]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
Start new release.
[youtubedl] / youtube_dl / extractor / common.py
1 # coding: utf-8
2 from __future__ import unicode_literals
3
4 import base64
5 import datetime
6 import hashlib
7 import json
8 import netrc
9 import os
10 import random
11 import re
12 import socket
13 import sys
14 import time
15 import math
16
17 from ..compat import (
18 compat_cookiejar,
19 compat_cookies,
20 compat_etree_fromstring,
21 compat_getpass,
22 compat_integer_types,
23 compat_http_client,
24 compat_os_name,
25 compat_str,
26 compat_urllib_error,
27 compat_urllib_parse_unquote,
28 compat_urllib_parse_urlencode,
29 compat_urllib_request,
30 compat_urlparse,
31 compat_xml_parse_error,
32 )
33 from ..downloader.f4m import (
34 get_base_url,
35 remove_encrypted_media,
36 )
37 from ..utils import (
38 NO_DEFAULT,
39 age_restricted,
40 base_url,
41 bug_reports_message,
42 clean_html,
43 compiled_regex_type,
44 determine_ext,
45 determine_protocol,
46 error_to_compat_str,
47 ExtractorError,
48 extract_attributes,
49 fix_xml_ampersands,
50 float_or_none,
51 GeoRestrictedError,
52 GeoUtils,
53 int_or_none,
54 js_to_json,
55 JSON_LD_RE,
56 mimetype2ext,
57 orderedSet,
58 parse_codecs,
59 parse_duration,
60 parse_iso8601,
61 parse_m3u8_attributes,
62 RegexNotFoundError,
63 sanitized_Request,
64 sanitize_filename,
65 unescapeHTML,
66 unified_strdate,
67 unified_timestamp,
68 update_Request,
69 update_url_query,
70 urljoin,
71 url_basename,
72 url_or_none,
73 xpath_element,
74 xpath_text,
75 xpath_with_ns,
76 )
77
78
79 class InfoExtractor(object):
80 """Information Extractor class.
81
82 Information extractors are the classes that, given a URL, extract
83 information about the video (or videos) the URL refers to. This
84 information includes the real video URL, the video title, author and
85 others. The information is stored in a dictionary which is then
86 passed to the YoutubeDL. The YoutubeDL processes this
87 information possibly downloading the video to the file system, among
88 other possible outcomes.
89
90 The type field determines the type of the result.
91 By far the most common value (and the default if _type is missing) is
92 "video", which indicates a single video.
93
94 For a video, the dictionaries must include the following fields:
95
96 id: Video identifier.
97 title: Video title, unescaped.
98
99 Additionally, it must contain either a formats entry or a url one:
100
101 formats: A list of dictionaries for each format available, ordered
102 from worst to best quality.
103
104 Potential fields:
105 * url Mandatory. The URL of the video file
106 * manifest_url
107 The URL of the manifest file in case of
108 fragmented media (DASH, hls, hds)
109 * ext Will be calculated from URL if missing
110 * format A human-readable description of the format
111 ("mp4 container with h264/opus").
112 Calculated from the format_id, width, height.
113 and format_note fields if missing.
114 * format_id A short description of the format
115 ("mp4_h264_opus" or "19").
116 Technically optional, but strongly recommended.
117 * format_note Additional info about the format
118 ("3D" or "DASH video")
119 * width Width of the video, if known
120 * height Height of the video, if known
121 * resolution Textual description of width and height
122 * tbr Average bitrate of audio and video in KBit/s
123 * abr Average audio bitrate in KBit/s
124 * acodec Name of the audio codec in use
125 * asr Audio sampling rate in Hertz
126 * vbr Average video bitrate in KBit/s
127 * fps Frame rate
128 * vcodec Name of the video codec in use
129 * container Name of the container format
130 * filesize The number of bytes, if known in advance
131 * filesize_approx An estimate for the number of bytes
132 * player_url SWF Player URL (used for rtmpdump).
133 * protocol The protocol that will be used for the actual
134 download, lower-case.
135 "http", "https", "rtsp", "rtmp", "rtmpe",
136 "m3u8", "m3u8_native" or "http_dash_segments".
137 * fragment_base_url
138 Base URL for fragments. Each fragment's path
139 value (if present) will be relative to
140 this URL.
141 * fragments A list of fragments of a fragmented media.
142 Each fragment entry must contain either an url
143 or a path. If an url is present it should be
144 considered by a client. Otherwise both path and
145 fragment_base_url must be present. Here is
146 the list of all potential fields:
147 * "url" - fragment's URL
148 * "path" - fragment's path relative to
149 fragment_base_url
150 * "duration" (optional, int or float)
151 * "filesize" (optional, int)
152 * preference Order number of this format. If this field is
153 present and not None, the formats get sorted
154 by this field, regardless of all other values.
155 -1 for default (order by other properties),
156 -2 or smaller for less than default.
157 < -1000 to hide the format (if there is
158 another one which is strictly better)
159 * language Language code, e.g. "de" or "en-US".
160 * language_preference Is this in the language mentioned in
161 the URL?
162 10 if it's what the URL is about,
163 -1 for default (don't know),
164 -10 otherwise, other values reserved for now.
165 * quality Order number of the video quality of this
166 format, irrespective of the file format.
167 -1 for default (order by other properties),
168 -2 or smaller for less than default.
169 * source_preference Order number for this video source
170 (quality takes higher priority)
171 -1 for default (order by other properties),
172 -2 or smaller for less than default.
173 * http_headers A dictionary of additional HTTP headers
174 to add to the request.
175 * stretched_ratio If given and not 1, indicates that the
176 video's pixels are not square.
177 width : height ratio as float.
178 * no_resume The server does not support resuming the
179 (HTTP or RTMP) download. Boolean.
180 * downloader_options A dictionary of downloader options as
181 described in FileDownloader
182
183 url: Final video URL.
184 ext: Video filename extension.
185 format: The video format, defaults to ext (used for --get-format)
186 player_url: SWF Player URL (used for rtmpdump).
187
188 The following fields are optional:
189
190 alt_title: A secondary title of the video.
191 display_id An alternative identifier for the video, not necessarily
192 unique, but available before title. Typically, id is
193 something like "4234987", title "Dancing naked mole rats",
194 and display_id "dancing-naked-mole-rats"
195 thumbnails: A list of dictionaries, with the following entries:
196 * "id" (optional, string) - Thumbnail format ID
197 * "url"
198 * "preference" (optional, int) - quality of the image
199 * "width" (optional, int)
200 * "height" (optional, int)
201 * "resolution" (optional, string "{width}x{height"},
202 deprecated)
203 * "filesize" (optional, int)
204 thumbnail: Full URL to a video thumbnail image.
205 description: Full video description.
206 uploader: Full name of the video uploader.
207 license: License name the video is licensed under.
208 creator: The creator of the video.
209 release_date: The date (YYYYMMDD) when the video was released.
210 timestamp: UNIX timestamp of the moment the video became available.
211 upload_date: Video upload date (YYYYMMDD).
212 If not explicitly set, calculated from timestamp.
213 uploader_id: Nickname or id of the video uploader.
214 uploader_url: Full URL to a personal webpage of the video uploader.
215 channel: Full name of the channel the video is uploaded on.
216 Note that channel fields may or may not repeat uploader
217 fields. This depends on a particular extractor.
218 channel_id: Id of the channel.
219 channel_url: Full URL to a channel webpage.
220 location: Physical location where the video was filmed.
221 subtitles: The available subtitles as a dictionary in the format
222 {tag: subformats}. "tag" is usually a language code, and
223 "subformats" is a list sorted from lower to higher
224 preference, each element is a dictionary with the "ext"
225 entry and one of:
226 * "data": The subtitles file contents
227 * "url": A URL pointing to the subtitles file
228 "ext" will be calculated from URL if missing
229 automatic_captions: Like 'subtitles', used by the YoutubeIE for
230 automatically generated captions
231 duration: Length of the video in seconds, as an integer or float.
232 view_count: How many users have watched the video on the platform.
233 like_count: Number of positive ratings of the video
234 dislike_count: Number of negative ratings of the video
235 repost_count: Number of reposts of the video
236 average_rating: Average rating give by users, the scale used depends on the webpage
237 comment_count: Number of comments on the video
238 comments: A list of comments, each with one or more of the following
239 properties (all but one of text or html optional):
240 * "author" - human-readable name of the comment author
241 * "author_id" - user ID of the comment author
242 * "id" - Comment ID
243 * "html" - Comment as HTML
244 * "text" - Plain text of the comment
245 * "timestamp" - UNIX timestamp of comment
246 * "parent" - ID of the comment this one is replying to.
247 Set to "root" to indicate that this is a
248 comment to the original video.
249 age_limit: Age restriction for the video, as an integer (years)
250 webpage_url: The URL to the video webpage, if given to youtube-dl it
251 should allow to get the same result again. (It will be set
252 by YoutubeDL if it's missing)
253 categories: A list of categories that the video falls in, for example
254 ["Sports", "Berlin"]
255 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
256 is_live: True, False, or None (=unknown). Whether this video is a
257 live stream that goes on instead of a fixed-length video.
258 start_time: Time in seconds where the reproduction should start, as
259 specified in the URL.
260 end_time: Time in seconds where the reproduction should end, as
261 specified in the URL.
262 chapters: A list of dictionaries, with the following entries:
263 * "start_time" - The start time of the chapter in seconds
264 * "end_time" - The end time of the chapter in seconds
265 * "title" (optional, string)
266
267 The following fields should only be used when the video belongs to some logical
268 chapter or section:
269
270 chapter: Name or title of the chapter the video belongs to.
271 chapter_number: Number of the chapter the video belongs to, as an integer.
272 chapter_id: Id of the chapter the video belongs to, as a unicode string.
273
274 The following fields should only be used when the video is an episode of some
275 series, programme or podcast:
276
277 series: Title of the series or programme the video episode belongs to.
278 season: Title of the season the video episode belongs to.
279 season_number: Number of the season the video episode belongs to, as an integer.
280 season_id: Id of the season the video episode belongs to, as a unicode string.
281 episode: Title of the video episode. Unlike mandatory video title field,
282 this field should denote the exact title of the video episode
283 without any kind of decoration.
284 episode_number: Number of the video episode within a season, as an integer.
285 episode_id: Id of the video episode, as a unicode string.
286
287 The following fields should only be used when the media is a track or a part of
288 a music album:
289
290 track: Title of the track.
291 track_number: Number of the track within an album or a disc, as an integer.
292 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
293 as a unicode string.
294 artist: Artist(s) of the track.
295 genre: Genre(s) of the track.
296 album: Title of the album the track belongs to.
297 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
298 album_artist: List of all artists appeared on the album (e.g.
299 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
300 and compilations).
301 disc_number: Number of the disc or other physical medium the track belongs to,
302 as an integer.
303 release_year: Year (YYYY) when the album was released.
304
305 Unless mentioned otherwise, the fields should be Unicode strings.
306
307 Unless mentioned otherwise, None is equivalent to absence of information.
308
309
310 _type "playlist" indicates multiple videos.
311 There must be a key "entries", which is a list, an iterable, or a PagedList
312 object, each element of which is a valid dictionary by this specification.
313
314 Additionally, playlists can have "id", "title", "description", "uploader",
315 "uploader_id", "uploader_url" attributes with the same semantics as videos
316 (see above).
317
318
319 _type "multi_video" indicates that there are multiple videos that
320 form a single show, for examples multiple acts of an opera or TV episode.
321 It must have an entries key like a playlist and contain all the keys
322 required for a video at the same time.
323
324
325 _type "url" indicates that the video must be extracted from another
326 location, possibly by a different extractor. Its only required key is:
327 "url" - the next URL to extract.
328 The key "ie_key" can be set to the class name (minus the trailing "IE",
329 e.g. "Youtube") if the extractor class is known in advance.
330 Additionally, the dictionary may have any properties of the resolved entity
331 known in advance, for example "title" if the title of the referred video is
332 known ahead of time.
333
334
335 _type "url_transparent" entities have the same specification as "url", but
336 indicate that the given additional information is more precise than the one
337 associated with the resolved URL.
338 This is useful when a site employs a video service that hosts the video and
339 its technical metadata, but that video service does not embed a useful
340 title, description etc.
341
342
343 Subclasses of this one should re-define the _real_initialize() and
344 _real_extract() methods and define a _VALID_URL regexp.
345 Probably, they should also be added to the list of extractors.
346
347 _GEO_BYPASS attribute may be set to False in order to disable
348 geo restriction bypass mechanisms for a particular extractor.
349 Though it won't disable explicit geo restriction bypass based on
350 country code provided with geo_bypass_country.
351
352 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
353 countries for this extractor. One of these countries will be used by
354 geo restriction bypass mechanism right away in order to bypass
355 geo restriction, of course, if the mechanism is not disabled.
356
357 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
358 IP blocks in CIDR notation for this extractor. One of these IP blocks
359 will be used by geo restriction bypass mechanism similarly
360 to _GEO_COUNTRIES.
361
362 Finally, the _WORKING attribute should be set to False for broken IEs
363 in order to warn the users and skip the tests.
364 """
365
366 _ready = False
367 _downloader = None
368 _x_forwarded_for_ip = None
369 _GEO_BYPASS = True
370 _GEO_COUNTRIES = None
371 _GEO_IP_BLOCKS = None
372 _WORKING = True
373
374 def __init__(self, downloader=None):
375 """Constructor. Receives an optional downloader."""
376 self._ready = False
377 self._x_forwarded_for_ip = None
378 self.set_downloader(downloader)
379
380 @classmethod
381 def suitable(cls, url):
382 """Receives a URL and returns True if suitable for this IE."""
383
384 # This does not use has/getattr intentionally - we want to know whether
385 # we have cached the regexp for *this* class, whereas getattr would also
386 # match the superclass
387 if '_VALID_URL_RE' not in cls.__dict__:
388 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
389 return cls._VALID_URL_RE.match(url) is not None
390
391 @classmethod
392 def _match_id(cls, url):
393 if '_VALID_URL_RE' not in cls.__dict__:
394 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
395 m = cls._VALID_URL_RE.match(url)
396 assert m
397 return compat_str(m.group('id'))
398
399 @classmethod
400 def working(cls):
401 """Getter method for _WORKING."""
402 return cls._WORKING
403
404 def initialize(self):
405 """Initializes an instance (authentication, etc)."""
406 self._initialize_geo_bypass({
407 'countries': self._GEO_COUNTRIES,
408 'ip_blocks': self._GEO_IP_BLOCKS,
409 })
410 if not self._ready:
411 self._real_initialize()
412 self._ready = True
413
414 def _initialize_geo_bypass(self, geo_bypass_context):
415 """
416 Initialize geo restriction bypass mechanism.
417
418 This method is used to initialize geo bypass mechanism based on faking
419 X-Forwarded-For HTTP header. A random country from provided country list
420 is selected and a random IP belonging to this country is generated. This
421 IP will be passed as X-Forwarded-For HTTP header in all subsequent
422 HTTP requests.
423
424 This method will be used for initial geo bypass mechanism initialization
425 during the instance initialization with _GEO_COUNTRIES and
426 _GEO_IP_BLOCKS.
427
428 You may also manually call it from extractor's code if geo bypass
429 information is not available beforehand (e.g. obtained during
430 extraction) or due to some other reason. In this case you should pass
431 this information in geo bypass context passed as first argument. It may
432 contain following fields:
433
434 countries: List of geo unrestricted countries (similar
435 to _GEO_COUNTRIES)
436 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
437 (similar to _GEO_IP_BLOCKS)
438
439 """
440 if not self._x_forwarded_for_ip:
441
442 # Geo bypass mechanism is explicitly disabled by user
443 if not self._downloader.params.get('geo_bypass', True):
444 return
445
446 if not geo_bypass_context:
447 geo_bypass_context = {}
448
449 # Backward compatibility: previously _initialize_geo_bypass
450 # expected a list of countries, some 3rd party code may still use
451 # it this way
452 if isinstance(geo_bypass_context, (list, tuple)):
453 geo_bypass_context = {
454 'countries': geo_bypass_context,
455 }
456
457 # The whole point of geo bypass mechanism is to fake IP
458 # as X-Forwarded-For HTTP header based on some IP block or
459 # country code.
460
461 # Path 1: bypassing based on IP block in CIDR notation
462
463 # Explicit IP block specified by user, use it right away
464 # regardless of whether extractor is geo bypassable or not
465 ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
466
467 # Otherwise use random IP block from geo bypass context but only
468 # if extractor is known as geo bypassable
469 if not ip_block:
470 ip_blocks = geo_bypass_context.get('ip_blocks')
471 if self._GEO_BYPASS and ip_blocks:
472 ip_block = random.choice(ip_blocks)
473
474 if ip_block:
475 self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
476 if self._downloader.params.get('verbose', False):
477 self._downloader.to_screen(
478 '[debug] Using fake IP %s as X-Forwarded-For.'
479 % self._x_forwarded_for_ip)
480 return
481
482 # Path 2: bypassing based on country code
483
484 # Explicit country code specified by user, use it right away
485 # regardless of whether extractor is geo bypassable or not
486 country = self._downloader.params.get('geo_bypass_country', None)
487
488 # Otherwise use random country code from geo bypass context but
489 # only if extractor is known as geo bypassable
490 if not country:
491 countries = geo_bypass_context.get('countries')
492 if self._GEO_BYPASS and countries:
493 country = random.choice(countries)
494
495 if country:
496 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
497 if self._downloader.params.get('verbose', False):
498 self._downloader.to_screen(
499 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
500 % (self._x_forwarded_for_ip, country.upper()))
501
502 def extract(self, url):
503 """Extracts URL information and returns it in list of dicts."""
504 try:
505 for _ in range(2):
506 try:
507 self.initialize()
508 ie_result = self._real_extract(url)
509 if self._x_forwarded_for_ip:
510 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
511 return ie_result
512 except GeoRestrictedError as e:
513 if self.__maybe_fake_ip_and_retry(e.countries):
514 continue
515 raise
516 except ExtractorError:
517 raise
518 except compat_http_client.IncompleteRead as e:
519 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
520 except (KeyError, StopIteration) as e:
521 raise ExtractorError('An extractor error has occurred.', cause=e)
522
523 def __maybe_fake_ip_and_retry(self, countries):
524 if (not self._downloader.params.get('geo_bypass_country', None) and
525 self._GEO_BYPASS and
526 self._downloader.params.get('geo_bypass', True) and
527 not self._x_forwarded_for_ip and
528 countries):
529 country_code = random.choice(countries)
530 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
531 if self._x_forwarded_for_ip:
532 self.report_warning(
533 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
534 % (self._x_forwarded_for_ip, country_code.upper()))
535 return True
536 return False
537
538 def set_downloader(self, downloader):
539 """Sets the downloader for this IE."""
540 self._downloader = downloader
541
542 def _real_initialize(self):
543 """Real initialization process. Redefine in subclasses."""
544 pass
545
546 def _real_extract(self, url):
547 """Real extraction process. Redefine in subclasses."""
548 pass
549
550 @classmethod
551 def ie_key(cls):
552 """A string for getting the InfoExtractor with get_info_extractor"""
553 return compat_str(cls.__name__[:-2])
554
555 @property
556 def IE_NAME(self):
557 return compat_str(type(self).__name__[:-2])
558
559 @staticmethod
560 def __can_accept_status_code(err, expected_status):
561 assert isinstance(err, compat_urllib_error.HTTPError)
562 if expected_status is None:
563 return False
564 if isinstance(expected_status, compat_integer_types):
565 return err.code == expected_status
566 elif isinstance(expected_status, (list, tuple)):
567 return err.code in expected_status
568 elif callable(expected_status):
569 return expected_status(err.code) is True
570 else:
571 assert False
572
573 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
574 """
575 Return the response handle.
576
577 See _download_webpage docstring for arguments specification.
578 """
579 if note is None:
580 self.report_download_webpage(video_id)
581 elif note is not False:
582 if video_id is None:
583 self.to_screen('%s' % (note,))
584 else:
585 self.to_screen('%s: %s' % (video_id, note))
586
587 # Some sites check X-Forwarded-For HTTP header in order to figure out
588 # the origin of the client behind proxy. This allows bypassing geo
589 # restriction by faking this header's value to IP that belongs to some
590 # geo unrestricted country. We will do so once we encounter any
591 # geo restriction error.
592 if self._x_forwarded_for_ip:
593 if 'X-Forwarded-For' not in headers:
594 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
595
596 if isinstance(url_or_request, compat_urllib_request.Request):
597 url_or_request = update_Request(
598 url_or_request, data=data, headers=headers, query=query)
599 else:
600 if query:
601 url_or_request = update_url_query(url_or_request, query)
602 if data is not None or headers:
603 url_or_request = sanitized_Request(url_or_request, data, headers)
604 try:
605 return self._downloader.urlopen(url_or_request)
606 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
607 if isinstance(err, compat_urllib_error.HTTPError):
608 if self.__can_accept_status_code(err, expected_status):
609 # Retain reference to error to prevent file object from
610 # being closed before it can be read. Works around the
611 # effects of <https://bugs.python.org/issue15002>
612 # introduced in Python 3.4.1.
613 err.fp._error = err
614 return err.fp
615
616 if errnote is False:
617 return False
618 if errnote is None:
619 errnote = 'Unable to download webpage'
620
621 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
622 if fatal:
623 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
624 else:
625 self._downloader.report_warning(errmsg)
626 return False
627
628 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):
629 """
630 Return a tuple (page content as string, URL handle).
631
632 See _download_webpage docstring for arguments specification.
633 """
634 # Strip hashes from the URL (#1038)
635 if isinstance(url_or_request, (compat_str, str)):
636 url_or_request = url_or_request.partition('#')[0]
637
638 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
639 if urlh is False:
640 assert not fatal
641 return False
642 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
643 return (content, urlh)
644
645 @staticmethod
646 def _guess_encoding_from_content(content_type, webpage_bytes):
647 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
648 if m:
649 encoding = m.group(1)
650 else:
651 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
652 webpage_bytes[:1024])
653 if m:
654 encoding = m.group(1).decode('ascii')
655 elif webpage_bytes.startswith(b'\xff\xfe'):
656 encoding = 'utf-16'
657 else:
658 encoding = 'utf-8'
659
660 return encoding
661
662 def __check_blocked(self, content):
663 first_block = content[:512]
664 if ('<title>Access to this site is blocked</title>' in content and
665 'Websense' in first_block):
666 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
667 blocked_iframe = self._html_search_regex(
668 r'<iframe src="([^"]+)"', content,
669 'Websense information URL', default=None)
670 if blocked_iframe:
671 msg += ' Visit %s for more details' % blocked_iframe
672 raise ExtractorError(msg, expected=True)
673 if '<title>The URL you requested has been blocked</title>' in first_block:
674 msg = (
675 'Access to this webpage has been blocked by Indian censorship. '
676 'Use a VPN or proxy server (with --proxy) to route around it.')
677 block_msg = self._html_search_regex(
678 r'</h1><p>(.*?)</p>',
679 content, 'block message', default=None)
680 if block_msg:
681 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
682 raise ExtractorError(msg, expected=True)
683 if ('<title>TTK :: Š”Š¾ŃŃ‚ŃƒŠæ Šŗ рŠµŃŃƒŃ€ŃŃƒ Š¾Š³Ń€Š°Š½ŠøчŠµŠ½</title>' in content and
684 'blocklist.rkn.gov.ru' in content):
685 raise ExtractorError(
686 'Access to this webpage has been blocked by decision of the Russian government. '
687 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
688 expected=True)
689
690 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
691 content_type = urlh.headers.get('Content-Type', '')
692 webpage_bytes = urlh.read()
693 if prefix is not None:
694 webpage_bytes = prefix + webpage_bytes
695 if not encoding:
696 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
697 if self._downloader.params.get('dump_intermediate_pages', False):
698 self.to_screen('Dumping request to ' + urlh.geturl())
699 dump = base64.b64encode(webpage_bytes).decode('ascii')
700 self._downloader.to_screen(dump)
701 if self._downloader.params.get('write_pages', False):
702 basen = '%s_%s' % (video_id, urlh.geturl())
703 if len(basen) > 240:
704 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
705 basen = basen[:240 - len(h)] + h
706 raw_filename = basen + '.dump'
707 filename = sanitize_filename(raw_filename, restricted=True)
708 self.to_screen('Saving request to ' + filename)
709 # Working around MAX_PATH limitation on Windows (see
710 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
711 if compat_os_name == 'nt':
712 absfilepath = os.path.abspath(filename)
713 if len(absfilepath) > 259:
714 filename = '\\\\?\\' + absfilepath
715 with open(filename, 'wb') as outf:
716 outf.write(webpage_bytes)
717
718 try:
719 content = webpage_bytes.decode(encoding, 'replace')
720 except LookupError:
721 content = webpage_bytes.decode('utf-8', 'replace')
722
723 self.__check_blocked(content)
724
725 return content
726
727 def _download_webpage(
728 self, url_or_request, video_id, note=None, errnote=None,
729 fatal=True, tries=1, timeout=5, encoding=None, data=None,
730 headers={}, query={}, expected_status=None):
731 """
732 Return the data of the page as a string.
733
734 Arguments:
735 url_or_request -- plain text URL as a string or
736 a compat_urllib_request.Requestobject
737 video_id -- Video/playlist/item identifier (string)
738
739 Keyword arguments:
740 note -- note printed before downloading (string)
741 errnote -- note printed in case of an error (string)
742 fatal -- flag denoting whether error should be considered fatal,
743 i.e. whether it should cause ExtractionError to be raised,
744 otherwise a warning will be reported and extraction continued
745 tries -- number of tries
746 timeout -- sleep interval between tries
747 encoding -- encoding for a page content decoding, guessed automatically
748 when not explicitly specified
749 data -- POST data (bytes)
750 headers -- HTTP headers (dict)
751 query -- URL query (dict)
752 expected_status -- allows to accept failed HTTP requests (non 2xx
753 status code) by explicitly specifying a set of accepted status
754 codes. Can be any of the following entities:
755 - an integer type specifying an exact failed status code to
756 accept
757 - a list or a tuple of integer types specifying a list of
758 failed status codes to accept
759 - a callable accepting an actual failed status code and
760 returning True if it should be accepted
761 Note that this argument does not affect success status codes (2xx)
762 which are always accepted.
763 """
764
765 success = False
766 try_count = 0
767 while success is False:
768 try:
769 res = self._download_webpage_handle(
770 url_or_request, video_id, note, errnote, fatal,
771 encoding=encoding, data=data, headers=headers, query=query,
772 expected_status=expected_status)
773 success = True
774 except compat_http_client.IncompleteRead as e:
775 try_count += 1
776 if try_count >= tries:
777 raise e
778 self._sleep(timeout, video_id)
779 if res is False:
780 return res
781 else:
782 content, _ = res
783 return content
784
785 def _download_xml_handle(
786 self, url_or_request, video_id, note='Downloading XML',
787 errnote='Unable to download XML', transform_source=None,
788 fatal=True, encoding=None, data=None, headers={}, query={},
789 expected_status=None):
790 """
791 Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle).
792
793 See _download_webpage docstring for arguments specification.
794 """
795 res = self._download_webpage_handle(
796 url_or_request, video_id, note, errnote, fatal=fatal,
797 encoding=encoding, data=data, headers=headers, query=query,
798 expected_status=expected_status)
799 if res is False:
800 return res
801 xml_string, urlh = res
802 return self._parse_xml(
803 xml_string, video_id, transform_source=transform_source,
804 fatal=fatal), urlh
805
806 def _download_xml(
807 self, url_or_request, video_id,
808 note='Downloading XML', errnote='Unable to download XML',
809 transform_source=None, fatal=True, encoding=None,
810 data=None, headers={}, query={}, expected_status=None):
811 """
812 Return the xml as an xml.etree.ElementTree.Element.
813
814 See _download_webpage docstring for arguments specification.
815 """
816 res = self._download_xml_handle(
817 url_or_request, video_id, note=note, errnote=errnote,
818 transform_source=transform_source, fatal=fatal, encoding=encoding,
819 data=data, headers=headers, query=query,
820 expected_status=expected_status)
821 return res if res is False else res[0]
822
823 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
824 if transform_source:
825 xml_string = transform_source(xml_string)
826 try:
827 return compat_etree_fromstring(xml_string.encode('utf-8'))
828 except compat_xml_parse_error as ve:
829 errmsg = '%s: Failed to parse XML ' % video_id
830 if fatal:
831 raise ExtractorError(errmsg, cause=ve)
832 else:
833 self.report_warning(errmsg + str(ve))
834
835 def _download_json_handle(
836 self, url_or_request, video_id, note='Downloading JSON metadata',
837 errnote='Unable to download JSON metadata', transform_source=None,
838 fatal=True, encoding=None, data=None, headers={}, query={},
839 expected_status=None):
840 """
841 Return a tuple (JSON object, URL handle).
842
843 See _download_webpage docstring for arguments specification.
844 """
845 res = self._download_webpage_handle(
846 url_or_request, video_id, note, errnote, fatal=fatal,
847 encoding=encoding, data=data, headers=headers, query=query,
848 expected_status=expected_status)
849 if res is False:
850 return res
851 json_string, urlh = res
852 return self._parse_json(
853 json_string, video_id, transform_source=transform_source,
854 fatal=fatal), urlh
855
856 def _download_json(
857 self, url_or_request, video_id, note='Downloading JSON metadata',
858 errnote='Unable to download JSON metadata', transform_source=None,
859 fatal=True, encoding=None, data=None, headers={}, query={},
860 expected_status=None):
861 """
862 Return the JSON object as a dict.
863
864 See _download_webpage docstring for arguments specification.
865 """
866 res = self._download_json_handle(
867 url_or_request, video_id, note=note, errnote=errnote,
868 transform_source=transform_source, fatal=fatal, encoding=encoding,
869 data=data, headers=headers, query=query,
870 expected_status=expected_status)
871 return res if res is False else res[0]
872
873 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
874 if transform_source:
875 json_string = transform_source(json_string)
876 try:
877 return json.loads(json_string)
878 except ValueError as ve:
879 errmsg = '%s: Failed to parse JSON ' % video_id
880 if fatal:
881 raise ExtractorError(errmsg, cause=ve)
882 else:
883 self.report_warning(errmsg + str(ve))
884
885 def report_warning(self, msg, video_id=None):
886 idstr = '' if video_id is None else '%s: ' % video_id
887 self._downloader.report_warning(
888 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
889
890 def to_screen(self, msg):
891 """Print msg to screen, prefixing it with '[ie_name]'"""
892 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
893
894 def report_extraction(self, id_or_name):
895 """Report information extraction."""
896 self.to_screen('%s: Extracting information' % id_or_name)
897
898 def report_download_webpage(self, video_id):
899 """Report webpage download."""
900 self.to_screen('%s: Downloading webpage' % video_id)
901
902 def report_age_confirmation(self):
903 """Report attempt to confirm age."""
904 self.to_screen('Confirming age')
905
906 def report_login(self):
907 """Report attempt to log in."""
908 self.to_screen('Logging in')
909
910 @staticmethod
911 def raise_login_required(msg='This video is only available for registered users'):
912 raise ExtractorError(
913 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
914 expected=True)
915
916 @staticmethod
917 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
918 raise GeoRestrictedError(msg, countries=countries)
919
920 # Methods for following #608
921 @staticmethod
922 def url_result(url, ie=None, video_id=None, video_title=None):
923 """Returns a URL that points to a page that should be processed"""
924 # TODO: ie should be the class used for getting the info
925 video_info = {'_type': 'url',
926 'url': url,
927 'ie_key': ie}
928 if video_id is not None:
929 video_info['id'] = video_id
930 if video_title is not None:
931 video_info['title'] = video_title
932 return video_info
933
934 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
935 urls = orderedSet(
936 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
937 for m in matches)
938 return self.playlist_result(
939 urls, playlist_id=playlist_id, playlist_title=playlist_title)
940
941 @staticmethod
942 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
943 """Returns a playlist"""
944 video_info = {'_type': 'playlist',
945 'entries': entries}
946 if playlist_id:
947 video_info['id'] = playlist_id
948 if playlist_title:
949 video_info['title'] = playlist_title
950 if playlist_description:
951 video_info['description'] = playlist_description
952 return video_info
953
954 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
955 """
956 Perform a regex search on the given string, using a single or a list of
957 patterns returning the first matching group.
958 In case of failure return a default value or raise a WARNING or a
959 RegexNotFoundError, depending on fatal, specifying the field name.
960 """
961 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
962 mobj = re.search(pattern, string, flags)
963 else:
964 for p in pattern:
965 mobj = re.search(p, string, flags)
966 if mobj:
967 break
968
969 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
970 _name = '\033[0;34m%s\033[0m' % name
971 else:
972 _name = name
973
974 if mobj:
975 if group is None:
976 # return the first matching group
977 return next(g for g in mobj.groups() if g is not None)
978 else:
979 return mobj.group(group)
980 elif default is not NO_DEFAULT:
981 return default
982 elif fatal:
983 raise RegexNotFoundError('Unable to extract %s' % _name)
984 else:
985 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
986 return None
987
988 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
989 """
990 Like _search_regex, but strips HTML tags and unescapes entities.
991 """
992 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
993 if res:
994 return clean_html(res).strip()
995 else:
996 return res
997
998 def _get_netrc_login_info(self, netrc_machine=None):
999 username = None
1000 password = None
1001 netrc_machine = netrc_machine or self._NETRC_MACHINE
1002
1003 if self._downloader.params.get('usenetrc', False):
1004 try:
1005 info = netrc.netrc().authenticators(netrc_machine)
1006 if info is not None:
1007 username = info[0]
1008 password = info[2]
1009 else:
1010 raise netrc.NetrcParseError(
1011 'No authenticators for %s' % netrc_machine)
1012 except (IOError, netrc.NetrcParseError) as err:
1013 self._downloader.report_warning(
1014 'parsing .netrc: %s' % error_to_compat_str(err))
1015
1016 return username, password
1017
1018 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
1019 """
1020 Get the login info as (username, password)
1021 First look for the manually specified credentials using username_option
1022 and password_option as keys in params dictionary. If no such credentials
1023 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1024 value.
1025 If there's no info available, return (None, None)
1026 """
1027 if self._downloader is None:
1028 return (None, None)
1029
1030 downloader_params = self._downloader.params
1031
1032 # Attempt to use provided username and password or .netrc data
1033 if downloader_params.get(username_option) is not None:
1034 username = downloader_params[username_option]
1035 password = downloader_params[password_option]
1036 else:
1037 username, password = self._get_netrc_login_info(netrc_machine)
1038
1039 return username, password
1040
1041 def _get_tfa_info(self, note='two-factor verification code'):
1042 """
1043 Get the two-factor authentication info
1044 TODO - asking the user will be required for sms/phone verify
1045 currently just uses the command line option
1046 If there's no info available, return None
1047 """
1048 if self._downloader is None:
1049 return None
1050 downloader_params = self._downloader.params
1051
1052 if downloader_params.get('twofactor') is not None:
1053 return downloader_params['twofactor']
1054
1055 return compat_getpass('Type %s and press [Return]: ' % note)
1056
1057 # Helper functions for extracting OpenGraph info
1058 @staticmethod
1059 def _og_regexes(prop):
1060 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
1061 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
1062 % {'prop': re.escape(prop)})
1063 template = r'<meta[^>]+?%s[^>]+?%s'
1064 return [
1065 template % (property_re, content_re),
1066 template % (content_re, property_re),
1067 ]
1068
1069 @staticmethod
1070 def _meta_regex(prop):
1071 return r'''(?isx)<meta
1072 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1073 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1074
1075 def _og_search_property(self, prop, html, name=None, **kargs):
1076 if not isinstance(prop, (list, tuple)):
1077 prop = [prop]
1078 if name is None:
1079 name = 'OpenGraph %s' % prop[0]
1080 og_regexes = []
1081 for p in prop:
1082 og_regexes.extend(self._og_regexes(p))
1083 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1084 if escaped is None:
1085 return None
1086 return unescapeHTML(escaped)
1087
1088 def _og_search_thumbnail(self, html, **kargs):
1089 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
1090
1091 def _og_search_description(self, html, **kargs):
1092 return self._og_search_property('description', html, fatal=False, **kargs)
1093
1094 def _og_search_title(self, html, **kargs):
1095 return self._og_search_property('title', html, **kargs)
1096
1097 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
1098 regexes = self._og_regexes('video') + self._og_regexes('video:url')
1099 if secure:
1100 regexes = self._og_regexes('video:secure_url') + regexes
1101 return self._html_search_regex(regexes, html, name, **kargs)
1102
1103 def _og_search_url(self, html, **kargs):
1104 return self._og_search_property('url', html, **kargs)
1105
1106 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1107 if not isinstance(name, (list, tuple)):
1108 name = [name]
1109 if display_name is None:
1110 display_name = name[0]
1111 return self._html_search_regex(
1112 [self._meta_regex(n) for n in name],
1113 html, display_name, fatal=fatal, group='content', **kwargs)
1114
1115 def _dc_search_uploader(self, html):
1116 return self._html_search_meta('dc.creator', html, 'uploader')
1117
1118 def _rta_search(self, html):
1119 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1120 if re.search(r'(?ix)<meta\s+name="rating"\s+'
1121 r' content="RTA-5042-1996-1400-1577-RTA"',
1122 html):
1123 return 18
1124 return 0
1125
1126 def _media_rating_search(self, html):
1127 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1128 rating = self._html_search_meta('rating', html)
1129
1130 if not rating:
1131 return None
1132
1133 RATING_TABLE = {
1134 'safe for kids': 0,
1135 'general': 8,
1136 '14 years': 14,
1137 'mature': 17,
1138 'restricted': 19,
1139 }
1140 return RATING_TABLE.get(rating.lower())
1141
1142 def _family_friendly_search(self, html):
1143 # See http://schema.org/VideoObject
1144 family_friendly = self._html_search_meta(
1145 'isFamilyFriendly', html, default=None)
1146
1147 if not family_friendly:
1148 return None
1149
1150 RATING_TABLE = {
1151 '1': 0,
1152 'true': 0,
1153 '0': 18,
1154 'false': 18,
1155 }
1156 return RATING_TABLE.get(family_friendly.lower())
1157
1158 def _twitter_search_player(self, html):
1159 return self._html_search_meta('twitter:player', html,
1160 'twitter card player')
1161
1162 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1163 json_ld = self._search_regex(
1164 JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
1165 default = kwargs.get('default', NO_DEFAULT)
1166 if not json_ld:
1167 return default if default is not NO_DEFAULT else {}
1168 # JSON-LD may be malformed and thus `fatal` should be respected.
1169 # At the same time `default` may be passed that assumes `fatal=False`
1170 # for _search_regex. Let's simulate the same behavior here as well.
1171 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
1172 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
1173
1174 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
1175 if isinstance(json_ld, compat_str):
1176 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
1177 if not json_ld:
1178 return {}
1179 info = {}
1180 if not isinstance(json_ld, (list, tuple, dict)):
1181 return info
1182 if isinstance(json_ld, dict):
1183 json_ld = [json_ld]
1184
1185 INTERACTION_TYPE_MAP = {
1186 'CommentAction': 'comment',
1187 'AgreeAction': 'like',
1188 'DisagreeAction': 'dislike',
1189 'LikeAction': 'like',
1190 'DislikeAction': 'dislike',
1191 'ListenAction': 'view',
1192 'WatchAction': 'view',
1193 'ViewAction': 'view',
1194 }
1195
1196 def extract_interaction_statistic(e):
1197 interaction_statistic = e.get('interactionStatistic')
1198 if not isinstance(interaction_statistic, list):
1199 return
1200 for is_e in interaction_statistic:
1201 if not isinstance(is_e, dict):
1202 continue
1203 if is_e.get('@type') != 'InteractionCounter':
1204 continue
1205 interaction_type = is_e.get('interactionType')
1206 if not isinstance(interaction_type, compat_str):
1207 continue
1208 interaction_count = int_or_none(is_e.get('userInteractionCount'))
1209 if interaction_count is None:
1210 continue
1211 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
1212 if not count_kind:
1213 continue
1214 count_key = '%s_count' % count_kind
1215 if info.get(count_key) is not None:
1216 continue
1217 info[count_key] = interaction_count
1218
1219 def extract_video_object(e):
1220 assert e['@type'] == 'VideoObject'
1221 info.update({
1222 'url': url_or_none(e.get('contentUrl')),
1223 'title': unescapeHTML(e.get('name')),
1224 'description': unescapeHTML(e.get('description')),
1225 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
1226 'duration': parse_duration(e.get('duration')),
1227 'timestamp': unified_timestamp(e.get('uploadDate')),
1228 'filesize': float_or_none(e.get('contentSize')),
1229 'tbr': int_or_none(e.get('bitrate')),
1230 'width': int_or_none(e.get('width')),
1231 'height': int_or_none(e.get('height')),
1232 'view_count': int_or_none(e.get('interactionCount')),
1233 })
1234 extract_interaction_statistic(e)
1235
1236 for e in json_ld:
1237 if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
1238 item_type = e.get('@type')
1239 if expected_type is not None and expected_type != item_type:
1240 return info
1241 if item_type in ('TVEpisode', 'Episode'):
1242 episode_name = unescapeHTML(e.get('name'))
1243 info.update({
1244 'episode': episode_name,
1245 'episode_number': int_or_none(e.get('episodeNumber')),
1246 'description': unescapeHTML(e.get('description')),
1247 })
1248 if not info.get('title') and episode_name:
1249 info['title'] = episode_name
1250 part_of_season = e.get('partOfSeason')
1251 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1252 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1253 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1254 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1255 info['series'] = unescapeHTML(part_of_series.get('name'))
1256 elif item_type == 'Movie':
1257 info.update({
1258 'title': unescapeHTML(e.get('name')),
1259 'description': unescapeHTML(e.get('description')),
1260 'duration': parse_duration(e.get('duration')),
1261 'timestamp': unified_timestamp(e.get('dateCreated')),
1262 })
1263 elif item_type in ('Article', 'NewsArticle'):
1264 info.update({
1265 'timestamp': parse_iso8601(e.get('datePublished')),
1266 'title': unescapeHTML(e.get('headline')),
1267 'description': unescapeHTML(e.get('articleBody')),
1268 })
1269 elif item_type == 'VideoObject':
1270 extract_video_object(e)
1271 continue
1272 video = e.get('video')
1273 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1274 extract_video_object(video)
1275 break
1276 return dict((k, v) for k, v in info.items() if v is not None)
1277
1278 @staticmethod
1279 def _hidden_inputs(html):
1280 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1281 hidden_inputs = {}
1282 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1283 attrs = extract_attributes(input)
1284 if not input:
1285 continue
1286 if attrs.get('type') not in ('hidden', 'submit'):
1287 continue
1288 name = attrs.get('name') or attrs.get('id')
1289 value = attrs.get('value')
1290 if name and value is not None:
1291 hidden_inputs[name] = value
1292 return hidden_inputs
1293
1294 def _form_hidden_inputs(self, form_id, html):
1295 form = self._search_regex(
1296 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1297 html, '%s form' % form_id, group='form')
1298 return self._hidden_inputs(form)
1299
1300 def _sort_formats(self, formats, field_preference=None):
1301 if not formats:
1302 raise ExtractorError('No video formats found')
1303
1304 for f in formats:
1305 # Automatically determine tbr when missing based on abr and vbr (improves
1306 # formats sorting in some cases)
1307 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1308 f['tbr'] = f['abr'] + f['vbr']
1309
1310 def _formats_key(f):
1311 # TODO remove the following workaround
1312 from ..utils import determine_ext
1313 if not f.get('ext') and 'url' in f:
1314 f['ext'] = determine_ext(f['url'])
1315
1316 if isinstance(field_preference, (list, tuple)):
1317 return tuple(
1318 f.get(field)
1319 if f.get(field) is not None
1320 else ('' if field == 'format_id' else -1)
1321 for field in field_preference)
1322
1323 preference = f.get('preference')
1324 if preference is None:
1325 preference = 0
1326 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1327 preference -= 0.5
1328
1329 protocol = f.get('protocol') or determine_protocol(f)
1330 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1331
1332 if f.get('vcodec') == 'none': # audio only
1333 preference -= 50
1334 if self._downloader.params.get('prefer_free_formats'):
1335 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1336 else:
1337 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1338 ext_preference = 0
1339 try:
1340 audio_ext_preference = ORDER.index(f['ext'])
1341 except ValueError:
1342 audio_ext_preference = -1
1343 else:
1344 if f.get('acodec') == 'none': # video only
1345 preference -= 40
1346 if self._downloader.params.get('prefer_free_formats'):
1347 ORDER = ['flv', 'mp4', 'webm']
1348 else:
1349 ORDER = ['webm', 'flv', 'mp4']
1350 try:
1351 ext_preference = ORDER.index(f['ext'])
1352 except ValueError:
1353 ext_preference = -1
1354 audio_ext_preference = 0
1355
1356 return (
1357 preference,
1358 f.get('language_preference') if f.get('language_preference') is not None else -1,
1359 f.get('quality') if f.get('quality') is not None else -1,
1360 f.get('tbr') if f.get('tbr') is not None else -1,
1361 f.get('filesize') if f.get('filesize') is not None else -1,
1362 f.get('vbr') if f.get('vbr') is not None else -1,
1363 f.get('height') if f.get('height') is not None else -1,
1364 f.get('width') if f.get('width') is not None else -1,
1365 proto_preference,
1366 ext_preference,
1367 f.get('abr') if f.get('abr') is not None else -1,
1368 audio_ext_preference,
1369 f.get('fps') if f.get('fps') is not None else -1,
1370 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1371 f.get('source_preference') if f.get('source_preference') is not None else -1,
1372 f.get('format_id') if f.get('format_id') is not None else '',
1373 )
1374 formats.sort(key=_formats_key)
1375
1376 def _check_formats(self, formats, video_id):
1377 if formats:
1378 formats[:] = filter(
1379 lambda f: self._is_valid_url(
1380 f['url'], video_id,
1381 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1382 formats)
1383
1384 @staticmethod
1385 def _remove_duplicate_formats(formats):
1386 format_urls = set()
1387 unique_formats = []
1388 for f in formats:
1389 if f['url'] not in format_urls:
1390 format_urls.add(f['url'])
1391 unique_formats.append(f)
1392 formats[:] = unique_formats
1393
1394 def _is_valid_url(self, url, video_id, item='video', headers={}):
1395 url = self._proto_relative_url(url, scheme='http:')
1396 # For now assume non HTTP(S) URLs always valid
1397 if not (url.startswith('http://') or url.startswith('https://')):
1398 return True
1399 try:
1400 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1401 return True
1402 except ExtractorError as e:
1403 if isinstance(e.cause, compat_urllib_error.URLError):
1404 self.to_screen(
1405 '%s: %s URL is invalid, skipping' % (video_id, item))
1406 return False
1407 raise
1408
1409 def http_scheme(self):
1410 """ Either "http:" or "https:", depending on the user's preferences """
1411 return (
1412 'http:'
1413 if self._downloader.params.get('prefer_insecure', False)
1414 else 'https:')
1415
1416 def _proto_relative_url(self, url, scheme=None):
1417 if url is None:
1418 return url
1419 if url.startswith('//'):
1420 if scheme is None:
1421 scheme = self.http_scheme()
1422 return scheme + url
1423 else:
1424 return url
1425
1426 def _sleep(self, timeout, video_id, msg_template=None):
1427 if msg_template is None:
1428 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1429 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1430 self.to_screen(msg)
1431 time.sleep(timeout)
1432
1433 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1434 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1435 fatal=True, m3u8_id=None):
1436 manifest = self._download_xml(
1437 manifest_url, video_id, 'Downloading f4m manifest',
1438 'Unable to download f4m manifest',
1439 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1440 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1441 transform_source=transform_source,
1442 fatal=fatal)
1443
1444 if manifest is False:
1445 return []
1446
1447 return self._parse_f4m_formats(
1448 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1449 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1450
1451 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1452 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1453 fatal=True, m3u8_id=None):
1454 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1455 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1456 if akamai_pv is not None and ';' in akamai_pv.text:
1457 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1458 if playerVerificationChallenge.strip() != '':
1459 return []
1460
1461 formats = []
1462 manifest_version = '1.0'
1463 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1464 if not media_nodes:
1465 manifest_version = '2.0'
1466 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1467 # Remove unsupported DRM protected media from final formats
1468 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1469 media_nodes = remove_encrypted_media(media_nodes)
1470 if not media_nodes:
1471 return formats
1472
1473 manifest_base_url = get_base_url(manifest)
1474
1475 bootstrap_info = xpath_element(
1476 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1477 'bootstrap info', default=None)
1478
1479 vcodec = None
1480 mime_type = xpath_text(
1481 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1482 'base URL', default=None)
1483 if mime_type and mime_type.startswith('audio/'):
1484 vcodec = 'none'
1485
1486 for i, media_el in enumerate(media_nodes):
1487 tbr = int_or_none(media_el.attrib.get('bitrate'))
1488 width = int_or_none(media_el.attrib.get('width'))
1489 height = int_or_none(media_el.attrib.get('height'))
1490 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1491 # If <bootstrapInfo> is present, the specified f4m is a
1492 # stream-level manifest, and only set-level manifests may refer to
1493 # external resources. See section 11.4 and section 4 of F4M spec
1494 if bootstrap_info is None:
1495 media_url = None
1496 # @href is introduced in 2.0, see section 11.6 of F4M spec
1497 if manifest_version == '2.0':
1498 media_url = media_el.attrib.get('href')
1499 if media_url is None:
1500 media_url = media_el.attrib.get('url')
1501 if not media_url:
1502 continue
1503 manifest_url = (
1504 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1505 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1506 # If media_url is itself a f4m manifest do the recursive extraction
1507 # since bitrates in parent manifest (this one) and media_url manifest
1508 # may differ leading to inability to resolve the format by requested
1509 # bitrate in f4m downloader
1510 ext = determine_ext(manifest_url)
1511 if ext == 'f4m':
1512 f4m_formats = self._extract_f4m_formats(
1513 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1514 transform_source=transform_source, fatal=fatal)
1515 # Sometimes stream-level manifest contains single media entry that
1516 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1517 # At the same time parent's media entry in set-level manifest may
1518 # contain it. We will copy it from parent in such cases.
1519 if len(f4m_formats) == 1:
1520 f = f4m_formats[0]
1521 f.update({
1522 'tbr': f.get('tbr') or tbr,
1523 'width': f.get('width') or width,
1524 'height': f.get('height') or height,
1525 'format_id': f.get('format_id') if not tbr else format_id,
1526 'vcodec': vcodec,
1527 })
1528 formats.extend(f4m_formats)
1529 continue
1530 elif ext == 'm3u8':
1531 formats.extend(self._extract_m3u8_formats(
1532 manifest_url, video_id, 'mp4', preference=preference,
1533 m3u8_id=m3u8_id, fatal=fatal))
1534 continue
1535 formats.append({
1536 'format_id': format_id,
1537 'url': manifest_url,
1538 'manifest_url': manifest_url,
1539 'ext': 'flv' if bootstrap_info is not None else None,
1540 'protocol': 'f4m',
1541 'tbr': tbr,
1542 'width': width,
1543 'height': height,
1544 'vcodec': vcodec,
1545 'preference': preference,
1546 })
1547 return formats
1548
1549 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1550 return {
1551 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1552 'url': m3u8_url,
1553 'ext': ext,
1554 'protocol': 'm3u8',
1555 'preference': preference - 100 if preference else -100,
1556 'resolution': 'multiple',
1557 'format_note': 'Quality selection URL',
1558 }
1559
1560 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1561 entry_protocol='m3u8', preference=None,
1562 m3u8_id=None, note=None, errnote=None,
1563 fatal=True, live=False):
1564 res = self._download_webpage_handle(
1565 m3u8_url, video_id,
1566 note=note or 'Downloading m3u8 information',
1567 errnote=errnote or 'Failed to download m3u8 information',
1568 fatal=fatal)
1569
1570 if res is False:
1571 return []
1572
1573 m3u8_doc, urlh = res
1574 m3u8_url = urlh.geturl()
1575
1576 return self._parse_m3u8_formats(
1577 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1578 preference=preference, m3u8_id=m3u8_id, live=live)
1579
1580 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1581 entry_protocol='m3u8', preference=None,
1582 m3u8_id=None, live=False):
1583 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1584 return []
1585
1586 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1587 return []
1588
1589 formats = []
1590
1591 format_url = lambda u: (
1592 u
1593 if re.match(r'^https?://', u)
1594 else compat_urlparse.urljoin(m3u8_url, u))
1595
1596 # References:
1597 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1598 # 2. https://github.com/rg3/youtube-dl/issues/12211
1599
1600 # We should try extracting formats only from master playlists [1, 4.3.4],
1601 # i.e. playlists that describe available qualities. On the other hand
1602 # media playlists [1, 4.3.3] should be returned as is since they contain
1603 # just the media without qualities renditions.
1604 # Fortunately, master playlist can be easily distinguished from media
1605 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1606 # master playlist tags MUST NOT appear in a media playist and vice versa.
1607 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1608 # media playlist and MUST NOT appear in master playlist thus we can
1609 # clearly detect media playlist with this criterion.
1610
1611 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1612 return [{
1613 'url': m3u8_url,
1614 'format_id': m3u8_id,
1615 'ext': ext,
1616 'protocol': entry_protocol,
1617 'preference': preference,
1618 }]
1619
1620 groups = {}
1621 last_stream_inf = {}
1622
1623 def extract_media(x_media_line):
1624 media = parse_m3u8_attributes(x_media_line)
1625 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1626 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1627 if not (media_type and group_id and name):
1628 return
1629 groups.setdefault(group_id, []).append(media)
1630 if media_type not in ('VIDEO', 'AUDIO'):
1631 return
1632 media_url = media.get('URI')
1633 if media_url:
1634 format_id = []
1635 for v in (m3u8_id, group_id, name):
1636 if v:
1637 format_id.append(v)
1638 f = {
1639 'format_id': '-'.join(format_id),
1640 'url': format_url(media_url),
1641 'manifest_url': m3u8_url,
1642 'language': media.get('LANGUAGE'),
1643 'ext': ext,
1644 'protocol': entry_protocol,
1645 'preference': preference,
1646 }
1647 if media_type == 'AUDIO':
1648 f['vcodec'] = 'none'
1649 formats.append(f)
1650
1651 def build_stream_name():
1652 # Despite specification does not mention NAME attribute for
1653 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1654 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1655 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1656 stream_name = last_stream_inf.get('NAME')
1657 if stream_name:
1658 return stream_name
1659 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1660 # from corresponding rendition group
1661 stream_group_id = last_stream_inf.get('VIDEO')
1662 if not stream_group_id:
1663 return
1664 stream_group = groups.get(stream_group_id)
1665 if not stream_group:
1666 return stream_group_id
1667 rendition = stream_group[0]
1668 return rendition.get('NAME') or stream_group_id
1669
1670 for line in m3u8_doc.splitlines():
1671 if line.startswith('#EXT-X-STREAM-INF:'):
1672 last_stream_inf = parse_m3u8_attributes(line)
1673 elif line.startswith('#EXT-X-MEDIA:'):
1674 extract_media(line)
1675 elif line.startswith('#') or not line.strip():
1676 continue
1677 else:
1678 tbr = float_or_none(
1679 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1680 last_stream_inf.get('BANDWIDTH'), scale=1000)
1681 format_id = []
1682 if m3u8_id:
1683 format_id.append(m3u8_id)
1684 stream_name = build_stream_name()
1685 # Bandwidth of live streams may differ over time thus making
1686 # format_id unpredictable. So it's better to keep provided
1687 # format_id intact.
1688 if not live:
1689 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1690 manifest_url = format_url(line.strip())
1691 f = {
1692 'format_id': '-'.join(format_id),
1693 'url': manifest_url,
1694 'manifest_url': m3u8_url,
1695 'tbr': tbr,
1696 'ext': ext,
1697 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1698 'protocol': entry_protocol,
1699 'preference': preference,
1700 }
1701 resolution = last_stream_inf.get('RESOLUTION')
1702 if resolution:
1703 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1704 if mobj:
1705 f['width'] = int(mobj.group('width'))
1706 f['height'] = int(mobj.group('height'))
1707 # Unified Streaming Platform
1708 mobj = re.search(
1709 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1710 if mobj:
1711 abr, vbr = mobj.groups()
1712 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1713 f.update({
1714 'vbr': vbr,
1715 'abr': abr,
1716 })
1717 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1718 f.update(codecs)
1719 audio_group_id = last_stream_inf.get('AUDIO')
1720 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1721 # references a rendition group MUST have a CODECS attribute.
1722 # However, this is not always respected, for example, [2]
1723 # contains EXT-X-STREAM-INF tag which references AUDIO
1724 # rendition group but does not have CODECS and despite
1725 # referencing an audio group it represents a complete
1726 # (with audio and video) format. So, for such cases we will
1727 # ignore references to rendition groups and treat them
1728 # as complete formats.
1729 if audio_group_id and codecs and f.get('vcodec') != 'none':
1730 audio_group = groups.get(audio_group_id)
1731 if audio_group and audio_group[0].get('URI'):
1732 # TODO: update acodec for audio only formats with
1733 # the same GROUP-ID
1734 f['acodec'] = 'none'
1735 formats.append(f)
1736 last_stream_inf = {}
1737 return formats
1738
1739 @staticmethod
1740 def _xpath_ns(path, namespace=None):
1741 if not namespace:
1742 return path
1743 out = []
1744 for c in path.split('/'):
1745 if not c or c == '.':
1746 out.append(c)
1747 else:
1748 out.append('{%s}%s' % (namespace, c))
1749 return '/'.join(out)
1750
1751 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1752 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1753
1754 if smil is False:
1755 assert not fatal
1756 return []
1757
1758 namespace = self._parse_smil_namespace(smil)
1759
1760 return self._parse_smil_formats(
1761 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1762
1763 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1764 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1765 if smil is False:
1766 return {}
1767 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1768
1769 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1770 return self._download_xml(
1771 smil_url, video_id, 'Downloading SMIL file',
1772 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1773
1774 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1775 namespace = self._parse_smil_namespace(smil)
1776
1777 formats = self._parse_smil_formats(
1778 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1779 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1780
1781 video_id = os.path.splitext(url_basename(smil_url))[0]
1782 title = None
1783 description = None
1784 upload_date = None
1785 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1786 name = meta.attrib.get('name')
1787 content = meta.attrib.get('content')
1788 if not name or not content:
1789 continue
1790 if not title and name == 'title':
1791 title = content
1792 elif not description and name in ('description', 'abstract'):
1793 description = content
1794 elif not upload_date and name == 'date':
1795 upload_date = unified_strdate(content)
1796
1797 thumbnails = [{
1798 'id': image.get('type'),
1799 'url': image.get('src'),
1800 'width': int_or_none(image.get('width')),
1801 'height': int_or_none(image.get('height')),
1802 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1803
1804 return {
1805 'id': video_id,
1806 'title': title or video_id,
1807 'description': description,
1808 'upload_date': upload_date,
1809 'thumbnails': thumbnails,
1810 'formats': formats,
1811 'subtitles': subtitles,
1812 }
1813
1814 def _parse_smil_namespace(self, smil):
1815 return self._search_regex(
1816 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1817
1818 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1819 base = smil_url
1820 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1821 b = meta.get('base') or meta.get('httpBase')
1822 if b:
1823 base = b
1824 break
1825
1826 formats = []
1827 rtmp_count = 0
1828 http_count = 0
1829 m3u8_count = 0
1830
1831 srcs = []
1832 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1833 for medium in media:
1834 src = medium.get('src')
1835 if not src or src in srcs:
1836 continue
1837 srcs.append(src)
1838
1839 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1840 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1841 width = int_or_none(medium.get('width'))
1842 height = int_or_none(medium.get('height'))
1843 proto = medium.get('proto')
1844 ext = medium.get('ext')
1845 src_ext = determine_ext(src)
1846 streamer = medium.get('streamer') or base
1847
1848 if proto == 'rtmp' or streamer.startswith('rtmp'):
1849 rtmp_count += 1
1850 formats.append({
1851 'url': streamer,
1852 'play_path': src,
1853 'ext': 'flv',
1854 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1855 'tbr': bitrate,
1856 'filesize': filesize,
1857 'width': width,
1858 'height': height,
1859 })
1860 if transform_rtmp_url:
1861 streamer, src = transform_rtmp_url(streamer, src)
1862 formats[-1].update({
1863 'url': streamer,
1864 'play_path': src,
1865 })
1866 continue
1867
1868 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1869 src_url = src_url.strip()
1870
1871 if proto == 'm3u8' or src_ext == 'm3u8':
1872 m3u8_formats = self._extract_m3u8_formats(
1873 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1874 if len(m3u8_formats) == 1:
1875 m3u8_count += 1
1876 m3u8_formats[0].update({
1877 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1878 'tbr': bitrate,
1879 'width': width,
1880 'height': height,
1881 })
1882 formats.extend(m3u8_formats)
1883 elif src_ext == 'f4m':
1884 f4m_url = src_url
1885 if not f4m_params:
1886 f4m_params = {
1887 'hdcore': '3.2.0',
1888 'plugin': 'flowplayer-3.2.0.1',
1889 }
1890 f4m_url += '&' if '?' in f4m_url else '?'
1891 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1892 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1893 elif src_ext == 'mpd':
1894 formats.extend(self._extract_mpd_formats(
1895 src_url, video_id, mpd_id='dash', fatal=False))
1896 elif re.search(r'\.ism/[Mm]anifest', src_url):
1897 formats.extend(self._extract_ism_formats(
1898 src_url, video_id, ism_id='mss', fatal=False))
1899 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1900 http_count += 1
1901 formats.append({
1902 'url': src_url,
1903 'ext': ext or src_ext or 'flv',
1904 'format_id': 'http-%d' % (bitrate or http_count),
1905 'tbr': bitrate,
1906 'filesize': filesize,
1907 'width': width,
1908 'height': height,
1909 })
1910
1911 return formats
1912
1913 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1914 urls = []
1915 subtitles = {}
1916 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1917 src = textstream.get('src')
1918 if not src or src in urls:
1919 continue
1920 urls.append(src)
1921 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1922 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1923 subtitles.setdefault(lang, []).append({
1924 'url': src,
1925 'ext': ext,
1926 })
1927 return subtitles
1928
1929 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1930 xspf = self._download_xml(
1931 xspf_url, playlist_id, 'Downloading xpsf playlist',
1932 'Unable to download xspf manifest', fatal=fatal)
1933 if xspf is False:
1934 return []
1935 return self._parse_xspf(
1936 xspf, playlist_id, xspf_url=xspf_url,
1937 xspf_base_url=base_url(xspf_url))
1938
1939 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1940 NS_MAP = {
1941 'xspf': 'http://xspf.org/ns/0/',
1942 's1': 'http://static.streamone.nl/player/ns/0',
1943 }
1944
1945 entries = []
1946 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1947 title = xpath_text(
1948 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1949 description = xpath_text(
1950 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1951 thumbnail = xpath_text(
1952 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1953 duration = float_or_none(
1954 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1955
1956 formats = []
1957 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
1958 format_url = urljoin(xspf_base_url, location.text)
1959 if not format_url:
1960 continue
1961 formats.append({
1962 'url': format_url,
1963 'manifest_url': xspf_url,
1964 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1965 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1966 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1967 })
1968 self._sort_formats(formats)
1969
1970 entries.append({
1971 'id': playlist_id,
1972 'title': title,
1973 'description': description,
1974 'thumbnail': thumbnail,
1975 'duration': duration,
1976 'formats': formats,
1977 })
1978 return entries
1979
1980 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1981 res = self._download_xml_handle(
1982 mpd_url, video_id,
1983 note=note or 'Downloading MPD manifest',
1984 errnote=errnote or 'Failed to download MPD manifest',
1985 fatal=fatal)
1986 if res is False:
1987 return []
1988 mpd_doc, urlh = res
1989 mpd_base_url = base_url(urlh.geturl())
1990
1991 return self._parse_mpd_formats(
1992 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
1993 formats_dict=formats_dict, mpd_url=mpd_url)
1994
1995 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1996 """
1997 Parse formats from MPD manifest.
1998 References:
1999 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2000 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2001 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2002 """
2003 if mpd_doc.get('type') == 'dynamic':
2004 return []
2005
2006 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2007
2008 def _add_ns(path):
2009 return self._xpath_ns(path, namespace)
2010
2011 def is_drm_protected(element):
2012 return element.find(_add_ns('ContentProtection')) is not None
2013
2014 def extract_multisegment_info(element, ms_parent_info):
2015 ms_info = ms_parent_info.copy()
2016
2017 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2018 # common attributes and elements. We will only extract relevant
2019 # for us.
2020 def extract_common(source):
2021 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2022 if segment_timeline is not None:
2023 s_e = segment_timeline.findall(_add_ns('S'))
2024 if s_e:
2025 ms_info['total_number'] = 0
2026 ms_info['s'] = []
2027 for s in s_e:
2028 r = int(s.get('r', 0))
2029 ms_info['total_number'] += 1 + r
2030 ms_info['s'].append({
2031 't': int(s.get('t', 0)),
2032 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2033 'd': int(s.attrib['d']),
2034 'r': r,
2035 })
2036 start_number = source.get('startNumber')
2037 if start_number:
2038 ms_info['start_number'] = int(start_number)
2039 timescale = source.get('timescale')
2040 if timescale:
2041 ms_info['timescale'] = int(timescale)
2042 segment_duration = source.get('duration')
2043 if segment_duration:
2044 ms_info['segment_duration'] = float(segment_duration)
2045
2046 def extract_Initialization(source):
2047 initialization = source.find(_add_ns('Initialization'))
2048 if initialization is not None:
2049 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2050
2051 segment_list = element.find(_add_ns('SegmentList'))
2052 if segment_list is not None:
2053 extract_common(segment_list)
2054 extract_Initialization(segment_list)
2055 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2056 if segment_urls_e:
2057 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2058 else:
2059 segment_template = element.find(_add_ns('SegmentTemplate'))
2060 if segment_template is not None:
2061 extract_common(segment_template)
2062 media = segment_template.get('media')
2063 if media:
2064 ms_info['media'] = media
2065 initialization = segment_template.get('initialization')
2066 if initialization:
2067 ms_info['initialization'] = initialization
2068 else:
2069 extract_Initialization(segment_template)
2070 return ms_info
2071
2072 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2073 formats = []
2074 for period in mpd_doc.findall(_add_ns('Period')):
2075 period_duration = parse_duration(period.get('duration')) or mpd_duration
2076 period_ms_info = extract_multisegment_info(period, {
2077 'start_number': 1,
2078 'timescale': 1,
2079 })
2080 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2081 if is_drm_protected(adaptation_set):
2082 continue
2083 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2084 for representation in adaptation_set.findall(_add_ns('Representation')):
2085 if is_drm_protected(representation):
2086 continue
2087 representation_attrib = adaptation_set.attrib.copy()
2088 representation_attrib.update(representation.attrib)
2089 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2090 mime_type = representation_attrib['mimeType']
2091 content_type = mime_type.split('/')[0]
2092 if content_type == 'text':
2093 # TODO implement WebVTT downloading
2094 pass
2095 elif content_type in ('video', 'audio'):
2096 base_url = ''
2097 for element in (representation, adaptation_set, period, mpd_doc):
2098 base_url_e = element.find(_add_ns('BaseURL'))
2099 if base_url_e is not None:
2100 base_url = base_url_e.text + base_url
2101 if re.match(r'^https?://', base_url):
2102 break
2103 if mpd_base_url and not re.match(r'^https?://', base_url):
2104 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2105 mpd_base_url += '/'
2106 base_url = mpd_base_url + base_url
2107 representation_id = representation_attrib.get('id')
2108 lang = representation_attrib.get('lang')
2109 url_el = representation.find(_add_ns('BaseURL'))
2110 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2111 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2112 f = {
2113 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2114 'url': base_url,
2115 'manifest_url': mpd_url,
2116 'ext': mimetype2ext(mime_type),
2117 'width': int_or_none(representation_attrib.get('width')),
2118 'height': int_or_none(representation_attrib.get('height')),
2119 'tbr': float_or_none(bandwidth, 1000),
2120 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2121 'fps': int_or_none(representation_attrib.get('frameRate')),
2122 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2123 'format_note': 'DASH %s' % content_type,
2124 'filesize': filesize,
2125 'container': mimetype2ext(mime_type) + '_dash',
2126 }
2127 f.update(parse_codecs(representation_attrib.get('codecs')))
2128 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2129
2130 def prepare_template(template_name, identifiers):
2131 tmpl = representation_ms_info[template_name]
2132 # First of, % characters outside $...$ templates
2133 # must be escaped by doubling for proper processing
2134 # by % operator string formatting used further (see
2135 # https://github.com/rg3/youtube-dl/issues/16867).
2136 t = ''
2137 in_template = False
2138 for c in tmpl:
2139 t += c
2140 if c == '$':
2141 in_template = not in_template
2142 elif c == '%' and not in_template:
2143 t += c
2144 # Next, $...$ templates are translated to their
2145 # %(...) counterparts to be used with % operator
2146 t = t.replace('$RepresentationID$', representation_id)
2147 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2148 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2149 t.replace('$$', '$')
2150 return t
2151
2152 # @initialization is a regular template like @media one
2153 # so it should be handled just the same way (see
2154 # https://github.com/rg3/youtube-dl/issues/11605)
2155 if 'initialization' in representation_ms_info:
2156 initialization_template = prepare_template(
2157 'initialization',
2158 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2159 # $Time$ shall not be included for @initialization thus
2160 # only $Bandwidth$ remains
2161 ('Bandwidth', ))
2162 representation_ms_info['initialization_url'] = initialization_template % {
2163 'Bandwidth': bandwidth,
2164 }
2165
2166 def location_key(location):
2167 return 'url' if re.match(r'^https?://', location) else 'path'
2168
2169 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2170
2171 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2172 media_location_key = location_key(media_template)
2173
2174 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2175 # can't be used at the same time
2176 if '%(Number' in media_template and 's' not in representation_ms_info:
2177 segment_duration = None
2178 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2179 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2180 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2181 representation_ms_info['fragments'] = [{
2182 media_location_key: media_template % {
2183 'Number': segment_number,
2184 'Bandwidth': bandwidth,
2185 },
2186 'duration': segment_duration,
2187 } for segment_number in range(
2188 representation_ms_info['start_number'],
2189 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2190 else:
2191 # $Number*$ or $Time$ in media template with S list available
2192 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2193 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2194 representation_ms_info['fragments'] = []
2195 segment_time = 0
2196 segment_d = None
2197 segment_number = representation_ms_info['start_number']
2198
2199 def add_segment_url():
2200 segment_url = media_template % {
2201 'Time': segment_time,
2202 'Bandwidth': bandwidth,
2203 'Number': segment_number,
2204 }
2205 representation_ms_info['fragments'].append({
2206 media_location_key: segment_url,
2207 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2208 })
2209
2210 for num, s in enumerate(representation_ms_info['s']):
2211 segment_time = s.get('t') or segment_time
2212 segment_d = s['d']
2213 add_segment_url()
2214 segment_number += 1
2215 for r in range(s.get('r', 0)):
2216 segment_time += segment_d
2217 add_segment_url()
2218 segment_number += 1
2219 segment_time += segment_d
2220 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2221 # No media template
2222 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2223 # or any YouTube dashsegments video
2224 fragments = []
2225 segment_index = 0
2226 timescale = representation_ms_info['timescale']
2227 for s in representation_ms_info['s']:
2228 duration = float_or_none(s['d'], timescale)
2229 for r in range(s.get('r', 0) + 1):
2230 segment_uri = representation_ms_info['segment_urls'][segment_index]
2231 fragments.append({
2232 location_key(segment_uri): segment_uri,
2233 'duration': duration,
2234 })
2235 segment_index += 1
2236 representation_ms_info['fragments'] = fragments
2237 elif 'segment_urls' in representation_ms_info:
2238 # Segment URLs with no SegmentTimeline
2239 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2240 # https://github.com/rg3/youtube-dl/pull/14844
2241 fragments = []
2242 segment_duration = float_or_none(
2243 representation_ms_info['segment_duration'],
2244 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2245 for segment_url in representation_ms_info['segment_urls']:
2246 fragment = {
2247 location_key(segment_url): segment_url,
2248 }
2249 if segment_duration:
2250 fragment['duration'] = segment_duration
2251 fragments.append(fragment)
2252 representation_ms_info['fragments'] = fragments
2253 # NB: MPD manifest may contain direct URLs to unfragmented media.
2254 # No fragments key is present in this case.
2255 if 'fragments' in representation_ms_info:
2256 f.update({
2257 'fragment_base_url': base_url,
2258 'fragments': [],
2259 'protocol': 'http_dash_segments',
2260 })
2261 if 'initialization_url' in representation_ms_info:
2262 initialization_url = representation_ms_info['initialization_url']
2263 if not f.get('url'):
2264 f['url'] = initialization_url
2265 f['fragments'].append({location_key(initialization_url): initialization_url})
2266 f['fragments'].extend(representation_ms_info['fragments'])
2267 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2268 # is not necessarily unique within a Period thus formats with
2269 # the same `format_id` are quite possible. There are numerous examples
2270 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2271 # https://github.com/rg3/youtube-dl/issues/13919)
2272 full_info = formats_dict.get(representation_id, {}).copy()
2273 full_info.update(f)
2274 formats.append(full_info)
2275 else:
2276 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2277 return formats
2278
2279 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2280 res = self._download_xml_handle(
2281 ism_url, video_id,
2282 note=note or 'Downloading ISM manifest',
2283 errnote=errnote or 'Failed to download ISM manifest',
2284 fatal=fatal)
2285 if res is False:
2286 return []
2287 ism_doc, urlh = res
2288
2289 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2290
2291 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2292 """
2293 Parse formats from ISM manifest.
2294 References:
2295 1. [MS-SSTR]: Smooth Streaming Protocol,
2296 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2297 """
2298 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2299 return []
2300
2301 duration = int(ism_doc.attrib['Duration'])
2302 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2303
2304 formats = []
2305 for stream in ism_doc.findall('StreamIndex'):
2306 stream_type = stream.get('Type')
2307 if stream_type not in ('video', 'audio'):
2308 continue
2309 url_pattern = stream.attrib['Url']
2310 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2311 stream_name = stream.get('Name')
2312 for track in stream.findall('QualityLevel'):
2313 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2314 # TODO: add support for WVC1 and WMAP
2315 if fourcc not in ('H264', 'AVC1', 'AACL'):
2316 self.report_warning('%s is not a supported codec' % fourcc)
2317 continue
2318 tbr = int(track.attrib['Bitrate']) // 1000
2319 # [1] does not mention Width and Height attributes. However,
2320 # they're often present while MaxWidth and MaxHeight are
2321 # missing, so should be used as fallbacks
2322 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2323 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2324 sampling_rate = int_or_none(track.get('SamplingRate'))
2325
2326 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2327 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2328
2329 fragments = []
2330 fragment_ctx = {
2331 'time': 0,
2332 }
2333 stream_fragments = stream.findall('c')
2334 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2335 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2336 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2337 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2338 if not fragment_ctx['duration']:
2339 try:
2340 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2341 except IndexError:
2342 next_fragment_time = duration
2343 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2344 for _ in range(fragment_repeat):
2345 fragments.append({
2346 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2347 'duration': fragment_ctx['duration'] / stream_timescale,
2348 })
2349 fragment_ctx['time'] += fragment_ctx['duration']
2350
2351 format_id = []
2352 if ism_id:
2353 format_id.append(ism_id)
2354 if stream_name:
2355 format_id.append(stream_name)
2356 format_id.append(compat_str(tbr))
2357
2358 formats.append({
2359 'format_id': '-'.join(format_id),
2360 'url': ism_url,
2361 'manifest_url': ism_url,
2362 'ext': 'ismv' if stream_type == 'video' else 'isma',
2363 'width': width,
2364 'height': height,
2365 'tbr': tbr,
2366 'asr': sampling_rate,
2367 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2368 'acodec': 'none' if stream_type == 'video' else fourcc,
2369 'protocol': 'ism',
2370 'fragments': fragments,
2371 '_download_params': {
2372 'duration': duration,
2373 'timescale': stream_timescale,
2374 'width': width or 0,
2375 'height': height or 0,
2376 'fourcc': fourcc,
2377 'codec_private_data': track.get('CodecPrivateData'),
2378 'sampling_rate': sampling_rate,
2379 'channels': int_or_none(track.get('Channels', 2)),
2380 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2381 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2382 },
2383 })
2384 return formats
2385
2386 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2387 def absolute_url(item_url):
2388 return urljoin(base_url, item_url)
2389
2390 def parse_content_type(content_type):
2391 if not content_type:
2392 return {}
2393 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2394 if ctr:
2395 mimetype, codecs = ctr.groups()
2396 f = parse_codecs(codecs)
2397 f['ext'] = mimetype2ext(mimetype)
2398 return f
2399 return {}
2400
2401 def _media_formats(src, cur_media_type, type_info={}):
2402 full_url = absolute_url(src)
2403 ext = type_info.get('ext') or determine_ext(full_url)
2404 if ext == 'm3u8':
2405 is_plain_url = False
2406 formats = self._extract_m3u8_formats(
2407 full_url, video_id, ext='mp4',
2408 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2409 preference=preference, fatal=False)
2410 elif ext == 'mpd':
2411 is_plain_url = False
2412 formats = self._extract_mpd_formats(
2413 full_url, video_id, mpd_id=mpd_id, fatal=False)
2414 else:
2415 is_plain_url = True
2416 formats = [{
2417 'url': full_url,
2418 'vcodec': 'none' if cur_media_type == 'audio' else None,
2419 }]
2420 return is_plain_url, formats
2421
2422 entries = []
2423 # amp-video and amp-audio are very similar to their HTML5 counterparts
2424 # so we wll include them right here (see
2425 # https://www.ampproject.org/docs/reference/components/amp-video)
2426 media_tags = [(media_tag, media_type, '')
2427 for media_tag, media_type
2428 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2429 media_tags.extend(re.findall(
2430 # We only allow video|audio followed by a whitespace or '>'.
2431 # Allowing more characters may end up in significant slow down (see
2432 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2433 # http://www.porntrex.com/maps/videositemap.xml).
2434 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2435 for media_tag, media_type, media_content in media_tags:
2436 media_info = {
2437 'formats': [],
2438 'subtitles': {},
2439 }
2440 media_attributes = extract_attributes(media_tag)
2441 src = media_attributes.get('src')
2442 if src:
2443 _, formats = _media_formats(src, media_type)
2444 media_info['formats'].extend(formats)
2445 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2446 if media_content:
2447 for source_tag in re.findall(r'<source[^>]+>', media_content):
2448 source_attributes = extract_attributes(source_tag)
2449 src = source_attributes.get('src')
2450 if not src:
2451 continue
2452 f = parse_content_type(source_attributes.get('type'))
2453 is_plain_url, formats = _media_formats(src, media_type, f)
2454 if is_plain_url:
2455 # res attribute is not standard but seen several times
2456 # in the wild
2457 f.update({
2458 'height': int_or_none(source_attributes.get('res')),
2459 'format_id': source_attributes.get('label'),
2460 })
2461 f.update(formats[0])
2462 media_info['formats'].append(f)
2463 else:
2464 media_info['formats'].extend(formats)
2465 for track_tag in re.findall(r'<track[^>]+>', media_content):
2466 track_attributes = extract_attributes(track_tag)
2467 kind = track_attributes.get('kind')
2468 if not kind or kind in ('subtitles', 'captions'):
2469 src = track_attributes.get('src')
2470 if not src:
2471 continue
2472 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2473 media_info['subtitles'].setdefault(lang, []).append({
2474 'url': absolute_url(src),
2475 })
2476 for f in media_info['formats']:
2477 f.setdefault('http_headers', {})['Referer'] = base_url
2478 if media_info['formats'] or media_info['subtitles']:
2479 entries.append(media_info)
2480 return entries
2481
2482 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2483 formats = []
2484 hdcore_sign = 'hdcore=3.7.0'
2485 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2486 hds_host = hosts.get('hds')
2487 if hds_host:
2488 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2489 if 'hdcore=' not in f4m_url:
2490 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2491 f4m_formats = self._extract_f4m_formats(
2492 f4m_url, video_id, f4m_id='hds', fatal=False)
2493 for entry in f4m_formats:
2494 entry.update({'extra_param_to_segment_url': hdcore_sign})
2495 formats.extend(f4m_formats)
2496 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2497 hls_host = hosts.get('hls')
2498 if hls_host:
2499 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2500 formats.extend(self._extract_m3u8_formats(
2501 m3u8_url, video_id, 'mp4', 'm3u8_native',
2502 m3u8_id='hls', fatal=False))
2503 return formats
2504
2505 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2506 query = compat_urlparse.urlparse(url).query
2507 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2508 mobj = re.search(
2509 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2510 url_base = mobj.group('url')
2511 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2512 formats = []
2513
2514 def manifest_url(manifest):
2515 m_url = '%s/%s' % (http_base_url, manifest)
2516 if query:
2517 m_url += '?%s' % query
2518 return m_url
2519
2520 if 'm3u8' not in skip_protocols:
2521 formats.extend(self._extract_m3u8_formats(
2522 manifest_url('playlist.m3u8'), video_id, 'mp4',
2523 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2524 if 'f4m' not in skip_protocols:
2525 formats.extend(self._extract_f4m_formats(
2526 manifest_url('manifest.f4m'),
2527 video_id, f4m_id='hds', fatal=False))
2528 if 'dash' not in skip_protocols:
2529 formats.extend(self._extract_mpd_formats(
2530 manifest_url('manifest.mpd'),
2531 video_id, mpd_id='dash', fatal=False))
2532 if re.search(r'(?:/smil:|\.smil)', url_base):
2533 if 'smil' not in skip_protocols:
2534 rtmp_formats = self._extract_smil_formats(
2535 manifest_url('jwplayer.smil'),
2536 video_id, fatal=False)
2537 for rtmp_format in rtmp_formats:
2538 rtsp_format = rtmp_format.copy()
2539 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2540 del rtsp_format['play_path']
2541 del rtsp_format['ext']
2542 rtsp_format.update({
2543 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2544 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2545 'protocol': 'rtsp',
2546 })
2547 formats.extend([rtmp_format, rtsp_format])
2548 else:
2549 for protocol in ('rtmp', 'rtsp'):
2550 if protocol not in skip_protocols:
2551 formats.append({
2552 'url': '%s:%s' % (protocol, url_base),
2553 'format_id': protocol,
2554 'protocol': protocol,
2555 })
2556 return formats
2557
2558 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2559 mobj = re.search(
2560 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2561 webpage)
2562 if mobj:
2563 try:
2564 jwplayer_data = self._parse_json(mobj.group('options'),
2565 video_id=video_id,
2566 transform_source=transform_source)
2567 except ExtractorError:
2568 pass
2569 else:
2570 if isinstance(jwplayer_data, dict):
2571 return jwplayer_data
2572
2573 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2574 jwplayer_data = self._find_jwplayer_data(
2575 webpage, video_id, transform_source=js_to_json)
2576 return self._parse_jwplayer_data(
2577 jwplayer_data, video_id, *args, **kwargs)
2578
2579 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2580 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2581 # JWPlayer backward compatibility: flattened playlists
2582 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2583 if 'playlist' not in jwplayer_data:
2584 jwplayer_data = {'playlist': [jwplayer_data]}
2585
2586 entries = []
2587
2588 # JWPlayer backward compatibility: single playlist item
2589 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2590 if not isinstance(jwplayer_data['playlist'], list):
2591 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2592
2593 for video_data in jwplayer_data['playlist']:
2594 # JWPlayer backward compatibility: flattened sources
2595 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2596 if 'sources' not in video_data:
2597 video_data['sources'] = [video_data]
2598
2599 this_video_id = video_id or video_data['mediaid']
2600
2601 formats = self._parse_jwplayer_formats(
2602 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2603 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2604
2605 subtitles = {}
2606 tracks = video_data.get('tracks')
2607 if tracks and isinstance(tracks, list):
2608 for track in tracks:
2609 if not isinstance(track, dict):
2610 continue
2611 track_kind = track.get('kind')
2612 if not track_kind or not isinstance(track_kind, compat_str):
2613 continue
2614 if track_kind.lower() not in ('captions', 'subtitles'):
2615 continue
2616 track_url = urljoin(base_url, track.get('file'))
2617 if not track_url:
2618 continue
2619 subtitles.setdefault(track.get('label') or 'en', []).append({
2620 'url': self._proto_relative_url(track_url)
2621 })
2622
2623 entry = {
2624 'id': this_video_id,
2625 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2626 'description': video_data.get('description'),
2627 'thumbnail': self._proto_relative_url(video_data.get('image')),
2628 'timestamp': int_or_none(video_data.get('pubdate')),
2629 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2630 'subtitles': subtitles,
2631 }
2632 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2633 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2634 entry.update({
2635 '_type': 'url_transparent',
2636 'url': formats[0]['url'],
2637 })
2638 else:
2639 self._sort_formats(formats)
2640 entry['formats'] = formats
2641 entries.append(entry)
2642 if len(entries) == 1:
2643 return entries[0]
2644 else:
2645 return self.playlist_result(entries)
2646
2647 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2648 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2649 urls = []
2650 formats = []
2651 for source in jwplayer_sources_data:
2652 if not isinstance(source, dict):
2653 continue
2654 source_url = self._proto_relative_url(source.get('file'))
2655 if not source_url:
2656 continue
2657 if base_url:
2658 source_url = compat_urlparse.urljoin(base_url, source_url)
2659 if source_url in urls:
2660 continue
2661 urls.append(source_url)
2662 source_type = source.get('type') or ''
2663 ext = mimetype2ext(source_type) or determine_ext(source_url)
2664 if source_type == 'hls' or ext == 'm3u8':
2665 formats.extend(self._extract_m3u8_formats(
2666 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2667 m3u8_id=m3u8_id, fatal=False))
2668 elif source_type == 'dash' or ext == 'mpd':
2669 formats.extend(self._extract_mpd_formats(
2670 source_url, video_id, mpd_id=mpd_id, fatal=False))
2671 elif ext == 'smil':
2672 formats.extend(self._extract_smil_formats(
2673 source_url, video_id, fatal=False))
2674 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2675 elif source_type.startswith('audio') or ext in (
2676 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2677 formats.append({
2678 'url': source_url,
2679 'vcodec': 'none',
2680 'ext': ext,
2681 })
2682 else:
2683 height = int_or_none(source.get('height'))
2684 if height is None:
2685 # Often no height is provided but there is a label in
2686 # format like "1080p", "720p SD", or 1080.
2687 height = int_or_none(self._search_regex(
2688 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2689 'height', default=None))
2690 a_format = {
2691 'url': source_url,
2692 'width': int_or_none(source.get('width')),
2693 'height': height,
2694 'tbr': int_or_none(source.get('bitrate')),
2695 'ext': ext,
2696 }
2697 if source_url.startswith('rtmp'):
2698 a_format['ext'] = 'flv'
2699 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2700 # of jwplayer.flash.swf
2701 rtmp_url_parts = re.split(
2702 r'((?:mp4|mp3|flv):)', source_url, 1)
2703 if len(rtmp_url_parts) == 3:
2704 rtmp_url, prefix, play_path = rtmp_url_parts
2705 a_format.update({
2706 'url': rtmp_url,
2707 'play_path': prefix + play_path,
2708 })
2709 if rtmp_params:
2710 a_format.update(rtmp_params)
2711 formats.append(a_format)
2712 return formats
2713
2714 def _live_title(self, name):
2715 """ Generate the title for a live video """
2716 now = datetime.datetime.now()
2717 now_str = now.strftime('%Y-%m-%d %H:%M')
2718 return name + ' ' + now_str
2719
2720 def _int(self, v, name, fatal=False, **kwargs):
2721 res = int_or_none(v, **kwargs)
2722 if 'get_attr' in kwargs:
2723 print(getattr(v, kwargs['get_attr']))
2724 if res is None:
2725 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2726 if fatal:
2727 raise ExtractorError(msg)
2728 else:
2729 self._downloader.report_warning(msg)
2730 return res
2731
2732 def _float(self, v, name, fatal=False, **kwargs):
2733 res = float_or_none(v, **kwargs)
2734 if res is None:
2735 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2736 if fatal:
2737 raise ExtractorError(msg)
2738 else:
2739 self._downloader.report_warning(msg)
2740 return res
2741
2742 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2743 path='/', secure=False, discard=False, rest={}, **kwargs):
2744 cookie = compat_cookiejar.Cookie(
2745 0, name, value, port, port is not None, domain, True,
2746 domain.startswith('.'), path, True, secure, expire_time,
2747 discard, None, None, rest)
2748 self._downloader.cookiejar.set_cookie(cookie)
2749
2750 def _get_cookies(self, url):
2751 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2752 req = sanitized_Request(url)
2753 self._downloader.cookiejar.add_cookie_header(req)
2754 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2755
2756 def get_testcases(self, include_onlymatching=False):
2757 t = getattr(self, '_TEST', None)
2758 if t:
2759 assert not hasattr(self, '_TESTS'), \
2760 '%s has _TEST and _TESTS' % type(self).__name__
2761 tests = [t]
2762 else:
2763 tests = getattr(self, '_TESTS', [])
2764 for t in tests:
2765 if not include_onlymatching and t.get('only_matching', False):
2766 continue
2767 t['name'] = type(self).__name__[:-len('IE')]
2768 yield t
2769
2770 def is_suitable(self, age_limit):
2771 """ Test whether the extractor is generally suitable for the given
2772 age limit (i.e. pornographic sites are not, all others usually are) """
2773
2774 any_restricted = False
2775 for tc in self.get_testcases(include_onlymatching=False):
2776 if tc.get('playlist', []):
2777 tc = tc['playlist'][0]
2778 is_restricted = age_restricted(
2779 tc.get('info_dict', {}).get('age_limit'), age_limit)
2780 if not is_restricted:
2781 return True
2782 any_restricted = any_restricted or is_restricted
2783 return not any_restricted
2784
2785 def extract_subtitles(self, *args, **kwargs):
2786 if (self._downloader.params.get('writesubtitles', False) or
2787 self._downloader.params.get('listsubtitles')):
2788 return self._get_subtitles(*args, **kwargs)
2789 return {}
2790
2791 def _get_subtitles(self, *args, **kwargs):
2792 raise NotImplementedError('This method must be implemented by subclasses')
2793
2794 @staticmethod
2795 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2796 """ Merge subtitle items for one language. Items with duplicated URLs
2797 will be dropped. """
2798 list1_urls = set([item['url'] for item in subtitle_list1])
2799 ret = list(subtitle_list1)
2800 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2801 return ret
2802
2803 @classmethod
2804 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2805 """ Merge two subtitle dictionaries, language by language. """
2806 ret = dict(subtitle_dict1)
2807 for lang in subtitle_dict2:
2808 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2809 return ret
2810
2811 def extract_automatic_captions(self, *args, **kwargs):
2812 if (self._downloader.params.get('writeautomaticsub', False) or
2813 self._downloader.params.get('listsubtitles')):
2814 return self._get_automatic_captions(*args, **kwargs)
2815 return {}
2816
2817 def _get_automatic_captions(self, *args, **kwargs):
2818 raise NotImplementedError('This method must be implemented by subclasses')
2819
2820 def mark_watched(self, *args, **kwargs):
2821 if (self._downloader.params.get('mark_watched', False) and
2822 (self._get_login_info()[0] is not None or
2823 self._downloader.params.get('cookiefile') is not None)):
2824 self._mark_watched(*args, **kwargs)
2825
2826 def _mark_watched(self, *args, **kwargs):
2827 raise NotImplementedError('This method must be implemented by subclasses')
2828
2829 def geo_verification_headers(self):
2830 headers = {}
2831 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2832 if geo_verification_proxy:
2833 headers['Ytdl-request-proxy'] = geo_verification_proxy
2834 return headers
2835
2836 def _generic_id(self, url):
2837 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2838
2839 def _generic_title(self, url):
2840 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2841
2842
2843 class SearchInfoExtractor(InfoExtractor):
2844 """
2845 Base class for paged search queries extractors.
2846 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2847 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2848 """
2849
2850 @classmethod
2851 def _make_valid_url(cls):
2852 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2853
2854 @classmethod
2855 def suitable(cls, url):
2856 return re.match(cls._make_valid_url(), url) is not None
2857
2858 def _real_extract(self, query):
2859 mobj = re.match(self._make_valid_url(), query)
2860 if mobj is None:
2861 raise ExtractorError('Invalid search query "%s"' % query)
2862
2863 prefix = mobj.group('prefix')
2864 query = mobj.group('query')
2865 if prefix == '':
2866 return self._get_n_results(query, 1)
2867 elif prefix == 'all':
2868 return self._get_n_results(query, self._MAX_RESULTS)
2869 else:
2870 n = int(prefix)
2871 if n <= 0:
2872 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2873 elif n > self._MAX_RESULTS:
2874 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2875 n = self._MAX_RESULTS
2876 return self._get_n_results(query, n)
2877
2878 def _get_n_results(self, query, n):
2879 """Get a specified number of results for a query"""
2880 raise NotImplementedError('This method must be implemented by subclasses')
2881
2882 @property
2883 def SEARCH_KEY(self):
2884 return self._SEARCH_KEY