]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
Update changelog.
[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 info.update({
1243 'episode': unescapeHTML(e.get('name')),
1244 'episode_number': int_or_none(e.get('episodeNumber')),
1245 'description': unescapeHTML(e.get('description')),
1246 })
1247 part_of_season = e.get('partOfSeason')
1248 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1249 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1250 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1251 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1252 info['series'] = unescapeHTML(part_of_series.get('name'))
1253 elif item_type in ('Article', 'NewsArticle'):
1254 info.update({
1255 'timestamp': parse_iso8601(e.get('datePublished')),
1256 'title': unescapeHTML(e.get('headline')),
1257 'description': unescapeHTML(e.get('articleBody')),
1258 })
1259 elif item_type == 'VideoObject':
1260 extract_video_object(e)
1261 continue
1262 video = e.get('video')
1263 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1264 extract_video_object(video)
1265 break
1266 return dict((k, v) for k, v in info.items() if v is not None)
1267
1268 @staticmethod
1269 def _hidden_inputs(html):
1270 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1271 hidden_inputs = {}
1272 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1273 attrs = extract_attributes(input)
1274 if not input:
1275 continue
1276 if attrs.get('type') not in ('hidden', 'submit'):
1277 continue
1278 name = attrs.get('name') or attrs.get('id')
1279 value = attrs.get('value')
1280 if name and value is not None:
1281 hidden_inputs[name] = value
1282 return hidden_inputs
1283
1284 def _form_hidden_inputs(self, form_id, html):
1285 form = self._search_regex(
1286 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1287 html, '%s form' % form_id, group='form')
1288 return self._hidden_inputs(form)
1289
1290 def _sort_formats(self, formats, field_preference=None):
1291 if not formats:
1292 raise ExtractorError('No video formats found')
1293
1294 for f in formats:
1295 # Automatically determine tbr when missing based on abr and vbr (improves
1296 # formats sorting in some cases)
1297 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1298 f['tbr'] = f['abr'] + f['vbr']
1299
1300 def _formats_key(f):
1301 # TODO remove the following workaround
1302 from ..utils import determine_ext
1303 if not f.get('ext') and 'url' in f:
1304 f['ext'] = determine_ext(f['url'])
1305
1306 if isinstance(field_preference, (list, tuple)):
1307 return tuple(
1308 f.get(field)
1309 if f.get(field) is not None
1310 else ('' if field == 'format_id' else -1)
1311 for field in field_preference)
1312
1313 preference = f.get('preference')
1314 if preference is None:
1315 preference = 0
1316 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1317 preference -= 0.5
1318
1319 protocol = f.get('protocol') or determine_protocol(f)
1320 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1321
1322 if f.get('vcodec') == 'none': # audio only
1323 preference -= 50
1324 if self._downloader.params.get('prefer_free_formats'):
1325 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1326 else:
1327 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1328 ext_preference = 0
1329 try:
1330 audio_ext_preference = ORDER.index(f['ext'])
1331 except ValueError:
1332 audio_ext_preference = -1
1333 else:
1334 if f.get('acodec') == 'none': # video only
1335 preference -= 40
1336 if self._downloader.params.get('prefer_free_formats'):
1337 ORDER = ['flv', 'mp4', 'webm']
1338 else:
1339 ORDER = ['webm', 'flv', 'mp4']
1340 try:
1341 ext_preference = ORDER.index(f['ext'])
1342 except ValueError:
1343 ext_preference = -1
1344 audio_ext_preference = 0
1345
1346 return (
1347 preference,
1348 f.get('language_preference') if f.get('language_preference') is not None else -1,
1349 f.get('quality') if f.get('quality') is not None else -1,
1350 f.get('tbr') if f.get('tbr') is not None else -1,
1351 f.get('filesize') if f.get('filesize') is not None else -1,
1352 f.get('vbr') if f.get('vbr') is not None else -1,
1353 f.get('height') if f.get('height') is not None else -1,
1354 f.get('width') if f.get('width') is not None else -1,
1355 proto_preference,
1356 ext_preference,
1357 f.get('abr') if f.get('abr') is not None else -1,
1358 audio_ext_preference,
1359 f.get('fps') if f.get('fps') is not None else -1,
1360 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1361 f.get('source_preference') if f.get('source_preference') is not None else -1,
1362 f.get('format_id') if f.get('format_id') is not None else '',
1363 )
1364 formats.sort(key=_formats_key)
1365
1366 def _check_formats(self, formats, video_id):
1367 if formats:
1368 formats[:] = filter(
1369 lambda f: self._is_valid_url(
1370 f['url'], video_id,
1371 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1372 formats)
1373
1374 @staticmethod
1375 def _remove_duplicate_formats(formats):
1376 format_urls = set()
1377 unique_formats = []
1378 for f in formats:
1379 if f['url'] not in format_urls:
1380 format_urls.add(f['url'])
1381 unique_formats.append(f)
1382 formats[:] = unique_formats
1383
1384 def _is_valid_url(self, url, video_id, item='video', headers={}):
1385 url = self._proto_relative_url(url, scheme='http:')
1386 # For now assume non HTTP(S) URLs always valid
1387 if not (url.startswith('http://') or url.startswith('https://')):
1388 return True
1389 try:
1390 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1391 return True
1392 except ExtractorError as e:
1393 if isinstance(e.cause, compat_urllib_error.URLError):
1394 self.to_screen(
1395 '%s: %s URL is invalid, skipping' % (video_id, item))
1396 return False
1397 raise
1398
1399 def http_scheme(self):
1400 """ Either "http:" or "https:", depending on the user's preferences """
1401 return (
1402 'http:'
1403 if self._downloader.params.get('prefer_insecure', False)
1404 else 'https:')
1405
1406 def _proto_relative_url(self, url, scheme=None):
1407 if url is None:
1408 return url
1409 if url.startswith('//'):
1410 if scheme is None:
1411 scheme = self.http_scheme()
1412 return scheme + url
1413 else:
1414 return url
1415
1416 def _sleep(self, timeout, video_id, msg_template=None):
1417 if msg_template is None:
1418 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1419 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1420 self.to_screen(msg)
1421 time.sleep(timeout)
1422
1423 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1424 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1425 fatal=True, m3u8_id=None):
1426 manifest = self._download_xml(
1427 manifest_url, video_id, 'Downloading f4m manifest',
1428 'Unable to download f4m manifest',
1429 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1430 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1431 transform_source=transform_source,
1432 fatal=fatal)
1433
1434 if manifest is False:
1435 return []
1436
1437 return self._parse_f4m_formats(
1438 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1439 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1440
1441 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1442 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1443 fatal=True, m3u8_id=None):
1444 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1445 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1446 if akamai_pv is not None and ';' in akamai_pv.text:
1447 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1448 if playerVerificationChallenge.strip() != '':
1449 return []
1450
1451 formats = []
1452 manifest_version = '1.0'
1453 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1454 if not media_nodes:
1455 manifest_version = '2.0'
1456 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1457 # Remove unsupported DRM protected media from final formats
1458 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1459 media_nodes = remove_encrypted_media(media_nodes)
1460 if not media_nodes:
1461 return formats
1462
1463 manifest_base_url = get_base_url(manifest)
1464
1465 bootstrap_info = xpath_element(
1466 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1467 'bootstrap info', default=None)
1468
1469 vcodec = None
1470 mime_type = xpath_text(
1471 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1472 'base URL', default=None)
1473 if mime_type and mime_type.startswith('audio/'):
1474 vcodec = 'none'
1475
1476 for i, media_el in enumerate(media_nodes):
1477 tbr = int_or_none(media_el.attrib.get('bitrate'))
1478 width = int_or_none(media_el.attrib.get('width'))
1479 height = int_or_none(media_el.attrib.get('height'))
1480 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1481 # If <bootstrapInfo> is present, the specified f4m is a
1482 # stream-level manifest, and only set-level manifests may refer to
1483 # external resources. See section 11.4 and section 4 of F4M spec
1484 if bootstrap_info is None:
1485 media_url = None
1486 # @href is introduced in 2.0, see section 11.6 of F4M spec
1487 if manifest_version == '2.0':
1488 media_url = media_el.attrib.get('href')
1489 if media_url is None:
1490 media_url = media_el.attrib.get('url')
1491 if not media_url:
1492 continue
1493 manifest_url = (
1494 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1495 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1496 # If media_url is itself a f4m manifest do the recursive extraction
1497 # since bitrates in parent manifest (this one) and media_url manifest
1498 # may differ leading to inability to resolve the format by requested
1499 # bitrate in f4m downloader
1500 ext = determine_ext(manifest_url)
1501 if ext == 'f4m':
1502 f4m_formats = self._extract_f4m_formats(
1503 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1504 transform_source=transform_source, fatal=fatal)
1505 # Sometimes stream-level manifest contains single media entry that
1506 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1507 # At the same time parent's media entry in set-level manifest may
1508 # contain it. We will copy it from parent in such cases.
1509 if len(f4m_formats) == 1:
1510 f = f4m_formats[0]
1511 f.update({
1512 'tbr': f.get('tbr') or tbr,
1513 'width': f.get('width') or width,
1514 'height': f.get('height') or height,
1515 'format_id': f.get('format_id') if not tbr else format_id,
1516 'vcodec': vcodec,
1517 })
1518 formats.extend(f4m_formats)
1519 continue
1520 elif ext == 'm3u8':
1521 formats.extend(self._extract_m3u8_formats(
1522 manifest_url, video_id, 'mp4', preference=preference,
1523 m3u8_id=m3u8_id, fatal=fatal))
1524 continue
1525 formats.append({
1526 'format_id': format_id,
1527 'url': manifest_url,
1528 'manifest_url': manifest_url,
1529 'ext': 'flv' if bootstrap_info is not None else None,
1530 'protocol': 'f4m',
1531 'tbr': tbr,
1532 'width': width,
1533 'height': height,
1534 'vcodec': vcodec,
1535 'preference': preference,
1536 })
1537 return formats
1538
1539 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1540 return {
1541 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1542 'url': m3u8_url,
1543 'ext': ext,
1544 'protocol': 'm3u8',
1545 'preference': preference - 100 if preference else -100,
1546 'resolution': 'multiple',
1547 'format_note': 'Quality selection URL',
1548 }
1549
1550 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1551 entry_protocol='m3u8', preference=None,
1552 m3u8_id=None, note=None, errnote=None,
1553 fatal=True, live=False):
1554 res = self._download_webpage_handle(
1555 m3u8_url, video_id,
1556 note=note or 'Downloading m3u8 information',
1557 errnote=errnote or 'Failed to download m3u8 information',
1558 fatal=fatal)
1559
1560 if res is False:
1561 return []
1562
1563 m3u8_doc, urlh = res
1564 m3u8_url = urlh.geturl()
1565
1566 return self._parse_m3u8_formats(
1567 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1568 preference=preference, m3u8_id=m3u8_id, live=live)
1569
1570 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1571 entry_protocol='m3u8', preference=None,
1572 m3u8_id=None, live=False):
1573 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1574 return []
1575
1576 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1577 return []
1578
1579 formats = []
1580
1581 format_url = lambda u: (
1582 u
1583 if re.match(r'^https?://', u)
1584 else compat_urlparse.urljoin(m3u8_url, u))
1585
1586 # References:
1587 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1588 # 2. https://github.com/rg3/youtube-dl/issues/12211
1589
1590 # We should try extracting formats only from master playlists [1, 4.3.4],
1591 # i.e. playlists that describe available qualities. On the other hand
1592 # media playlists [1, 4.3.3] should be returned as is since they contain
1593 # just the media without qualities renditions.
1594 # Fortunately, master playlist can be easily distinguished from media
1595 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1596 # master playlist tags MUST NOT appear in a media playist and vice versa.
1597 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1598 # media playlist and MUST NOT appear in master playlist thus we can
1599 # clearly detect media playlist with this criterion.
1600
1601 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1602 return [{
1603 'url': m3u8_url,
1604 'format_id': m3u8_id,
1605 'ext': ext,
1606 'protocol': entry_protocol,
1607 'preference': preference,
1608 }]
1609
1610 groups = {}
1611 last_stream_inf = {}
1612
1613 def extract_media(x_media_line):
1614 media = parse_m3u8_attributes(x_media_line)
1615 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1616 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1617 if not (media_type and group_id and name):
1618 return
1619 groups.setdefault(group_id, []).append(media)
1620 if media_type not in ('VIDEO', 'AUDIO'):
1621 return
1622 media_url = media.get('URI')
1623 if media_url:
1624 format_id = []
1625 for v in (m3u8_id, group_id, name):
1626 if v:
1627 format_id.append(v)
1628 f = {
1629 'format_id': '-'.join(format_id),
1630 'url': format_url(media_url),
1631 'manifest_url': m3u8_url,
1632 'language': media.get('LANGUAGE'),
1633 'ext': ext,
1634 'protocol': entry_protocol,
1635 'preference': preference,
1636 }
1637 if media_type == 'AUDIO':
1638 f['vcodec'] = 'none'
1639 formats.append(f)
1640
1641 def build_stream_name():
1642 # Despite specification does not mention NAME attribute for
1643 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1644 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1645 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1646 stream_name = last_stream_inf.get('NAME')
1647 if stream_name:
1648 return stream_name
1649 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1650 # from corresponding rendition group
1651 stream_group_id = last_stream_inf.get('VIDEO')
1652 if not stream_group_id:
1653 return
1654 stream_group = groups.get(stream_group_id)
1655 if not stream_group:
1656 return stream_group_id
1657 rendition = stream_group[0]
1658 return rendition.get('NAME') or stream_group_id
1659
1660 for line in m3u8_doc.splitlines():
1661 if line.startswith('#EXT-X-STREAM-INF:'):
1662 last_stream_inf = parse_m3u8_attributes(line)
1663 elif line.startswith('#EXT-X-MEDIA:'):
1664 extract_media(line)
1665 elif line.startswith('#') or not line.strip():
1666 continue
1667 else:
1668 tbr = float_or_none(
1669 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1670 last_stream_inf.get('BANDWIDTH'), scale=1000)
1671 format_id = []
1672 if m3u8_id:
1673 format_id.append(m3u8_id)
1674 stream_name = build_stream_name()
1675 # Bandwidth of live streams may differ over time thus making
1676 # format_id unpredictable. So it's better to keep provided
1677 # format_id intact.
1678 if not live:
1679 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1680 manifest_url = format_url(line.strip())
1681 f = {
1682 'format_id': '-'.join(format_id),
1683 'url': manifest_url,
1684 'manifest_url': m3u8_url,
1685 'tbr': tbr,
1686 'ext': ext,
1687 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1688 'protocol': entry_protocol,
1689 'preference': preference,
1690 }
1691 resolution = last_stream_inf.get('RESOLUTION')
1692 if resolution:
1693 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1694 if mobj:
1695 f['width'] = int(mobj.group('width'))
1696 f['height'] = int(mobj.group('height'))
1697 # Unified Streaming Platform
1698 mobj = re.search(
1699 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1700 if mobj:
1701 abr, vbr = mobj.groups()
1702 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1703 f.update({
1704 'vbr': vbr,
1705 'abr': abr,
1706 })
1707 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1708 f.update(codecs)
1709 audio_group_id = last_stream_inf.get('AUDIO')
1710 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1711 # references a rendition group MUST have a CODECS attribute.
1712 # However, this is not always respected, for example, [2]
1713 # contains EXT-X-STREAM-INF tag which references AUDIO
1714 # rendition group but does not have CODECS and despite
1715 # referencing an audio group it represents a complete
1716 # (with audio and video) format. So, for such cases we will
1717 # ignore references to rendition groups and treat them
1718 # as complete formats.
1719 if audio_group_id and codecs and f.get('vcodec') != 'none':
1720 audio_group = groups.get(audio_group_id)
1721 if audio_group and audio_group[0].get('URI'):
1722 # TODO: update acodec for audio only formats with
1723 # the same GROUP-ID
1724 f['acodec'] = 'none'
1725 formats.append(f)
1726 last_stream_inf = {}
1727 return formats
1728
1729 @staticmethod
1730 def _xpath_ns(path, namespace=None):
1731 if not namespace:
1732 return path
1733 out = []
1734 for c in path.split('/'):
1735 if not c or c == '.':
1736 out.append(c)
1737 else:
1738 out.append('{%s}%s' % (namespace, c))
1739 return '/'.join(out)
1740
1741 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1742 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1743
1744 if smil is False:
1745 assert not fatal
1746 return []
1747
1748 namespace = self._parse_smil_namespace(smil)
1749
1750 return self._parse_smil_formats(
1751 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1752
1753 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1754 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1755 if smil is False:
1756 return {}
1757 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1758
1759 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1760 return self._download_xml(
1761 smil_url, video_id, 'Downloading SMIL file',
1762 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1763
1764 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1765 namespace = self._parse_smil_namespace(smil)
1766
1767 formats = self._parse_smil_formats(
1768 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1769 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1770
1771 video_id = os.path.splitext(url_basename(smil_url))[0]
1772 title = None
1773 description = None
1774 upload_date = None
1775 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1776 name = meta.attrib.get('name')
1777 content = meta.attrib.get('content')
1778 if not name or not content:
1779 continue
1780 if not title and name == 'title':
1781 title = content
1782 elif not description and name in ('description', 'abstract'):
1783 description = content
1784 elif not upload_date and name == 'date':
1785 upload_date = unified_strdate(content)
1786
1787 thumbnails = [{
1788 'id': image.get('type'),
1789 'url': image.get('src'),
1790 'width': int_or_none(image.get('width')),
1791 'height': int_or_none(image.get('height')),
1792 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1793
1794 return {
1795 'id': video_id,
1796 'title': title or video_id,
1797 'description': description,
1798 'upload_date': upload_date,
1799 'thumbnails': thumbnails,
1800 'formats': formats,
1801 'subtitles': subtitles,
1802 }
1803
1804 def _parse_smil_namespace(self, smil):
1805 return self._search_regex(
1806 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1807
1808 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1809 base = smil_url
1810 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1811 b = meta.get('base') or meta.get('httpBase')
1812 if b:
1813 base = b
1814 break
1815
1816 formats = []
1817 rtmp_count = 0
1818 http_count = 0
1819 m3u8_count = 0
1820
1821 srcs = []
1822 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1823 for medium in media:
1824 src = medium.get('src')
1825 if not src or src in srcs:
1826 continue
1827 srcs.append(src)
1828
1829 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1830 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1831 width = int_or_none(medium.get('width'))
1832 height = int_or_none(medium.get('height'))
1833 proto = medium.get('proto')
1834 ext = medium.get('ext')
1835 src_ext = determine_ext(src)
1836 streamer = medium.get('streamer') or base
1837
1838 if proto == 'rtmp' or streamer.startswith('rtmp'):
1839 rtmp_count += 1
1840 formats.append({
1841 'url': streamer,
1842 'play_path': src,
1843 'ext': 'flv',
1844 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1845 'tbr': bitrate,
1846 'filesize': filesize,
1847 'width': width,
1848 'height': height,
1849 })
1850 if transform_rtmp_url:
1851 streamer, src = transform_rtmp_url(streamer, src)
1852 formats[-1].update({
1853 'url': streamer,
1854 'play_path': src,
1855 })
1856 continue
1857
1858 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1859 src_url = src_url.strip()
1860
1861 if proto == 'm3u8' or src_ext == 'm3u8':
1862 m3u8_formats = self._extract_m3u8_formats(
1863 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1864 if len(m3u8_formats) == 1:
1865 m3u8_count += 1
1866 m3u8_formats[0].update({
1867 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1868 'tbr': bitrate,
1869 'width': width,
1870 'height': height,
1871 })
1872 formats.extend(m3u8_formats)
1873 elif src_ext == 'f4m':
1874 f4m_url = src_url
1875 if not f4m_params:
1876 f4m_params = {
1877 'hdcore': '3.2.0',
1878 'plugin': 'flowplayer-3.2.0.1',
1879 }
1880 f4m_url += '&' if '?' in f4m_url else '?'
1881 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1882 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1883 elif src_ext == 'mpd':
1884 formats.extend(self._extract_mpd_formats(
1885 src_url, video_id, mpd_id='dash', fatal=False))
1886 elif re.search(r'\.ism/[Mm]anifest', src_url):
1887 formats.extend(self._extract_ism_formats(
1888 src_url, video_id, ism_id='mss', fatal=False))
1889 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
1890 http_count += 1
1891 formats.append({
1892 'url': src_url,
1893 'ext': ext or src_ext or 'flv',
1894 'format_id': 'http-%d' % (bitrate or http_count),
1895 'tbr': bitrate,
1896 'filesize': filesize,
1897 'width': width,
1898 'height': height,
1899 })
1900
1901 return formats
1902
1903 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1904 urls = []
1905 subtitles = {}
1906 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1907 src = textstream.get('src')
1908 if not src or src in urls:
1909 continue
1910 urls.append(src)
1911 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1912 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1913 subtitles.setdefault(lang, []).append({
1914 'url': src,
1915 'ext': ext,
1916 })
1917 return subtitles
1918
1919 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
1920 xspf = self._download_xml(
1921 xspf_url, playlist_id, 'Downloading xpsf playlist',
1922 'Unable to download xspf manifest', fatal=fatal)
1923 if xspf is False:
1924 return []
1925 return self._parse_xspf(
1926 xspf, playlist_id, xspf_url=xspf_url,
1927 xspf_base_url=base_url(xspf_url))
1928
1929 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
1930 NS_MAP = {
1931 'xspf': 'http://xspf.org/ns/0/',
1932 's1': 'http://static.streamone.nl/player/ns/0',
1933 }
1934
1935 entries = []
1936 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1937 title = xpath_text(
1938 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1939 description = xpath_text(
1940 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1941 thumbnail = xpath_text(
1942 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1943 duration = float_or_none(
1944 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1945
1946 formats = []
1947 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
1948 format_url = urljoin(xspf_base_url, location.text)
1949 if not format_url:
1950 continue
1951 formats.append({
1952 'url': format_url,
1953 'manifest_url': xspf_url,
1954 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1955 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1956 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1957 })
1958 self._sort_formats(formats)
1959
1960 entries.append({
1961 'id': playlist_id,
1962 'title': title,
1963 'description': description,
1964 'thumbnail': thumbnail,
1965 'duration': duration,
1966 'formats': formats,
1967 })
1968 return entries
1969
1970 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1971 res = self._download_xml_handle(
1972 mpd_url, video_id,
1973 note=note or 'Downloading MPD manifest',
1974 errnote=errnote or 'Failed to download MPD manifest',
1975 fatal=fatal)
1976 if res is False:
1977 return []
1978 mpd_doc, urlh = res
1979 mpd_base_url = base_url(urlh.geturl())
1980
1981 return self._parse_mpd_formats(
1982 mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
1983 formats_dict=formats_dict, mpd_url=mpd_url)
1984
1985 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1986 """
1987 Parse formats from MPD manifest.
1988 References:
1989 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1990 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1991 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1992 """
1993 if mpd_doc.get('type') == 'dynamic':
1994 return []
1995
1996 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1997
1998 def _add_ns(path):
1999 return self._xpath_ns(path, namespace)
2000
2001 def is_drm_protected(element):
2002 return element.find(_add_ns('ContentProtection')) is not None
2003
2004 def extract_multisegment_info(element, ms_parent_info):
2005 ms_info = ms_parent_info.copy()
2006
2007 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2008 # common attributes and elements. We will only extract relevant
2009 # for us.
2010 def extract_common(source):
2011 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2012 if segment_timeline is not None:
2013 s_e = segment_timeline.findall(_add_ns('S'))
2014 if s_e:
2015 ms_info['total_number'] = 0
2016 ms_info['s'] = []
2017 for s in s_e:
2018 r = int(s.get('r', 0))
2019 ms_info['total_number'] += 1 + r
2020 ms_info['s'].append({
2021 't': int(s.get('t', 0)),
2022 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2023 'd': int(s.attrib['d']),
2024 'r': r,
2025 })
2026 start_number = source.get('startNumber')
2027 if start_number:
2028 ms_info['start_number'] = int(start_number)
2029 timescale = source.get('timescale')
2030 if timescale:
2031 ms_info['timescale'] = int(timescale)
2032 segment_duration = source.get('duration')
2033 if segment_duration:
2034 ms_info['segment_duration'] = float(segment_duration)
2035
2036 def extract_Initialization(source):
2037 initialization = source.find(_add_ns('Initialization'))
2038 if initialization is not None:
2039 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2040
2041 segment_list = element.find(_add_ns('SegmentList'))
2042 if segment_list is not None:
2043 extract_common(segment_list)
2044 extract_Initialization(segment_list)
2045 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2046 if segment_urls_e:
2047 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2048 else:
2049 segment_template = element.find(_add_ns('SegmentTemplate'))
2050 if segment_template is not None:
2051 extract_common(segment_template)
2052 media = segment_template.get('media')
2053 if media:
2054 ms_info['media'] = media
2055 initialization = segment_template.get('initialization')
2056 if initialization:
2057 ms_info['initialization'] = initialization
2058 else:
2059 extract_Initialization(segment_template)
2060 return ms_info
2061
2062 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2063 formats = []
2064 for period in mpd_doc.findall(_add_ns('Period')):
2065 period_duration = parse_duration(period.get('duration')) or mpd_duration
2066 period_ms_info = extract_multisegment_info(period, {
2067 'start_number': 1,
2068 'timescale': 1,
2069 })
2070 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2071 if is_drm_protected(adaptation_set):
2072 continue
2073 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2074 for representation in adaptation_set.findall(_add_ns('Representation')):
2075 if is_drm_protected(representation):
2076 continue
2077 representation_attrib = adaptation_set.attrib.copy()
2078 representation_attrib.update(representation.attrib)
2079 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2080 mime_type = representation_attrib['mimeType']
2081 content_type = mime_type.split('/')[0]
2082 if content_type == 'text':
2083 # TODO implement WebVTT downloading
2084 pass
2085 elif content_type in ('video', 'audio'):
2086 base_url = ''
2087 for element in (representation, adaptation_set, period, mpd_doc):
2088 base_url_e = element.find(_add_ns('BaseURL'))
2089 if base_url_e is not None:
2090 base_url = base_url_e.text + base_url
2091 if re.match(r'^https?://', base_url):
2092 break
2093 if mpd_base_url and not re.match(r'^https?://', base_url):
2094 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2095 mpd_base_url += '/'
2096 base_url = mpd_base_url + base_url
2097 representation_id = representation_attrib.get('id')
2098 lang = representation_attrib.get('lang')
2099 url_el = representation.find(_add_ns('BaseURL'))
2100 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2101 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2102 f = {
2103 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2104 'url': base_url,
2105 'manifest_url': mpd_url,
2106 'ext': mimetype2ext(mime_type),
2107 'width': int_or_none(representation_attrib.get('width')),
2108 'height': int_or_none(representation_attrib.get('height')),
2109 'tbr': float_or_none(bandwidth, 1000),
2110 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2111 'fps': int_or_none(representation_attrib.get('frameRate')),
2112 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2113 'format_note': 'DASH %s' % content_type,
2114 'filesize': filesize,
2115 'container': mimetype2ext(mime_type) + '_dash',
2116 }
2117 f.update(parse_codecs(representation_attrib.get('codecs')))
2118 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2119
2120 def prepare_template(template_name, identifiers):
2121 tmpl = representation_ms_info[template_name]
2122 # First of, % characters outside $...$ templates
2123 # must be escaped by doubling for proper processing
2124 # by % operator string formatting used further (see
2125 # https://github.com/rg3/youtube-dl/issues/16867).
2126 t = ''
2127 in_template = False
2128 for c in tmpl:
2129 t += c
2130 if c == '$':
2131 in_template = not in_template
2132 elif c == '%' and not in_template:
2133 t += c
2134 # Next, $...$ templates are translated to their
2135 # %(...) counterparts to be used with % operator
2136 t = t.replace('$RepresentationID$', representation_id)
2137 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2138 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2139 t.replace('$$', '$')
2140 return t
2141
2142 # @initialization is a regular template like @media one
2143 # so it should be handled just the same way (see
2144 # https://github.com/rg3/youtube-dl/issues/11605)
2145 if 'initialization' in representation_ms_info:
2146 initialization_template = prepare_template(
2147 'initialization',
2148 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2149 # $Time$ shall not be included for @initialization thus
2150 # only $Bandwidth$ remains
2151 ('Bandwidth', ))
2152 representation_ms_info['initialization_url'] = initialization_template % {
2153 'Bandwidth': bandwidth,
2154 }
2155
2156 def location_key(location):
2157 return 'url' if re.match(r'^https?://', location) else 'path'
2158
2159 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2160
2161 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2162 media_location_key = location_key(media_template)
2163
2164 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2165 # can't be used at the same time
2166 if '%(Number' in media_template and 's' not in representation_ms_info:
2167 segment_duration = None
2168 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2169 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2170 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2171 representation_ms_info['fragments'] = [{
2172 media_location_key: media_template % {
2173 'Number': segment_number,
2174 'Bandwidth': bandwidth,
2175 },
2176 'duration': segment_duration,
2177 } for segment_number in range(
2178 representation_ms_info['start_number'],
2179 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2180 else:
2181 # $Number*$ or $Time$ in media template with S list available
2182 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2183 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2184 representation_ms_info['fragments'] = []
2185 segment_time = 0
2186 segment_d = None
2187 segment_number = representation_ms_info['start_number']
2188
2189 def add_segment_url():
2190 segment_url = media_template % {
2191 'Time': segment_time,
2192 'Bandwidth': bandwidth,
2193 'Number': segment_number,
2194 }
2195 representation_ms_info['fragments'].append({
2196 media_location_key: segment_url,
2197 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2198 })
2199
2200 for num, s in enumerate(representation_ms_info['s']):
2201 segment_time = s.get('t') or segment_time
2202 segment_d = s['d']
2203 add_segment_url()
2204 segment_number += 1
2205 for r in range(s.get('r', 0)):
2206 segment_time += segment_d
2207 add_segment_url()
2208 segment_number += 1
2209 segment_time += segment_d
2210 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2211 # No media template
2212 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2213 # or any YouTube dashsegments video
2214 fragments = []
2215 segment_index = 0
2216 timescale = representation_ms_info['timescale']
2217 for s in representation_ms_info['s']:
2218 duration = float_or_none(s['d'], timescale)
2219 for r in range(s.get('r', 0) + 1):
2220 segment_uri = representation_ms_info['segment_urls'][segment_index]
2221 fragments.append({
2222 location_key(segment_uri): segment_uri,
2223 'duration': duration,
2224 })
2225 segment_index += 1
2226 representation_ms_info['fragments'] = fragments
2227 elif 'segment_urls' in representation_ms_info:
2228 # Segment URLs with no SegmentTimeline
2229 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2230 # https://github.com/rg3/youtube-dl/pull/14844
2231 fragments = []
2232 segment_duration = float_or_none(
2233 representation_ms_info['segment_duration'],
2234 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2235 for segment_url in representation_ms_info['segment_urls']:
2236 fragment = {
2237 location_key(segment_url): segment_url,
2238 }
2239 if segment_duration:
2240 fragment['duration'] = segment_duration
2241 fragments.append(fragment)
2242 representation_ms_info['fragments'] = fragments
2243 # NB: MPD manifest may contain direct URLs to unfragmented media.
2244 # No fragments key is present in this case.
2245 if 'fragments' in representation_ms_info:
2246 f.update({
2247 'fragment_base_url': base_url,
2248 'fragments': [],
2249 'protocol': 'http_dash_segments',
2250 })
2251 if 'initialization_url' in representation_ms_info:
2252 initialization_url = representation_ms_info['initialization_url']
2253 if not f.get('url'):
2254 f['url'] = initialization_url
2255 f['fragments'].append({location_key(initialization_url): initialization_url})
2256 f['fragments'].extend(representation_ms_info['fragments'])
2257 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2258 # is not necessarily unique within a Period thus formats with
2259 # the same `format_id` are quite possible. There are numerous examples
2260 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2261 # https://github.com/rg3/youtube-dl/issues/13919)
2262 full_info = formats_dict.get(representation_id, {}).copy()
2263 full_info.update(f)
2264 formats.append(full_info)
2265 else:
2266 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2267 return formats
2268
2269 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2270 res = self._download_xml_handle(
2271 ism_url, video_id,
2272 note=note or 'Downloading ISM manifest',
2273 errnote=errnote or 'Failed to download ISM manifest',
2274 fatal=fatal)
2275 if res is False:
2276 return []
2277 ism_doc, urlh = res
2278
2279 return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
2280
2281 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2282 """
2283 Parse formats from ISM manifest.
2284 References:
2285 1. [MS-SSTR]: Smooth Streaming Protocol,
2286 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2287 """
2288 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2289 return []
2290
2291 duration = int(ism_doc.attrib['Duration'])
2292 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2293
2294 formats = []
2295 for stream in ism_doc.findall('StreamIndex'):
2296 stream_type = stream.get('Type')
2297 if stream_type not in ('video', 'audio'):
2298 continue
2299 url_pattern = stream.attrib['Url']
2300 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2301 stream_name = stream.get('Name')
2302 for track in stream.findall('QualityLevel'):
2303 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2304 # TODO: add support for WVC1 and WMAP
2305 if fourcc not in ('H264', 'AVC1', 'AACL'):
2306 self.report_warning('%s is not a supported codec' % fourcc)
2307 continue
2308 tbr = int(track.attrib['Bitrate']) // 1000
2309 # [1] does not mention Width and Height attributes. However,
2310 # they're often present while MaxWidth and MaxHeight are
2311 # missing, so should be used as fallbacks
2312 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2313 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2314 sampling_rate = int_or_none(track.get('SamplingRate'))
2315
2316 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2317 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2318
2319 fragments = []
2320 fragment_ctx = {
2321 'time': 0,
2322 }
2323 stream_fragments = stream.findall('c')
2324 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2325 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2326 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2327 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2328 if not fragment_ctx['duration']:
2329 try:
2330 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2331 except IndexError:
2332 next_fragment_time = duration
2333 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2334 for _ in range(fragment_repeat):
2335 fragments.append({
2336 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2337 'duration': fragment_ctx['duration'] / stream_timescale,
2338 })
2339 fragment_ctx['time'] += fragment_ctx['duration']
2340
2341 format_id = []
2342 if ism_id:
2343 format_id.append(ism_id)
2344 if stream_name:
2345 format_id.append(stream_name)
2346 format_id.append(compat_str(tbr))
2347
2348 formats.append({
2349 'format_id': '-'.join(format_id),
2350 'url': ism_url,
2351 'manifest_url': ism_url,
2352 'ext': 'ismv' if stream_type == 'video' else 'isma',
2353 'width': width,
2354 'height': height,
2355 'tbr': tbr,
2356 'asr': sampling_rate,
2357 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2358 'acodec': 'none' if stream_type == 'video' else fourcc,
2359 'protocol': 'ism',
2360 'fragments': fragments,
2361 '_download_params': {
2362 'duration': duration,
2363 'timescale': stream_timescale,
2364 'width': width or 0,
2365 'height': height or 0,
2366 'fourcc': fourcc,
2367 'codec_private_data': track.get('CodecPrivateData'),
2368 'sampling_rate': sampling_rate,
2369 'channels': int_or_none(track.get('Channels', 2)),
2370 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2371 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2372 },
2373 })
2374 return formats
2375
2376 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2377 def absolute_url(item_url):
2378 return urljoin(base_url, item_url)
2379
2380 def parse_content_type(content_type):
2381 if not content_type:
2382 return {}
2383 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2384 if ctr:
2385 mimetype, codecs = ctr.groups()
2386 f = parse_codecs(codecs)
2387 f['ext'] = mimetype2ext(mimetype)
2388 return f
2389 return {}
2390
2391 def _media_formats(src, cur_media_type, type_info={}):
2392 full_url = absolute_url(src)
2393 ext = type_info.get('ext') or determine_ext(full_url)
2394 if ext == 'm3u8':
2395 is_plain_url = False
2396 formats = self._extract_m3u8_formats(
2397 full_url, video_id, ext='mp4',
2398 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2399 preference=preference, fatal=False)
2400 elif ext == 'mpd':
2401 is_plain_url = False
2402 formats = self._extract_mpd_formats(
2403 full_url, video_id, mpd_id=mpd_id, fatal=False)
2404 else:
2405 is_plain_url = True
2406 formats = [{
2407 'url': full_url,
2408 'vcodec': 'none' if cur_media_type == 'audio' else None,
2409 }]
2410 return is_plain_url, formats
2411
2412 entries = []
2413 # amp-video and amp-audio are very similar to their HTML5 counterparts
2414 # so we wll include them right here (see
2415 # https://www.ampproject.org/docs/reference/components/amp-video)
2416 media_tags = [(media_tag, media_type, '')
2417 for media_tag, media_type
2418 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2419 media_tags.extend(re.findall(
2420 # We only allow video|audio followed by a whitespace or '>'.
2421 # Allowing more characters may end up in significant slow down (see
2422 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2423 # http://www.porntrex.com/maps/videositemap.xml).
2424 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2425 for media_tag, media_type, media_content in media_tags:
2426 media_info = {
2427 'formats': [],
2428 'subtitles': {},
2429 }
2430 media_attributes = extract_attributes(media_tag)
2431 src = media_attributes.get('src')
2432 if src:
2433 _, formats = _media_formats(src, media_type)
2434 media_info['formats'].extend(formats)
2435 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2436 if media_content:
2437 for source_tag in re.findall(r'<source[^>]+>', media_content):
2438 source_attributes = extract_attributes(source_tag)
2439 src = source_attributes.get('src')
2440 if not src:
2441 continue
2442 f = parse_content_type(source_attributes.get('type'))
2443 is_plain_url, formats = _media_formats(src, media_type, f)
2444 if is_plain_url:
2445 # res attribute is not standard but seen several times
2446 # in the wild
2447 f.update({
2448 'height': int_or_none(source_attributes.get('res')),
2449 'format_id': source_attributes.get('label'),
2450 })
2451 f.update(formats[0])
2452 media_info['formats'].append(f)
2453 else:
2454 media_info['formats'].extend(formats)
2455 for track_tag in re.findall(r'<track[^>]+>', media_content):
2456 track_attributes = extract_attributes(track_tag)
2457 kind = track_attributes.get('kind')
2458 if not kind or kind in ('subtitles', 'captions'):
2459 src = track_attributes.get('src')
2460 if not src:
2461 continue
2462 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2463 media_info['subtitles'].setdefault(lang, []).append({
2464 'url': absolute_url(src),
2465 })
2466 for f in media_info['formats']:
2467 f.setdefault('http_headers', {})['Referer'] = base_url
2468 if media_info['formats'] or media_info['subtitles']:
2469 entries.append(media_info)
2470 return entries
2471
2472 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2473 formats = []
2474 hdcore_sign = 'hdcore=3.7.0'
2475 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2476 hds_host = hosts.get('hds')
2477 if hds_host:
2478 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2479 if 'hdcore=' not in f4m_url:
2480 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2481 f4m_formats = self._extract_f4m_formats(
2482 f4m_url, video_id, f4m_id='hds', fatal=False)
2483 for entry in f4m_formats:
2484 entry.update({'extra_param_to_segment_url': hdcore_sign})
2485 formats.extend(f4m_formats)
2486 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2487 hls_host = hosts.get('hls')
2488 if hls_host:
2489 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2490 formats.extend(self._extract_m3u8_formats(
2491 m3u8_url, video_id, 'mp4', 'm3u8_native',
2492 m3u8_id='hls', fatal=False))
2493 return formats
2494
2495 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2496 query = compat_urlparse.urlparse(url).query
2497 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2498 mobj = re.search(
2499 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
2500 url_base = mobj.group('url')
2501 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
2502 formats = []
2503
2504 def manifest_url(manifest):
2505 m_url = '%s/%s' % (http_base_url, manifest)
2506 if query:
2507 m_url += '?%s' % query
2508 return m_url
2509
2510 if 'm3u8' not in skip_protocols:
2511 formats.extend(self._extract_m3u8_formats(
2512 manifest_url('playlist.m3u8'), video_id, 'mp4',
2513 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2514 if 'f4m' not in skip_protocols:
2515 formats.extend(self._extract_f4m_formats(
2516 manifest_url('manifest.f4m'),
2517 video_id, f4m_id='hds', fatal=False))
2518 if 'dash' not in skip_protocols:
2519 formats.extend(self._extract_mpd_formats(
2520 manifest_url('manifest.mpd'),
2521 video_id, mpd_id='dash', fatal=False))
2522 if re.search(r'(?:/smil:|\.smil)', url_base):
2523 if 'smil' not in skip_protocols:
2524 rtmp_formats = self._extract_smil_formats(
2525 manifest_url('jwplayer.smil'),
2526 video_id, fatal=False)
2527 for rtmp_format in rtmp_formats:
2528 rtsp_format = rtmp_format.copy()
2529 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2530 del rtsp_format['play_path']
2531 del rtsp_format['ext']
2532 rtsp_format.update({
2533 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2534 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2535 'protocol': 'rtsp',
2536 })
2537 formats.extend([rtmp_format, rtsp_format])
2538 else:
2539 for protocol in ('rtmp', 'rtsp'):
2540 if protocol not in skip_protocols:
2541 formats.append({
2542 'url': '%s:%s' % (protocol, url_base),
2543 'format_id': protocol,
2544 'protocol': protocol,
2545 })
2546 return formats
2547
2548 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2549 mobj = re.search(
2550 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2551 webpage)
2552 if mobj:
2553 try:
2554 jwplayer_data = self._parse_json(mobj.group('options'),
2555 video_id=video_id,
2556 transform_source=transform_source)
2557 except ExtractorError:
2558 pass
2559 else:
2560 if isinstance(jwplayer_data, dict):
2561 return jwplayer_data
2562
2563 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2564 jwplayer_data = self._find_jwplayer_data(
2565 webpage, video_id, transform_source=js_to_json)
2566 return self._parse_jwplayer_data(
2567 jwplayer_data, video_id, *args, **kwargs)
2568
2569 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2570 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2571 # JWPlayer backward compatibility: flattened playlists
2572 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2573 if 'playlist' not in jwplayer_data:
2574 jwplayer_data = {'playlist': [jwplayer_data]}
2575
2576 entries = []
2577
2578 # JWPlayer backward compatibility: single playlist item
2579 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2580 if not isinstance(jwplayer_data['playlist'], list):
2581 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2582
2583 for video_data in jwplayer_data['playlist']:
2584 # JWPlayer backward compatibility: flattened sources
2585 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2586 if 'sources' not in video_data:
2587 video_data['sources'] = [video_data]
2588
2589 this_video_id = video_id or video_data['mediaid']
2590
2591 formats = self._parse_jwplayer_formats(
2592 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2593 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2594
2595 subtitles = {}
2596 tracks = video_data.get('tracks')
2597 if tracks and isinstance(tracks, list):
2598 for track in tracks:
2599 if not isinstance(track, dict):
2600 continue
2601 track_kind = track.get('kind')
2602 if not track_kind or not isinstance(track_kind, compat_str):
2603 continue
2604 if track_kind.lower() not in ('captions', 'subtitles'):
2605 continue
2606 track_url = urljoin(base_url, track.get('file'))
2607 if not track_url:
2608 continue
2609 subtitles.setdefault(track.get('label') or 'en', []).append({
2610 'url': self._proto_relative_url(track_url)
2611 })
2612
2613 entry = {
2614 'id': this_video_id,
2615 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2616 'description': video_data.get('description'),
2617 'thumbnail': self._proto_relative_url(video_data.get('image')),
2618 'timestamp': int_or_none(video_data.get('pubdate')),
2619 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2620 'subtitles': subtitles,
2621 }
2622 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2623 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2624 entry.update({
2625 '_type': 'url_transparent',
2626 'url': formats[0]['url'],
2627 })
2628 else:
2629 self._sort_formats(formats)
2630 entry['formats'] = formats
2631 entries.append(entry)
2632 if len(entries) == 1:
2633 return entries[0]
2634 else:
2635 return self.playlist_result(entries)
2636
2637 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2638 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2639 urls = []
2640 formats = []
2641 for source in jwplayer_sources_data:
2642 if not isinstance(source, dict):
2643 continue
2644 source_url = self._proto_relative_url(source.get('file'))
2645 if not source_url:
2646 continue
2647 if base_url:
2648 source_url = compat_urlparse.urljoin(base_url, source_url)
2649 if source_url in urls:
2650 continue
2651 urls.append(source_url)
2652 source_type = source.get('type') or ''
2653 ext = mimetype2ext(source_type) or determine_ext(source_url)
2654 if source_type == 'hls' or ext == 'm3u8':
2655 formats.extend(self._extract_m3u8_formats(
2656 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2657 m3u8_id=m3u8_id, fatal=False))
2658 elif source_type == 'dash' or ext == 'mpd':
2659 formats.extend(self._extract_mpd_formats(
2660 source_url, video_id, mpd_id=mpd_id, fatal=False))
2661 elif ext == 'smil':
2662 formats.extend(self._extract_smil_formats(
2663 source_url, video_id, fatal=False))
2664 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2665 elif source_type.startswith('audio') or ext in (
2666 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2667 formats.append({
2668 'url': source_url,
2669 'vcodec': 'none',
2670 'ext': ext,
2671 })
2672 else:
2673 height = int_or_none(source.get('height'))
2674 if height is None:
2675 # Often no height is provided but there is a label in
2676 # format like "1080p", "720p SD", or 1080.
2677 height = int_or_none(self._search_regex(
2678 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2679 'height', default=None))
2680 a_format = {
2681 'url': source_url,
2682 'width': int_or_none(source.get('width')),
2683 'height': height,
2684 'tbr': int_or_none(source.get('bitrate')),
2685 'ext': ext,
2686 }
2687 if source_url.startswith('rtmp'):
2688 a_format['ext'] = 'flv'
2689 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2690 # of jwplayer.flash.swf
2691 rtmp_url_parts = re.split(
2692 r'((?:mp4|mp3|flv):)', source_url, 1)
2693 if len(rtmp_url_parts) == 3:
2694 rtmp_url, prefix, play_path = rtmp_url_parts
2695 a_format.update({
2696 'url': rtmp_url,
2697 'play_path': prefix + play_path,
2698 })
2699 if rtmp_params:
2700 a_format.update(rtmp_params)
2701 formats.append(a_format)
2702 return formats
2703
2704 def _live_title(self, name):
2705 """ Generate the title for a live video """
2706 now = datetime.datetime.now()
2707 now_str = now.strftime('%Y-%m-%d %H:%M')
2708 return name + ' ' + now_str
2709
2710 def _int(self, v, name, fatal=False, **kwargs):
2711 res = int_or_none(v, **kwargs)
2712 if 'get_attr' in kwargs:
2713 print(getattr(v, kwargs['get_attr']))
2714 if res is None:
2715 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2716 if fatal:
2717 raise ExtractorError(msg)
2718 else:
2719 self._downloader.report_warning(msg)
2720 return res
2721
2722 def _float(self, v, name, fatal=False, **kwargs):
2723 res = float_or_none(v, **kwargs)
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 _set_cookie(self, domain, name, value, expire_time=None, port=None,
2733 path='/', secure=False, discard=False, rest={}, **kwargs):
2734 cookie = compat_cookiejar.Cookie(
2735 0, name, value, port, port is not None, domain, True,
2736 domain.startswith('.'), path, True, secure, expire_time,
2737 discard, None, None, rest)
2738 self._downloader.cookiejar.set_cookie(cookie)
2739
2740 def _get_cookies(self, url):
2741 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2742 req = sanitized_Request(url)
2743 self._downloader.cookiejar.add_cookie_header(req)
2744 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2745
2746 def get_testcases(self, include_onlymatching=False):
2747 t = getattr(self, '_TEST', None)
2748 if t:
2749 assert not hasattr(self, '_TESTS'), \
2750 '%s has _TEST and _TESTS' % type(self).__name__
2751 tests = [t]
2752 else:
2753 tests = getattr(self, '_TESTS', [])
2754 for t in tests:
2755 if not include_onlymatching and t.get('only_matching', False):
2756 continue
2757 t['name'] = type(self).__name__[:-len('IE')]
2758 yield t
2759
2760 def is_suitable(self, age_limit):
2761 """ Test whether the extractor is generally suitable for the given
2762 age limit (i.e. pornographic sites are not, all others usually are) """
2763
2764 any_restricted = False
2765 for tc in self.get_testcases(include_onlymatching=False):
2766 if tc.get('playlist', []):
2767 tc = tc['playlist'][0]
2768 is_restricted = age_restricted(
2769 tc.get('info_dict', {}).get('age_limit'), age_limit)
2770 if not is_restricted:
2771 return True
2772 any_restricted = any_restricted or is_restricted
2773 return not any_restricted
2774
2775 def extract_subtitles(self, *args, **kwargs):
2776 if (self._downloader.params.get('writesubtitles', False) or
2777 self._downloader.params.get('listsubtitles')):
2778 return self._get_subtitles(*args, **kwargs)
2779 return {}
2780
2781 def _get_subtitles(self, *args, **kwargs):
2782 raise NotImplementedError('This method must be implemented by subclasses')
2783
2784 @staticmethod
2785 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2786 """ Merge subtitle items for one language. Items with duplicated URLs
2787 will be dropped. """
2788 list1_urls = set([item['url'] for item in subtitle_list1])
2789 ret = list(subtitle_list1)
2790 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2791 return ret
2792
2793 @classmethod
2794 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2795 """ Merge two subtitle dictionaries, language by language. """
2796 ret = dict(subtitle_dict1)
2797 for lang in subtitle_dict2:
2798 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2799 return ret
2800
2801 def extract_automatic_captions(self, *args, **kwargs):
2802 if (self._downloader.params.get('writeautomaticsub', False) or
2803 self._downloader.params.get('listsubtitles')):
2804 return self._get_automatic_captions(*args, **kwargs)
2805 return {}
2806
2807 def _get_automatic_captions(self, *args, **kwargs):
2808 raise NotImplementedError('This method must be implemented by subclasses')
2809
2810 def mark_watched(self, *args, **kwargs):
2811 if (self._downloader.params.get('mark_watched', False) and
2812 (self._get_login_info()[0] is not None or
2813 self._downloader.params.get('cookiefile') is not None)):
2814 self._mark_watched(*args, **kwargs)
2815
2816 def _mark_watched(self, *args, **kwargs):
2817 raise NotImplementedError('This method must be implemented by subclasses')
2818
2819 def geo_verification_headers(self):
2820 headers = {}
2821 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2822 if geo_verification_proxy:
2823 headers['Ytdl-request-proxy'] = geo_verification_proxy
2824 return headers
2825
2826 def _generic_id(self, url):
2827 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2828
2829 def _generic_title(self, url):
2830 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2831
2832
2833 class SearchInfoExtractor(InfoExtractor):
2834 """
2835 Base class for paged search queries extractors.
2836 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2837 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2838 """
2839
2840 @classmethod
2841 def _make_valid_url(cls):
2842 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2843
2844 @classmethod
2845 def suitable(cls, url):
2846 return re.match(cls._make_valid_url(), url) is not None
2847
2848 def _real_extract(self, query):
2849 mobj = re.match(self._make_valid_url(), query)
2850 if mobj is None:
2851 raise ExtractorError('Invalid search query "%s"' % query)
2852
2853 prefix = mobj.group('prefix')
2854 query = mobj.group('query')
2855 if prefix == '':
2856 return self._get_n_results(query, 1)
2857 elif prefix == 'all':
2858 return self._get_n_results(query, self._MAX_RESULTS)
2859 else:
2860 n = int(prefix)
2861 if n <= 0:
2862 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2863 elif n > self._MAX_RESULTS:
2864 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2865 n = self._MAX_RESULTS
2866 return self._get_n_results(query, n)
2867
2868 def _get_n_results(self, query, n):
2869 """Get a specified number of results for a query"""
2870 raise NotImplementedError('This method must be implemented by subclasses')
2871
2872 @property
2873 def SEARCH_KEY(self):
2874 return self._SEARCH_KEY