. """
3163 ret = dict(subtitle_dict1)
3164 for lang in subtitle_dict2:
3165 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
3168 def extract_automatic_captions(self, *args, **kwargs):
3169 if (self._downloader.params.get('writeautomaticsub', False)
3170 or self._downloader.params.get('listsubtitles')):
3171 return self._get_automatic_captions(*args, **kwargs)
3174 def _get_automatic_captions(self, *args, **kwargs):
3175 raise NotImplementedError('This method must be implemented by subclasses')
3177 def mark_watched(self, *args, **kwargs):
3178 if (self._downloader.params.get('mark_watched', False)
3179 and (self._get_login_info()[0] is not None
3180 or self._downloader.params.get('cookiefile') is not None)):
3181 self._mark_watched(*args, **kwargs)
3183 def _mark_watched(self, *args, **kwargs):
3184 raise NotImplementedError('This method must be implemented by subclasses')
3186 def geo_verification_headers(self):
3188 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
3189 if geo_verification_proxy:
3190 headers['Ytdl-request-proxy'] = geo_verification_proxy
3193 def _generic_id(self, url):
3194 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3196 def _generic_title(self, url):
3197 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3200 class SearchInfoExtractor(InfoExtractor):
3202 Base
class for paged search queries extractors
.
3203 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3204 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3208 def _make_valid_url(cls):
3209 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3212 def suitable(cls, url):
3213 return re.match(cls._make_valid_url(), url) is not None
3215 def _real_extract(self, query):
3216 mobj = re.match(self._make_valid_url(), query)
3218 raise ExtractorError('Invalid search query "%s"' % query)
3220 prefix = mobj.group('prefix')
3221 query = mobj.group('query')
3223 return self._get_n_results(query, 1)
3224 elif prefix == 'all':
3225 return self._get_n_results(query, self._MAX_RESULTS)
3229 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3230 elif n > self._MAX_RESULTS:
3231 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3232 n = self._MAX_RESULTS
3233 return self._get_n_results(query, n)
3235 def _get_n_results(self, query, n):
3236 """Get a specified number of results
for a query
"""
3237 raise NotImplementedError('This method must be implemented by subclasses')
3240 def SEARCH_KEY(self):
3241 return self._SEARCH_KEY