2 from __future__
import unicode_literals
17 from ..compat
import (
20 compat_etree_fromstring
,
26 compat_urllib_parse_unquote
,
27 compat_urllib_parse_urlencode
,
28 compat_urllib_request
,
30 compat_xml_parse_error
,
32 from ..downloader
.f4m
import (
34 remove_encrypted_media
,
59 parse_m3u8_attributes
,
76 class InfoExtractor(object):
77 """Information Extractor class.
79 Information extractors are the classes that, given a URL, extract
80 information about the video (or videos) the URL refers to. This
81 information includes the real video URL, the video title, author and
82 others. The information is stored in a dictionary which is then
83 passed to the YoutubeDL. The YoutubeDL processes this
84 information possibly downloading the video to the file system, among
85 other possible outcomes.
87 The type field determines the type of the result.
88 By far the most common value (and the default if _type is missing) is
89 "video", which indicates a single video.
91 For a video, the dictionaries must include the following fields:
94 title: Video title, unescaped.
96 Additionally, it must contain either a formats entry or a url one:
98 formats: A list of dictionaries for each format available, ordered
99 from worst to best quality.
102 * url Mandatory. The URL of the video file
104 The URL of the manifest file in case of
105 fragmented media (DASH, hls, hds)
106 * ext Will be calculated from URL if missing
107 * format A human-readable description of the format
108 ("mp4 container with h264/opus").
109 Calculated from the format_id, width, height.
110 and format_note fields if missing.
111 * format_id A short description of the format
112 ("mp4_h264_opus" or "19").
113 Technically optional, but strongly recommended.
114 * format_note Additional info about the format
115 ("3D" or "DASH video")
116 * width Width of the video, if known
117 * height Height of the video, if known
118 * resolution Textual description of width and height
119 * tbr Average bitrate of audio and video in KBit/s
120 * abr Average audio bitrate in KBit/s
121 * acodec Name of the audio codec in use
122 * asr Audio sampling rate in Hertz
123 * vbr Average video bitrate in KBit/s
125 * vcodec Name of the video codec in use
126 * container Name of the container format
127 * filesize The number of bytes, if known in advance
128 * filesize_approx An estimate for the number of bytes
129 * player_url SWF Player URL (used for rtmpdump).
130 * protocol The protocol that will be used for the actual
131 download, lower-case.
132 "http", "https", "rtsp", "rtmp", "rtmpe",
133 "m3u8", "m3u8_native" or "http_dash_segments".
135 Base URL for fragments. Each fragment's path
136 value (if present) will be relative to
138 * fragments A list of fragments of a fragmented media.
139 Each fragment entry must contain either an url
140 or a path. If an url is present it should be
141 considered by a client. Otherwise both path and
142 fragment_base_url must be present. Here is
143 the list of all potential fields:
144 * "url" - fragment's URL
145 * "path" - fragment's path relative to
147 * "duration" (optional, int or float)
148 * "filesize" (optional, int)
149 * preference Order number of this format. If this field is
150 present and not None, the formats get sorted
151 by this field, regardless of all other values.
152 -1 for default (order by other properties),
153 -2 or smaller for less than default.
154 < -1000 to hide the format (if there is
155 another one which is strictly better)
156 * language Language code, e.g. "de" or "en-US".
157 * language_preference Is this in the language mentioned in
159 10 if it's what the URL is about,
160 -1 for default (don't know),
161 -10 otherwise, other values reserved for now.
162 * quality Order number of the video quality of this
163 format, irrespective of the file format.
164 -1 for default (order by other properties),
165 -2 or smaller for less than default.
166 * source_preference Order number for this video source
167 (quality takes higher priority)
168 -1 for default (order by other properties),
169 -2 or smaller for less than default.
170 * http_headers A dictionary of additional HTTP headers
171 to add to the request.
172 * stretched_ratio If given and not 1, indicates that the
173 video's pixels are not square.
174 width : height ratio as float.
175 * no_resume The server does not support resuming the
176 (HTTP or RTMP) download. Boolean.
178 url: Final video URL.
179 ext: Video filename extension.
180 format: The video format, defaults to ext (used for --get-format)
181 player_url: SWF Player URL (used for rtmpdump).
183 The following fields are optional:
185 alt_title: A secondary title of the video.
186 display_id An alternative identifier for the video, not necessarily
187 unique, but available before title. Typically, id is
188 something like "4234987", title "Dancing naked mole rats",
189 and display_id "dancing-naked-mole-rats"
190 thumbnails: A list of dictionaries, with the following entries:
191 * "id" (optional, string) - Thumbnail format ID
193 * "preference" (optional, int) - quality of the image
194 * "width" (optional, int)
195 * "height" (optional, int)
196 * "resolution" (optional, string "{width}x{height"},
198 * "filesize" (optional, int)
199 thumbnail: Full URL to a video thumbnail image.
200 description: Full video description.
201 uploader: Full name of the video uploader.
202 license: License name the video is licensed under.
203 creator: The creator of the video.
204 release_date: The date (YYYYMMDD) when the video was released.
205 timestamp: UNIX timestamp of the moment the video became available.
206 upload_date: Video upload date (YYYYMMDD).
207 If not explicitly set, calculated from timestamp.
208 uploader_id: Nickname or id of the video uploader.
209 uploader_url: Full URL to a personal webpage of the video uploader.
210 location: Physical location where the video was filmed.
211 subtitles: The available subtitles as a dictionary in the format
212 {tag: subformats}. "tag" is usually a language code, and
213 "subformats" is a list sorted from lower to higher
214 preference, each element is a dictionary with the "ext"
216 * "data": The subtitles file contents
217 * "url": A URL pointing to the subtitles file
218 "ext" will be calculated from URL if missing
219 automatic_captions: Like 'subtitles', used by the YoutubeIE for
220 automatically generated captions
221 duration: Length of the video in seconds, as an integer or float.
222 view_count: How many users have watched the video on the platform.
223 like_count: Number of positive ratings of the video
224 dislike_count: Number of negative ratings of the video
225 repost_count: Number of reposts of the video
226 average_rating: Average rating give by users, the scale used depends on the webpage
227 comment_count: Number of comments on the video
228 comments: A list of comments, each with one or more of the following
229 properties (all but one of text or html optional):
230 * "author" - human-readable name of the comment author
231 * "author_id" - user ID of the comment author
233 * "html" - Comment as HTML
234 * "text" - Plain text of the comment
235 * "timestamp" - UNIX timestamp of comment
236 * "parent" - ID of the comment this one is replying to.
237 Set to "root" to indicate that this is a
238 comment to the original video.
239 age_limit: Age restriction for the video, as an integer (years)
240 webpage_url: The URL to the video webpage, if given to youtube-dl it
241 should allow to get the same result again. (It will be set
242 by YoutubeDL if it's missing)
243 categories: A list of categories that the video falls in, for example
245 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
246 is_live: True, False, or None (=unknown). Whether this video is a
247 live stream that goes on instead of a fixed-length video.
248 start_time: Time in seconds where the reproduction should start, as
249 specified in the URL.
250 end_time: Time in seconds where the reproduction should end, as
251 specified in the URL.
252 chapters: A list of dictionaries, with the following entries:
253 * "start_time" - The start time of the chapter in seconds
254 * "end_time" - The end time of the chapter in seconds
255 * "title" (optional, string)
257 The following fields should only be used when the video belongs to some logical
260 chapter: Name or title of the chapter the video belongs to.
261 chapter_number: Number of the chapter the video belongs to, as an integer.
262 chapter_id: Id of the chapter the video belongs to, as a unicode string.
264 The following fields should only be used when the video is an episode of some
265 series, programme or podcast:
267 series: Title of the series or programme the video episode belongs to.
268 season: Title of the season the video episode belongs to.
269 season_number: Number of the season the video episode belongs to, as an integer.
270 season_id: Id of the season the video episode belongs to, as a unicode string.
271 episode: Title of the video episode. Unlike mandatory video title field,
272 this field should denote the exact title of the video episode
273 without any kind of decoration.
274 episode_number: Number of the video episode within a season, as an integer.
275 episode_id: Id of the video episode, as a unicode string.
277 The following fields should only be used when the media is a track or a part of
280 track: Title of the track.
281 track_number: Number of the track within an album or a disc, as an integer.
282 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
284 artist: Artist(s) of the track.
285 genre: Genre(s) of the track.
286 album: Title of the album the track belongs to.
287 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
288 album_artist: List of all artists appeared on the album (e.g.
289 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
291 disc_number: Number of the disc or other physical medium the track belongs to,
293 release_year: Year (YYYY) when the album was released.
295 Unless mentioned otherwise, the fields should be Unicode strings.
297 Unless mentioned otherwise, None is equivalent to absence of information.
300 _type "playlist" indicates multiple videos.
301 There must be a key "entries", which is a list, an iterable, or a PagedList
302 object, each element of which is a valid dictionary by this specification.
304 Additionally, playlists can have "id", "title", "description", "uploader",
305 "uploader_id", "uploader_url" attributes with the same semantics as videos
309 _type "multi_video" indicates that there are multiple videos that
310 form a single show, for examples multiple acts of an opera or TV episode.
311 It must have an entries key like a playlist and contain all the keys
312 required for a video at the same time.
315 _type "url" indicates that the video must be extracted from another
316 location, possibly by a different extractor. Its only required key is:
317 "url" - the next URL to extract.
318 The key "ie_key" can be set to the class name (minus the trailing "IE",
319 e.g. "Youtube") if the extractor class is known in advance.
320 Additionally, the dictionary may have any properties of the resolved entity
321 known in advance, for example "title" if the title of the referred video is
325 _type "url_transparent" entities have the same specification as "url", but
326 indicate that the given additional information is more precise than the one
327 associated with the resolved URL.
328 This is useful when a site employs a video service that hosts the video and
329 its technical metadata, but that video service does not embed a useful
330 title, description etc.
333 Subclasses of this one should re-define the _real_initialize() and
334 _real_extract() methods and define a _VALID_URL regexp.
335 Probably, they should also be added to the list of extractors.
337 _GEO_BYPASS attribute may be set to False in order to disable
338 geo restriction bypass mechanisms for a particular extractor.
339 Though it won't disable explicit geo restriction bypass based on
340 country code provided with geo_bypass_country. (experimental)
342 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
343 countries for this extractor. One of these countries will be used by
344 geo restriction bypass mechanism right away in order to bypass
345 geo restriction, of course, if the mechanism is not disabled. (experimental)
347 NB: both these geo attributes are experimental and may change in future
348 or be completely removed.
350 Finally, the _WORKING attribute should be set to False for broken IEs
351 in order to warn the users and skip the tests.
356 _x_forwarded_for_ip
= None
358 _GEO_COUNTRIES
= None
361 def __init__(self
, downloader
=None):
362 """Constructor. Receives an optional downloader."""
364 self
._x
_forwarded
_for
_ip
= None
365 self
.set_downloader(downloader
)
368 def suitable(cls
, url
):
369 """Receives a URL and returns True if suitable for this IE."""
371 # This does not use has/getattr intentionally - we want to know whether
372 # we have cached the regexp for *this* class, whereas getattr would also
373 # match the superclass
374 if '_VALID_URL_RE' not in cls
.__dict
__:
375 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
376 return cls
._VALID
_URL
_RE
.match(url
) is not None
379 def _match_id(cls
, url
):
380 if '_VALID_URL_RE' not in cls
.__dict
__:
381 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
382 m
= cls
._VALID
_URL
_RE
.match(url
)
384 return compat_str(m
.group('id'))
388 """Getter method for _WORKING."""
391 def initialize(self
):
392 """Initializes an instance (authentication, etc)."""
393 self
._initialize
_geo
_bypass
(self
._GEO
_COUNTRIES
)
395 self
._real
_initialize
()
398 def _initialize_geo_bypass(self
, countries
):
400 Initialize geo restriction bypass mechanism.
402 This method is used to initialize geo bypass mechanism based on faking
403 X-Forwarded-For HTTP header. A random country from provided country list
404 is selected and a random IP belonging to this country is generated. This
405 IP will be passed as X-Forwarded-For HTTP header in all subsequent
408 This method will be used for initial geo bypass mechanism initialization
409 during the instance initialization with _GEO_COUNTRIES.
411 You may also manually call it from extractor's code if geo countries
412 information is not available beforehand (e.g. obtained during
413 extraction) or due to some another reason.
415 if not self
._x
_forwarded
_for
_ip
:
416 country_code
= self
._downloader
.params
.get('geo_bypass_country', None)
417 # If there is no explicit country for geo bypass specified and
418 # the extractor is known to be geo restricted let's fake IP
419 # as X-Forwarded-For right away.
420 if (not country_code
and
422 self
._downloader
.params
.get('geo_bypass', True) and
424 country_code
= random
.choice(countries
)
426 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
427 if self
._downloader
.params
.get('verbose', False):
428 self
._downloader
.to_screen(
429 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
430 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
432 def extract(self
, url
):
433 """Extracts URL information and returns it in list of dicts."""
438 ie_result
= self
._real
_extract
(url
)
439 if self
._x
_forwarded
_for
_ip
:
440 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
442 except GeoRestrictedError
as e
:
443 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
446 except ExtractorError
:
448 except compat_http_client
.IncompleteRead
as e
:
449 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
450 except (KeyError, StopIteration) as e
:
451 raise ExtractorError('An extractor error has occurred.', cause
=e
)
453 def __maybe_fake_ip_and_retry(self
, countries
):
454 if (not self
._downloader
.params
.get('geo_bypass_country', None) and
456 self
._downloader
.params
.get('geo_bypass', True) and
457 not self
._x
_forwarded
_for
_ip
and
459 country_code
= random
.choice(countries
)
460 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
461 if self
._x
_forwarded
_for
_ip
:
463 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
464 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
468 def set_downloader(self
, downloader
):
469 """Sets the downloader for this IE."""
470 self
._downloader
= downloader
472 def _real_initialize(self
):
473 """Real initialization process. Redefine in subclasses."""
476 def _real_extract(self
, url
):
477 """Real extraction process. Redefine in subclasses."""
482 """A string for getting the InfoExtractor with get_info_extractor"""
483 return compat_str(cls
.__name
__[:-2])
487 return compat_str(type(self
).__name
__[:-2])
489 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
490 """ Returns the response handle """
492 self
.report_download_webpage(video_id
)
493 elif note
is not False:
495 self
.to_screen('%s' % (note
,))
497 self
.to_screen('%s: %s' % (video_id
, note
))
499 # Some sites check X-Forwarded-For HTTP header in order to figure out
500 # the origin of the client behind proxy. This allows bypassing geo
501 # restriction by faking this header's value to IP that belongs to some
502 # geo unrestricted country. We will do so once we encounter any
503 # geo restriction error.
504 if self
._x
_forwarded
_for
_ip
:
505 if 'X-Forwarded-For' not in headers
:
506 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
508 if isinstance(url_or_request
, compat_urllib_request
.Request
):
509 url_or_request
= update_Request(
510 url_or_request
, data
=data
, headers
=headers
, query
=query
)
513 url_or_request
= update_url_query(url_or_request
, query
)
514 if data
is not None or headers
:
515 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
517 return self
._downloader
.urlopen(url_or_request
)
518 except (compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
) as err
:
522 errnote
= 'Unable to download webpage'
524 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
526 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
528 self
._downloader
.report_warning(errmsg
)
531 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, encoding
=None, data
=None, headers
={}, query={}
):
532 """ Returns a tuple (page content as string, URL handle) """
533 # Strip hashes from the URL (#1038)
534 if isinstance(url_or_request
, (compat_str
, str)):
535 url_or_request
= url_or_request
.partition('#')[0]
537 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
)
541 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
542 return (content
, urlh
)
545 def _guess_encoding_from_content(content_type
, webpage_bytes
):
546 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
548 encoding
= m
.group(1)
550 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
551 webpage_bytes[:1024])
553 encoding = m.group(1).decode('ascii')
554 elif webpage_bytes.startswith(b'\xff\xfe'):
561 def __check_blocked(self, content):
562 first_block = content[:512]
563 if ('<title>Access to this site is blocked</title>' in content and
564 'Websense' in first_block):
565 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
566 blocked_iframe = self._html_search_regex(
567 r'<iframe src="([^
"]+)"', content,
568 'Websense information URL
', default=None)
570 msg += ' Visit
%s for more details
' % blocked_iframe
571 raise ExtractorError(msg, expected=True)
572 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
574 'Access to this webpage has been blocked by Indian censorship
. '
575 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
576 block_msg = self._html_search_regex(
577 r'</h1
><p
>(.*?
)</p
>',
578 content, 'block message
', default=None)
580 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
581 raise ExtractorError(msg, expected=True)
582 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content and
583 'blocklist
.rkn
.gov
.ru
' in content):
584 raise ExtractorError(
585 'Access to this webpage has been blocked by decision of the Russian government
. '
586 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
589 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
590 content_type = urlh.headers.get('Content
-Type
', '')
591 webpage_bytes = urlh.read()
592 if prefix is not None:
593 webpage_bytes = prefix + webpage_bytes
595 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
596 if self._downloader.params.get('dump_intermediate_pages
', False):
597 self.to_screen('Dumping request to
' + urlh.geturl())
598 dump = base64.b64encode(webpage_bytes).decode('ascii
')
599 self._downloader.to_screen(dump)
600 if self._downloader.params.get('write_pages
', False):
601 basen = '%s_%s' % (video_id, urlh.geturl())
603 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
604 basen = basen[:240 - len(h)] + h
605 raw_filename = basen + '.dump
'
606 filename = sanitize_filename(raw_filename, restricted=True)
607 self.to_screen('Saving request to
' + filename)
608 # Working around MAX_PATH limitation on Windows (see
609 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
610 if compat_os_name == 'nt
':
611 absfilepath = os.path.abspath(filename)
612 if len(absfilepath) > 259:
613 filename = '\\\\?
\\' + absfilepath
614 with open(filename, 'wb
') as outf:
615 outf.write(webpage_bytes)
618 content = webpage_bytes.decode(encoding, 'replace
')
620 content = webpage_bytes.decode('utf
-8', 'replace
')
622 self.__check_blocked(content)
626 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}):
627 """ Returns the data of the page as a string """
630 while success is False:
632 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
634 except compat_http_client.IncompleteRead as e:
636 if try_count >= tries:
638 self._sleep(timeout, video_id)
645 def _download_xml(self, url_or_request, video_id,
646 note='Downloading XML
', errnote='Unable to download XML
',
647 transform_source=None, fatal=True, encoding=None,
648 data=None, headers={}, query={}):
649 """Return the xml as an xml.etree.ElementTree.Element"""
650 xml_string = self._download_webpage(
651 url_or_request, video_id, note, errnote, fatal=fatal,
652 encoding=encoding, data=data, headers=headers, query=query)
653 if xml_string is False:
655 return self._parse_xml(
656 xml_string, video_id, transform_source=transform_source,
659 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
661 xml_string = transform_source(xml_string)
663 return compat_etree_fromstring(xml_string.encode('utf
-8'))
664 except compat_xml_parse_error as ve:
665 errmsg = '%s: Failed to parse XML
' % video_id
667 raise ExtractorError(errmsg, cause=ve)
669 self.report_warning(errmsg + str(ve))
671 def _download_json(self, url_or_request, video_id,
672 note='Downloading JSON metadata
',
673 errnote='Unable to download JSON metadata
',
674 transform_source=None,
675 fatal=True, encoding=None, data=None, headers={}, query={}):
676 json_string = self._download_webpage(
677 url_or_request, video_id, note, errnote, fatal=fatal,
678 encoding=encoding, data=data, headers=headers, query=query)
679 if (not fatal) and json_string is False:
681 return self._parse_json(
682 json_string, video_id, transform_source=transform_source, fatal=fatal)
684 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
686 json_string = transform_source(json_string)
688 return json.loads(json_string)
689 except ValueError as ve:
690 errmsg = '%s: Failed to parse JSON
' % video_id
692 raise ExtractorError(errmsg, cause=ve)
694 self.report_warning(errmsg + str(ve))
696 def report_warning(self, msg, video_id=None):
697 idstr = '' if video_id is None else '%s: ' % video_id
698 self._downloader.report_warning(
699 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
701 def to_screen(self, msg):
702 """Print msg to screen, prefixing it with '[ie_name
]'"""
703 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
705 def report_extraction(self, id_or_name):
706 """Report information extraction."""
707 self.to_screen('%s: Extracting information
' % id_or_name)
709 def report_download_webpage(self, video_id):
710 """Report webpage download."""
711 self.to_screen('%s: Downloading webpage
' % video_id)
713 def report_age_confirmation(self):
714 """Report attempt to confirm age."""
715 self.to_screen('Confirming age
')
717 def report_login(self):
718 """Report attempt to log in."""
719 self.to_screen('Logging
in')
722 def raise_login_required(msg='This video
is only available
for registered users
'):
723 raise ExtractorError(
724 '%s. Use
--username
and --password
or --netrc to provide account credentials
.' % msg,
728 def raise_geo_restricted(msg='This video
is not available
from your location due to geo restriction
', countries=None):
729 raise GeoRestrictedError(msg, countries=countries)
731 # Methods for following #608
733 def url_result(url, ie=None, video_id=None, video_title=None):
734 """Returns a URL that points to a page that should be processed"""
735 # TODO: ie should be the class used for getting the info
736 video_info = {'_type
': 'url
',
739 if video_id is not None:
740 video_info['id'] = video_id
741 if video_title is not None:
742 video_info['title
'] = video_title
745 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
747 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
749 return self.playlist_result(
750 urls, playlist_id=playlist_id, playlist_title=playlist_title)
753 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
754 """Returns a playlist"""
755 video_info = {'_type
': 'playlist
',
758 video_info['id'] = playlist_id
760 video_info['title
'] = playlist_title
761 if playlist_description:
762 video_info['description
'] = playlist_description
765 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
767 Perform a regex search on the given string, using a single or a list of
768 patterns returning the first matching group.
769 In case of failure return a default value or raise a WARNING or a
770 RegexNotFoundError, depending on fatal, specifying the field name.
772 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
773 mobj = re.search(pattern, string, flags)
776 mobj = re.search(p, string, flags)
780 if not self._downloader.params.get('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
781 _name = '\033[0;34m
%s\033[0m
' % name
787 # return the first matching group
788 return next(g for g in mobj.groups() if g is not None)
790 return mobj.group(group)
791 elif default is not NO_DEFAULT:
794 raise RegexNotFoundError('Unable to extract
%s' % _name)
796 self._downloader.report_warning('unable to extract
%s' % _name + bug_reports_message())
799 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
801 Like _search_regex, but strips HTML tags and unescapes entities.
803 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
805 return clean_html(res).strip()
809 def _get_netrc_login_info(self, netrc_machine=None):
812 netrc_machine = netrc_machine or self._NETRC_MACHINE
814 if self._downloader.params.get('usenetrc
', False):
816 info = netrc.netrc().authenticators(netrc_machine)
821 raise netrc.NetrcParseError(
822 'No authenticators
for %s' % netrc_machine)
823 except (IOError, netrc.NetrcParseError) as err:
824 self._downloader.report_warning(
825 'parsing
.netrc
: %s' % error_to_compat_str(err))
827 return username, password
829 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
831 Get the login info as (username, password)
832 First look for the manually specified credentials using username_option
833 and password_option as keys in params dictionary. If no such credentials
834 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
836 If there's no info available
, return (None, None)
838 if self._downloader is None:
841 downloader_params = self._downloader.params
843 # Attempt to use provided username and password or .netrc data
844 if downloader_params.get(username_option) is not None:
845 username = downloader_params[username_option]
846 password = downloader_params[password_option]
848 username, password = self._get_netrc_login_info(netrc_machine)
850 return username, password
852 def _get_tfa_info(self, note='two-factor verification code'):
854 Get the two
-factor authentication info
855 TODO
- asking the user will be required
for sms
/phone verify
856 currently just uses the command line option
857 If there
's no info available, return None
859 if self._downloader is None:
861 downloader_params = self._downloader.params
863 if downloader_params.get('twofactor
') is not None:
864 return downloader_params['twofactor
']
866 return compat_getpass('Type
%s and press
[Return
]: ' % note)
868 # Helper functions for extracting OpenGraph info
870 def _og_regexes(prop):
871 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
872 property_re = (r'(?
:name|
property)=(?
:\'og
:%(prop)s\'|
"og:%(prop)s"|\s
*og
:%(prop)s\b)'
873 % {'prop': re.escape(prop)})
874 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
876 template % (property_re, content_re),
877 template % (content_re, property_re),
881 def _meta_regex(prop):
882 return r'''(?isx)<meta
883 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
884 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
886 def _og_search_property(self, prop, html, name=None, **kargs):
887 if not isinstance(prop, (list, tuple)):
890 name = 'OpenGraph
%s' % prop[0]
893 og_regexes.extend(self._og_regexes(p))
894 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
897 return unescapeHTML(escaped)
899 def _og_search_thumbnail(self, html, **kargs):
900 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
902 def _og_search_description(self, html, **kargs):
903 return self._og_search_property('description
', html, fatal=False, **kargs)
905 def _og_search_title(self, html, **kargs):
906 return self._og_search_property('title
', html, **kargs)
908 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
909 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
911 regexes = self._og_regexes('video
:secure_url
') + regexes
912 return self._html_search_regex(regexes, html, name, **kargs)
914 def _og_search_url(self, html, **kargs):
915 return self._og_search_property('url
', html, **kargs)
917 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
918 if not isinstance(name, (list, tuple)):
920 if display_name is None:
921 display_name = name[0]
922 return self._html_search_regex(
923 [self._meta_regex(n) for n in name],
924 html, display_name, fatal=fatal, group='content
', **kwargs)
926 def _dc_search_uploader(self, html):
927 return self._html_search_meta('dc
.creator
', html, 'uploader
')
929 def _rta_search(self, html):
930 # See http://www.rtalabel.org/index.php?content=howtofaq#single
931 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
932 r' content
="RTA-5042-1996-1400-1577-RTA"',
937 def _media_rating_search(self, html):
938 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
939 rating = self._html_search_meta('rating
', html)
951 return RATING_TABLE.get(rating.lower())
953 def _family_friendly_search(self, html):
954 # See http://schema.org/VideoObject
955 family_friendly = self._html_search_meta(
956 'isFamilyFriendly
', html, default=None)
958 if not family_friendly:
967 return RATING_TABLE.get(family_friendly.lower())
969 def _twitter_search_player(self, html):
970 return self._html_search_meta('twitter
:player
', html,
971 'twitter card player
')
973 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
974 json_ld = self._search_regex(
975 r'(?s
)<script
[^
>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
976 html, 'JSON-LD', group='json_ld', **kwargs)
977 default = kwargs.get('default', NO_DEFAULT)
979 return default if default is not NO_DEFAULT else {}
980 # JSON-LD may be malformed and thus `fatal` should be respected.
981 # At the same time `default` may be passed that assumes `fatal=False`
982 # for _search_regex. Let's simulate the same behavior here as well.
983 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
984 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
986 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
987 if isinstance(json_ld, compat_str):
988 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
992 if not isinstance(json_ld, (list, tuple, dict)):
994 if isinstance(json_ld, dict):
997 def extract_video_object(e):
998 assert e['@type'] == 'VideoObject'
1000 'url': e.get('contentUrl'),
1001 'title': unescapeHTML(e.get('name')),
1002 'description': unescapeHTML(e.get('description')),
1003 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
1004 'duration': parse_duration(e.get('duration')),
1005 'timestamp': unified_timestamp(e.get('uploadDate')),
1006 'filesize': float_or_none(e.get('contentSize')),
1007 'tbr': int_or_none(e.get('bitrate')),
1008 'width': int_or_none(e.get('width')),
1009 'height': int_or_none(e.get('height')),
1010 'view_count': int_or_none(e.get('interactionCount')),
1014 if e.get('@context') == 'http://schema.org':
1015 item_type = e.get('@type')
1016 if expected_type is not None and expected_type != item_type:
1018 if item_type in ('TVEpisode', 'Episode'):
1020 'episode': unescapeHTML(e.get('name')),
1021 'episode_number': int_or_none(e.get('episodeNumber')),
1022 'description': unescapeHTML(e.get('description')),
1024 part_of_season = e.get('partOfSeason')
1025 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1026 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1027 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1028 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1029 info['series'] = unescapeHTML(part_of_series.get('name'))
1030 elif item_type == 'Article':
1032 'timestamp': parse_iso8601(e.get('datePublished')),
1033 'title': unescapeHTML(e.get('headline')),
1034 'description': unescapeHTML(e.get('articleBody')),
1036 elif item_type == 'VideoObject':
1037 extract_video_object(e)
1039 video = e.get('video')
1040 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1041 extract_video_object(video)
1043 return dict((k, v) for k, v in info.items() if v is not None)
1046 def _hidden_inputs(html):
1047 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1049 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1050 attrs = extract_attributes(input)
1053 if attrs.get('type') not in ('hidden', 'submit'):
1055 name = attrs.get('name') or attrs.get('id')
1056 value = attrs.get('value')
1057 if name and value is not None:
1058 hidden_inputs[name] = value
1059 return hidden_inputs
1061 def _form_hidden_inputs(self, form_id, html):
1062 form = self._search_regex(
1063 r'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1064 html, '%s form
' % form_id, group='form
')
1065 return self._hidden_inputs(form)
1067 def _sort_formats(self, formats, field_preference=None):
1069 raise ExtractorError('No video formats found
')
1072 # Automatically determine tbr when missing based on abr and vbr (improves
1073 # formats sorting in some cases)
1074 if 'tbr
' not in f and f.get('abr
') is not None and f.get('vbr
') is not None:
1075 f['tbr
'] = f['abr
'] + f['vbr
']
1077 def _formats_key(f):
1078 # TODO remove the following workaround
1079 from ..utils import determine_ext
1080 if not f.get('ext
') and 'url
' in f:
1081 f['ext
'] = determine_ext(f['url
'])
1083 if isinstance(field_preference, (list, tuple)):
1086 if f.get(field) is not None
1087 else ('' if field == 'format_id
' else -1)
1088 for field in field_preference)
1090 preference = f.get('preference
')
1091 if preference is None:
1093 if f.get('ext
') in ['f4f
', 'f4m
']: # Not yet supported
1096 protocol = f.get('protocol
') or determine_protocol(f)
1097 proto_preference = 0 if protocol in ['http
', 'https
'] else (-0.5 if protocol == 'rtsp
' else -0.1)
1099 if f.get('vcodec
') == 'none
': # audio only
1101 if self._downloader.params.get('prefer_free_formats
'):
1102 ORDER = ['aac
', 'mp3
', 'm4a
', 'webm
', 'ogg
', 'opus
']
1104 ORDER = ['webm
', 'opus
', 'ogg
', 'mp3
', 'aac
', 'm4a
']
1107 audio_ext_preference = ORDER.index(f['ext
'])
1109 audio_ext_preference = -1
1111 if f.get('acodec
') == 'none
': # video only
1113 if self._downloader.params.get('prefer_free_formats
'):
1114 ORDER = ['flv
', 'mp4
', 'webm
']
1116 ORDER = ['webm
', 'flv
', 'mp4
']
1118 ext_preference = ORDER.index(f['ext
'])
1121 audio_ext_preference = 0
1125 f.get('language_preference
') if f.get('language_preference
') is not None else -1,
1126 f.get('quality
') if f.get('quality
') is not None else -1,
1127 f.get('tbr
') if f.get('tbr
') is not None else -1,
1128 f.get('filesize
') if f.get('filesize
') is not None else -1,
1129 f.get('vbr
') if f.get('vbr
') is not None else -1,
1130 f.get('height
') if f.get('height
') is not None else -1,
1131 f.get('width
') if f.get('width
') is not None else -1,
1134 f.get('abr
') if f.get('abr
') is not None else -1,
1135 audio_ext_preference,
1136 f.get('fps
') if f.get('fps
') is not None else -1,
1137 f.get('filesize_approx
') if f.get('filesize_approx
') is not None else -1,
1138 f.get('source_preference
') if f.get('source_preference
') is not None else -1,
1139 f.get('format_id
') if f.get('format_id
') is not None else '',
1141 formats.sort(key=_formats_key)
1143 def _check_formats(self, formats, video_id):
1145 formats[:] = filter(
1146 lambda f: self._is_valid_url(
1148 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1152 def _remove_duplicate_formats(formats):
1156 if f['url
'] not in format_urls:
1157 format_urls.add(f['url
'])
1158 unique_formats.append(f)
1159 formats[:] = unique_formats
1161 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1162 url = self._proto_relative_url(url, scheme='http
:')
1163 # For now assume non HTTP(S) URLs always valid
1164 if not (url.startswith('http
://') or url.startswith('https
://')):
1167 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1169 except ExtractorError as e:
1170 if isinstance(e.cause, compat_urllib_error.URLError):
1172 '%s: %s URL
is invalid
, skipping
' % (video_id, item))
1176 def http_scheme(self):
1177 """ Either "http:" or "https:", depending on the user's preferences
"""
1180 if self._downloader.params.get('prefer_insecure', False)
1183 def _proto_relative_url(self, url, scheme=None):
1186 if url.startswith('//'):
1188 scheme = self.http_scheme()
1193 def _sleep(self, timeout, video_id, msg_template=None):
1194 if msg_template is None:
1195 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1196 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1200 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1201 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1202 fatal=True, m3u8_id=None):
1203 manifest = self._download_xml(
1204 manifest_url, video_id, 'Downloading f4m manifest',
1205 'Unable to download f4m manifest',
1206 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1207 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1208 transform_source=transform_source,
1211 if manifest is False:
1214 return self._parse_f4m_formats(
1215 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1216 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1218 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1219 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1220 fatal=True, m3u8_id=None):
1221 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1222 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1223 if akamai_pv is not None and ';' in akamai_pv.text:
1224 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1225 if playerVerificationChallenge.strip() != '':
1229 manifest_version = '1.0'
1230 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1232 manifest_version = '2.0'
1233 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1234 # Remove unsupported DRM protected media from final formats
1235 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1236 media_nodes = remove_encrypted_media(media_nodes)
1240 manifest_base_url = get_base_url(manifest)
1242 bootstrap_info = xpath_element(
1243 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1244 'bootstrap info', default=None)
1247 mime_type = xpath_text(
1248 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1249 'base URL', default=None)
1250 if mime_type and mime_type.startswith('audio/'):
1253 for i, media_el in enumerate(media_nodes):
1254 tbr = int_or_none(media_el.attrib.get('bitrate'))
1255 width = int_or_none(media_el.attrib.get('width'))
1256 height = int_or_none(media_el.attrib.get('height'))
1257 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1258 # If <bootstrapInfo> is present, the specified f4m is a
1259 # stream-level manifest, and only set-level manifests may refer to
1260 # external resources. See section 11.4 and section 4 of F4M spec
1261 if bootstrap_info is None:
1263 # @href is introduced in 2.0, see section 11.6 of F4M spec
1264 if manifest_version == '2.0':
1265 media_url = media_el.attrib.get('href')
1266 if media_url is None:
1267 media_url = media_el.attrib.get('url')
1271 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1272 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1273 # If media_url is itself a f4m manifest do the recursive extraction
1274 # since bitrates in parent manifest (this one) and media_url manifest
1275 # may differ leading to inability to resolve the format by requested
1276 # bitrate in f4m downloader
1277 ext = determine_ext(manifest_url)
1279 f4m_formats = self._extract_f4m_formats(
1280 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1281 transform_source=transform_source, fatal=fatal)
1282 # Sometimes stream-level manifest contains single media entry that
1283 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1284 # At the same time parent's media entry in set-level manifest may
1285 # contain it. We will copy it from parent in such cases.
1286 if len(f4m_formats) == 1:
1289 'tbr': f.get('tbr') or tbr,
1290 'width': f.get('width') or width,
1291 'height': f.get('height') or height,
1292 'format_id': f.get('format_id') if not tbr else format_id,
1295 formats.extend(f4m_formats)
1298 formats.extend(self._extract_m3u8_formats(
1299 manifest_url, video_id, 'mp4', preference=preference,
1300 m3u8_id=m3u8_id, fatal=fatal))
1303 'format_id': format_id,
1304 'url': manifest_url,
1305 'manifest_url': manifest_url,
1306 'ext': 'flv' if bootstrap_info is not None else None,
1312 'preference': preference,
1316 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1318 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1322 'preference': preference - 100 if preference else -100,
1323 'resolution': 'multiple',
1324 'format_note': 'Quality selection URL',
1327 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1328 entry_protocol='m3u8', preference=None,
1329 m3u8_id=None, note=None, errnote=None,
1330 fatal=True, live=False):
1331 res = self._download_webpage_handle(
1333 note=note or 'Downloading m3u8 information',
1334 errnote=errnote or 'Failed to download m3u8 information',
1340 m3u8_doc, urlh = res
1341 m3u8_url = urlh.geturl()
1343 return self._parse_m3u8_formats(
1344 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1345 preference=preference, m3u8_id=m3u8_id, live=live)
1347 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1348 entry_protocol='m3u8', preference=None,
1349 m3u8_id=None, live=False):
1350 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1353 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1358 format_url = lambda u: (
1360 if re.match(r'^https?://', u)
1361 else compat_urlparse.urljoin(m3u8_url, u))
1364 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1365 # 2. https://github.com/rg3/youtube-dl/issues/12211
1367 # We should try extracting formats only from master playlists [1, 4.3.4],
1368 # i.e. playlists that describe available qualities. On the other hand
1369 # media playlists [1, 4.3.3] should be returned as is since they contain
1370 # just the media without qualities renditions.
1371 # Fortunately, master playlist can be easily distinguished from media
1372 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1373 # master playlist tags MUST NOT appear in a media playist and vice versa.
1374 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1375 # media playlist and MUST NOT appear in master playlist thus we can
1376 # clearly detect media playlist with this criterion.
1378 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1381 'format_id': m3u8_id,
1383 'protocol': entry_protocol,
1384 'preference': preference,
1388 last_stream_inf = {}
1390 def extract_media(x_media_line):
1391 media = parse_m3u8_attributes(x_media_line)
1392 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1393 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1394 if not (media_type and group_id and name):
1396 groups.setdefault(group_id, []).append(media)
1397 if media_type not in ('VIDEO', 'AUDIO'):
1399 media_url = media.get('URI')
1402 for v in (m3u8_id, group_id, name):
1406 'format_id': '-'.join(format_id),
1407 'url': format_url(media_url),
1408 'manifest_url': m3u8_url,
1409 'language': media.get('LANGUAGE'),
1411 'protocol': entry_protocol,
1412 'preference': preference,
1414 if media_type == 'AUDIO':
1415 f['vcodec'] = 'none'
1418 def build_stream_name():
1419 # Despite specification does not mention NAME attribute for
1420 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1421 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1422 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1423 stream_name = last_stream_inf.get('NAME')
1426 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1427 # from corresponding rendition group
1428 stream_group_id = last_stream_inf.get('VIDEO')
1429 if not stream_group_id:
1431 stream_group = groups.get(stream_group_id)
1432 if not stream_group:
1433 return stream_group_id
1434 rendition = stream_group[0]
1435 return rendition.get('NAME') or stream_group_id
1437 for line in m3u8_doc.splitlines():
1438 if line.startswith('#EXT-X-STREAM-INF:'):
1439 last_stream_inf = parse_m3u8_attributes(line)
1440 elif line.startswith('#EXT-X-MEDIA:'):
1442 elif line.startswith('#') or not line.strip():
1445 tbr = float_or_none(
1446 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1447 last_stream_inf.get('BANDWIDTH'), scale=1000)
1450 format_id.append(m3u8_id)
1451 stream_name = build_stream_name()
1452 # Bandwidth of live streams may differ over time thus making
1453 # format_id unpredictable. So it's better to keep provided
1456 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1457 manifest_url = format_url(line.strip())
1459 'format_id': '-'.join(format_id),
1460 'url': manifest_url,
1461 'manifest_url': m3u8_url,
1464 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1465 'protocol': entry_protocol,
1466 'preference': preference,
1468 resolution = last_stream_inf.get('RESOLUTION')
1470 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1472 f['width'] = int(mobj.group('width'))
1473 f['height'] = int(mobj.group('height'))
1474 # Unified Streaming Platform
1476 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1478 abr, vbr = mobj.groups()
1479 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1484 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1486 audio_group_id = last_stream_inf.get('AUDIO')
1487 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1488 # references a rendition group MUST have a CODECS attribute.
1489 # However, this is not always respected, for example, [2]
1490 # contains EXT-X-STREAM-INF tag which references AUDIO
1491 # rendition group but does not have CODECS and despite
1492 # referencing audio group an audio group, it represents
1493 # a complete (with audio and video) format. So, for such cases
1494 # we will ignore references to rendition groups and treat them
1495 # as complete formats.
1496 if audio_group_id and codecs and f.get('vcodec') != 'none':
1497 audio_group = groups.get(audio_group_id)
1498 if audio_group and audio_group[0].get('URI'):
1499 # TODO: update acodec for audio only formats with
1501 f['acodec'] = 'none'
1503 last_stream_inf = {}
1507 def _xpath_ns(path, namespace=None):
1511 for c in path.split('/'):
1512 if not c or c == '.':
1515 out.append('{%s}%s' % (namespace, c))
1516 return '/'.join(out)
1518 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1519 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1525 namespace = self._parse_smil_namespace(smil)
1527 return self._parse_smil_formats(
1528 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1530 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1531 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1534 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1536 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1537 return self._download_xml(
1538 smil_url, video_id, 'Downloading SMIL file',
1539 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1541 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1542 namespace = self._parse_smil_namespace(smil)
1544 formats = self._parse_smil_formats(
1545 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1546 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1548 video_id = os.path.splitext(url_basename(smil_url))[0]
1552 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1553 name = meta.attrib.get('name')
1554 content = meta.attrib.get('content')
1555 if not name or not content:
1557 if not title and name == 'title':
1559 elif not description and name in ('description', 'abstract'):
1560 description = content
1561 elif not upload_date and name == 'date':
1562 upload_date = unified_strdate(content)
1565 'id': image.get('type'),
1566 'url': image.get('src'),
1567 'width': int_or_none(image.get('width')),
1568 'height': int_or_none(image.get('height')),
1569 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1573 'title': title or video_id,
1574 'description': description,
1575 'upload_date': upload_date,
1576 'thumbnails': thumbnails,
1578 'subtitles': subtitles,
1581 def _parse_smil_namespace(self, smil):
1582 return self._search_regex(
1583 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1585 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1587 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1588 b = meta.get('base') or meta.get('httpBase')
1599 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1600 for medium in media:
1601 src = medium.get('src')
1602 if not src or src in srcs:
1606 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1607 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1608 width = int_or_none(medium.get('width'))
1609 height = int_or_none(medium.get('height'))
1610 proto = medium.get('proto')
1611 ext = medium.get('ext')
1612 src_ext = determine_ext(src)
1613 streamer = medium.get('streamer') or base
1615 if proto == 'rtmp' or streamer.startswith('rtmp'):
1621 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1623 'filesize': filesize,
1627 if transform_rtmp_url:
1628 streamer, src = transform_rtmp_url(streamer, src)
1629 formats[-1].update({
1635 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1636 src_url = src_url.strip()
1638 if proto == 'm3u8' or src_ext == 'm3u8':
1639 m3u8_formats = self._extract_m3u8_formats(
1640 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1641 if len(m3u8_formats) == 1:
1643 m3u8_formats[0].update({
1644 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1649 formats.extend(m3u8_formats)
1652 if src_ext == 'f4m':
1657 'plugin': 'flowplayer-3.2.0.1',
1659 f4m_url += '&' if '?' in f4m_url else '?'
1660 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1661 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1664 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1668 'ext': ext or src_ext or 'flv',
1669 'format_id': 'http-%d' % (bitrate or http_count),
1671 'filesize': filesize,
1679 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1682 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1683 src = textstream.get('src')
1684 if not src or src in urls:
1687 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1688 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1689 subtitles.setdefault(lang, []).append({
1695 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1696 xspf = self._download_xml(
1697 playlist_url, playlist_id, 'Downloading xpsf playlist',
1698 'Unable to download xspf manifest', fatal=fatal)
1701 return self._parse_xspf(xspf, playlist_id)
1703 def _parse_xspf(self, playlist, playlist_id):
1705 'xspf': 'http://xspf.org/ns/0/',
1706 's1': 'http://static.streamone.nl/player/ns/0',
1710 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1712 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1713 description = xpath_text(
1714 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1715 thumbnail = xpath_text(
1716 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1717 duration = float_or_none(
1718 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1721 'url': location.text,
1722 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1723 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1724 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1725 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1726 self._sort_formats(formats)
1731 'description': description,
1732 'thumbnail': thumbnail,
1733 'duration': duration,
1738 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1739 res = self._download_webpage_handle(
1741 note=note or 'Downloading MPD manifest',
1742 errnote=errnote or 'Failed to download MPD manifest',
1747 mpd_base_url = base_url(urlh.geturl())
1749 return self._parse_mpd_formats(
1750 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1751 formats_dict=formats_dict, mpd_url=mpd_url)
1753 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1755 Parse formats
from MPD manifest
.
1757 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
1758 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
1759 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
1761 if mpd_doc.get('type') == 'dynamic':
1764 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1767 return self._xpath_ns(path, namespace)
1769 def is_drm_protected(element):
1770 return element.find(_add_ns('ContentProtection')) is not None
1772 def extract_multisegment_info(element, ms_parent_info):
1773 ms_info = ms_parent_info.copy()
1775 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1776 # common attributes and elements. We will only extract relevant
1778 def extract_common(source):
1779 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1780 if segment_timeline is not None:
1781 s_e = segment_timeline.findall(_add_ns('S'))
1783 ms_info['total_number'] = 0
1786 r = int(s.get('r', 0))
1787 ms_info['total_number'] += 1 + r
1788 ms_info['s'].append({
1789 't': int(s.get('t', 0)),
1790 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1791 'd': int(s.attrib['d']),
1794 start_number = source.get('startNumber')
1796 ms_info['start_number'] = int(start_number)
1797 timescale = source.get('timescale')
1799 ms_info['timescale'] = int(timescale)
1800 segment_duration = source.get('duration')
1801 if segment_duration:
1802 ms_info['segment_duration'] = float(segment_duration)
1804 def extract_Initialization(source):
1805 initialization = source.find(_add_ns('Initialization'))
1806 if initialization is not None:
1807 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1809 segment_list = element.find(_add_ns('SegmentList'))
1810 if segment_list is not None:
1811 extract_common(segment_list)
1812 extract_Initialization(segment_list)
1813 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1815 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1817 segment_template = element.find(_add_ns('SegmentTemplate'))
1818 if segment_template is not None:
1819 extract_common(segment_template)
1820 media = segment_template.get('media')
1822 ms_info['media'] = media
1823 initialization = segment_template.get('initialization')
1825 ms_info['initialization'] = initialization
1827 extract_Initialization(segment_template)
1830 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1832 for period in mpd_doc.findall(_add_ns('Period')):
1833 period_duration = parse_duration(period.get('duration')) or mpd_duration
1834 period_ms_info = extract_multisegment_info(period, {
1838 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1839 if is_drm_protected(adaptation_set):
1841 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1842 for representation in adaptation_set.findall(_add_ns('Representation')):
1843 if is_drm_protected(representation):
1845 representation_attrib = adaptation_set.attrib.copy()
1846 representation_attrib.update(representation.attrib)
1847 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1848 mime_type = representation_attrib['mimeType']
1849 content_type = mime_type.split('/')[0]
1850 if content_type == 'text':
1851 # TODO implement WebVTT downloading
1853 elif content_type in ('video', 'audio'):
1855 for element in (representation, adaptation_set, period, mpd_doc):
1856 base_url_e = element.find(_add_ns('BaseURL'))
1857 if base_url_e is not None:
1858 base_url = base_url_e.text + base_url
1859 if re.match(r'^https?://', base_url):
1861 if mpd_base_url and not re.match(r'^https?://', base_url):
1862 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1864 base_url = mpd_base_url + base_url
1865 representation_id = representation_attrib.get('id')
1866 lang = representation_attrib.get('lang')
1867 url_el = representation.find(_add_ns('BaseURL'))
1868 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1869 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1871 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1873 'manifest_url': mpd_url,
1874 'ext': mimetype2ext(mime_type),
1875 'width': int_or_none(representation_attrib.get('width')),
1876 'height': int_or_none(representation_attrib.get('height')),
1877 'tbr': float_or_none(bandwidth, 1000),
1878 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1879 'fps': int_or_none(representation_attrib.get('frameRate')),
1880 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1881 'format_note': 'DASH %s' % content_type,
1882 'filesize': filesize,
1884 f.update(parse_codecs(representation_attrib.get('codecs')))
1885 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1887 def prepare_template(template_name, identifiers):
1888 t = representation_ms_info[template_name]
1889 t = t.replace('$RepresentationID$', representation_id)
1890 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1891 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1892 t.replace('$$', '$')
1895 # @initialization is a regular template like @media one
1896 # so it should be handled just the same way (see
1897 # https://github.com/rg3/youtube-dl/issues/11605)
1898 if 'initialization' in representation_ms_info:
1899 initialization_template = prepare_template(
1901 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1902 # $Time$ shall not be included for @initialization thus
1903 # only $Bandwidth$ remains
1905 representation_ms_info['initialization_url'] = initialization_template % {
1906 'Bandwidth': bandwidth,
1909 def location_key(location):
1910 return 'url' if re.match(r'^https?://', location) else 'path'
1912 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1914 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1915 media_location_key = location_key(media_template)
1917 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1918 # can't be used at the same time
1919 if '%(Number' in media_template and 's' not in representation_ms_info:
1920 segment_duration = None
1921 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
1922 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1923 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1924 representation_ms_info['fragments'] = [{
1925 media_location_key: media_template % {
1926 'Number': segment_number,
1927 'Bandwidth': bandwidth,
1929 'duration': segment_duration,
1930 } for segment_number in range(
1931 representation_ms_info['start_number'],
1932 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1934 # $Number*$ or $Time$ in media template with S list available
1935 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1936 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1937 representation_ms_info['fragments'] = []
1940 segment_number = representation_ms_info['start_number']
1942 def add_segment_url():
1943 segment_url = media_template % {
1944 'Time': segment_time,
1945 'Bandwidth': bandwidth,
1946 'Number': segment_number,
1948 representation_ms_info['fragments'].append({
1949 media_location_key: segment_url,
1950 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1953 for num, s in enumerate(representation_ms_info['s']):
1954 segment_time = s.get('t') or segment_time
1958 for r in range(s.get('r', 0)):
1959 segment_time += segment_d
1962 segment_time += segment_d
1963 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1965 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1966 # or any YouTube dashsegments video
1969 timescale = representation_ms_info['timescale']
1970 for s in representation_ms_info['s']:
1971 duration = float_or_none(s['d'], timescale)
1972 for r in range(s.get('r', 0) + 1):
1973 segment_uri = representation_ms_info['segment_urls'][segment_index]
1975 location_key(segment_uri): segment_uri,
1976 'duration': duration,
1979 representation_ms_info['fragments'] = fragments
1980 elif 'segment_urls' in representation_ms_info:
1981 # Segment URLs with no SegmentTimeline
1982 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
1983 # https://github.com/rg3/youtube-dl/pull/14844
1985 segment_duration = float_or_none(
1986 representation_ms_info['segment_duration'],
1987 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
1988 for segment_url in representation_ms_info['segment_urls']:
1990 location_key(segment_url): segment_url,
1992 if segment_duration:
1993 fragment['duration'] = segment_duration
1994 fragments.append(fragment)
1995 representation_ms_info['fragments'] = fragments
1996 # NB: MPD manifest may contain direct URLs to unfragmented media.
1997 # No fragments key is present in this case.
1998 if 'fragments' in representation_ms_info:
2000 'fragment_base_url': base_url,
2002 'protocol': 'http_dash_segments',
2004 if 'initialization_url' in representation_ms_info:
2005 initialization_url = representation_ms_info['initialization_url']
2006 if not f.get('url'):
2007 f['url'] = initialization_url
2008 f['fragments'].append({location_key(initialization_url): initialization_url})
2009 f['fragments'].extend(representation_ms_info['fragments'])
2011 existing_format = next(
2012 fo for fo in formats
2013 if fo['format_id'] == representation_id)
2014 except StopIteration:
2015 full_info = formats_dict.get(representation_id, {}).copy()
2017 formats.append(full_info)
2019 existing_format.update(f)
2021 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2024 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2025 res = self._download_webpage_handle(
2027 note=note or 'Downloading ISM manifest',
2028 errnote=errnote or 'Failed to download ISM manifest',
2034 return self._parse_ism_formats(
2035 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
2037 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2039 Parse formats
from ISM manifest
.
2041 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2042 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2044 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2047 duration = int(ism_doc.attrib['Duration'])
2048 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2051 for stream in ism_doc.findall('StreamIndex'):
2052 stream_type = stream.get('Type')
2053 if stream_type not in ('video', 'audio'):
2055 url_pattern = stream.attrib['Url']
2056 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2057 stream_name = stream.get('Name')
2058 for track in stream.findall('QualityLevel'):
2059 fourcc = track.get('FourCC')
2060 # TODO: add support for WVC1 and WMAP
2061 if fourcc not in ('H264', 'AVC1', 'AACL'):
2062 self.report_warning('%s is not a supported codec' % fourcc)
2064 tbr = int(track.attrib['Bitrate']) // 1000
2065 # [1] does not mention Width and Height attributes. However,
2066 # they're often present while MaxWidth and MaxHeight are
2067 # missing, so should be used as fallbacks
2068 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2069 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2070 sampling_rate = int_or_none(track.get('SamplingRate'))
2072 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2073 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2079 stream_fragments = stream.findall('c')
2080 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2081 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2082 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2083 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2084 if not fragment_ctx['duration']:
2086 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2088 next_fragment_time = duration
2089 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2090 for _ in range(fragment_repeat):
2092 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2093 'duration': fragment_ctx['duration'] / stream_timescale,
2095 fragment_ctx['time'] += fragment_ctx['duration']
2099 format_id.append(ism_id)
2101 format_id.append(stream_name)
2102 format_id.append(compat_str(tbr))
2105 'format_id': '-'.join(format_id),
2107 'manifest_url': ism_url,
2108 'ext': 'ismv' if stream_type == 'video' else 'isma',
2112 'asr': sampling_rate,
2113 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2114 'acodec': 'none' if stream_type == 'video' else fourcc,
2116 'fragments': fragments,
2117 '_download_params': {
2118 'duration': duration,
2119 'timescale': stream_timescale,
2120 'width': width or 0,
2121 'height': height or 0,
2123 'codec_private_data': track.get('CodecPrivateData'),
2124 'sampling_rate': sampling_rate,
2125 'channels': int_or_none(track.get('Channels', 2)),
2126 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2127 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2132 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2133 def absolute_url(video_url):
2134 return compat_urlparse.urljoin(base_url, video_url)
2136 def parse_content_type(content_type):
2137 if not content_type:
2139 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2141 mimetype, codecs = ctr.groups()
2142 f = parse_codecs(codecs)
2143 f['ext'] = mimetype2ext(mimetype)
2147 def _media_formats(src, cur_media_type, type_info={}):
2148 full_url = absolute_url(src)
2149 ext = type_info.get('ext') or determine_ext(full_url)
2151 is_plain_url = False
2152 formats = self._extract_m3u8_formats(
2153 full_url, video_id, ext='mp4',
2154 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2155 preference=preference, fatal=False)
2157 is_plain_url = False
2158 formats = self._extract_mpd_formats(
2159 full_url, video_id, mpd_id=mpd_id, fatal=False)
2164 'vcodec': 'none' if cur_media_type == 'audio' else None,
2166 return is_plain_url, formats
2169 # amp-video and amp-audio are very similar to their HTML5 counterparts
2170 # so we wll include them right here (see
2171 # https://www.ampproject.org/docs/reference/components/amp-video)
2172 media_tags = [(media_tag, media_type, '')
2173 for media_tag, media_type
2174 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2175 media_tags.extend(re.findall(
2176 # We only allow video|audio followed by a whitespace or '>'.
2177 # Allowing more characters may end up in significant slow down (see
2178 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2179 # http://www.porntrex.com/maps/videositemap.xml).
2180 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2181 for media_tag, media_type, media_content in media_tags:
2186 media_attributes = extract_attributes(media_tag)
2187 src = media_attributes.get('src')
2189 _, formats = _media_formats(src, media_type)
2190 media_info['formats'].extend(formats)
2191 media_info['thumbnail'] = media_attributes.get('poster')
2193 for source_tag in re.findall(r'<source[^>]+>', media_content):
2194 source_attributes = extract_attributes(source_tag)
2195 src = source_attributes.get('src')
2198 f = parse_content_type(source_attributes.get('type'))
2199 is_plain_url, formats = _media_formats(src, media_type, f)
2201 # res attribute is not standard but seen several times
2204 'height': int_or_none(source_attributes.get('res')),
2205 'format_id': source_attributes.get('label'),
2207 f.update(formats[0])
2208 media_info['formats'].append(f)
2210 media_info['formats'].extend(formats)
2211 for track_tag in re.findall(r'<track[^>]+>', media_content):
2212 track_attributes = extract_attributes(track_tag)
2213 kind = track_attributes.get('kind')
2214 if not kind or kind in ('subtitles', 'captions'):
2215 src = track_attributes.get('src')
2218 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2219 media_info['subtitles'].setdefault(lang, []).append({
2220 'url': absolute_url(src),
2222 if media_info['formats'] or media_info['subtitles']:
2223 entries.append(media_info)
2226 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2228 hdcore_sign = 'hdcore=3.7.0'
2229 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2230 hds_host = hosts.get('hds')
2232 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2233 if 'hdcore=' not in f4m_url:
2234 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2235 f4m_formats = self._extract_f4m_formats(
2236 f4m_url, video_id, f4m_id='hds', fatal=False)
2237 for entry in f4m_formats:
2238 entry.update({'extra_param_to_segment_url': hdcore_sign})
2239 formats.extend(f4m_formats)
2240 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2241 hls_host = hosts.get('hls')
2243 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2244 formats.extend(self._extract_m3u8_formats(
2245 m3u8_url, video_id, 'mp4', 'm3u8_native',
2246 m3u8_id='hls', fatal=False))
2249 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2250 query = compat_urlparse.urlparse(url).query
2251 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2252 url_base = self._search_regex(
2253 r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url')
2254 http_base_url = '%s:%s' % ('http', url_base)
2257 def manifest_url(manifest):
2258 m_url = '%s/%s' % (http_base_url, manifest)
2260 m_url += '?%s' % query
2263 if 'm3u8' not in skip_protocols:
2264 formats.extend(self._extract_m3u8_formats(
2265 manifest_url('playlist.m3u8'), video_id, 'mp4',
2266 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2267 if 'f4m' not in skip_protocols:
2268 formats.extend(self._extract_f4m_formats(
2269 manifest_url('manifest.f4m'),
2270 video_id, f4m_id='hds', fatal=False))
2271 if 'dash' not in skip_protocols:
2272 formats.extend(self._extract_mpd_formats(
2273 manifest_url('manifest.mpd'),
2274 video_id, mpd_id='dash', fatal=False))
2275 if re.search(r'(?:/smil:|\.smil)', url_base):
2276 if 'smil' not in skip_protocols:
2277 rtmp_formats = self._extract_smil_formats(
2278 manifest_url('jwplayer.smil'),
2279 video_id, fatal=False)
2280 for rtmp_format in rtmp_formats:
2281 rtsp_format = rtmp_format.copy()
2282 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2283 del rtsp_format['play_path']
2284 del rtsp_format['ext']
2285 rtsp_format.update({
2286 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2287 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2290 formats.extend([rtmp_format, rtsp_format])
2292 for protocol in ('rtmp', 'rtsp'):
2293 if protocol not in skip_protocols:
2295 'url': '%s:%s' % (protocol, url_base),
2296 'format_id': protocol,
2297 'protocol': protocol,
2301 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2303 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2307 jwplayer_data = self._parse_json(mobj.group('options'),
2309 transform_source=transform_source)
2310 except ExtractorError:
2313 if isinstance(jwplayer_data, dict):
2314 return jwplayer_data
2316 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2317 jwplayer_data = self._find_jwplayer_data(
2318 webpage, video_id, transform_source=js_to_json)
2319 return self._parse_jwplayer_data(
2320 jwplayer_data, video_id, *args, **kwargs)
2322 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2323 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2324 # JWPlayer backward compatibility: flattened playlists
2325 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2326 if 'playlist' not in jwplayer_data:
2327 jwplayer_data = {'playlist': [jwplayer_data]}
2331 # JWPlayer backward compatibility: single playlist item
2332 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2333 if not isinstance(jwplayer_data['playlist'], list):
2334 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2336 for video_data in jwplayer_data['playlist']:
2337 # JWPlayer backward compatibility: flattened sources
2338 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2339 if 'sources' not in video_data:
2340 video_data['sources'] = [video_data]
2342 this_video_id = video_id or video_data['mediaid']
2344 formats = self._parse_jwplayer_formats(
2345 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2346 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2349 tracks = video_data.get('tracks')
2350 if tracks and isinstance(tracks, list):
2351 for track in tracks:
2352 if not isinstance(track, dict):
2354 if track.get('kind') != 'captions':
2356 track_url = urljoin(base_url, track.get('file'))
2359 subtitles.setdefault(track.get('label') or 'en', []).append({
2360 'url': self._proto_relative_url(track_url)
2364 'id': this_video_id,
2365 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2366 'description': video_data.get('description'),
2367 'thumbnail': self._proto_relative_url(video_data.get('image')),
2368 'timestamp': int_or_none(video_data.get('pubdate')),
2369 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2370 'subtitles': subtitles,
2372 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2373 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2375 '_type': 'url_transparent',
2376 'url': formats[0]['url'],
2379 self._sort_formats(formats)
2380 entry['formats'] = formats
2381 entries.append(entry)
2382 if len(entries) == 1:
2385 return self.playlist_result(entries)
2387 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2388 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2391 for source in jwplayer_sources_data:
2392 if not isinstance(source, dict):
2394 source_url = self._proto_relative_url(source.get('file'))
2398 source_url = compat_urlparse.urljoin(base_url, source_url)
2399 if source_url in urls:
2401 urls.append(source_url)
2402 source_type = source.get('type') or ''
2403 ext = mimetype2ext(source_type) or determine_ext(source_url)
2404 if source_type == 'hls' or ext == 'm3u8':
2405 formats.extend(self._extract_m3u8_formats(
2406 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2407 m3u8_id=m3u8_id, fatal=False))
2409 formats.extend(self._extract_mpd_formats(
2410 source_url, video_id, mpd_id=mpd_id, fatal=False))
2412 formats.extend(self._extract_smil_formats(
2413 source_url, video_id, fatal=False))
2414 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2415 elif source_type.startswith('audio') or ext in (
2416 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2423 height = int_or_none(source.get('height'))
2425 # Often no height is provided but there is a label in
2426 # format like "1080p", "720p SD", or 1080.
2427 height = int_or_none(self._search_regex(
2428 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2429 'height', default=None))
2432 'width': int_or_none(source.get('width')),
2434 'tbr': int_or_none(source.get('bitrate')),
2437 if source_url.startswith('rtmp'):
2438 a_format['ext'] = 'flv'
2439 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2440 # of jwplayer.flash.swf
2441 rtmp_url_parts = re.split(
2442 r'((?:mp4|mp3|flv):)', source_url, 1)
2443 if len(rtmp_url_parts) == 3:
2444 rtmp_url, prefix, play_path = rtmp_url_parts
2447 'play_path': prefix + play_path,
2450 a_format.update(rtmp_params)
2451 formats.append(a_format)
2454 def _live_title(self, name):
2455 """ Generate the title
for a live video
"""
2456 now = datetime.datetime.now()
2457 now_str = now.strftime('%Y-%m-%d %H:%M')
2458 return name + ' ' + now_str
2460 def _int(self, v, name, fatal=False, **kwargs):
2461 res = int_or_none(v, **kwargs)
2462 if 'get_attr' in kwargs:
2463 print(getattr(v, kwargs['get_attr']))
2465 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2467 raise ExtractorError(msg)
2469 self._downloader.report_warning(msg)
2472 def _float(self, v, name, fatal=False, **kwargs):
2473 res = float_or_none(v, **kwargs)
2475 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2477 raise ExtractorError(msg)
2479 self._downloader.report_warning(msg)
2482 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2483 path='/', secure=False, discard=False, rest={}, **kwargs):
2484 cookie = compat_cookiejar.Cookie(
2485 0, name, value, port, port is not None, domain, True,
2486 domain.startswith('.'), path, True, secure, expire_time,
2487 discard, None, None, rest)
2488 self._downloader.cookiejar.set_cookie(cookie)
2490 def _get_cookies(self, url):
2491 """ Return a compat_cookies
.SimpleCookie
with the cookies
for the url
"""
2492 req = sanitized_Request(url)
2493 self._downloader.cookiejar.add_cookie_header(req)
2494 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2496 def get_testcases(self, include_onlymatching=False):
2497 t = getattr(self, '_TEST', None)
2499 assert not hasattr(self, '_TESTS'), \
2500 '%s has _TEST and _TESTS' % type(self).__name__
2503 tests = getattr(self, '_TESTS', [])
2505 if not include_onlymatching and t.get('only_matching', False):
2507 t['name'] = type(self).__name__[:-len('IE')]
2510 def is_suitable(self, age_limit):
2511 """ Test whether the extractor
is generally suitable
for the given
2512 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
2514 any_restricted = False
2515 for tc in self.get_testcases(include_onlymatching=False):
2516 if tc.get('playlist', []):
2517 tc = tc['playlist'][0]
2518 is_restricted = age_restricted(
2519 tc.get('info_dict', {}).get('age_limit'), age_limit)
2520 if not is_restricted:
2522 any_restricted = any_restricted or is_restricted
2523 return not any_restricted
2525 def extract_subtitles(self, *args, **kwargs):
2526 if (self._downloader.params.get('writesubtitles', False) or
2527 self._downloader.params.get('listsubtitles')):
2528 return self._get_subtitles(*args, **kwargs)
2531 def _get_subtitles(self, *args, **kwargs):
2532 raise NotImplementedError('This method must be implemented by subclasses')
2535 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2536 """ Merge subtitle items
for one language
. Items
with duplicated URLs
2537 will be dropped
. """
2538 list1_urls = set([item['url'] for item in subtitle_list1])
2539 ret = list(subtitle_list1)
2540 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2544 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2545 """ Merge two subtitle dictionaries
, language by language
. """
2546 ret = dict(subtitle_dict1)
2547 for lang in subtitle_dict2:
2548 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2551 def extract_automatic_captions(self, *args, **kwargs):
2552 if (self._downloader.params.get('writeautomaticsub', False) or
2553 self._downloader.params.get('listsubtitles')):
2554 return self._get_automatic_captions(*args, **kwargs)
2557 def _get_automatic_captions(self, *args, **kwargs):
2558 raise NotImplementedError('This method must be implemented by subclasses')
2560 def mark_watched(self, *args, **kwargs):
2561 if (self._downloader.params.get('mark_watched', False) and
2562 (self._get_login_info()[0] is not None or
2563 self._downloader.params.get('cookiefile') is not None)):
2564 self._mark_watched(*args, **kwargs)
2566 def _mark_watched(self, *args, **kwargs):
2567 raise NotImplementedError('This method must be implemented by subclasses')
2569 def geo_verification_headers(self):
2571 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2572 if geo_verification_proxy:
2573 headers['Ytdl-request-proxy'] = geo_verification_proxy
2576 def _generic_id(self, url):
2577 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2579 def _generic_title(self, url):
2580 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2583 class SearchInfoExtractor(InfoExtractor):
2585 Base
class for paged search queries extractors
.
2586 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
2587 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
2591 def _make_valid_url(cls):
2592 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2595 def suitable(cls, url):
2596 return re.match(cls._make_valid_url(), url) is not None
2598 def _real_extract(self, query):
2599 mobj = re.match(self._make_valid_url(), query)
2601 raise ExtractorError('Invalid search query "%s"' % query)
2603 prefix = mobj.group('prefix')
2604 query = mobj.group('query')
2606 return self._get_n_results(query, 1)
2607 elif prefix == 'all':
2608 return self._get_n_results(query, self._MAX_RESULTS)
2612 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2613 elif n > self._MAX_RESULTS:
2614 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2615 n = self._MAX_RESULTS
2616 return self._get_n_results(query, n)
2618 def _get_n_results(self, query, n):
2619 """Get a specified number of results
for a query
"""
2620 raise NotImplementedError('This method must be implemented by subclasses')
2623 def SEARCH_KEY(self):
2624 return self._SEARCH_KEY