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
.f4m
import (
37 remove_encrypted_media
,
65 parse_m3u8_attributes
,
86 class InfoExtractor(object):
87 """Information Extractor class.
89 Information extractors are the classes that, given a URL, extract
90 information about the video (or videos) the URL refers to. This
91 information includes the real video URL, the video title, author and
92 others. The information is stored in a dictionary which is then
93 passed to the YoutubeDL. The YoutubeDL processes this
94 information possibly downloading the video to the file system, among
95 other possible outcomes.
97 The type field determines the type of the result.
98 By far the most common value (and the default if _type is missing) is
99 "video", which indicates a single video.
101 For a video, the dictionaries must include the following fields:
103 id: Video identifier.
104 title: Video title, unescaped.
106 Additionally, it must contain either a formats entry or a url one:
108 formats: A list of dictionaries for each format available, ordered
109 from worst to best quality.
112 * url The mandatory URL representing the media:
113 for plain file media - HTTP URL of this file,
115 for HLS - URL of the M3U8 media playlist,
116 for HDS - URL of the F4M manifest,
118 - HTTP URL to plain file media (in case of
120 - URL of the MPD manifest or base URL
121 representing the media if MPD manifest
122 is parsed from a string (in case of
124 for MSS - URL of the ISM manifest.
126 The URL of the manifest file in case of
128 for HLS - URL of the M3U8 master playlist,
129 for HDS - URL of the F4M manifest,
130 for DASH - URL of the MPD manifest,
131 for MSS - URL of the ISM manifest.
132 * ext Will be calculated from URL if missing
133 * format A human-readable description of the format
134 ("mp4 container with h264/opus").
135 Calculated from the format_id, width, height.
136 and format_note fields if missing.
137 * format_id A short description of the format
138 ("mp4_h264_opus" or "19").
139 Technically optional, but strongly recommended.
140 * format_note Additional info about the format
141 ("3D" or "DASH video")
142 * width Width of the video, if known
143 * height Height of the video, if known
144 * resolution Textual description of width and height
145 * tbr Average bitrate of audio and video in KBit/s
146 * abr Average audio bitrate in KBit/s
147 * acodec Name of the audio codec in use
148 * asr Audio sampling rate in Hertz
149 * vbr Average video bitrate in KBit/s
151 * vcodec Name of the video codec in use
152 * container Name of the container format
153 * filesize The number of bytes, if known in advance
154 * filesize_approx An estimate for the number of bytes
155 * player_url SWF Player URL (used for rtmpdump).
156 * protocol The protocol that will be used for the actual
157 download, lower-case.
158 "http", "https", "rtsp", "rtmp", "rtmpe",
159 "m3u8", "m3u8_native" or "http_dash_segments".
161 Base URL for fragments. Each fragment's path
162 value (if present) will be relative to
164 * fragments A list of fragments of a fragmented media.
165 Each fragment entry must contain either an url
166 or a path. If an url is present it should be
167 considered by a client. Otherwise both path and
168 fragment_base_url must be present. Here is
169 the list of all potential fields:
170 * "url" - fragment's URL
171 * "path" - fragment's path relative to
173 * "duration" (optional, int or float)
174 * "filesize" (optional, int)
175 * preference Order number of this format. If this field is
176 present and not None, the formats get sorted
177 by this field, regardless of all other values.
178 -1 for default (order by other properties),
179 -2 or smaller for less than default.
180 < -1000 to hide the format (if there is
181 another one which is strictly better)
182 * language Language code, e.g. "de" or "en-US".
183 * language_preference Is this in the language mentioned in
185 10 if it's what the URL is about,
186 -1 for default (don't know),
187 -10 otherwise, other values reserved for now.
188 * quality Order number of the video quality of this
189 format, irrespective of the file format.
190 -1 for default (order by other properties),
191 -2 or smaller for less than default.
192 * source_preference Order number for this video source
193 (quality takes higher priority)
194 -1 for default (order by other properties),
195 -2 or smaller for less than default.
196 * http_headers A dictionary of additional HTTP headers
197 to add to the request.
198 * stretched_ratio If given and not 1, indicates that the
199 video's pixels are not square.
200 width : height ratio as float.
201 * no_resume The server does not support resuming the
202 (HTTP or RTMP) download. Boolean.
203 * downloader_options A dictionary of downloader options as
204 described in FileDownloader
206 url: Final video URL.
207 ext: Video filename extension.
208 format: The video format, defaults to ext (used for --get-format)
209 player_url: SWF Player URL (used for rtmpdump).
211 The following fields are optional:
213 alt_title: A secondary title of the video.
214 display_id An alternative identifier for the video, not necessarily
215 unique, but available before title. Typically, id is
216 something like "4234987", title "Dancing naked mole rats",
217 and display_id "dancing-naked-mole-rats"
218 thumbnails: A list of dictionaries, with the following entries:
219 * "id" (optional, string) - Thumbnail format ID
221 * "preference" (optional, int) - quality of the image
222 * "width" (optional, int)
223 * "height" (optional, int)
224 * "resolution" (optional, string "{width}x{height}",
226 * "filesize" (optional, int)
227 thumbnail: Full URL to a video thumbnail image.
228 description: Full video description.
229 uploader: Full name of the video uploader.
230 license: License name the video is licensed under.
231 creator: The creator of the video.
232 release_date: The date (YYYYMMDD) when the video was released.
233 timestamp: UNIX timestamp of the moment the video became available.
234 upload_date: Video upload date (YYYYMMDD).
235 If not explicitly set, calculated from timestamp.
236 uploader_id: Nickname or id of the video uploader.
237 uploader_url: Full URL to a personal webpage of the video uploader.
238 channel: Full name of the channel the video is uploaded on.
239 Note that channel fields may or may not repeat uploader
240 fields. This depends on a particular extractor.
241 channel_id: Id of the channel.
242 channel_url: Full URL to a channel webpage.
243 location: Physical location where the video was filmed.
244 subtitles: The available subtitles as a dictionary in the format
245 {tag: subformats}. "tag" is usually a language code, and
246 "subformats" is a list sorted from lower to higher
247 preference, each element is a dictionary with the "ext"
249 * "data": The subtitles file contents
250 * "url": A URL pointing to the subtitles file
251 "ext" will be calculated from URL if missing
252 automatic_captions: Like 'subtitles', used by the YoutubeIE for
253 automatically generated captions
254 duration: Length of the video in seconds, as an integer or float.
255 view_count: How many users have watched the video on the platform.
256 like_count: Number of positive ratings of the video
257 dislike_count: Number of negative ratings of the video
258 repost_count: Number of reposts of the video
259 average_rating: Average rating give by users, the scale used depends on the webpage
260 comment_count: Number of comments on the video
261 comments: A list of comments, each with one or more of the following
262 properties (all but one of text or html optional):
263 * "author" - human-readable name of the comment author
264 * "author_id" - user ID of the comment author
266 * "html" - Comment as HTML
267 * "text" - Plain text of the comment
268 * "timestamp" - UNIX timestamp of comment
269 * "parent" - ID of the comment this one is replying to.
270 Set to "root" to indicate that this is a
271 comment to the original video.
272 age_limit: Age restriction for the video, as an integer (years)
273 webpage_url: The URL to the video webpage, if given to youtube-dl it
274 should allow to get the same result again. (It will be set
275 by YoutubeDL if it's missing)
276 categories: A list of categories that the video falls in, for example
278 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
279 is_live: True, False, or None (=unknown). Whether this video is a
280 live stream that goes on instead of a fixed-length video.
281 start_time: Time in seconds where the reproduction should start, as
282 specified in the URL.
283 end_time: Time in seconds where the reproduction should end, as
284 specified in the URL.
285 chapters: A list of dictionaries, with the following entries:
286 * "start_time" - The start time of the chapter in seconds
287 * "end_time" - The end time of the chapter in seconds
288 * "title" (optional, string)
290 The following fields should only be used when the video belongs to some logical
293 chapter: Name or title of the chapter the video belongs to.
294 chapter_number: Number of the chapter the video belongs to, as an integer.
295 chapter_id: Id of the chapter the video belongs to, as a unicode string.
297 The following fields should only be used when the video is an episode of some
298 series, programme or podcast:
300 series: Title of the series or programme the video episode belongs to.
301 season: Title of the season the video episode belongs to.
302 season_number: Number of the season the video episode belongs to, as an integer.
303 season_id: Id of the season the video episode belongs to, as a unicode string.
304 episode: Title of the video episode. Unlike mandatory video title field,
305 this field should denote the exact title of the video episode
306 without any kind of decoration.
307 episode_number: Number of the video episode within a season, as an integer.
308 episode_id: Id of the video episode, as a unicode string.
310 The following fields should only be used when the media is a track or a part of
313 track: Title of the track.
314 track_number: Number of the track within an album or a disc, as an integer.
315 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
317 artist: Artist(s) of the track.
318 genre: Genre(s) of the track.
319 album: Title of the album the track belongs to.
320 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
321 album_artist: List of all artists appeared on the album (e.g.
322 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
324 disc_number: Number of the disc or other physical medium the track belongs to,
326 release_year: Year (YYYY) when the album was released.
328 Unless mentioned otherwise, the fields should be Unicode strings.
330 Unless mentioned otherwise, None is equivalent to absence of information.
333 _type "playlist" indicates multiple videos.
334 There must be a key "entries", which is a list, an iterable, or a PagedList
335 object, each element of which is a valid dictionary by this specification.
337 Additionally, playlists can have "id", "title", "description", "uploader",
338 "uploader_id", "uploader_url" attributes with the same semantics as videos
342 _type "multi_video" indicates that there are multiple videos that
343 form a single show, for examples multiple acts of an opera or TV episode.
344 It must have an entries key like a playlist and contain all the keys
345 required for a video at the same time.
348 _type "url" indicates that the video must be extracted from another
349 location, possibly by a different extractor. Its only required key is:
350 "url" - the next URL to extract.
351 The key "ie_key" can be set to the class name (minus the trailing "IE",
352 e.g. "Youtube") if the extractor class is known in advance.
353 Additionally, the dictionary may have any properties of the resolved entity
354 known in advance, for example "title" if the title of the referred video is
358 _type "url_transparent" entities have the same specification as "url", but
359 indicate that the given additional information is more precise than the one
360 associated with the resolved URL.
361 This is useful when a site employs a video service that hosts the video and
362 its technical metadata, but that video service does not embed a useful
363 title, description etc.
366 Subclasses of this one should re-define the _real_initialize() and
367 _real_extract() methods and define a _VALID_URL regexp.
368 Probably, they should also be added to the list of extractors.
370 _GEO_BYPASS attribute may be set to False in order to disable
371 geo restriction bypass mechanisms for a particular extractor.
372 Though it won't disable explicit geo restriction bypass based on
373 country code provided with geo_bypass_country.
375 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
376 countries for this extractor. One of these countries will be used by
377 geo restriction bypass mechanism right away in order to bypass
378 geo restriction, of course, if the mechanism is not disabled.
380 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
381 IP blocks in CIDR notation for this extractor. One of these IP blocks
382 will be used by geo restriction bypass mechanism similarly
385 Finally, the _WORKING attribute should be set to False for broken IEs
386 in order to warn the users and skip the tests.
391 _x_forwarded_for_ip
= None
393 _GEO_COUNTRIES
= None
394 _GEO_IP_BLOCKS
= None
397 def __init__(self
, downloader
=None):
398 """Constructor. Receives an optional downloader."""
400 self
._x
_forwarded
_for
_ip
= None
401 self
.set_downloader(downloader
)
404 def suitable(cls
, url
):
405 """Receives a URL and returns True if suitable for this IE."""
407 # This does not use has/getattr intentionally - we want to know whether
408 # we have cached the regexp for *this* class, whereas getattr would also
409 # match the superclass
410 if '_VALID_URL_RE' not in cls
.__dict
__:
411 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
412 return cls
._VALID
_URL
_RE
.match(url
) is not None
415 def _match_id(cls
, url
):
416 if '_VALID_URL_RE' not in cls
.__dict
__:
417 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
418 m
= cls
._VALID
_URL
_RE
.match(url
)
420 return compat_str(m
.group('id'))
424 """Getter method for _WORKING."""
427 def initialize(self
):
428 """Initializes an instance (authentication, etc)."""
429 self
._initialize
_geo
_bypass
({
430 'countries': self
._GEO
_COUNTRIES
,
431 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
434 self
._real
_initialize
()
437 def _initialize_geo_bypass(self
, geo_bypass_context
):
439 Initialize geo restriction bypass mechanism.
441 This method is used to initialize geo bypass mechanism based on faking
442 X-Forwarded-For HTTP header. A random country from provided country list
443 is selected and a random IP belonging to this country is generated. This
444 IP will be passed as X-Forwarded-For HTTP header in all subsequent
447 This method will be used for initial geo bypass mechanism initialization
448 during the instance initialization with _GEO_COUNTRIES and
451 You may also manually call it from extractor's code if geo bypass
452 information is not available beforehand (e.g. obtained during
453 extraction) or due to some other reason. In this case you should pass
454 this information in geo bypass context passed as first argument. It may
455 contain following fields:
457 countries: List of geo unrestricted countries (similar
459 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
460 (similar to _GEO_IP_BLOCKS)
463 if not self
._x
_forwarded
_for
_ip
:
465 # Geo bypass mechanism is explicitly disabled by user
466 if not self
._downloader
.params
.get('geo_bypass', True):
469 if not geo_bypass_context
:
470 geo_bypass_context
= {}
472 # Backward compatibility: previously _initialize_geo_bypass
473 # expected a list of countries, some 3rd party code may still use
475 if isinstance(geo_bypass_context
, (list, tuple)):
476 geo_bypass_context
= {
477 'countries': geo_bypass_context
,
480 # The whole point of geo bypass mechanism is to fake IP
481 # as X-Forwarded-For HTTP header based on some IP block or
484 # Path 1: bypassing based on IP block in CIDR notation
486 # Explicit IP block specified by user, use it right away
487 # regardless of whether extractor is geo bypassable or not
488 ip_block
= self
._downloader
.params
.get('geo_bypass_ip_block', None)
490 # Otherwise use random IP block from geo bypass context but only
491 # if extractor is known as geo bypassable
493 ip_blocks
= geo_bypass_context
.get('ip_blocks')
494 if self
._GEO
_BYPASS
and ip_blocks
:
495 ip_block
= random
.choice(ip_blocks
)
498 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
499 if self
._downloader
.params
.get('verbose', False):
500 self
._downloader
.to_screen(
501 '[debug] Using fake IP %s as X-Forwarded-For.'
502 % self
._x
_forwarded
_for
_ip
)
505 # Path 2: bypassing based on country code
507 # Explicit country code specified by user, use it right away
508 # regardless of whether extractor is geo bypassable or not
509 country
= self
._downloader
.params
.get('geo_bypass_country', None)
511 # Otherwise use random country code from geo bypass context but
512 # only if extractor is known as geo bypassable
514 countries
= geo_bypass_context
.get('countries')
515 if self
._GEO
_BYPASS
and countries
:
516 country
= random
.choice(countries
)
519 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
520 if self
._downloader
.params
.get('verbose', False):
521 self
._downloader
.to_screen(
522 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
523 % (self
._x
_forwarded
_for
_ip
, country
.upper()))
525 def extract(self
, url
):
526 """Extracts URL information and returns it in list of dicts."""
531 ie_result
= self
._real
_extract
(url
)
532 if self
._x
_forwarded
_for
_ip
:
533 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
535 except GeoRestrictedError
as e
:
536 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
539 except ExtractorError
:
541 except compat_http_client
.IncompleteRead
as e
:
542 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
543 except (KeyError, StopIteration) as e
:
544 raise ExtractorError('An extractor error has occurred.', cause
=e
)
546 def __maybe_fake_ip_and_retry(self
, countries
):
547 if (not self
._downloader
.params
.get('geo_bypass_country', None)
549 and self
._downloader
.params
.get('geo_bypass', True)
550 and not self
._x
_forwarded
_for
_ip
552 country_code
= random
.choice(countries
)
553 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
554 if self
._x
_forwarded
_for
_ip
:
556 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
557 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
561 def set_downloader(self
, downloader
):
562 """Sets the downloader for this IE."""
563 self
._downloader
= downloader
565 def _real_initialize(self
):
566 """Real initialization process. Redefine in subclasses."""
569 def _real_extract(self
, url
):
570 """Real extraction process. Redefine in subclasses."""
575 """A string for getting the InfoExtractor with get_info_extractor"""
576 return compat_str(cls
.__name
__[:-2])
580 return compat_str(type(self
).__name
__[:-2])
583 def __can_accept_status_code(err
, expected_status
):
584 assert isinstance(err
, compat_urllib_error
.HTTPError
)
585 if expected_status
is None:
587 if isinstance(expected_status
, compat_integer_types
):
588 return err
.code
== expected_status
589 elif isinstance(expected_status
, (list, tuple)):
590 return err
.code
in expected_status
591 elif callable(expected_status
):
592 return expected_status(err
.code
) is True
596 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
598 Return the response handle.
600 See _download_webpage docstring for arguments specification.
603 self
.report_download_webpage(video_id
)
604 elif note
is not False:
606 self
.to_screen('%s' % (note
,))
608 self
.to_screen('%s: %s' % (video_id
, note
))
610 # Some sites check X-Forwarded-For HTTP header in order to figure out
611 # the origin of the client behind proxy. This allows bypassing geo
612 # restriction by faking this header's value to IP that belongs to some
613 # geo unrestricted country. We will do so once we encounter any
614 # geo restriction error.
615 if self
._x
_forwarded
_for
_ip
:
616 if 'X-Forwarded-For' not in headers
:
617 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
619 if isinstance(url_or_request
, compat_urllib_request
.Request
):
620 url_or_request
= update_Request(
621 url_or_request
, data
=data
, headers
=headers
, query
=query
)
624 url_or_request
= update_url_query(url_or_request
, query
)
625 if data
is not None or headers
:
626 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
627 exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
628 if hasattr(ssl
, 'CertificateError'):
629 exceptions
.append(ssl
.CertificateError
)
631 return self
._downloader
.urlopen(url_or_request
)
632 except tuple(exceptions
) as err
:
633 if isinstance(err
, compat_urllib_error
.HTTPError
):
634 if self
.__can
_accept
_status
_code
(err
, expected_status
):
635 # Retain reference to error to prevent file object from
636 # being closed before it can be read. Works around the
637 # effects of <https://bugs.python.org/issue15002>
638 # introduced in Python 3.4.1.
645 errnote
= 'Unable to download webpage'
647 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
649 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
651 self
._downloader
.report_warning(errmsg
)
654 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):
656 Return a tuple (page content as string, URL handle).
658 See _download_webpage docstring for arguments specification.
660 # Strip hashes from the URL (#1038)
661 if isinstance(url_or_request
, (compat_str
, str)):
662 url_or_request
= url_or_request
.partition('#')[0]
664 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
668 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
669 return (content
, urlh
)
672 def _guess_encoding_from_content(content_type
, webpage_bytes
):
673 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
675 encoding
= m
.group(1)
677 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
678 webpage_bytes[:1024])
680 encoding = m.group(1).decode('ascii')
681 elif webpage_bytes.startswith(b'\xff\xfe'):
688 def __check_blocked(self, content):
689 first_block = content[:512]
690 if ('<title>Access to this site is blocked</title>' in content
691 and 'Websense' in first_block):
692 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
693 blocked_iframe = self._html_search_regex(
694 r'<iframe src="([^
"]+)"', content,
695 'Websense information URL
', default=None)
697 msg += ' Visit
%s for more details
' % blocked_iframe
698 raise ExtractorError(msg, expected=True)
699 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
701 'Access to this webpage has been blocked by Indian censorship
. '
702 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
703 block_msg = self._html_search_regex(
704 r'</h1
><p
>(.*?
)</p
>',
705 content, 'block message
', default=None)
707 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
708 raise ExtractorError(msg, expected=True)
709 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
710 and 'blocklist
.rkn
.gov
.ru
' in content):
711 raise ExtractorError(
712 'Access to this webpage has been blocked by decision of the Russian government
. '
713 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
716 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
717 content_type = urlh.headers.get('Content
-Type
', '')
718 webpage_bytes = urlh.read()
719 if prefix is not None:
720 webpage_bytes = prefix + webpage_bytes
722 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
723 if self._downloader.params.get('dump_intermediate_pages
', False):
724 self.to_screen('Dumping request to
' + urlh.geturl())
725 dump = base64.b64encode(webpage_bytes).decode('ascii
')
726 self._downloader.to_screen(dump)
727 if self._downloader.params.get('write_pages
', False):
728 basen = '%s_%s' % (video_id, urlh.geturl())
730 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
731 basen = basen[:240 - len(h)] + h
732 raw_filename = basen + '.dump
'
733 filename = sanitize_filename(raw_filename, restricted=True)
734 self.to_screen('Saving request to
' + filename)
735 # Working around MAX_PATH limitation on Windows (see
736 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
737 if compat_os_name == 'nt
':
738 absfilepath = os.path.abspath(filename)
739 if len(absfilepath) > 259:
740 filename = '\\\\?
\\' + absfilepath
741 with open(filename, 'wb
') as outf:
742 outf.write(webpage_bytes)
745 content = webpage_bytes.decode(encoding, 'replace
')
747 content = webpage_bytes.decode('utf
-8', 'replace
')
749 self.__check_blocked(content)
753 def _download_webpage(
754 self, url_or_request, video_id, note=None, errnote=None,
755 fatal=True, tries=1, timeout=5, encoding=None, data=None,
756 headers={}, query={}, expected_status=None):
758 Return the data of the page as a string.
761 url_or_request -- plain text URL as a string or
762 a compat_urllib_request.Requestobject
763 video_id -- Video/playlist/item identifier (string)
766 note -- note printed before downloading (string)
767 errnote -- note printed in case of an error (string)
768 fatal -- flag denoting whether error should be considered fatal,
769 i.e. whether it should cause ExtractionError to be raised,
770 otherwise a warning will be reported and extraction continued
771 tries -- number of tries
772 timeout -- sleep interval between tries
773 encoding -- encoding for a page content decoding, guessed automatically
774 when not explicitly specified
775 data -- POST data (bytes)
776 headers -- HTTP headers (dict)
777 query -- URL query (dict)
778 expected_status -- allows to accept failed HTTP requests (non 2xx
779 status code) by explicitly specifying a set of accepted status
780 codes. Can be any of the following entities:
781 - an integer type specifying an exact failed status code to
783 - a list or a tuple of integer types specifying a list of
784 failed status codes to accept
785 - a callable accepting an actual failed status code and
786 returning True if it should be accepted
787 Note that this argument does not affect success status codes (2xx)
788 which are always accepted.
793 while success is False:
795 res = self._download_webpage_handle(
796 url_or_request, video_id, note, errnote, fatal,
797 encoding=encoding, data=data, headers=headers, query=query,
798 expected_status=expected_status)
800 except compat_http_client.IncompleteRead as e:
802 if try_count >= tries:
804 self._sleep(timeout, video_id)
811 def _download_xml_handle(
812 self, url_or_request, video_id, note='Downloading XML
',
813 errnote='Unable to download XML
', transform_source=None,
814 fatal=True, encoding=None, data=None, headers={}, query={},
815 expected_status=None):
817 Return a tuple (xml as an compat_etree_Element, URL handle).
819 See _download_webpage docstring for arguments specification.
821 res = self._download_webpage_handle(
822 url_or_request, video_id, note, errnote, fatal=fatal,
823 encoding=encoding, data=data, headers=headers, query=query,
824 expected_status=expected_status)
827 xml_string, urlh = res
828 return self._parse_xml(
829 xml_string, video_id, transform_source=transform_source,
833 self, url_or_request, video_id,
834 note='Downloading XML
', errnote='Unable to download XML
',
835 transform_source=None, fatal=True, encoding=None,
836 data=None, headers={}, query={}, expected_status=None):
838 Return the xml as an compat_etree_Element.
840 See _download_webpage docstring for arguments specification.
842 res = self._download_xml_handle(
843 url_or_request, video_id, note=note, errnote=errnote,
844 transform_source=transform_source, fatal=fatal, encoding=encoding,
845 data=data, headers=headers, query=query,
846 expected_status=expected_status)
847 return res if res is False else res[0]
849 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
851 xml_string = transform_source(xml_string)
853 return compat_etree_fromstring(xml_string.encode('utf
-8'))
854 except compat_xml_parse_error as ve:
855 errmsg = '%s: Failed to parse XML
' % video_id
857 raise ExtractorError(errmsg, cause=ve)
859 self.report_warning(errmsg + str(ve))
861 def _download_json_handle(
862 self, url_or_request, video_id, note='Downloading JSON metadata
',
863 errnote='Unable to download JSON metadata
', transform_source=None,
864 fatal=True, encoding=None, data=None, headers={}, query={},
865 expected_status=None):
867 Return a tuple (JSON object, URL handle).
869 See _download_webpage docstring for arguments specification.
871 res = self._download_webpage_handle(
872 url_or_request, video_id, note, errnote, fatal=fatal,
873 encoding=encoding, data=data, headers=headers, query=query,
874 expected_status=expected_status)
877 json_string, urlh = res
878 return self._parse_json(
879 json_string, video_id, transform_source=transform_source,
883 self, url_or_request, video_id, note='Downloading JSON metadata
',
884 errnote='Unable to download JSON metadata
', transform_source=None,
885 fatal=True, encoding=None, data=None, headers={}, query={},
886 expected_status=None):
888 Return the JSON object as a dict.
890 See _download_webpage docstring for arguments specification.
892 res = self._download_json_handle(
893 url_or_request, video_id, note=note, errnote=errnote,
894 transform_source=transform_source, fatal=fatal, encoding=encoding,
895 data=data, headers=headers, query=query,
896 expected_status=expected_status)
897 return res if res is False else res[0]
899 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
901 json_string = transform_source(json_string)
903 return json.loads(json_string)
904 except ValueError as ve:
905 errmsg = '%s: Failed to parse JSON
' % video_id
907 raise ExtractorError(errmsg, cause=ve)
909 self.report_warning(errmsg + str(ve))
911 def report_warning(self, msg, video_id=None):
912 idstr = '' if video_id is None else '%s: ' % video_id
913 self._downloader.report_warning(
914 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
916 def to_screen(self, msg):
917 """Print msg to screen, prefixing it with '[ie_name
]'"""
918 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
920 def report_extraction(self, id_or_name):
921 """Report information extraction."""
922 self.to_screen('%s: Extracting information
' % id_or_name)
924 def report_download_webpage(self, video_id):
925 """Report webpage download."""
926 self.to_screen('%s: Downloading webpage
' % video_id)
928 def report_age_confirmation(self):
929 """Report attempt to confirm age."""
930 self.to_screen('Confirming age
')
932 def report_login(self):
933 """Report attempt to log in."""
934 self.to_screen('Logging
in')
937 def raise_login_required(msg='This video
is only available
for registered users
'):
938 raise ExtractorError(
939 '%s. Use
--username
and --password
or --netrc to provide account credentials
.' % msg,
943 def raise_geo_restricted(msg='This video
is not available
from your location due to geo restriction
', countries=None):
944 raise GeoRestrictedError(msg, countries=countries)
946 # Methods for following #608
948 def url_result(url, ie=None, video_id=None, video_title=None):
949 """Returns a URL that points to a page that should be processed"""
950 # TODO: ie should be the class used for getting the info
951 video_info = {'_type
': 'url
',
954 if video_id is not None:
955 video_info['id'] = video_id
956 if video_title is not None:
957 video_info['title
'] = video_title
960 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
962 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
964 return self.playlist_result(
965 urls, playlist_id=playlist_id, playlist_title=playlist_title)
968 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
969 """Returns a playlist"""
970 video_info = {'_type
': 'playlist
',
973 video_info['id'] = playlist_id
975 video_info['title
'] = playlist_title
976 if playlist_description:
977 video_info['description
'] = playlist_description
980 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
982 Perform a regex search on the given string, using a single or a list of
983 patterns returning the first matching group.
984 In case of failure return a default value or raise a WARNING or a
985 RegexNotFoundError, depending on fatal, specifying the field name.
987 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
988 mobj = re.search(pattern, string, flags)
991 mobj = re.search(p, string, flags)
995 if not self._downloader.params.get('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
996 _name = '\033[0;34m
%s\033[0m
' % name
1002 # return the first matching group
1003 return next(g for g in mobj.groups() if g is not None)
1005 return mobj.group(group)
1006 elif default is not NO_DEFAULT:
1009 raise RegexNotFoundError('Unable to extract
%s' % _name)
1011 self._downloader.report_warning('unable to extract
%s' % _name + bug_reports_message())
1014 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1016 Like _search_regex, but strips HTML tags and unescapes entities.
1018 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1020 return clean_html(res).strip()
1024 def _get_netrc_login_info(self, netrc_machine=None):
1027 netrc_machine = netrc_machine or self._NETRC_MACHINE
1029 if self._downloader.params.get('usenetrc
', False):
1031 info = netrc.netrc().authenticators(netrc_machine)
1032 if info is not None:
1036 raise netrc.NetrcParseError(
1037 'No authenticators
for %s' % netrc_machine)
1038 except (IOError, netrc.NetrcParseError) as err:
1039 self._downloader.report_warning(
1040 'parsing
.netrc
: %s' % error_to_compat_str(err))
1042 return username, password
1044 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1046 Get the login info as (username, password)
1047 First look for the manually specified credentials using username_option
1048 and password_option as keys in params dictionary. If no such credentials
1049 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1051 If there's no info available
, return (None, None)
1053 if self._downloader is None:
1056 downloader_params = self._downloader.params
1058 # Attempt to use provided username and password or .netrc data
1059 if downloader_params.get(username_option) is not None:
1060 username = downloader_params[username_option]
1061 password = downloader_params[password_option]
1063 username, password = self._get_netrc_login_info(netrc_machine)
1065 return username, password
1067 def _get_tfa_info(self, note='two-factor verification code'):
1069 Get the two
-factor authentication info
1070 TODO
- asking the user will be required
for sms
/phone verify
1071 currently just uses the command line option
1072 If there
's no info available, return None
1074 if self._downloader is None:
1076 downloader_params = self._downloader.params
1078 if downloader_params.get('twofactor
') is not None:
1079 return downloader_params['twofactor
']
1081 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1083 # Helper functions for extracting OpenGraph info
1085 def _og_regexes(prop):
1086 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1087 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1088 % {'prop': re.escape(prop)})
1089 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1091 template % (property_re, content_re),
1092 template % (content_re, property_re),
1096 def _meta_regex(prop):
1097 return r'''(?isx)<meta
1098 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1099 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1101 def _og_search_property(self, prop, html, name=None, **kargs):
1102 if not isinstance(prop, (list, tuple)):
1105 name = 'OpenGraph
%s' % prop[0]
1108 og_regexes.extend(self._og_regexes(p))
1109 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1112 return unescapeHTML(escaped)
1114 def _og_search_thumbnail(self, html, **kargs):
1115 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1117 def _og_search_description(self, html, **kargs):
1118 return self._og_search_property('description
', html, fatal=False, **kargs)
1120 def _og_search_title(self, html, **kargs):
1121 return self._og_search_property('title
', html, **kargs)
1123 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1124 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1126 regexes = self._og_regexes('video
:secure_url
') + regexes
1127 return self._html_search_regex(regexes, html, name, **kargs)
1129 def _og_search_url(self, html, **kargs):
1130 return self._og_search_property('url
', html, **kargs)
1132 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1133 if not isinstance(name, (list, tuple)):
1135 if display_name is None:
1136 display_name = name[0]
1137 return self._html_search_regex(
1138 [self._meta_regex(n) for n in name],
1139 html, display_name, fatal=fatal, group='content
', **kwargs)
1141 def _dc_search_uploader(self, html):
1142 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1144 def _rta_search(self, html):
1145 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1146 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1147 r' content
="RTA-5042-1996-1400-1577-RTA"',
1152 def _media_rating_search(self, html):
1153 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1154 rating = self._html_search_meta('rating
', html)
1166 return RATING_TABLE.get(rating.lower())
1168 def _family_friendly_search(self, html):
1169 # See http://schema.org/VideoObject
1170 family_friendly = self._html_search_meta(
1171 'isFamilyFriendly
', html, default=None)
1173 if not family_friendly:
1182 return RATING_TABLE.get(family_friendly.lower())
1184 def _twitter_search_player(self, html):
1185 return self._html_search_meta('twitter
:player
', html,
1186 'twitter card player
')
1188 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1189 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1190 default = kwargs.get('default
', NO_DEFAULT)
1191 # JSON-LD may be malformed and thus `fatal` should be respected.
1192 # At the same time `default` may be passed that assumes `fatal=False`
1193 # for _search_regex. Let's simulate the same behavior here
as well
.
1194 fatal
= kwargs
.get('fatal', True) if default
== NO_DEFAULT
else False
1196 for mobj
in json_ld_list
:
1197 json_ld_item
= self
._parse
_json
(
1198 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1199 if not json_ld_item
:
1201 if isinstance(json_ld_item
, dict):
1202 json_ld
.append(json_ld_item
)
1203 elif isinstance(json_ld_item
, (list, tuple)):
1204 json_ld
.extend(json_ld_item
)
1206 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1209 if default
is not NO_DEFAULT
:
1212 raise RegexNotFoundError('Unable to extract JSON-LD')
1214 self
._downloader
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1217 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1218 if isinstance(json_ld
, compat_str
):
1219 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1223 if not isinstance(json_ld
, (list, tuple, dict)):
1225 if isinstance(json_ld
, dict):
1228 INTERACTION_TYPE_MAP
= {
1229 'CommentAction': 'comment',
1230 'AgreeAction': 'like',
1231 'DisagreeAction': 'dislike',
1232 'LikeAction': 'like',
1233 'DislikeAction': 'dislike',
1234 'ListenAction': 'view',
1235 'WatchAction': 'view',
1236 'ViewAction': 'view',
1239 def extract_interaction_statistic(e
):
1240 interaction_statistic
= e
.get('interactionStatistic')
1241 if not isinstance(interaction_statistic
, list):
1243 for is_e
in interaction_statistic
:
1244 if not isinstance(is_e
, dict):
1246 if is_e
.get('@type') != 'InteractionCounter':
1248 interaction_type
= is_e
.get('interactionType')
1249 if not isinstance(interaction_type
, compat_str
):
1251 interaction_count
= int_or_none(is_e
.get('userInteractionCount'))
1252 if interaction_count
is None:
1254 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1257 count_key
= '%s_count' % count_kind
1258 if info
.get(count_key
) is not None:
1260 info
[count_key
] = interaction_count
1262 def extract_video_object(e
):
1263 assert e
['@type'] == 'VideoObject'
1265 'url': url_or_none(e
.get('contentUrl')),
1266 'title': unescapeHTML(e
.get('name')),
1267 'description': unescapeHTML(e
.get('description')),
1268 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1269 'duration': parse_duration(e
.get('duration')),
1270 'timestamp': unified_timestamp(e
.get('uploadDate')),
1271 'filesize': float_or_none(e
.get('contentSize')),
1272 'tbr': int_or_none(e
.get('bitrate')),
1273 'width': int_or_none(e
.get('width')),
1274 'height': int_or_none(e
.get('height')),
1275 'view_count': int_or_none(e
.get('interactionCount')),
1277 extract_interaction_statistic(e
)
1281 item_type
= e
.get('@type')
1282 if expected_type
is not None and expected_type
!= item_type
:
1284 if item_type
in ('TVEpisode', 'Episode'):
1285 episode_name
= unescapeHTML(e
.get('name'))
1287 'episode': episode_name
,
1288 'episode_number': int_or_none(e
.get('episodeNumber')),
1289 'description': unescapeHTML(e
.get('description')),
1291 if not info
.get('title') and episode_name
:
1292 info
['title'] = episode_name
1293 part_of_season
= e
.get('partOfSeason')
1294 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1296 'season': unescapeHTML(part_of_season
.get('name')),
1297 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1299 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1300 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1301 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1302 elif item_type
== 'Movie':
1304 'title': unescapeHTML(e
.get('name')),
1305 'description': unescapeHTML(e
.get('description')),
1306 'duration': parse_duration(e
.get('duration')),
1307 'timestamp': unified_timestamp(e
.get('dateCreated')),
1309 elif item_type
in ('Article', 'NewsArticle'):
1311 'timestamp': parse_iso8601(e
.get('datePublished')),
1312 'title': unescapeHTML(e
.get('headline')),
1313 'description': unescapeHTML(e
.get('articleBody')),
1315 elif item_type
== 'VideoObject':
1316 extract_video_object(e
)
1317 if expected_type
is None:
1321 video
= e
.get('video')
1322 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1323 extract_video_object(video
)
1324 if expected_type
is None:
1328 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1331 def _hidden_inputs(html
):
1332 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1334 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1335 attrs
= extract_attributes(input)
1338 if attrs
.get('type') not in ('hidden', 'submit'):
1340 name
= attrs
.get('name') or attrs
.get('id')
1341 value
= attrs
.get('value')
1342 if name
and value
is not None:
1343 hidden_inputs
[name
] = value
1344 return hidden_inputs
1346 def _form_hidden_inputs(self
, form_id
, html
):
1347 form
= self
._search
_regex
(
1348 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1349 html, '%s form
' % form_id, group='form
')
1350 return self._hidden_inputs(form)
1352 def _sort_formats(self, formats, field_preference=None):
1354 raise ExtractorError('No video formats found
')
1357 # Automatically determine tbr when missing based on abr and vbr (improves
1358 # formats sorting in some cases)
1359 if 'tbr
' not in f and f.get('abr
') is not None and f.get('vbr
') is not None:
1360 f['tbr
'] = f['abr
'] + f['vbr
']
1362 def _formats_key(f):
1363 # TODO remove the following workaround
1364 from ..utils import determine_ext
1365 if not f.get('ext
') and 'url
' in f:
1366 f['ext
'] = determine_ext(f['url
'])
1368 if isinstance(field_preference, (list, tuple)):
1371 if f.get(field) is not None
1372 else ('' if field == 'format_id
' else -1)
1373 for field in field_preference)
1375 preference = f.get('preference
')
1376 if preference is None:
1378 if f.get('ext
') in ['f4f
', 'f4m
']: # Not yet supported
1381 protocol = f.get('protocol
') or determine_protocol(f)
1382 proto_preference = 0 if protocol in ['http
', 'https
'] else (-0.5 if protocol == 'rtsp
' else -0.1)
1384 if f.get('vcodec
') == 'none
': # audio only
1386 if self._downloader.params.get('prefer_free_formats
'):
1387 ORDER = ['aac
', 'mp3
', 'm4a
', 'webm
', 'ogg
', 'opus
']
1389 ORDER = ['webm
', 'opus
', 'ogg
', 'mp3
', 'aac
', 'm4a
']
1392 audio_ext_preference = ORDER.index(f['ext
'])
1394 audio_ext_preference = -1
1396 if f.get('acodec
') == 'none
': # video only
1398 if self._downloader.params.get('prefer_free_formats
'):
1399 ORDER = ['flv
', 'mp4
', 'webm
']
1401 ORDER = ['webm
', 'flv
', 'mp4
']
1403 ext_preference = ORDER.index(f['ext
'])
1406 audio_ext_preference = 0
1410 f.get('language_preference
') if f.get('language_preference
') is not None else -1,
1411 f.get('quality
') if f.get('quality
') is not None else -1,
1412 f.get('tbr
') if f.get('tbr
') is not None else -1,
1413 f.get('filesize
') if f.get('filesize
') is not None else -1,
1414 f.get('vbr
') if f.get('vbr
') is not None else -1,
1415 f.get('height
') if f.get('height
') is not None else -1,
1416 f.get('width
') if f.get('width
') is not None else -1,
1419 f.get('abr
') if f.get('abr
') is not None else -1,
1420 audio_ext_preference,
1421 f.get('fps
') if f.get('fps
') is not None else -1,
1422 f.get('filesize_approx
') if f.get('filesize_approx
') is not None else -1,
1423 f.get('source_preference
') if f.get('source_preference
') is not None else -1,
1424 f.get('format_id
') if f.get('format_id
') is not None else '',
1426 formats.sort(key=_formats_key)
1428 def _check_formats(self, formats, video_id):
1430 formats[:] = filter(
1431 lambda f: self._is_valid_url(
1433 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1437 def _remove_duplicate_formats(formats):
1441 if f['url
'] not in format_urls:
1442 format_urls.add(f['url
'])
1443 unique_formats.append(f)
1444 formats[:] = unique_formats
1446 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1447 url = self._proto_relative_url(url, scheme='http
:')
1448 # For now assume non HTTP(S) URLs always valid
1449 if not (url.startswith('http
://') or url.startswith('https
://')):
1452 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1454 except ExtractorError:
1456 '%s: %s URL
is invalid
, skipping
' % (video_id, item))
1459 def http_scheme(self):
1460 """ Either "http:" or "https:", depending on the user's preferences
"""
1463 if self._downloader.params.get('prefer_insecure', False)
1466 def _proto_relative_url(self, url, scheme=None):
1469 if url.startswith('//'):
1471 scheme = self.http_scheme()
1476 def _sleep(self, timeout, video_id, msg_template=None):
1477 if msg_template is None:
1478 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1479 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1483 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1484 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1485 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1486 manifest = self._download_xml(
1487 manifest_url, video_id, 'Downloading f4m manifest',
1488 'Unable to download f4m manifest',
1489 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1490 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1491 transform_source=transform_source,
1492 fatal=fatal, data=data, headers=headers, query=query)
1494 if manifest is False:
1497 return self._parse_f4m_formats(
1498 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1499 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1501 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1502 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1503 fatal=True, m3u8_id=None):
1504 if not isinstance(manifest, compat_etree_Element) and not fatal:
1507 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1508 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1509 if akamai_pv is not None and ';' in akamai_pv.text:
1510 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1511 if playerVerificationChallenge.strip() != '':
1515 manifest_version = '1.0'
1516 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1518 manifest_version = '2.0'
1519 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1520 # Remove unsupported DRM protected media from final formats
1521 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1522 media_nodes = remove_encrypted_media(media_nodes)
1526 manifest_base_url = get_base_url(manifest)
1528 bootstrap_info = xpath_element(
1529 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1530 'bootstrap info', default=None)
1533 mime_type = xpath_text(
1534 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1535 'base URL', default=None)
1536 if mime_type and mime_type.startswith('audio/'):
1539 for i, media_el in enumerate(media_nodes):
1540 tbr = int_or_none(media_el.attrib.get('bitrate'))
1541 width = int_or_none(media_el.attrib.get('width'))
1542 height = int_or_none(media_el.attrib.get('height'))
1543 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1544 # If <bootstrapInfo> is present, the specified f4m is a
1545 # stream-level manifest, and only set-level manifests may refer to
1546 # external resources. See section 11.4 and section 4 of F4M spec
1547 if bootstrap_info is None:
1549 # @href is introduced in 2.0, see section 11.6 of F4M spec
1550 if manifest_version == '2.0':
1551 media_url = media_el.attrib.get('href')
1552 if media_url is None:
1553 media_url = media_el.attrib.get('url')
1557 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1558 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1559 # If media_url is itself a f4m manifest do the recursive extraction
1560 # since bitrates in parent manifest (this one) and media_url manifest
1561 # may differ leading to inability to resolve the format by requested
1562 # bitrate in f4m downloader
1563 ext = determine_ext(manifest_url)
1565 f4m_formats = self._extract_f4m_formats(
1566 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1567 transform_source=transform_source, fatal=fatal)
1568 # Sometimes stream-level manifest contains single media entry that
1569 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1570 # At the same time parent's media entry in set-level manifest may
1571 # contain it. We will copy it from parent in such cases.
1572 if len(f4m_formats) == 1:
1575 'tbr': f.get('tbr') or tbr,
1576 'width': f.get('width') or width,
1577 'height': f.get('height') or height,
1578 'format_id': f.get('format_id') if not tbr else format_id,
1581 formats.extend(f4m_formats)
1584 formats.extend(self._extract_m3u8_formats(
1585 manifest_url, video_id, 'mp4', preference=preference,
1586 m3u8_id=m3u8_id, fatal=fatal))
1589 'format_id': format_id,
1590 'url': manifest_url,
1591 'manifest_url': manifest_url,
1592 'ext': 'flv' if bootstrap_info is not None else None,
1598 'preference': preference,
1602 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1604 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1608 'preference': preference - 100 if preference else -100,
1609 'resolution': 'multiple',
1610 'format_note': 'Quality selection URL',
1613 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1614 entry_protocol='m3u8', preference=None,
1615 m3u8_id=None, note=None, errnote=None,
1616 fatal=True, live=False, data=None, headers={},
1618 res = self._download_webpage_handle(
1620 note=note or 'Downloading m3u8 information',
1621 errnote=errnote or 'Failed to download m3u8 information',
1622 fatal=fatal, data=data, headers=headers, query=query)
1627 m3u8_doc, urlh = res
1628 m3u8_url = urlh.geturl()
1630 return self._parse_m3u8_formats(
1631 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1632 preference=preference, m3u8_id=m3u8_id, live=live)
1634 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1635 entry_protocol='m3u8', preference=None,
1636 m3u8_id=None, live=False):
1637 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1640 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1645 format_url = lambda u: (
1647 if re.match(r'^https?://', u)
1648 else compat_urlparse.urljoin(m3u8_url, u))
1651 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1652 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1653 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1655 # We should try extracting formats only from master playlists [1, 4.3.4],
1656 # i.e. playlists that describe available qualities. On the other hand
1657 # media playlists [1, 4.3.3] should be returned as is since they contain
1658 # just the media without qualities renditions.
1659 # Fortunately, master playlist can be easily distinguished from media
1660 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1661 # master playlist tags MUST NOT appear in a media playist and vice versa.
1662 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1663 # media playlist and MUST NOT appear in master playlist thus we can
1664 # clearly detect media playlist with this criterion.
1666 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1669 'format_id': m3u8_id,
1671 'protocol': entry_protocol,
1672 'preference': preference,
1676 last_stream_inf = {}
1678 def extract_media(x_media_line):
1679 media = parse_m3u8_attributes(x_media_line)
1680 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1681 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1682 if not (media_type and group_id and name):
1684 groups.setdefault(group_id, []).append(media)
1685 if media_type not in ('VIDEO', 'AUDIO'):
1687 media_url = media.get('URI')
1690 for v in (m3u8_id, group_id, name):
1694 'format_id': '-'.join(format_id),
1695 'url': format_url(media_url),
1696 'manifest_url': m3u8_url,
1697 'language': media.get('LANGUAGE'),
1699 'protocol': entry_protocol,
1700 'preference': preference,
1702 if media_type == 'AUDIO':
1703 f['vcodec'] = 'none'
1706 def build_stream_name():
1707 # Despite specification does not mention NAME attribute for
1708 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1709 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1710 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1711 stream_name = last_stream_inf.get('NAME')
1714 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1715 # from corresponding rendition group
1716 stream_group_id = last_stream_inf.get('VIDEO')
1717 if not stream_group_id:
1719 stream_group = groups.get(stream_group_id)
1720 if not stream_group:
1721 return stream_group_id
1722 rendition = stream_group[0]
1723 return rendition.get('NAME') or stream_group_id
1725 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
1726 # chance to detect video only formats when EXT-X-STREAM-INF tags
1727 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
1728 for line in m3u8_doc.splitlines():
1729 if line.startswith('#EXT-X-MEDIA:'):
1732 for line in m3u8_doc.splitlines():
1733 if line.startswith('#EXT-X-STREAM-INF:'):
1734 last_stream_inf = parse_m3u8_attributes(line)
1735 elif line.startswith('#') or not line.strip():
1738 tbr = float_or_none(
1739 last_stream_inf.get('AVERAGE-BANDWIDTH')
1740 or last_stream_inf.get('BANDWIDTH'), scale=1000)
1743 format_id.append(m3u8_id)
1744 stream_name = build_stream_name()
1745 # Bandwidth of live streams may differ over time thus making
1746 # format_id unpredictable. So it's better to keep provided
1749 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1750 manifest_url = format_url(line.strip())
1752 'format_id': '-'.join(format_id),
1753 'url': manifest_url,
1754 'manifest_url': m3u8_url,
1757 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1758 'protocol': entry_protocol,
1759 'preference': preference,
1761 resolution = last_stream_inf.get('RESOLUTION')
1763 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1765 f['width'] = int(mobj.group('width'))
1766 f['height'] = int(mobj.group('height'))
1767 # Unified Streaming Platform
1769 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1771 abr, vbr = mobj.groups()
1772 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1777 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1779 audio_group_id = last_stream_inf.get('AUDIO')
1780 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1781 # references a rendition group MUST have a CODECS attribute.
1782 # However, this is not always respected, for example, [2]
1783 # contains EXT-X-STREAM-INF tag which references AUDIO
1784 # rendition group but does not have CODECS and despite
1785 # referencing an audio group it represents a complete
1786 # (with audio and video) format. So, for such cases we will
1787 # ignore references to rendition groups and treat them
1788 # as complete formats.
1789 if audio_group_id and codecs and f.get('vcodec') != 'none':
1790 audio_group = groups.get(audio_group_id)
1791 if audio_group and audio_group[0].get('URI'):
1792 # TODO: update acodec for audio only formats with
1794 f['acodec'] = 'none'
1798 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
1801 del http_f['manifest_url']
1803 'format_id': f['format_id'].replace('hls-', 'http-'),
1805 'url': progressive_uri,
1807 formats.append(http_f)
1809 last_stream_inf = {}
1813 def _xpath_ns(path, namespace=None):
1817 for c in path.split('/'):
1818 if not c or c == '.':
1821 out.append('{%s}%s' % (namespace, c))
1822 return '/'.join(out)
1824 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1825 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1831 namespace = self._parse_smil_namespace(smil)
1833 return self._parse_smil_formats(
1834 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1836 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1837 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1840 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1842 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1843 return self._download_xml(
1844 smil_url, video_id, 'Downloading SMIL file',
1845 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1847 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1848 namespace = self._parse_smil_namespace(smil)
1850 formats = self._parse_smil_formats(
1851 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1852 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1854 video_id = os.path.splitext(url_basename(smil_url))[0]
1858 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1859 name = meta.attrib.get('name')
1860 content = meta.attrib.get('content')
1861 if not name or not content:
1863 if not title and name == 'title':
1865 elif not description and name in ('description', 'abstract'):
1866 description = content
1867 elif not upload_date and name == 'date':
1868 upload_date = unified_strdate(content)
1871 'id': image.get('type'),
1872 'url': image.get('src'),
1873 'width': int_or_none(image.get('width')),
1874 'height': int_or_none(image.get('height')),
1875 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1879 'title': title or video_id,
1880 'description': description,
1881 'upload_date': upload_date,
1882 'thumbnails': thumbnails,
1884 'subtitles': subtitles,
1887 def _parse_smil_namespace(self, smil):
1888 return self._search_regex(
1889 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1891 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1893 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1894 b = meta.get('base') or meta.get('httpBase')
1905 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1906 for medium in media:
1907 src = medium.get('src')
1908 if not src or src in srcs:
1912 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1913 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1914 width = int_or_none(medium.get('width'))
1915 height = int_or_none(medium.get('height'))
1916 proto = medium.get('proto')
1917 ext = medium.get('ext')
1918 src_ext = determine_ext(src)
1919 streamer = medium.get('streamer') or base
1921 if proto == 'rtmp' or streamer.startswith('rtmp'):
1927 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1929 'filesize': filesize,
1933 if transform_rtmp_url:
1934 streamer, src = transform_rtmp_url(streamer, src)
1935 formats[-1].update({
1941 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1942 src_url = src_url.strip()
1944 if proto == 'm3u8' or src_ext == 'm3u8':
1945 m3u8_formats = self._extract_m3u8_formats(
1946 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1947 if len(m3u8_formats) == 1:
1949 m3u8_formats[0].update({
1950 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1955 formats.extend(m3u8_formats)
1956 elif src_ext == 'f4m':
1961 'plugin': 'flowplayer-3.2.0.1',
1963 f4m_url += '&' if '?' in f4m_url else '?'
1964 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1965 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1966 elif src_ext == 'mpd':
1967 formats.extend(self._extract_mpd_formats(
1968 src_url, video_id, mpd_id='dash', fatal=False))
1969 elif re.search(r'\.ism/[Mm]anifest', src_url):
1970 formats.extend(self._extract_ism_formats(
1971 src_url, video_id, ism_id='mss', fatal=False))
1972 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1976 'ext': ext or src_ext or 'flv',
1977 'format_id': 'http-%d' % (bitrate or http_count),
1979 'filesize': filesize,
1986 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1989 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1990 src = textstream.get('src')
1991 if not src or src in urls:
1994 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1995 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1996 subtitles.setdefault(lang, []).append({
2002 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2003 xspf = self._download_xml(
2004 xspf_url, playlist_id, 'Downloading xpsf playlist',
2005 'Unable to download xspf manifest', fatal=fatal)
2008 return self._parse_xspf(
2009 xspf, playlist_id, xspf_url=xspf_url,
2010 xspf_base_url=base_url(xspf_url))
2012 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2014 'xspf': 'http://xspf.org/ns/0/',
2015 's1': 'http://static.streamone.nl/player/ns/0',
2019 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2021 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2022 description = xpath_text(
2023 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2024 thumbnail = xpath_text(
2025 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2026 duration = float_or_none(
2027 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2030 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2031 format_url = urljoin(xspf_base_url, location.text)
2036 'manifest_url': xspf_url,
2037 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2038 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2039 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2041 self._sort_formats(formats)
2046 'description': description,
2047 'thumbnail': thumbnail,
2048 'duration': duration,
2053 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
2054 res = self._download_xml_handle(
2056 note=note or 'Downloading MPD manifest',
2057 errnote=errnote or 'Failed to download MPD manifest',
2058 fatal=fatal, data=data, headers=headers, query=query)
2064 mpd_base_url = base_url(urlh.geturl())
2066 return self._parse_mpd_formats(
2067 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
2068 formats_dict=formats_dict, mpd_url=mpd_url)
2070 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
2072 Parse formats
from MPD manifest
.
2074 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2075 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2076 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2078 if mpd_doc.get('type') == 'dynamic':
2081 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2084 return self._xpath_ns(path, namespace)
2086 def is_drm_protected(element):
2087 return element.find(_add_ns('ContentProtection')) is not None
2089 def extract_multisegment_info(element, ms_parent_info):
2090 ms_info = ms_parent_info.copy()
2092 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2093 # common attributes and elements. We will only extract relevant
2095 def extract_common(source):
2096 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2097 if segment_timeline is not None:
2098 s_e = segment_timeline.findall(_add_ns('S'))
2100 ms_info['total_number'] = 0
2103 r = int(s.get('r', 0))
2104 ms_info['total_number'] += 1 + r
2105 ms_info['s'].append({
2106 't': int(s.get('t', 0)),
2107 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2108 'd': int(s.attrib['d']),
2111 start_number = source.get('startNumber')
2113 ms_info['start_number'] = int(start_number)
2114 timescale = source.get('timescale')
2116 ms_info['timescale'] = int(timescale)
2117 segment_duration = source.get('duration')
2118 if segment_duration:
2119 ms_info['segment_duration'] = float(segment_duration)
2121 def extract_Initialization(source):
2122 initialization = source.find(_add_ns('Initialization'))
2123 if initialization is not None:
2124 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2126 segment_list = element.find(_add_ns('SegmentList'))
2127 if segment_list is not None:
2128 extract_common(segment_list)
2129 extract_Initialization(segment_list)
2130 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2132 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2134 segment_template = element.find(_add_ns('SegmentTemplate'))
2135 if segment_template is not None:
2136 extract_common(segment_template)
2137 media = segment_template.get('media')
2139 ms_info['media'] = media
2140 initialization = segment_template.get('initialization')
2142 ms_info['initialization'] = initialization
2144 extract_Initialization(segment_template)
2147 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2149 for period in mpd_doc.findall(_add_ns('Period')):
2150 period_duration = parse_duration(period.get('duration')) or mpd_duration
2151 period_ms_info = extract_multisegment_info(period, {
2155 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2156 if is_drm_protected(adaptation_set):
2158 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2159 for representation in adaptation_set.findall(_add_ns('Representation')):
2160 if is_drm_protected(representation):
2162 representation_attrib = adaptation_set.attrib.copy()
2163 representation_attrib.update(representation.attrib)
2164 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2165 mime_type = representation_attrib['mimeType']
2166 content_type = mime_type.split('/')[0]
2167 if content_type == 'text':
2168 # TODO implement WebVTT downloading
2170 elif content_type in ('video', 'audio'):
2172 for element in (representation, adaptation_set, period, mpd_doc):
2173 base_url_e = element.find(_add_ns('BaseURL'))
2174 if base_url_e is not None:
2175 base_url = base_url_e.text + base_url
2176 if re.match(r'^https?://', base_url):
2178 if mpd_base_url and not re.match(r'^https?://', base_url):
2179 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2181 base_url = mpd_base_url + base_url
2182 representation_id = representation_attrib.get('id')
2183 lang = representation_attrib.get('lang')
2184 url_el = representation.find(_add_ns('BaseURL'))
2185 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2186 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2188 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2189 'manifest_url': mpd_url,
2190 'ext': mimetype2ext(mime_type),
2191 'width': int_or_none(representation_attrib.get('width')),
2192 'height': int_or_none(representation_attrib.get('height')),
2193 'tbr': float_or_none(bandwidth, 1000),
2194 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2195 'fps': int_or_none(representation_attrib.get('frameRate')),
2196 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2197 'format_note': 'DASH %s' % content_type,
2198 'filesize': filesize,
2199 'container': mimetype2ext(mime_type) + '_dash',
2201 f.update(parse_codecs(representation_attrib.get('codecs')))
2202 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2204 def prepare_template(template_name, identifiers):
2205 tmpl = representation_ms_info[template_name]
2206 # First of, % characters outside $...$ templates
2207 # must be escaped by doubling for proper processing
2208 # by % operator string formatting used further (see
2209 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2215 in_template = not in_template
2216 elif c == '%' and not in_template:
2218 # Next, $...$ templates are translated to their
2219 # %(...) counterparts to be used with % operator
2220 t = t.replace('$RepresentationID$', representation_id)
2221 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2222 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2223 t.replace('$$', '$')
2226 # @initialization is a regular template like @media one
2227 # so it should be handled just the same way (see
2228 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2229 if 'initialization' in representation_ms_info:
2230 initialization_template = prepare_template(
2232 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2233 # $Time$ shall not be included for @initialization thus
2234 # only $Bandwidth$ remains
2236 representation_ms_info['initialization_url'] = initialization_template % {
2237 'Bandwidth': bandwidth,
2240 def location_key(location):
2241 return 'url' if re.match(r'^https?://', location) else 'path'
2243 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2245 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2246 media_location_key = location_key(media_template)
2248 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2249 # can't be used at the same time
2250 if '%(Number' in media_template and 's' not in representation_ms_info:
2251 segment_duration = None
2252 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2253 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2254 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2255 representation_ms_info['fragments'] = [{
2256 media_location_key: media_template % {
2257 'Number': segment_number,
2258 'Bandwidth': bandwidth,
2260 'duration': segment_duration,
2261 } for segment_number in range(
2262 representation_ms_info['start_number'],
2263 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2265 # $Number*$ or $Time$ in media template with S list available
2266 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2267 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2268 representation_ms_info['fragments'] = []
2271 segment_number = representation_ms_info['start_number']
2273 def add_segment_url():
2274 segment_url = media_template % {
2275 'Time': segment_time,
2276 'Bandwidth': bandwidth,
2277 'Number': segment_number,
2279 representation_ms_info['fragments'].append({
2280 media_location_key: segment_url,
2281 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2284 for num, s in enumerate(representation_ms_info['s']):
2285 segment_time = s.get('t') or segment_time
2289 for r in range(s.get('r', 0)):
2290 segment_time += segment_d
2293 segment_time += segment_d
2294 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2296 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2297 # or any YouTube dashsegments video
2300 timescale = representation_ms_info['timescale']
2301 for s in representation_ms_info['s']:
2302 duration = float_or_none(s['d'], timescale)
2303 for r in range(s.get('r', 0) + 1):
2304 segment_uri = representation_ms_info['segment_urls'][segment_index]
2306 location_key(segment_uri): segment_uri,
2307 'duration': duration,
2310 representation_ms_info['fragments'] = fragments
2311 elif 'segment_urls' in representation_ms_info:
2312 # Segment URLs with no SegmentTimeline
2313 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2314 # https://github.com/ytdl-org/youtube-dl/pull/14844
2316 segment_duration = float_or_none(
2317 representation_ms_info['segment_duration'],
2318 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2319 for segment_url in representation_ms_info['segment_urls']:
2321 location_key(segment_url): segment_url,
2323 if segment_duration:
2324 fragment['duration'] = segment_duration
2325 fragments.append(fragment)
2326 representation_ms_info['fragments'] = fragments
2327 # If there is a fragments key available then we correctly recognized fragmented media.
2328 # Otherwise we will assume unfragmented media with direct access. Technically, such
2329 # assumption is not necessarily correct since we may simply have no support for
2330 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2331 if 'fragments' in representation_ms_info:
2333 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2334 'url': mpd_url or base_url,
2335 'fragment_base_url': base_url,
2337 'protocol': 'http_dash_segments',
2339 if 'initialization_url' in representation_ms_info:
2340 initialization_url = representation_ms_info['initialization_url']
2341 if not f.get('url'):
2342 f['url'] = initialization_url
2343 f['fragments'].append({location_key(initialization_url): initialization_url})
2344 f['fragments'].extend(representation_ms_info['fragments'])
2346 # Assuming direct URL to unfragmented media.
2349 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2350 # is not necessarily unique within a Period thus formats with
2351 # the same `format_id` are quite possible. There are numerous examples
2352 # of such manifests (see https://github.com/ytdl-org/youtube-dl/issues/15111,
2353 # https://github.com/ytdl-org/youtube-dl/issues/13919)
2354 full_info = formats_dict.get(representation_id, {}).copy()
2356 formats.append(full_info)
2358 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2361 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2362 res = self._download_xml_handle(
2364 note=note or 'Downloading ISM manifest',
2365 errnote=errnote or 'Failed to download ISM manifest',
2366 fatal=fatal, data=data, headers=headers, query=query)
2373 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2375 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2377 Parse formats
from ISM manifest
.
2379 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2380 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2382 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2385 duration = int(ism_doc.attrib['Duration'])
2386 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2389 for stream in ism_doc.findall('StreamIndex'):
2390 stream_type = stream.get('Type')
2391 if stream_type not in ('video', 'audio'):
2393 url_pattern = stream.attrib['Url']
2394 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2395 stream_name = stream.get('Name')
2396 for track in stream.findall('QualityLevel'):
2397 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2398 # TODO: add support for WVC1 and WMAP
2399 if fourcc not in ('H264', 'AVC1', 'AACL'):
2400 self.report_warning('%s is not a supported codec' % fourcc)
2402 tbr = int(track.attrib['Bitrate']) // 1000
2403 # [1] does not mention Width and Height attributes. However,
2404 # they're often present while MaxWidth and MaxHeight are
2405 # missing, so should be used as fallbacks
2406 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2407 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2408 sampling_rate = int_or_none(track.get('SamplingRate'))
2410 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2411 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2417 stream_fragments = stream.findall('c')
2418 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2419 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2420 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2421 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2422 if not fragment_ctx['duration']:
2424 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2426 next_fragment_time = duration
2427 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2428 for _ in range(fragment_repeat):
2430 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2431 'duration': fragment_ctx['duration'] / stream_timescale,
2433 fragment_ctx['time'] += fragment_ctx['duration']
2437 format_id.append(ism_id)
2439 format_id.append(stream_name)
2440 format_id.append(compat_str(tbr))
2443 'format_id': '-'.join(format_id),
2445 'manifest_url': ism_url,
2446 'ext': 'ismv' if stream_type == 'video' else 'isma',
2450 'asr': sampling_rate,
2451 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2452 'acodec': 'none' if stream_type == 'video' else fourcc,
2454 'fragments': fragments,
2455 '_download_params': {
2456 'duration': duration,
2457 'timescale': stream_timescale,
2458 'width': width or 0,
2459 'height': height or 0,
2461 'codec_private_data': track.get('CodecPrivateData'),
2462 'sampling_rate': sampling_rate,
2463 'channels': int_or_none(track.get('Channels', 2)),
2464 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2465 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2470 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2471 def absolute_url(item_url):
2472 return urljoin(base_url, item_url)
2474 def parse_content_type(content_type):
2475 if not content_type:
2477 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2479 mimetype, codecs = ctr.groups()
2480 f = parse_codecs(codecs)
2481 f['ext'] = mimetype2ext(mimetype)
2485 def _media_formats(src, cur_media_type, type_info={}):
2486 full_url = absolute_url(src)
2487 ext = type_info.get('ext') or determine_ext(full_url)
2489 is_plain_url = False
2490 formats = self._extract_m3u8_formats(
2491 full_url, video_id, ext='mp4',
2492 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2493 preference=preference, fatal=False)
2495 is_plain_url = False
2496 formats = self._extract_mpd_formats(
2497 full_url, video_id, mpd_id=mpd_id, fatal=False)
2502 'vcodec': 'none' if cur_media_type == 'audio' else None,
2504 return is_plain_url, formats
2507 # amp-video and amp-audio are very similar to their HTML5 counterparts
2508 # so we wll include them right here (see
2509 # https://www.ampproject.org/docs/reference/components/amp-video)
2510 media_tags = [(media_tag, media_type, '')
2511 for media_tag, media_type
2512 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2513 media_tags.extend(re.findall(
2514 # We only allow video|audio followed by a whitespace or '>'.
2515 # Allowing more characters may end up in significant slow down (see
2516 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2517 # http://www.porntrex.com/maps/videositemap.xml).
2518 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2519 for media_tag, media_type, media_content in media_tags:
2524 media_attributes = extract_attributes(media_tag)
2525 src = strip_or_none(media_attributes.get('src'))
2527 _, formats = _media_formats(src, media_type)
2528 media_info['formats'].extend(formats)
2529 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2531 for source_tag in re.findall(r'<source[^>]+>', media_content):
2532 s_attr = extract_attributes(source_tag)
2533 # data-video-src and data-src are non standard but seen
2534 # several times in the wild
2535 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2538 f = parse_content_type(s_attr.get('type'))
2539 is_plain_url, formats = _media_formats(src, media_type, f)
2541 # width, height, res, label and title attributes are
2542 # all not standard but seen several times in the wild
2545 for lbl in ('label', 'title')
2546 if str_or_none(s_attr.get(lbl))
2548 width = int_or_none(s_attr.get('width'))
2549 height = (int_or_none(s_attr.get('height'))
2550 or int_or_none(s_attr.get('res')))
2551 if not width or not height:
2553 resolution = parse_resolution(lbl)
2556 width = width or resolution.get('width')
2557 height = height or resolution.get('height')
2559 tbr = parse_bitrate(lbl)
2568 'format_id': s_attr.get('label') or s_attr.get('title'),
2570 f.update(formats[0])
2571 media_info['formats'].append(f)
2573 media_info['formats'].extend(formats)
2574 for track_tag in re.findall(r'<track[^>]+>', media_content):
2575 track_attributes = extract_attributes(track_tag)
2576 kind = track_attributes.get('kind')
2577 if not kind or kind in ('subtitles', 'captions'):
2578 src = strip_or_none(track_attributes.get('src'))
2581 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2582 media_info['subtitles'].setdefault(lang, []).append({
2583 'url': absolute_url(src),
2585 for f in media_info['formats']:
2586 f.setdefault('http_headers', {})['Referer'] = base_url
2587 if media_info['formats'] or media_info['subtitles']:
2588 entries.append(media_info)
2591 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2593 hdcore_sign = 'hdcore=3.7.0'
2594 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2595 hds_host = hosts.get('hds')
2597 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2598 if 'hdcore=' not in f4m_url:
2599 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2600 f4m_formats = self._extract_f4m_formats(
2601 f4m_url, video_id, f4m_id='hds', fatal=False)
2602 for entry in f4m_formats:
2603 entry.update({'extra_param_to_segment_url': hdcore_sign})
2604 formats.extend(f4m_formats)
2605 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2606 hls_host = hosts.get('hls')
2608 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2609 formats.extend(self._extract_m3u8_formats(
2610 m3u8_url, video_id, 'mp4', 'm3u8_native',
2611 m3u8_id='hls', fatal=False))
2614 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2615 query = compat_urlparse.urlparse(url).query
2616 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2618 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2619 url_base = mobj.group('url')
2620 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2623 def manifest_url(manifest):
2624 m_url = '%s/%s' % (http_base_url, manifest)
2626 m_url += '?%s' % query
2629 if 'm3u8' not in skip_protocols:
2630 formats.extend(self._extract_m3u8_formats(
2631 manifest_url('playlist.m3u8'), video_id, 'mp4',
2632 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2633 if 'f4m' not in skip_protocols:
2634 formats.extend(self._extract_f4m_formats(
2635 manifest_url('manifest.f4m'),
2636 video_id, f4m_id='hds', fatal=False))
2637 if 'dash' not in skip_protocols:
2638 formats.extend(self._extract_mpd_formats(
2639 manifest_url('manifest.mpd'),
2640 video_id, mpd_id='dash', fatal=False))
2641 if re.search(r'(?:/smil:|\.smil)', url_base):
2642 if 'smil' not in skip_protocols:
2643 rtmp_formats = self._extract_smil_formats(
2644 manifest_url('jwplayer.smil'),
2645 video_id, fatal=False)
2646 for rtmp_format in rtmp_formats:
2647 rtsp_format = rtmp_format.copy()
2648 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2649 del rtsp_format['play_path']
2650 del rtsp_format['ext']
2651 rtsp_format.update({
2652 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2653 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2656 formats.extend([rtmp_format, rtsp_format])
2658 for protocol in ('rtmp', 'rtsp'):
2659 if protocol not in skip_protocols:
2661 'url': '%s:%s' % (protocol, url_base),
2662 'format_id': protocol,
2663 'protocol': protocol,
2667 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2669 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2673 jwplayer_data = self._parse_json(mobj.group('options'),
2675 transform_source=transform_source)
2676 except ExtractorError:
2679 if isinstance(jwplayer_data, dict):
2680 return jwplayer_data
2682 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2683 jwplayer_data = self._find_jwplayer_data(
2684 webpage, video_id, transform_source=js_to_json)
2685 return self._parse_jwplayer_data(
2686 jwplayer_data, video_id, *args, **kwargs)
2688 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2689 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2690 # JWPlayer backward compatibility: flattened playlists
2691 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2692 if 'playlist' not in jwplayer_data:
2693 jwplayer_data = {'playlist': [jwplayer_data]}
2697 # JWPlayer backward compatibility: single playlist item
2698 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2699 if not isinstance(jwplayer_data['playlist'], list):
2700 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2702 for video_data in jwplayer_data['playlist']:
2703 # JWPlayer backward compatibility: flattened sources
2704 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2705 if 'sources' not in video_data:
2706 video_data['sources'] = [video_data]
2708 this_video_id = video_id or video_data['mediaid']
2710 formats = self._parse_jwplayer_formats(
2711 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2712 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2715 tracks = video_data.get('tracks')
2716 if tracks and isinstance(tracks, list):
2717 for track in tracks:
2718 if not isinstance(track, dict):
2720 track_kind = track.get('kind')
2721 if not track_kind or not isinstance(track_kind, compat_str):
2723 if track_kind.lower() not in ('captions', 'subtitles'):
2725 track_url = urljoin(base_url, track.get('file'))
2728 subtitles.setdefault(track.get('label') or 'en', []).append({
2729 'url': self._proto_relative_url(track_url)
2733 'id': this_video_id,
2734 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2735 'description': clean_html(video_data.get('description')),
2736 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
2737 'timestamp': int_or_none(video_data.get('pubdate')),
2738 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2739 'subtitles': subtitles,
2741 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2742 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2744 '_type': 'url_transparent',
2745 'url': formats[0]['url'],
2748 self._sort_formats(formats)
2749 entry['formats'] = formats
2750 entries.append(entry)
2751 if len(entries) == 1:
2754 return self.playlist_result(entries)
2756 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2757 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2760 for source in jwplayer_sources_data:
2761 if not isinstance(source, dict):
2763 source_url = urljoin(
2764 base_url, self._proto_relative_url(source.get('file')))
2765 if not source_url or source_url in urls:
2767 urls.append(source_url)
2768 source_type = source.get('type') or ''
2769 ext = mimetype2ext(source_type) or determine_ext(source_url)
2770 if source_type == 'hls' or ext == 'm3u8':
2771 formats.extend(self._extract_m3u8_formats(
2772 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2773 m3u8_id=m3u8_id, fatal=False))
2774 elif source_type == 'dash' or ext == 'mpd':
2775 formats.extend(self._extract_mpd_formats(
2776 source_url, video_id, mpd_id=mpd_id, fatal=False))
2778 formats.extend(self._extract_smil_formats(
2779 source_url, video_id, fatal=False))
2780 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2781 elif source_type.startswith('audio') or ext in (
2782 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2789 height = int_or_none(source.get('height'))
2791 # Often no height is provided but there is a label in
2792 # format like "1080p", "720p SD", or 1080.
2793 height = int_or_none(self._search_regex(
2794 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2795 'height', default=None))
2798 'width': int_or_none(source.get('width')),
2800 'tbr': int_or_none(source.get('bitrate')),
2803 if source_url.startswith('rtmp'):
2804 a_format['ext'] = 'flv'
2805 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2806 # of jwplayer.flash.swf
2807 rtmp_url_parts = re.split(
2808 r'((?:mp4|mp3|flv):)', source_url, 1)
2809 if len(rtmp_url_parts) == 3:
2810 rtmp_url, prefix, play_path = rtmp_url_parts
2813 'play_path': prefix + play_path,
2816 a_format.update(rtmp_params)
2817 formats.append(a_format)
2820 def _live_title(self, name):
2821 """ Generate the title
for a live video
"""
2822 now = datetime.datetime.now()
2823 now_str = now.strftime('%Y-%m-%d %H:%M')
2824 return name + ' ' + now_str
2826 def _int(self, v, name, fatal=False, **kwargs):
2827 res = int_or_none(v, **kwargs)
2828 if 'get_attr' in kwargs:
2829 print(getattr(v, kwargs['get_attr']))
2831 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2833 raise ExtractorError(msg)
2835 self._downloader.report_warning(msg)
2838 def _float(self, v, name, fatal=False, **kwargs):
2839 res = float_or_none(v, **kwargs)
2841 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2843 raise ExtractorError(msg)
2845 self._downloader.report_warning(msg)
2848 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2849 path='/', secure=False, discard=False, rest={}, **kwargs):
2850 cookie = compat_cookiejar_Cookie(
2851 0, name, value, port, port is not None, domain, True,
2852 domain.startswith('.'), path, True, secure, expire_time,
2853 discard, None, None, rest)
2854 self._downloader.cookiejar.set_cookie(cookie)
2856 def _get_cookies(self, url):
2857 """ Return a compat_cookies
.SimpleCookie
with the cookies
for the url
"""
2858 req = sanitized_Request(url)
2859 self._downloader.cookiejar.add_cookie_header(req)
2860 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2862 def _apply_first_set_cookie_header(self, url_handle, cookie):
2864 Apply first Set
-Cookie header instead of the last
. Experimental
.
2866 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
2867 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
2868 than
second (new
). However
, as of RFC6265 the newer one cookie
2869 should be
set into cookie store what actually happens
.
2870 We will workaround this issue by resetting the cookie to
2871 the first one manually
.
2872 1. https
://new
.vk
.com
/
2873 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
2874 3. https
://learning
.oreilly
.com
/
2876 for header, cookies in url_handle.headers.items():
2877 if header.lower() != 'set-cookie':
2879 if sys.version_info[0] >= 3:
2880 cookies = cookies.encode('iso-8859-1')
2881 cookies = cookies.decode('utf-8')
2882 cookie_value = re.search(
2883 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
2885 value, domain = cookie_value.groups()
2886 self._set_cookie(domain, cookie, value)
2889 def get_testcases(self, include_onlymatching=False):
2890 t = getattr(self, '_TEST', None)
2892 assert not hasattr(self, '_TESTS'), \
2893 '%s has _TEST and _TESTS' % type(self).__name__
2896 tests = getattr(self, '_TESTS', [])
2898 if not include_onlymatching and t.get('only_matching', False):
2900 t['name'] = type(self).__name__[:-len('IE')]
2903 def is_suitable(self, age_limit):
2904 """ Test whether the extractor
is generally suitable
for the given
2905 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
2907 any_restricted = False
2908 for tc in self.get_testcases(include_onlymatching=False):
2909 if tc.get('playlist', []):
2910 tc = tc['playlist'][0]
2911 is_restricted = age_restricted(
2912 tc.get('info_dict', {}).get('age_limit'), age_limit)
2913 if not is_restricted:
2915 any_restricted = any_restricted or is_restricted
2916 return not any_restricted
2918 def extract_subtitles(self, *args, **kwargs):
2919 if (self._downloader.params.get('writesubtitles', False)
2920 or self._downloader.params.get('listsubtitles')):
2921 return self._get_subtitles(*args, **kwargs)
2924 def _get_subtitles(self, *args, **kwargs):
2925 raise NotImplementedError('This method must be implemented by subclasses')
2928 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2929 """ Merge subtitle items
for one language
. Items
with duplicated URLs
2930 will be dropped
. """
2931 list1_urls = set([item['url'] for item in subtitle_list1])
2932 ret = list(subtitle_list1)
2933 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2937 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2938 """ Merge two subtitle dictionaries
, language by language
. """
2939 ret = dict(subtitle_dict1)
2940 for lang in subtitle_dict2:
2941 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2944 def extract_automatic_captions(self, *args, **kwargs):
2945 if (self._downloader.params.get('writeautomaticsub', False)
2946 or self._downloader.params.get('listsubtitles')):
2947 return self._get_automatic_captions(*args, **kwargs)
2950 def _get_automatic_captions(self, *args, **kwargs):
2951 raise NotImplementedError('This method must be implemented by subclasses')
2953 def mark_watched(self, *args, **kwargs):
2954 if (self._downloader.params.get('mark_watched', False)
2955 and (self._get_login_info()[0] is not None
2956 or self._downloader.params.get('cookiefile') is not None)):
2957 self._mark_watched(*args, **kwargs)
2959 def _mark_watched(self, *args, **kwargs):
2960 raise NotImplementedError('This method must be implemented by subclasses')
2962 def geo_verification_headers(self):
2964 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2965 if geo_verification_proxy:
2966 headers['Ytdl-request-proxy'] = geo_verification_proxy
2969 def _generic_id(self, url):
2970 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2972 def _generic_title(self, url):
2973 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2976 class SearchInfoExtractor(InfoExtractor):
2978 Base
class for paged search queries extractors
.
2979 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
2980 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
2984 def _make_valid_url(cls):
2985 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2988 def suitable(cls, url):
2989 return re.match(cls._make_valid_url(), url) is not None
2991 def _real_extract(self, query):
2992 mobj = re.match(self._make_valid_url(), query)
2994 raise ExtractorError('Invalid search query "%s"' % query)
2996 prefix = mobj.group('prefix')
2997 query = mobj.group('query')
2999 return self._get_n_results(query, 1)
3000 elif prefix == 'all':
3001 return self._get_n_results(query, self._MAX_RESULTS)
3005 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3006 elif n > self._MAX_RESULTS:
3007 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3008 n = self._MAX_RESULTS
3009 return self._get_n_results(query, n)
3011 def _get_n_results(self, query, n):
3012 """Get a specified number of results
for a query
"""
3013 raise NotImplementedError('This method must be implemented by subclasses')
3016 def SEARCH_KEY(self):
3017 return self._SEARCH_KEY