]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
5b6a09c0b7f732ef5e9361775e42719af0deb8b7
[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_http_client,
23 compat_os_name,
24 compat_str,
25 compat_urllib_error,
26 compat_urllib_parse_unquote,
27 compat_urllib_parse_urlencode,
28 compat_urllib_request,
29 compat_urlparse,
30 compat_xml_parse_error,
31 )
32 from ..downloader.f4m import (
33 get_base_url,
34 remove_encrypted_media,
35 )
36 from ..utils import (
37 NO_DEFAULT,
38 age_restricted,
39 base_url,
40 bug_reports_message,
41 clean_html,
42 compiled_regex_type,
43 determine_ext,
44 determine_protocol,
45 error_to_compat_str,
46 ExtractorError,
47 extract_attributes,
48 fix_xml_ampersands,
49 float_or_none,
50 GeoRestrictedError,
51 GeoUtils,
52 int_or_none,
53 js_to_json,
54 mimetype2ext,
55 orderedSet,
56 parse_codecs,
57 parse_duration,
58 parse_iso8601,
59 parse_m3u8_attributes,
60 RegexNotFoundError,
61 sanitized_Request,
62 sanitize_filename,
63 unescapeHTML,
64 unified_strdate,
65 unified_timestamp,
66 update_Request,
67 update_url_query,
68 urljoin,
69 url_basename,
70 xpath_element,
71 xpath_text,
72 xpath_with_ns,
73 )
74
75
76 class InfoExtractor(object):
77 """Information Extractor class.
78
79 Information extractors are the classes that, given a URL, extract
80 information about the video (or videos) the URL refers to. This
81 information includes the real video URL, the video title, author and
82 others. The information is stored in a dictionary which is then
83 passed to the YoutubeDL. The YoutubeDL processes this
84 information possibly downloading the video to the file system, among
85 other possible outcomes.
86
87 The type field determines the type of the result.
88 By far the most common value (and the default if _type is missing) is
89 "video", which indicates a single video.
90
91 For a video, the dictionaries must include the following fields:
92
93 id: Video identifier.
94 title: Video title, unescaped.
95
96 Additionally, it must contain either a formats entry or a url one:
97
98 formats: A list of dictionaries for each format available, ordered
99 from worst to best quality.
100
101 Potential fields:
102 * url Mandatory. The URL of the video file
103 * manifest_url
104 The URL of the manifest file in case of
105 fragmented media (DASH, hls, hds)
106 * ext Will be calculated from URL if missing
107 * format A human-readable description of the format
108 ("mp4 container with h264/opus").
109 Calculated from the format_id, width, height.
110 and format_note fields if missing.
111 * format_id A short description of the format
112 ("mp4_h264_opus" or "19").
113 Technically optional, but strongly recommended.
114 * format_note Additional info about the format
115 ("3D" or "DASH video")
116 * width Width of the video, if known
117 * height Height of the video, if known
118 * resolution Textual description of width and height
119 * tbr Average bitrate of audio and video in KBit/s
120 * abr Average audio bitrate in KBit/s
121 * acodec Name of the audio codec in use
122 * asr Audio sampling rate in Hertz
123 * vbr Average video bitrate in KBit/s
124 * fps Frame rate
125 * vcodec Name of the video codec in use
126 * container Name of the container format
127 * filesize The number of bytes, if known in advance
128 * filesize_approx An estimate for the number of bytes
129 * player_url SWF Player URL (used for rtmpdump).
130 * protocol The protocol that will be used for the actual
131 download, lower-case.
132 "http", "https", "rtsp", "rtmp", "rtmpe",
133 "m3u8", "m3u8_native" or "http_dash_segments".
134 * fragment_base_url
135 Base URL for fragments. Each fragment's path
136 value (if present) will be relative to
137 this URL.
138 * fragments A list of fragments of a fragmented media.
139 Each fragment entry must contain either an url
140 or a path. If an url is present it should be
141 considered by a client. Otherwise both path and
142 fragment_base_url must be present. Here is
143 the list of all potential fields:
144 * "url" - fragment's URL
145 * "path" - fragment's path relative to
146 fragment_base_url
147 * "duration" (optional, int or float)
148 * "filesize" (optional, int)
149 * preference Order number of this format. If this field is
150 present and not None, the formats get sorted
151 by this field, regardless of all other values.
152 -1 for default (order by other properties),
153 -2 or smaller for less than default.
154 < -1000 to hide the format (if there is
155 another one which is strictly better)
156 * language Language code, e.g. "de" or "en-US".
157 * language_preference Is this in the language mentioned in
158 the URL?
159 10 if it's what the URL is about,
160 -1 for default (don't know),
161 -10 otherwise, other values reserved for now.
162 * quality Order number of the video quality of this
163 format, irrespective of the file format.
164 -1 for default (order by other properties),
165 -2 or smaller for less than default.
166 * source_preference Order number for this video source
167 (quality takes higher priority)
168 -1 for default (order by other properties),
169 -2 or smaller for less than default.
170 * http_headers A dictionary of additional HTTP headers
171 to add to the request.
172 * stretched_ratio If given and not 1, indicates that the
173 video's pixels are not square.
174 width : height ratio as float.
175 * no_resume The server does not support resuming the
176 (HTTP or RTMP) download. Boolean.
177
178 url: Final video URL.
179 ext: Video filename extension.
180 format: The video format, defaults to ext (used for --get-format)
181 player_url: SWF Player URL (used for rtmpdump).
182
183 The following fields are optional:
184
185 alt_title: A secondary title of the video.
186 display_id An alternative identifier for the video, not necessarily
187 unique, but available before title. Typically, id is
188 something like "4234987", title "Dancing naked mole rats",
189 and display_id "dancing-naked-mole-rats"
190 thumbnails: A list of dictionaries, with the following entries:
191 * "id" (optional, string) - Thumbnail format ID
192 * "url"
193 * "preference" (optional, int) - quality of the image
194 * "width" (optional, int)
195 * "height" (optional, int)
196 * "resolution" (optional, string "{width}x{height"},
197 deprecated)
198 * "filesize" (optional, int)
199 thumbnail: Full URL to a video thumbnail image.
200 description: Full video description.
201 uploader: Full name of the video uploader.
202 license: License name the video is licensed under.
203 creator: The creator of the video.
204 release_date: The date (YYYYMMDD) when the video was released.
205 timestamp: UNIX timestamp of the moment the video became available.
206 upload_date: Video upload date (YYYYMMDD).
207 If not explicitly set, calculated from timestamp.
208 uploader_id: Nickname or id of the video uploader.
209 uploader_url: Full URL to a personal webpage of the video uploader.
210 location: Physical location where the video was filmed.
211 subtitles: The available subtitles as a dictionary in the format
212 {tag: subformats}. "tag" is usually a language code, and
213 "subformats" is a list sorted from lower to higher
214 preference, each element is a dictionary with the "ext"
215 entry and one of:
216 * "data": The subtitles file contents
217 * "url": A URL pointing to the subtitles file
218 "ext" will be calculated from URL if missing
219 automatic_captions: Like 'subtitles', used by the YoutubeIE for
220 automatically generated captions
221 duration: Length of the video in seconds, as an integer or float.
222 view_count: How many users have watched the video on the platform.
223 like_count: Number of positive ratings of the video
224 dislike_count: Number of negative ratings of the video
225 repost_count: Number of reposts of the video
226 average_rating: Average rating give by users, the scale used depends on the webpage
227 comment_count: Number of comments on the video
228 comments: A list of comments, each with one or more of the following
229 properties (all but one of text or html optional):
230 * "author" - human-readable name of the comment author
231 * "author_id" - user ID of the comment author
232 * "id" - Comment ID
233 * "html" - Comment as HTML
234 * "text" - Plain text of the comment
235 * "timestamp" - UNIX timestamp of comment
236 * "parent" - ID of the comment this one is replying to.
237 Set to "root" to indicate that this is a
238 comment to the original video.
239 age_limit: Age restriction for the video, as an integer (years)
240 webpage_url: The URL to the video webpage, if given to youtube-dl it
241 should allow to get the same result again. (It will be set
242 by YoutubeDL if it's missing)
243 categories: A list of categories that the video falls in, for example
244 ["Sports", "Berlin"]
245 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
246 is_live: True, False, or None (=unknown). Whether this video is a
247 live stream that goes on instead of a fixed-length video.
248 start_time: Time in seconds where the reproduction should start, as
249 specified in the URL.
250 end_time: Time in seconds where the reproduction should end, as
251 specified in the URL.
252 chapters: A list of dictionaries, with the following entries:
253 * "start_time" - The start time of the chapter in seconds
254 * "end_time" - The end time of the chapter in seconds
255 * "title" (optional, string)
256
257 The following fields should only be used when the video belongs to some logical
258 chapter or section:
259
260 chapter: Name or title of the chapter the video belongs to.
261 chapter_number: Number of the chapter the video belongs to, as an integer.
262 chapter_id: Id of the chapter the video belongs to, as a unicode string.
263
264 The following fields should only be used when the video is an episode of some
265 series, programme or podcast:
266
267 series: Title of the series or programme the video episode belongs to.
268 season: Title of the season the video episode belongs to.
269 season_number: Number of the season the video episode belongs to, as an integer.
270 season_id: Id of the season the video episode belongs to, as a unicode string.
271 episode: Title of the video episode. Unlike mandatory video title field,
272 this field should denote the exact title of the video episode
273 without any kind of decoration.
274 episode_number: Number of the video episode within a season, as an integer.
275 episode_id: Id of the video episode, as a unicode string.
276
277 The following fields should only be used when the media is a track or a part of
278 a music album:
279
280 track: Title of the track.
281 track_number: Number of the track within an album or a disc, as an integer.
282 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
283 as a unicode string.
284 artist: Artist(s) of the track.
285 genre: Genre(s) of the track.
286 album: Title of the album the track belongs to.
287 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
288 album_artist: List of all artists appeared on the album (e.g.
289 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
290 and compilations).
291 disc_number: Number of the disc or other physical medium the track belongs to,
292 as an integer.
293 release_year: Year (YYYY) when the album was released.
294
295 Unless mentioned otherwise, the fields should be Unicode strings.
296
297 Unless mentioned otherwise, None is equivalent to absence of information.
298
299
300 _type "playlist" indicates multiple videos.
301 There must be a key "entries", which is a list, an iterable, or a PagedList
302 object, each element of which is a valid dictionary by this specification.
303
304 Additionally, playlists can have "id", "title", "description", "uploader",
305 "uploader_id", "uploader_url" attributes with the same semantics as videos
306 (see above).
307
308
309 _type "multi_video" indicates that there are multiple videos that
310 form a single show, for examples multiple acts of an opera or TV episode.
311 It must have an entries key like a playlist and contain all the keys
312 required for a video at the same time.
313
314
315 _type "url" indicates that the video must be extracted from another
316 location, possibly by a different extractor. Its only required key is:
317 "url" - the next URL to extract.
318 The key "ie_key" can be set to the class name (minus the trailing "IE",
319 e.g. "Youtube") if the extractor class is known in advance.
320 Additionally, the dictionary may have any properties of the resolved entity
321 known in advance, for example "title" if the title of the referred video is
322 known ahead of time.
323
324
325 _type "url_transparent" entities have the same specification as "url", but
326 indicate that the given additional information is more precise than the one
327 associated with the resolved URL.
328 This is useful when a site employs a video service that hosts the video and
329 its technical metadata, but that video service does not embed a useful
330 title, description etc.
331
332
333 Subclasses of this one should re-define the _real_initialize() and
334 _real_extract() methods and define a _VALID_URL regexp.
335 Probably, they should also be added to the list of extractors.
336
337 _GEO_BYPASS attribute may be set to False in order to disable
338 geo restriction bypass mechanisms for a particular extractor.
339 Though it won't disable explicit geo restriction bypass based on
340 country code provided with geo_bypass_country. (experimental)
341
342 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
343 countries for this extractor. One of these countries will be used by
344 geo restriction bypass mechanism right away in order to bypass
345 geo restriction, of course, if the mechanism is not disabled. (experimental)
346
347 NB: both these geo attributes are experimental and may change in future
348 or be completely removed.
349
350 Finally, the _WORKING attribute should be set to False for broken IEs
351 in order to warn the users and skip the tests.
352 """
353
354 _ready = False
355 _downloader = None
356 _x_forwarded_for_ip = None
357 _GEO_BYPASS = True
358 _GEO_COUNTRIES = None
359 _WORKING = True
360
361 def __init__(self, downloader=None):
362 """Constructor. Receives an optional downloader."""
363 self._ready = False
364 self._x_forwarded_for_ip = None
365 self.set_downloader(downloader)
366
367 @classmethod
368 def suitable(cls, url):
369 """Receives a URL and returns True if suitable for this IE."""
370
371 # This does not use has/getattr intentionally - we want to know whether
372 # we have cached the regexp for *this* class, whereas getattr would also
373 # match the superclass
374 if '_VALID_URL_RE' not in cls.__dict__:
375 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
376 return cls._VALID_URL_RE.match(url) is not None
377
378 @classmethod
379 def _match_id(cls, url):
380 if '_VALID_URL_RE' not in cls.__dict__:
381 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
382 m = cls._VALID_URL_RE.match(url)
383 assert m
384 return compat_str(m.group('id'))
385
386 @classmethod
387 def working(cls):
388 """Getter method for _WORKING."""
389 return cls._WORKING
390
391 def initialize(self):
392 """Initializes an instance (authentication, etc)."""
393 self._initialize_geo_bypass(self._GEO_COUNTRIES)
394 if not self._ready:
395 self._real_initialize()
396 self._ready = True
397
398 def _initialize_geo_bypass(self, countries):
399 """
400 Initialize geo restriction bypass mechanism.
401
402 This method is used to initialize geo bypass mechanism based on faking
403 X-Forwarded-For HTTP header. A random country from provided country list
404 is selected and a random IP belonging to this country is generated. This
405 IP will be passed as X-Forwarded-For HTTP header in all subsequent
406 HTTP requests.
407
408 This method will be used for initial geo bypass mechanism initialization
409 during the instance initialization with _GEO_COUNTRIES.
410
411 You may also manually call it from extractor's code if geo countries
412 information is not available beforehand (e.g. obtained during
413 extraction) or due to some another reason.
414 """
415 if not self._x_forwarded_for_ip:
416 country_code = self._downloader.params.get('geo_bypass_country', None)
417 # If there is no explicit country for geo bypass specified and
418 # the extractor is known to be geo restricted let's fake IP
419 # as X-Forwarded-For right away.
420 if (not country_code and
421 self._GEO_BYPASS and
422 self._downloader.params.get('geo_bypass', True) and
423 countries):
424 country_code = random.choice(countries)
425 if country_code:
426 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
427 if self._downloader.params.get('verbose', False):
428 self._downloader.to_screen(
429 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
430 % (self._x_forwarded_for_ip, country_code.upper()))
431
432 def extract(self, url):
433 """Extracts URL information and returns it in list of dicts."""
434 try:
435 for _ in range(2):
436 try:
437 self.initialize()
438 ie_result = self._real_extract(url)
439 if self._x_forwarded_for_ip:
440 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
441 return ie_result
442 except GeoRestrictedError as e:
443 if self.__maybe_fake_ip_and_retry(e.countries):
444 continue
445 raise
446 except ExtractorError:
447 raise
448 except compat_http_client.IncompleteRead as e:
449 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
450 except (KeyError, StopIteration) as e:
451 raise ExtractorError('An extractor error has occurred.', cause=e)
452
453 def __maybe_fake_ip_and_retry(self, countries):
454 if (not self._downloader.params.get('geo_bypass_country', None) and
455 self._GEO_BYPASS and
456 self._downloader.params.get('geo_bypass', True) and
457 not self._x_forwarded_for_ip and
458 countries):
459 country_code = random.choice(countries)
460 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
461 if self._x_forwarded_for_ip:
462 self.report_warning(
463 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
464 % (self._x_forwarded_for_ip, country_code.upper()))
465 return True
466 return False
467
468 def set_downloader(self, downloader):
469 """Sets the downloader for this IE."""
470 self._downloader = downloader
471
472 def _real_initialize(self):
473 """Real initialization process. Redefine in subclasses."""
474 pass
475
476 def _real_extract(self, url):
477 """Real extraction process. Redefine in subclasses."""
478 pass
479
480 @classmethod
481 def ie_key(cls):
482 """A string for getting the InfoExtractor with get_info_extractor"""
483 return compat_str(cls.__name__[:-2])
484
485 @property
486 def IE_NAME(self):
487 return compat_str(type(self).__name__[:-2])
488
489 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
490 """ Returns the response handle """
491 if note is None:
492 self.report_download_webpage(video_id)
493 elif note is not False:
494 if video_id is None:
495 self.to_screen('%s' % (note,))
496 else:
497 self.to_screen('%s: %s' % (video_id, note))
498
499 # Some sites check X-Forwarded-For HTTP header in order to figure out
500 # the origin of the client behind proxy. This allows bypassing geo
501 # restriction by faking this header's value to IP that belongs to some
502 # geo unrestricted country. We will do so once we encounter any
503 # geo restriction error.
504 if self._x_forwarded_for_ip:
505 if 'X-Forwarded-For' not in headers:
506 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
507
508 if isinstance(url_or_request, compat_urllib_request.Request):
509 url_or_request = update_Request(
510 url_or_request, data=data, headers=headers, query=query)
511 else:
512 if query:
513 url_or_request = update_url_query(url_or_request, query)
514 if data is not None or headers:
515 url_or_request = sanitized_Request(url_or_request, data, headers)
516 try:
517 return self._downloader.urlopen(url_or_request)
518 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
519 if errnote is False:
520 return False
521 if errnote is None:
522 errnote = 'Unable to download webpage'
523
524 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
525 if fatal:
526 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
527 else:
528 self._downloader.report_warning(errmsg)
529 return False
530
531 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
532 """ Returns a tuple (page content as string, URL handle) """
533 # Strip hashes from the URL (#1038)
534 if isinstance(url_or_request, (compat_str, str)):
535 url_or_request = url_or_request.partition('#')[0]
536
537 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
538 if urlh is False:
539 assert not fatal
540 return False
541 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
542 return (content, urlh)
543
544 @staticmethod
545 def _guess_encoding_from_content(content_type, webpage_bytes):
546 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
547 if m:
548 encoding = m.group(1)
549 else:
550 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
551 webpage_bytes[:1024])
552 if m:
553 encoding = m.group(1).decode('ascii')
554 elif webpage_bytes.startswith(b'\xff\xfe'):
555 encoding = 'utf-16'
556 else:
557 encoding = 'utf-8'
558
559 return encoding
560
561 def __check_blocked(self, content):
562 first_block = content[:512]
563 if ('<title>Access to this site is blocked</title>' in content and
564 'Websense' in first_block):
565 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
566 blocked_iframe = self._html_search_regex(
567 r'<iframe src="([^"]+)"', content,
568 'Websense information URL', default=None)
569 if blocked_iframe:
570 msg += ' Visit %s for more details' % blocked_iframe
571 raise ExtractorError(msg, expected=True)
572 if '<title>The URL you requested has been blocked</title>' in first_block:
573 msg = (
574 'Access to this webpage has been blocked by Indian censorship. '
575 'Use a VPN or proxy server (with --proxy) to route around it.')
576 block_msg = self._html_search_regex(
577 r'</h1><p>(.*?)</p>',
578 content, 'block message', default=None)
579 if block_msg:
580 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
581 raise ExtractorError(msg, expected=True)
582 if ('<title>TTK :: Š”Š¾ŃŃ‚ŃƒŠæ Šŗ рŠµŃŃƒŃ€ŃŃƒ Š¾Š³Ń€Š°Š½ŠøчŠµŠ½</title>' in content and
583 'blocklist.rkn.gov.ru' in content):
584 raise ExtractorError(
585 'Access to this webpage has been blocked by decision of the Russian government. '
586 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
587 expected=True)
588
589 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
590 content_type = urlh.headers.get('Content-Type', '')
591 webpage_bytes = urlh.read()
592 if prefix is not None:
593 webpage_bytes = prefix + webpage_bytes
594 if not encoding:
595 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
596 if self._downloader.params.get('dump_intermediate_pages', False):
597 self.to_screen('Dumping request to ' + urlh.geturl())
598 dump = base64.b64encode(webpage_bytes).decode('ascii')
599 self._downloader.to_screen(dump)
600 if self._downloader.params.get('write_pages', False):
601 basen = '%s_%s' % (video_id, urlh.geturl())
602 if len(basen) > 240:
603 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
604 basen = basen[:240 - len(h)] + h
605 raw_filename = basen + '.dump'
606 filename = sanitize_filename(raw_filename, restricted=True)
607 self.to_screen('Saving request to ' + filename)
608 # Working around MAX_PATH limitation on Windows (see
609 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
610 if compat_os_name == 'nt':
611 absfilepath = os.path.abspath(filename)
612 if len(absfilepath) > 259:
613 filename = '\\\\?\\' + absfilepath
614 with open(filename, 'wb') as outf:
615 outf.write(webpage_bytes)
616
617 try:
618 content = webpage_bytes.decode(encoding, 'replace')
619 except LookupError:
620 content = webpage_bytes.decode('utf-8', 'replace')
621
622 self.__check_blocked(content)
623
624 return content
625
626 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}):
627 """ Returns the data of the page as a string """
628 success = False
629 try_count = 0
630 while success is False:
631 try:
632 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
633 success = True
634 except compat_http_client.IncompleteRead as e:
635 try_count += 1
636 if try_count >= tries:
637 raise e
638 self._sleep(timeout, video_id)
639 if res is False:
640 return res
641 else:
642 content, _ = res
643 return content
644
645 def _download_xml(self, url_or_request, video_id,
646 note='Downloading XML', errnote='Unable to download XML',
647 transform_source=None, fatal=True, encoding=None,
648 data=None, headers={}, query={}):
649 """Return the xml as an xml.etree.ElementTree.Element"""
650 xml_string = self._download_webpage(
651 url_or_request, video_id, note, errnote, fatal=fatal,
652 encoding=encoding, data=data, headers=headers, query=query)
653 if xml_string is False:
654 return xml_string
655 return self._parse_xml(
656 xml_string, video_id, transform_source=transform_source,
657 fatal=fatal)
658
659 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
660 if transform_source:
661 xml_string = transform_source(xml_string)
662 try:
663 return compat_etree_fromstring(xml_string.encode('utf-8'))
664 except compat_xml_parse_error as ve:
665 errmsg = '%s: Failed to parse XML ' % video_id
666 if fatal:
667 raise ExtractorError(errmsg, cause=ve)
668 else:
669 self.report_warning(errmsg + str(ve))
670
671 def _download_json(self, url_or_request, video_id,
672 note='Downloading JSON metadata',
673 errnote='Unable to download JSON metadata',
674 transform_source=None,
675 fatal=True, encoding=None, data=None, headers={}, query={}):
676 json_string = self._download_webpage(
677 url_or_request, video_id, note, errnote, fatal=fatal,
678 encoding=encoding, data=data, headers=headers, query=query)
679 if (not fatal) and json_string is False:
680 return None
681 return self._parse_json(
682 json_string, video_id, transform_source=transform_source, fatal=fatal)
683
684 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
685 if transform_source:
686 json_string = transform_source(json_string)
687 try:
688 return json.loads(json_string)
689 except ValueError as ve:
690 errmsg = '%s: Failed to parse JSON ' % video_id
691 if fatal:
692 raise ExtractorError(errmsg, cause=ve)
693 else:
694 self.report_warning(errmsg + str(ve))
695
696 def report_warning(self, msg, video_id=None):
697 idstr = '' if video_id is None else '%s: ' % video_id
698 self._downloader.report_warning(
699 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
700
701 def to_screen(self, msg):
702 """Print msg to screen, prefixing it with '[ie_name]'"""
703 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
704
705 def report_extraction(self, id_or_name):
706 """Report information extraction."""
707 self.to_screen('%s: Extracting information' % id_or_name)
708
709 def report_download_webpage(self, video_id):
710 """Report webpage download."""
711 self.to_screen('%s: Downloading webpage' % video_id)
712
713 def report_age_confirmation(self):
714 """Report attempt to confirm age."""
715 self.to_screen('Confirming age')
716
717 def report_login(self):
718 """Report attempt to log in."""
719 self.to_screen('Logging in')
720
721 @staticmethod
722 def raise_login_required(msg='This video is only available for registered users'):
723 raise ExtractorError(
724 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
725 expected=True)
726
727 @staticmethod
728 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
729 raise GeoRestrictedError(msg, countries=countries)
730
731 # Methods for following #608
732 @staticmethod
733 def url_result(url, ie=None, video_id=None, video_title=None):
734 """Returns a URL that points to a page that should be processed"""
735 # TODO: ie should be the class used for getting the info
736 video_info = {'_type': 'url',
737 'url': url,
738 'ie_key': ie}
739 if video_id is not None:
740 video_info['id'] = video_id
741 if video_title is not None:
742 video_info['title'] = video_title
743 return video_info
744
745 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
746 urls = orderedSet(
747 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
748 for m in matches)
749 return self.playlist_result(
750 urls, playlist_id=playlist_id, playlist_title=playlist_title)
751
752 @staticmethod
753 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
754 """Returns a playlist"""
755 video_info = {'_type': 'playlist',
756 'entries': entries}
757 if playlist_id:
758 video_info['id'] = playlist_id
759 if playlist_title:
760 video_info['title'] = playlist_title
761 if playlist_description:
762 video_info['description'] = playlist_description
763 return video_info
764
765 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
766 """
767 Perform a regex search on the given string, using a single or a list of
768 patterns returning the first matching group.
769 In case of failure return a default value or raise a WARNING or a
770 RegexNotFoundError, depending on fatal, specifying the field name.
771 """
772 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
773 mobj = re.search(pattern, string, flags)
774 else:
775 for p in pattern:
776 mobj = re.search(p, string, flags)
777 if mobj:
778 break
779
780 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
781 _name = '\033[0;34m%s\033[0m' % name
782 else:
783 _name = name
784
785 if mobj:
786 if group is None:
787 # return the first matching group
788 return next(g for g in mobj.groups() if g is not None)
789 else:
790 return mobj.group(group)
791 elif default is not NO_DEFAULT:
792 return default
793 elif fatal:
794 raise RegexNotFoundError('Unable to extract %s' % _name)
795 else:
796 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
797 return None
798
799 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
800 """
801 Like _search_regex, but strips HTML tags and unescapes entities.
802 """
803 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
804 if res:
805 return clean_html(res).strip()
806 else:
807 return res
808
809 def _get_netrc_login_info(self, netrc_machine=None):
810 username = None
811 password = None
812 netrc_machine = netrc_machine or self._NETRC_MACHINE
813
814 if self._downloader.params.get('usenetrc', False):
815 try:
816 info = netrc.netrc().authenticators(netrc_machine)
817 if info is not None:
818 username = info[0]
819 password = info[2]
820 else:
821 raise netrc.NetrcParseError(
822 'No authenticators for %s' % netrc_machine)
823 except (IOError, netrc.NetrcParseError) as err:
824 self._downloader.report_warning(
825 'parsing .netrc: %s' % error_to_compat_str(err))
826
827 return username, password
828
829 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
830 """
831 Get the login info as (username, password)
832 First look for the manually specified credentials using username_option
833 and password_option as keys in params dictionary. If no such credentials
834 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
835 value.
836 If there's no info available, return (None, None)
837 """
838 if self._downloader is None:
839 return (None, None)
840
841 downloader_params = self._downloader.params
842
843 # Attempt to use provided username and password or .netrc data
844 if downloader_params.get(username_option) is not None:
845 username = downloader_params[username_option]
846 password = downloader_params[password_option]
847 else:
848 username, password = self._get_netrc_login_info(netrc_machine)
849
850 return username, password
851
852 def _get_tfa_info(self, note='two-factor verification code'):
853 """
854 Get the two-factor authentication info
855 TODO - asking the user will be required for sms/phone verify
856 currently just uses the command line option
857 If there's no info available, return None
858 """
859 if self._downloader is None:
860 return None
861 downloader_params = self._downloader.params
862
863 if downloader_params.get('twofactor') is not None:
864 return downloader_params['twofactor']
865
866 return compat_getpass('Type %s and press [Return]: ' % note)
867
868 # Helper functions for extracting OpenGraph info
869 @staticmethod
870 def _og_regexes(prop):
871 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
872 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
873 % {'prop': re.escape(prop)})
874 template = r'<meta[^>]+?%s[^>]+?%s'
875 return [
876 template % (property_re, content_re),
877 template % (content_re, property_re),
878 ]
879
880 @staticmethod
881 def _meta_regex(prop):
882 return r'''(?isx)<meta
883 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
884 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
885
886 def _og_search_property(self, prop, html, name=None, **kargs):
887 if not isinstance(prop, (list, tuple)):
888 prop = [prop]
889 if name is None:
890 name = 'OpenGraph %s' % prop[0]
891 og_regexes = []
892 for p in prop:
893 og_regexes.extend(self._og_regexes(p))
894 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
895 if escaped is None:
896 return None
897 return unescapeHTML(escaped)
898
899 def _og_search_thumbnail(self, html, **kargs):
900 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
901
902 def _og_search_description(self, html, **kargs):
903 return self._og_search_property('description', html, fatal=False, **kargs)
904
905 def _og_search_title(self, html, **kargs):
906 return self._og_search_property('title', html, **kargs)
907
908 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
909 regexes = self._og_regexes('video') + self._og_regexes('video:url')
910 if secure:
911 regexes = self._og_regexes('video:secure_url') + regexes
912 return self._html_search_regex(regexes, html, name, **kargs)
913
914 def _og_search_url(self, html, **kargs):
915 return self._og_search_property('url', html, **kargs)
916
917 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
918 if not isinstance(name, (list, tuple)):
919 name = [name]
920 if display_name is None:
921 display_name = name[0]
922 return self._html_search_regex(
923 [self._meta_regex(n) for n in name],
924 html, display_name, fatal=fatal, group='content', **kwargs)
925
926 def _dc_search_uploader(self, html):
927 return self._html_search_meta('dc.creator', html, 'uploader')
928
929 def _rta_search(self, html):
930 # See http://www.rtalabel.org/index.php?content=howtofaq#single
931 if re.search(r'(?ix)<meta\s+name="rating"\s+'
932 r' content="RTA-5042-1996-1400-1577-RTA"',
933 html):
934 return 18
935 return 0
936
937 def _media_rating_search(self, html):
938 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
939 rating = self._html_search_meta('rating', html)
940
941 if not rating:
942 return None
943
944 RATING_TABLE = {
945 'safe for kids': 0,
946 'general': 8,
947 '14 years': 14,
948 'mature': 17,
949 'restricted': 19,
950 }
951 return RATING_TABLE.get(rating.lower())
952
953 def _family_friendly_search(self, html):
954 # See http://schema.org/VideoObject
955 family_friendly = self._html_search_meta(
956 'isFamilyFriendly', html, default=None)
957
958 if not family_friendly:
959 return None
960
961 RATING_TABLE = {
962 '1': 0,
963 'true': 0,
964 '0': 18,
965 'false': 18,
966 }
967 return RATING_TABLE.get(family_friendly.lower())
968
969 def _twitter_search_player(self, html):
970 return self._html_search_meta('twitter:player', html,
971 'twitter card player')
972
973 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
974 json_ld = self._search_regex(
975 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
976 html, 'JSON-LD', group='json_ld', **kwargs)
977 default = kwargs.get('default', NO_DEFAULT)
978 if not json_ld:
979 return default if default is not NO_DEFAULT else {}
980 # JSON-LD may be malformed and thus `fatal` should be respected.
981 # At the same time `default` may be passed that assumes `fatal=False`
982 # for _search_regex. Let's simulate the same behavior here as well.
983 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
984 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
985
986 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
987 if isinstance(json_ld, compat_str):
988 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
989 if not json_ld:
990 return {}
991 info = {}
992 if not isinstance(json_ld, (list, tuple, dict)):
993 return info
994 if isinstance(json_ld, dict):
995 json_ld = [json_ld]
996
997 def extract_video_object(e):
998 assert e['@type'] == 'VideoObject'
999 info.update({
1000 'url': e.get('contentUrl'),
1001 'title': unescapeHTML(e.get('name')),
1002 'description': unescapeHTML(e.get('description')),
1003 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
1004 'duration': parse_duration(e.get('duration')),
1005 'timestamp': unified_timestamp(e.get('uploadDate')),
1006 'filesize': float_or_none(e.get('contentSize')),
1007 'tbr': int_or_none(e.get('bitrate')),
1008 'width': int_or_none(e.get('width')),
1009 'height': int_or_none(e.get('height')),
1010 'view_count': int_or_none(e.get('interactionCount')),
1011 })
1012
1013 for e in json_ld:
1014 if e.get('@context') == 'http://schema.org':
1015 item_type = e.get('@type')
1016 if expected_type is not None and expected_type != item_type:
1017 return info
1018 if item_type in ('TVEpisode', 'Episode'):
1019 info.update({
1020 'episode': unescapeHTML(e.get('name')),
1021 'episode_number': int_or_none(e.get('episodeNumber')),
1022 'description': unescapeHTML(e.get('description')),
1023 })
1024 part_of_season = e.get('partOfSeason')
1025 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1026 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1027 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1028 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1029 info['series'] = unescapeHTML(part_of_series.get('name'))
1030 elif item_type == 'Article':
1031 info.update({
1032 'timestamp': parse_iso8601(e.get('datePublished')),
1033 'title': unescapeHTML(e.get('headline')),
1034 'description': unescapeHTML(e.get('articleBody')),
1035 })
1036 elif item_type == 'VideoObject':
1037 extract_video_object(e)
1038 continue
1039 video = e.get('video')
1040 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1041 extract_video_object(video)
1042 break
1043 return dict((k, v) for k, v in info.items() if v is not None)
1044
1045 @staticmethod
1046 def _hidden_inputs(html):
1047 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1048 hidden_inputs = {}
1049 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1050 attrs = extract_attributes(input)
1051 if not input:
1052 continue
1053 if attrs.get('type') not in ('hidden', 'submit'):
1054 continue
1055 name = attrs.get('name') or attrs.get('id')
1056 value = attrs.get('value')
1057 if name and value is not None:
1058 hidden_inputs[name] = value
1059 return hidden_inputs
1060
1061 def _form_hidden_inputs(self, form_id, html):
1062 form = self._search_regex(
1063 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1064 html, '%s form' % form_id, group='form')
1065 return self._hidden_inputs(form)
1066
1067 def _sort_formats(self, formats, field_preference=None):
1068 if not formats:
1069 raise ExtractorError('No video formats found')
1070
1071 for f in formats:
1072 # Automatically determine tbr when missing based on abr and vbr (improves
1073 # formats sorting in some cases)
1074 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1075 f['tbr'] = f['abr'] + f['vbr']
1076
1077 def _formats_key(f):
1078 # TODO remove the following workaround
1079 from ..utils import determine_ext
1080 if not f.get('ext') and 'url' in f:
1081 f['ext'] = determine_ext(f['url'])
1082
1083 if isinstance(field_preference, (list, tuple)):
1084 return tuple(
1085 f.get(field)
1086 if f.get(field) is not None
1087 else ('' if field == 'format_id' else -1)
1088 for field in field_preference)
1089
1090 preference = f.get('preference')
1091 if preference is None:
1092 preference = 0
1093 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1094 preference -= 0.5
1095
1096 protocol = f.get('protocol') or determine_protocol(f)
1097 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1098
1099 if f.get('vcodec') == 'none': # audio only
1100 preference -= 50
1101 if self._downloader.params.get('prefer_free_formats'):
1102 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1103 else:
1104 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1105 ext_preference = 0
1106 try:
1107 audio_ext_preference = ORDER.index(f['ext'])
1108 except ValueError:
1109 audio_ext_preference = -1
1110 else:
1111 if f.get('acodec') == 'none': # video only
1112 preference -= 40
1113 if self._downloader.params.get('prefer_free_formats'):
1114 ORDER = ['flv', 'mp4', 'webm']
1115 else:
1116 ORDER = ['webm', 'flv', 'mp4']
1117 try:
1118 ext_preference = ORDER.index(f['ext'])
1119 except ValueError:
1120 ext_preference = -1
1121 audio_ext_preference = 0
1122
1123 return (
1124 preference,
1125 f.get('language_preference') if f.get('language_preference') is not None else -1,
1126 f.get('quality') if f.get('quality') is not None else -1,
1127 f.get('tbr') if f.get('tbr') is not None else -1,
1128 f.get('filesize') if f.get('filesize') is not None else -1,
1129 f.get('vbr') if f.get('vbr') is not None else -1,
1130 f.get('height') if f.get('height') is not None else -1,
1131 f.get('width') if f.get('width') is not None else -1,
1132 proto_preference,
1133 ext_preference,
1134 f.get('abr') if f.get('abr') is not None else -1,
1135 audio_ext_preference,
1136 f.get('fps') if f.get('fps') is not None else -1,
1137 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1138 f.get('source_preference') if f.get('source_preference') is not None else -1,
1139 f.get('format_id') if f.get('format_id') is not None else '',
1140 )
1141 formats.sort(key=_formats_key)
1142
1143 def _check_formats(self, formats, video_id):
1144 if formats:
1145 formats[:] = filter(
1146 lambda f: self._is_valid_url(
1147 f['url'], video_id,
1148 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1149 formats)
1150
1151 @staticmethod
1152 def _remove_duplicate_formats(formats):
1153 format_urls = set()
1154 unique_formats = []
1155 for f in formats:
1156 if f['url'] not in format_urls:
1157 format_urls.add(f['url'])
1158 unique_formats.append(f)
1159 formats[:] = unique_formats
1160
1161 def _is_valid_url(self, url, video_id, item='video', headers={}):
1162 url = self._proto_relative_url(url, scheme='http:')
1163 # For now assume non HTTP(S) URLs always valid
1164 if not (url.startswith('http://') or url.startswith('https://')):
1165 return True
1166 try:
1167 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1168 return True
1169 except ExtractorError as e:
1170 if isinstance(e.cause, compat_urllib_error.URLError):
1171 self.to_screen(
1172 '%s: %s URL is invalid, skipping' % (video_id, item))
1173 return False
1174 raise
1175
1176 def http_scheme(self):
1177 """ Either "http:" or "https:", depending on the user's preferences """
1178 return (
1179 'http:'
1180 if self._downloader.params.get('prefer_insecure', False)
1181 else 'https:')
1182
1183 def _proto_relative_url(self, url, scheme=None):
1184 if url is None:
1185 return url
1186 if url.startswith('//'):
1187 if scheme is None:
1188 scheme = self.http_scheme()
1189 return scheme + url
1190 else:
1191 return url
1192
1193 def _sleep(self, timeout, video_id, msg_template=None):
1194 if msg_template is None:
1195 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1196 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1197 self.to_screen(msg)
1198 time.sleep(timeout)
1199
1200 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1201 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1202 fatal=True, m3u8_id=None):
1203 manifest = self._download_xml(
1204 manifest_url, video_id, 'Downloading f4m manifest',
1205 'Unable to download f4m manifest',
1206 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1207 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1208 transform_source=transform_source,
1209 fatal=fatal)
1210
1211 if manifest is False:
1212 return []
1213
1214 return self._parse_f4m_formats(
1215 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1216 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1217
1218 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1219 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1220 fatal=True, m3u8_id=None):
1221 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1222 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1223 if akamai_pv is not None and ';' in akamai_pv.text:
1224 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1225 if playerVerificationChallenge.strip() != '':
1226 return []
1227
1228 formats = []
1229 manifest_version = '1.0'
1230 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1231 if not media_nodes:
1232 manifest_version = '2.0'
1233 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1234 # Remove unsupported DRM protected media from final formats
1235 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1236 media_nodes = remove_encrypted_media(media_nodes)
1237 if not media_nodes:
1238 return formats
1239
1240 manifest_base_url = get_base_url(manifest)
1241
1242 bootstrap_info = xpath_element(
1243 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1244 'bootstrap info', default=None)
1245
1246 vcodec = None
1247 mime_type = xpath_text(
1248 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1249 'base URL', default=None)
1250 if mime_type and mime_type.startswith('audio/'):
1251 vcodec = 'none'
1252
1253 for i, media_el in enumerate(media_nodes):
1254 tbr = int_or_none(media_el.attrib.get('bitrate'))
1255 width = int_or_none(media_el.attrib.get('width'))
1256 height = int_or_none(media_el.attrib.get('height'))
1257 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1258 # If <bootstrapInfo> is present, the specified f4m is a
1259 # stream-level manifest, and only set-level manifests may refer to
1260 # external resources. See section 11.4 and section 4 of F4M spec
1261 if bootstrap_info is None:
1262 media_url = None
1263 # @href is introduced in 2.0, see section 11.6 of F4M spec
1264 if manifest_version == '2.0':
1265 media_url = media_el.attrib.get('href')
1266 if media_url is None:
1267 media_url = media_el.attrib.get('url')
1268 if not media_url:
1269 continue
1270 manifest_url = (
1271 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1272 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1273 # If media_url is itself a f4m manifest do the recursive extraction
1274 # since bitrates in parent manifest (this one) and media_url manifest
1275 # may differ leading to inability to resolve the format by requested
1276 # bitrate in f4m downloader
1277 ext = determine_ext(manifest_url)
1278 if ext == 'f4m':
1279 f4m_formats = self._extract_f4m_formats(
1280 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1281 transform_source=transform_source, fatal=fatal)
1282 # Sometimes stream-level manifest contains single media entry that
1283 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1284 # At the same time parent's media entry in set-level manifest may
1285 # contain it. We will copy it from parent in such cases.
1286 if len(f4m_formats) == 1:
1287 f = f4m_formats[0]
1288 f.update({
1289 'tbr': f.get('tbr') or tbr,
1290 'width': f.get('width') or width,
1291 'height': f.get('height') or height,
1292 'format_id': f.get('format_id') if not tbr else format_id,
1293 'vcodec': vcodec,
1294 })
1295 formats.extend(f4m_formats)
1296 continue
1297 elif ext == 'm3u8':
1298 formats.extend(self._extract_m3u8_formats(
1299 manifest_url, video_id, 'mp4', preference=preference,
1300 m3u8_id=m3u8_id, fatal=fatal))
1301 continue
1302 formats.append({
1303 'format_id': format_id,
1304 'url': manifest_url,
1305 'manifest_url': manifest_url,
1306 'ext': 'flv' if bootstrap_info is not None else None,
1307 'protocol': 'f4m',
1308 'tbr': tbr,
1309 'width': width,
1310 'height': height,
1311 'vcodec': vcodec,
1312 'preference': preference,
1313 })
1314 return formats
1315
1316 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1317 return {
1318 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1319 'url': m3u8_url,
1320 'ext': ext,
1321 'protocol': 'm3u8',
1322 'preference': preference - 100 if preference else -100,
1323 'resolution': 'multiple',
1324 'format_note': 'Quality selection URL',
1325 }
1326
1327 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1328 entry_protocol='m3u8', preference=None,
1329 m3u8_id=None, note=None, errnote=None,
1330 fatal=True, live=False):
1331 res = self._download_webpage_handle(
1332 m3u8_url, video_id,
1333 note=note or 'Downloading m3u8 information',
1334 errnote=errnote or 'Failed to download m3u8 information',
1335 fatal=fatal)
1336
1337 if res is False:
1338 return []
1339
1340 m3u8_doc, urlh = res
1341 m3u8_url = urlh.geturl()
1342
1343 return self._parse_m3u8_formats(
1344 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1345 preference=preference, m3u8_id=m3u8_id, live=live)
1346
1347 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1348 entry_protocol='m3u8', preference=None,
1349 m3u8_id=None, live=False):
1350 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1351 return []
1352
1353 if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay
1354 return []
1355
1356 formats = []
1357
1358 format_url = lambda u: (
1359 u
1360 if re.match(r'^https?://', u)
1361 else compat_urlparse.urljoin(m3u8_url, u))
1362
1363 # References:
1364 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1365 # 2. https://github.com/rg3/youtube-dl/issues/12211
1366
1367 # We should try extracting formats only from master playlists [1, 4.3.4],
1368 # i.e. playlists that describe available qualities. On the other hand
1369 # media playlists [1, 4.3.3] should be returned as is since they contain
1370 # just the media without qualities renditions.
1371 # Fortunately, master playlist can be easily distinguished from media
1372 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1373 # master playlist tags MUST NOT appear in a media playist and vice versa.
1374 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1375 # media playlist and MUST NOT appear in master playlist thus we can
1376 # clearly detect media playlist with this criterion.
1377
1378 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1379 return [{
1380 'url': m3u8_url,
1381 'format_id': m3u8_id,
1382 'ext': ext,
1383 'protocol': entry_protocol,
1384 'preference': preference,
1385 }]
1386
1387 groups = {}
1388 last_stream_inf = {}
1389
1390 def extract_media(x_media_line):
1391 media = parse_m3u8_attributes(x_media_line)
1392 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1393 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1394 if not (media_type and group_id and name):
1395 return
1396 groups.setdefault(group_id, []).append(media)
1397 if media_type not in ('VIDEO', 'AUDIO'):
1398 return
1399 media_url = media.get('URI')
1400 if media_url:
1401 format_id = []
1402 for v in (m3u8_id, group_id, name):
1403 if v:
1404 format_id.append(v)
1405 f = {
1406 'format_id': '-'.join(format_id),
1407 'url': format_url(media_url),
1408 'manifest_url': m3u8_url,
1409 'language': media.get('LANGUAGE'),
1410 'ext': ext,
1411 'protocol': entry_protocol,
1412 'preference': preference,
1413 }
1414 if media_type == 'AUDIO':
1415 f['vcodec'] = 'none'
1416 formats.append(f)
1417
1418 def build_stream_name():
1419 # Despite specification does not mention NAME attribute for
1420 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1421 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1422 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1423 stream_name = last_stream_inf.get('NAME')
1424 if stream_name:
1425 return stream_name
1426 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1427 # from corresponding rendition group
1428 stream_group_id = last_stream_inf.get('VIDEO')
1429 if not stream_group_id:
1430 return
1431 stream_group = groups.get(stream_group_id)
1432 if not stream_group:
1433 return stream_group_id
1434 rendition = stream_group[0]
1435 return rendition.get('NAME') or stream_group_id
1436
1437 for line in m3u8_doc.splitlines():
1438 if line.startswith('#EXT-X-STREAM-INF:'):
1439 last_stream_inf = parse_m3u8_attributes(line)
1440 elif line.startswith('#EXT-X-MEDIA:'):
1441 extract_media(line)
1442 elif line.startswith('#') or not line.strip():
1443 continue
1444 else:
1445 tbr = float_or_none(
1446 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1447 last_stream_inf.get('BANDWIDTH'), scale=1000)
1448 format_id = []
1449 if m3u8_id:
1450 format_id.append(m3u8_id)
1451 stream_name = build_stream_name()
1452 # Bandwidth of live streams may differ over time thus making
1453 # format_id unpredictable. So it's better to keep provided
1454 # format_id intact.
1455 if not live:
1456 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1457 manifest_url = format_url(line.strip())
1458 f = {
1459 'format_id': '-'.join(format_id),
1460 'url': manifest_url,
1461 'manifest_url': m3u8_url,
1462 'tbr': tbr,
1463 'ext': ext,
1464 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1465 'protocol': entry_protocol,
1466 'preference': preference,
1467 }
1468 resolution = last_stream_inf.get('RESOLUTION')
1469 if resolution:
1470 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1471 if mobj:
1472 f['width'] = int(mobj.group('width'))
1473 f['height'] = int(mobj.group('height'))
1474 # Unified Streaming Platform
1475 mobj = re.search(
1476 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1477 if mobj:
1478 abr, vbr = mobj.groups()
1479 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1480 f.update({
1481 'vbr': vbr,
1482 'abr': abr,
1483 })
1484 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1485 f.update(codecs)
1486 audio_group_id = last_stream_inf.get('AUDIO')
1487 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1488 # references a rendition group MUST have a CODECS attribute.
1489 # However, this is not always respected, for example, [2]
1490 # contains EXT-X-STREAM-INF tag which references AUDIO
1491 # rendition group but does not have CODECS and despite
1492 # referencing audio group an audio group, it represents
1493 # a complete (with audio and video) format. So, for such cases
1494 # we will ignore references to rendition groups and treat them
1495 # as complete formats.
1496 if audio_group_id and codecs and f.get('vcodec') != 'none':
1497 audio_group = groups.get(audio_group_id)
1498 if audio_group and audio_group[0].get('URI'):
1499 # TODO: update acodec for audio only formats with
1500 # the same GROUP-ID
1501 f['acodec'] = 'none'
1502 formats.append(f)
1503 last_stream_inf = {}
1504 return formats
1505
1506 @staticmethod
1507 def _xpath_ns(path, namespace=None):
1508 if not namespace:
1509 return path
1510 out = []
1511 for c in path.split('/'):
1512 if not c or c == '.':
1513 out.append(c)
1514 else:
1515 out.append('{%s}%s' % (namespace, c))
1516 return '/'.join(out)
1517
1518 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1519 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1520
1521 if smil is False:
1522 assert not fatal
1523 return []
1524
1525 namespace = self._parse_smil_namespace(smil)
1526
1527 return self._parse_smil_formats(
1528 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1529
1530 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1531 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1532 if smil is False:
1533 return {}
1534 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1535
1536 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1537 return self._download_xml(
1538 smil_url, video_id, 'Downloading SMIL file',
1539 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1540
1541 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1542 namespace = self._parse_smil_namespace(smil)
1543
1544 formats = self._parse_smil_formats(
1545 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1546 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1547
1548 video_id = os.path.splitext(url_basename(smil_url))[0]
1549 title = None
1550 description = None
1551 upload_date = None
1552 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1553 name = meta.attrib.get('name')
1554 content = meta.attrib.get('content')
1555 if not name or not content:
1556 continue
1557 if not title and name == 'title':
1558 title = content
1559 elif not description and name in ('description', 'abstract'):
1560 description = content
1561 elif not upload_date and name == 'date':
1562 upload_date = unified_strdate(content)
1563
1564 thumbnails = [{
1565 'id': image.get('type'),
1566 'url': image.get('src'),
1567 'width': int_or_none(image.get('width')),
1568 'height': int_or_none(image.get('height')),
1569 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1570
1571 return {
1572 'id': video_id,
1573 'title': title or video_id,
1574 'description': description,
1575 'upload_date': upload_date,
1576 'thumbnails': thumbnails,
1577 'formats': formats,
1578 'subtitles': subtitles,
1579 }
1580
1581 def _parse_smil_namespace(self, smil):
1582 return self._search_regex(
1583 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1584
1585 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1586 base = smil_url
1587 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1588 b = meta.get('base') or meta.get('httpBase')
1589 if b:
1590 base = b
1591 break
1592
1593 formats = []
1594 rtmp_count = 0
1595 http_count = 0
1596 m3u8_count = 0
1597
1598 srcs = []
1599 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1600 for medium in media:
1601 src = medium.get('src')
1602 if not src or src in srcs:
1603 continue
1604 srcs.append(src)
1605
1606 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1607 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1608 width = int_or_none(medium.get('width'))
1609 height = int_or_none(medium.get('height'))
1610 proto = medium.get('proto')
1611 ext = medium.get('ext')
1612 src_ext = determine_ext(src)
1613 streamer = medium.get('streamer') or base
1614
1615 if proto == 'rtmp' or streamer.startswith('rtmp'):
1616 rtmp_count += 1
1617 formats.append({
1618 'url': streamer,
1619 'play_path': src,
1620 'ext': 'flv',
1621 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1622 'tbr': bitrate,
1623 'filesize': filesize,
1624 'width': width,
1625 'height': height,
1626 })
1627 if transform_rtmp_url:
1628 streamer, src = transform_rtmp_url(streamer, src)
1629 formats[-1].update({
1630 'url': streamer,
1631 'play_path': src,
1632 })
1633 continue
1634
1635 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1636 src_url = src_url.strip()
1637
1638 if proto == 'm3u8' or src_ext == 'm3u8':
1639 m3u8_formats = self._extract_m3u8_formats(
1640 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1641 if len(m3u8_formats) == 1:
1642 m3u8_count += 1
1643 m3u8_formats[0].update({
1644 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1645 'tbr': bitrate,
1646 'width': width,
1647 'height': height,
1648 })
1649 formats.extend(m3u8_formats)
1650 continue
1651
1652 if src_ext == 'f4m':
1653 f4m_url = src_url
1654 if not f4m_params:
1655 f4m_params = {
1656 'hdcore': '3.2.0',
1657 'plugin': 'flowplayer-3.2.0.1',
1658 }
1659 f4m_url += '&' if '?' in f4m_url else '?'
1660 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1661 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1662 continue
1663
1664 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1665 http_count += 1
1666 formats.append({
1667 'url': src_url,
1668 'ext': ext or src_ext or 'flv',
1669 'format_id': 'http-%d' % (bitrate or http_count),
1670 'tbr': bitrate,
1671 'filesize': filesize,
1672 'width': width,
1673 'height': height,
1674 })
1675 continue
1676
1677 return formats
1678
1679 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1680 urls = []
1681 subtitles = {}
1682 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1683 src = textstream.get('src')
1684 if not src or src in urls:
1685 continue
1686 urls.append(src)
1687 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1688 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1689 subtitles.setdefault(lang, []).append({
1690 'url': src,
1691 'ext': ext,
1692 })
1693 return subtitles
1694
1695 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1696 xspf = self._download_xml(
1697 playlist_url, playlist_id, 'Downloading xpsf playlist',
1698 'Unable to download xspf manifest', fatal=fatal)
1699 if xspf is False:
1700 return []
1701 return self._parse_xspf(xspf, playlist_id)
1702
1703 def _parse_xspf(self, playlist, playlist_id):
1704 NS_MAP = {
1705 'xspf': 'http://xspf.org/ns/0/',
1706 's1': 'http://static.streamone.nl/player/ns/0',
1707 }
1708
1709 entries = []
1710 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1711 title = xpath_text(
1712 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1713 description = xpath_text(
1714 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1715 thumbnail = xpath_text(
1716 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1717 duration = float_or_none(
1718 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1719
1720 formats = [{
1721 'url': location.text,
1722 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1723 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1724 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1725 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1726 self._sort_formats(formats)
1727
1728 entries.append({
1729 'id': playlist_id,
1730 'title': title,
1731 'description': description,
1732 'thumbnail': thumbnail,
1733 'duration': duration,
1734 'formats': formats,
1735 })
1736 return entries
1737
1738 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1739 res = self._download_webpage_handle(
1740 mpd_url, video_id,
1741 note=note or 'Downloading MPD manifest',
1742 errnote=errnote or 'Failed to download MPD manifest',
1743 fatal=fatal)
1744 if res is False:
1745 return []
1746 mpd, urlh = res
1747 mpd_base_url = base_url(urlh.geturl())
1748
1749 return self._parse_mpd_formats(
1750 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1751 formats_dict=formats_dict, mpd_url=mpd_url)
1752
1753 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1754 """
1755 Parse formats from MPD manifest.
1756 References:
1757 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1758 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1759 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1760 """
1761 if mpd_doc.get('type') == 'dynamic':
1762 return []
1763
1764 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1765
1766 def _add_ns(path):
1767 return self._xpath_ns(path, namespace)
1768
1769 def is_drm_protected(element):
1770 return element.find(_add_ns('ContentProtection')) is not None
1771
1772 def extract_multisegment_info(element, ms_parent_info):
1773 ms_info = ms_parent_info.copy()
1774
1775 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1776 # common attributes and elements. We will only extract relevant
1777 # for us.
1778 def extract_common(source):
1779 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1780 if segment_timeline is not None:
1781 s_e = segment_timeline.findall(_add_ns('S'))
1782 if s_e:
1783 ms_info['total_number'] = 0
1784 ms_info['s'] = []
1785 for s in s_e:
1786 r = int(s.get('r', 0))
1787 ms_info['total_number'] += 1 + r
1788 ms_info['s'].append({
1789 't': int(s.get('t', 0)),
1790 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1791 'd': int(s.attrib['d']),
1792 'r': r,
1793 })
1794 start_number = source.get('startNumber')
1795 if start_number:
1796 ms_info['start_number'] = int(start_number)
1797 timescale = source.get('timescale')
1798 if timescale:
1799 ms_info['timescale'] = int(timescale)
1800 segment_duration = source.get('duration')
1801 if segment_duration:
1802 ms_info['segment_duration'] = float(segment_duration)
1803
1804 def extract_Initialization(source):
1805 initialization = source.find(_add_ns('Initialization'))
1806 if initialization is not None:
1807 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1808
1809 segment_list = element.find(_add_ns('SegmentList'))
1810 if segment_list is not None:
1811 extract_common(segment_list)
1812 extract_Initialization(segment_list)
1813 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1814 if segment_urls_e:
1815 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1816 else:
1817 segment_template = element.find(_add_ns('SegmentTemplate'))
1818 if segment_template is not None:
1819 extract_common(segment_template)
1820 media = segment_template.get('media')
1821 if media:
1822 ms_info['media'] = media
1823 initialization = segment_template.get('initialization')
1824 if initialization:
1825 ms_info['initialization'] = initialization
1826 else:
1827 extract_Initialization(segment_template)
1828 return ms_info
1829
1830 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1831 formats = []
1832 for period in mpd_doc.findall(_add_ns('Period')):
1833 period_duration = parse_duration(period.get('duration')) or mpd_duration
1834 period_ms_info = extract_multisegment_info(period, {
1835 'start_number': 1,
1836 'timescale': 1,
1837 })
1838 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1839 if is_drm_protected(adaptation_set):
1840 continue
1841 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1842 for representation in adaptation_set.findall(_add_ns('Representation')):
1843 if is_drm_protected(representation):
1844 continue
1845 representation_attrib = adaptation_set.attrib.copy()
1846 representation_attrib.update(representation.attrib)
1847 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1848 mime_type = representation_attrib['mimeType']
1849 content_type = mime_type.split('/')[0]
1850 if content_type == 'text':
1851 # TODO implement WebVTT downloading
1852 pass
1853 elif content_type in ('video', 'audio'):
1854 base_url = ''
1855 for element in (representation, adaptation_set, period, mpd_doc):
1856 base_url_e = element.find(_add_ns('BaseURL'))
1857 if base_url_e is not None:
1858 base_url = base_url_e.text + base_url
1859 if re.match(r'^https?://', base_url):
1860 break
1861 if mpd_base_url and not re.match(r'^https?://', base_url):
1862 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1863 mpd_base_url += '/'
1864 base_url = mpd_base_url + base_url
1865 representation_id = representation_attrib.get('id')
1866 lang = representation_attrib.get('lang')
1867 url_el = representation.find(_add_ns('BaseURL'))
1868 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1869 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1870 f = {
1871 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1872 'url': base_url,
1873 'manifest_url': mpd_url,
1874 'ext': mimetype2ext(mime_type),
1875 'width': int_or_none(representation_attrib.get('width')),
1876 'height': int_or_none(representation_attrib.get('height')),
1877 'tbr': float_or_none(bandwidth, 1000),
1878 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1879 'fps': int_or_none(representation_attrib.get('frameRate')),
1880 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1881 'format_note': 'DASH %s' % content_type,
1882 'filesize': filesize,
1883 'container': mimetype2ext(mime_type) + '_dash',
1884 }
1885 f.update(parse_codecs(representation_attrib.get('codecs')))
1886 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1887
1888 def prepare_template(template_name, identifiers):
1889 t = representation_ms_info[template_name]
1890 t = t.replace('$RepresentationID$', representation_id)
1891 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1892 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1893 t.replace('$$', '$')
1894 return t
1895
1896 # @initialization is a regular template like @media one
1897 # so it should be handled just the same way (see
1898 # https://github.com/rg3/youtube-dl/issues/11605)
1899 if 'initialization' in representation_ms_info:
1900 initialization_template = prepare_template(
1901 'initialization',
1902 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1903 # $Time$ shall not be included for @initialization thus
1904 # only $Bandwidth$ remains
1905 ('Bandwidth', ))
1906 representation_ms_info['initialization_url'] = initialization_template % {
1907 'Bandwidth': bandwidth,
1908 }
1909
1910 def location_key(location):
1911 return 'url' if re.match(r'^https?://', location) else 'path'
1912
1913 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1914
1915 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1916 media_location_key = location_key(media_template)
1917
1918 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1919 # can't be used at the same time
1920 if '%(Number' in media_template and 's' not in representation_ms_info:
1921 segment_duration = None
1922 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
1923 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1924 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1925 representation_ms_info['fragments'] = [{
1926 media_location_key: media_template % {
1927 'Number': segment_number,
1928 'Bandwidth': bandwidth,
1929 },
1930 'duration': segment_duration,
1931 } for segment_number in range(
1932 representation_ms_info['start_number'],
1933 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1934 else:
1935 # $Number*$ or $Time$ in media template with S list available
1936 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1937 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1938 representation_ms_info['fragments'] = []
1939 segment_time = 0
1940 segment_d = None
1941 segment_number = representation_ms_info['start_number']
1942
1943 def add_segment_url():
1944 segment_url = media_template % {
1945 'Time': segment_time,
1946 'Bandwidth': bandwidth,
1947 'Number': segment_number,
1948 }
1949 representation_ms_info['fragments'].append({
1950 media_location_key: segment_url,
1951 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1952 })
1953
1954 for num, s in enumerate(representation_ms_info['s']):
1955 segment_time = s.get('t') or segment_time
1956 segment_d = s['d']
1957 add_segment_url()
1958 segment_number += 1
1959 for r in range(s.get('r', 0)):
1960 segment_time += segment_d
1961 add_segment_url()
1962 segment_number += 1
1963 segment_time += segment_d
1964 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1965 # No media template
1966 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1967 # or any YouTube dashsegments video
1968 fragments = []
1969 segment_index = 0
1970 timescale = representation_ms_info['timescale']
1971 for s in representation_ms_info['s']:
1972 duration = float_or_none(s['d'], timescale)
1973 for r in range(s.get('r', 0) + 1):
1974 segment_uri = representation_ms_info['segment_urls'][segment_index]
1975 fragments.append({
1976 location_key(segment_uri): segment_uri,
1977 'duration': duration,
1978 })
1979 segment_index += 1
1980 representation_ms_info['fragments'] = fragments
1981 elif 'segment_urls' in representation_ms_info:
1982 # Segment URLs with no SegmentTimeline
1983 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
1984 # https://github.com/rg3/youtube-dl/pull/14844
1985 fragments = []
1986 segment_duration = float_or_none(
1987 representation_ms_info['segment_duration'],
1988 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
1989 for segment_url in representation_ms_info['segment_urls']:
1990 fragment = {
1991 location_key(segment_url): segment_url,
1992 }
1993 if segment_duration:
1994 fragment['duration'] = segment_duration
1995 fragments.append(fragment)
1996 representation_ms_info['fragments'] = fragments
1997 # NB: MPD manifest may contain direct URLs to unfragmented media.
1998 # No fragments key is present in this case.
1999 if 'fragments' in representation_ms_info:
2000 f.update({
2001 'fragment_base_url': base_url,
2002 'fragments': [],
2003 'protocol': 'http_dash_segments',
2004 })
2005 if 'initialization_url' in representation_ms_info:
2006 initialization_url = representation_ms_info['initialization_url']
2007 if not f.get('url'):
2008 f['url'] = initialization_url
2009 f['fragments'].append({location_key(initialization_url): initialization_url})
2010 f['fragments'].extend(representation_ms_info['fragments'])
2011 # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
2012 # is not necessarily unique within a Period thus formats with
2013 # the same `format_id` are quite possible. There are numerous examples
2014 # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
2015 # https://github.com/rg3/youtube-dl/issues/13919)
2016 full_info = formats_dict.get(representation_id, {}).copy()
2017 full_info.update(f)
2018 formats.append(full_info)
2019 else:
2020 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2021 return formats
2022
2023 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2024 res = self._download_webpage_handle(
2025 ism_url, video_id,
2026 note=note or 'Downloading ISM manifest',
2027 errnote=errnote or 'Failed to download ISM manifest',
2028 fatal=fatal)
2029 if res is False:
2030 return []
2031 ism, urlh = res
2032
2033 return self._parse_ism_formats(
2034 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
2035
2036 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2037 """
2038 Parse formats from ISM manifest.
2039 References:
2040 1. [MS-SSTR]: Smooth Streaming Protocol,
2041 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2042 """
2043 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2044 return []
2045
2046 duration = int(ism_doc.attrib['Duration'])
2047 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2048
2049 formats = []
2050 for stream in ism_doc.findall('StreamIndex'):
2051 stream_type = stream.get('Type')
2052 if stream_type not in ('video', 'audio'):
2053 continue
2054 url_pattern = stream.attrib['Url']
2055 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2056 stream_name = stream.get('Name')
2057 for track in stream.findall('QualityLevel'):
2058 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2059 # TODO: add support for WVC1 and WMAP
2060 if fourcc not in ('H264', 'AVC1', 'AACL'):
2061 self.report_warning('%s is not a supported codec' % fourcc)
2062 continue
2063 tbr = int(track.attrib['Bitrate']) // 1000
2064 # [1] does not mention Width and Height attributes. However,
2065 # they're often present while MaxWidth and MaxHeight are
2066 # missing, so should be used as fallbacks
2067 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2068 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2069 sampling_rate = int_or_none(track.get('SamplingRate'))
2070
2071 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2072 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2073
2074 fragments = []
2075 fragment_ctx = {
2076 'time': 0,
2077 }
2078 stream_fragments = stream.findall('c')
2079 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2080 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2081 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2082 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2083 if not fragment_ctx['duration']:
2084 try:
2085 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2086 except IndexError:
2087 next_fragment_time = duration
2088 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2089 for _ in range(fragment_repeat):
2090 fragments.append({
2091 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2092 'duration': fragment_ctx['duration'] / stream_timescale,
2093 })
2094 fragment_ctx['time'] += fragment_ctx['duration']
2095
2096 format_id = []
2097 if ism_id:
2098 format_id.append(ism_id)
2099 if stream_name:
2100 format_id.append(stream_name)
2101 format_id.append(compat_str(tbr))
2102
2103 formats.append({
2104 'format_id': '-'.join(format_id),
2105 'url': ism_url,
2106 'manifest_url': ism_url,
2107 'ext': 'ismv' if stream_type == 'video' else 'isma',
2108 'width': width,
2109 'height': height,
2110 'tbr': tbr,
2111 'asr': sampling_rate,
2112 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2113 'acodec': 'none' if stream_type == 'video' else fourcc,
2114 'protocol': 'ism',
2115 'fragments': fragments,
2116 '_download_params': {
2117 'duration': duration,
2118 'timescale': stream_timescale,
2119 'width': width or 0,
2120 'height': height or 0,
2121 'fourcc': fourcc,
2122 'codec_private_data': track.get('CodecPrivateData'),
2123 'sampling_rate': sampling_rate,
2124 'channels': int_or_none(track.get('Channels', 2)),
2125 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2126 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2127 },
2128 })
2129 return formats
2130
2131 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2132 def absolute_url(video_url):
2133 return compat_urlparse.urljoin(base_url, video_url)
2134
2135 def parse_content_type(content_type):
2136 if not content_type:
2137 return {}
2138 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2139 if ctr:
2140 mimetype, codecs = ctr.groups()
2141 f = parse_codecs(codecs)
2142 f['ext'] = mimetype2ext(mimetype)
2143 return f
2144 return {}
2145
2146 def _media_formats(src, cur_media_type, type_info={}):
2147 full_url = absolute_url(src)
2148 ext = type_info.get('ext') or determine_ext(full_url)
2149 if ext == 'm3u8':
2150 is_plain_url = False
2151 formats = self._extract_m3u8_formats(
2152 full_url, video_id, ext='mp4',
2153 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2154 preference=preference, fatal=False)
2155 elif ext == 'mpd':
2156 is_plain_url = False
2157 formats = self._extract_mpd_formats(
2158 full_url, video_id, mpd_id=mpd_id, fatal=False)
2159 else:
2160 is_plain_url = True
2161 formats = [{
2162 'url': full_url,
2163 'vcodec': 'none' if cur_media_type == 'audio' else None,
2164 }]
2165 return is_plain_url, formats
2166
2167 entries = []
2168 # amp-video and amp-audio are very similar to their HTML5 counterparts
2169 # so we wll include them right here (see
2170 # https://www.ampproject.org/docs/reference/components/amp-video)
2171 media_tags = [(media_tag, media_type, '')
2172 for media_tag, media_type
2173 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2174 media_tags.extend(re.findall(
2175 # We only allow video|audio followed by a whitespace or '>'.
2176 # Allowing more characters may end up in significant slow down (see
2177 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2178 # http://www.porntrex.com/maps/videositemap.xml).
2179 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2180 for media_tag, media_type, media_content in media_tags:
2181 media_info = {
2182 'formats': [],
2183 'subtitles': {},
2184 }
2185 media_attributes = extract_attributes(media_tag)
2186 src = media_attributes.get('src')
2187 if src:
2188 _, formats = _media_formats(src, media_type)
2189 media_info['formats'].extend(formats)
2190 media_info['thumbnail'] = media_attributes.get('poster')
2191 if media_content:
2192 for source_tag in re.findall(r'<source[^>]+>', media_content):
2193 source_attributes = extract_attributes(source_tag)
2194 src = source_attributes.get('src')
2195 if not src:
2196 continue
2197 f = parse_content_type(source_attributes.get('type'))
2198 is_plain_url, formats = _media_formats(src, media_type, f)
2199 if is_plain_url:
2200 # res attribute is not standard but seen several times
2201 # in the wild
2202 f.update({
2203 'height': int_or_none(source_attributes.get('res')),
2204 'format_id': source_attributes.get('label'),
2205 })
2206 f.update(formats[0])
2207 media_info['formats'].append(f)
2208 else:
2209 media_info['formats'].extend(formats)
2210 for track_tag in re.findall(r'<track[^>]+>', media_content):
2211 track_attributes = extract_attributes(track_tag)
2212 kind = track_attributes.get('kind')
2213 if not kind or kind in ('subtitles', 'captions'):
2214 src = track_attributes.get('src')
2215 if not src:
2216 continue
2217 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2218 media_info['subtitles'].setdefault(lang, []).append({
2219 'url': absolute_url(src),
2220 })
2221 if media_info['formats'] or media_info['subtitles']:
2222 entries.append(media_info)
2223 return entries
2224
2225 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2226 formats = []
2227 hdcore_sign = 'hdcore=3.7.0'
2228 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2229 hds_host = hosts.get('hds')
2230 if hds_host:
2231 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2232 if 'hdcore=' not in f4m_url:
2233 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2234 f4m_formats = self._extract_f4m_formats(
2235 f4m_url, video_id, f4m_id='hds', fatal=False)
2236 for entry in f4m_formats:
2237 entry.update({'extra_param_to_segment_url': hdcore_sign})
2238 formats.extend(f4m_formats)
2239 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2240 hls_host = hosts.get('hls')
2241 if hls_host:
2242 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2243 formats.extend(self._extract_m3u8_formats(
2244 m3u8_url, video_id, 'mp4', 'm3u8_native',
2245 m3u8_id='hls', fatal=False))
2246 return formats
2247
2248 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2249 query = compat_urlparse.urlparse(url).query
2250 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2251 url_base = self._search_regex(
2252 r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url')
2253 http_base_url = '%s:%s' % ('http', url_base)
2254 formats = []
2255
2256 def manifest_url(manifest):
2257 m_url = '%s/%s' % (http_base_url, manifest)
2258 if query:
2259 m_url += '?%s' % query
2260 return m_url
2261
2262 if 'm3u8' not in skip_protocols:
2263 formats.extend(self._extract_m3u8_formats(
2264 manifest_url('playlist.m3u8'), video_id, 'mp4',
2265 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2266 if 'f4m' not in skip_protocols:
2267 formats.extend(self._extract_f4m_formats(
2268 manifest_url('manifest.f4m'),
2269 video_id, f4m_id='hds', fatal=False))
2270 if 'dash' not in skip_protocols:
2271 formats.extend(self._extract_mpd_formats(
2272 manifest_url('manifest.mpd'),
2273 video_id, mpd_id='dash', fatal=False))
2274 if re.search(r'(?:/smil:|\.smil)', url_base):
2275 if 'smil' not in skip_protocols:
2276 rtmp_formats = self._extract_smil_formats(
2277 manifest_url('jwplayer.smil'),
2278 video_id, fatal=False)
2279 for rtmp_format in rtmp_formats:
2280 rtsp_format = rtmp_format.copy()
2281 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2282 del rtsp_format['play_path']
2283 del rtsp_format['ext']
2284 rtsp_format.update({
2285 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2286 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2287 'protocol': 'rtsp',
2288 })
2289 formats.extend([rtmp_format, rtsp_format])
2290 else:
2291 for protocol in ('rtmp', 'rtsp'):
2292 if protocol not in skip_protocols:
2293 formats.append({
2294 'url': '%s:%s' % (protocol, url_base),
2295 'format_id': protocol,
2296 'protocol': protocol,
2297 })
2298 return formats
2299
2300 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2301 mobj = re.search(
2302 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2303 webpage)
2304 if mobj:
2305 try:
2306 jwplayer_data = self._parse_json(mobj.group('options'),
2307 video_id=video_id,
2308 transform_source=transform_source)
2309 except ExtractorError:
2310 pass
2311 else:
2312 if isinstance(jwplayer_data, dict):
2313 return jwplayer_data
2314
2315 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2316 jwplayer_data = self._find_jwplayer_data(
2317 webpage, video_id, transform_source=js_to_json)
2318 return self._parse_jwplayer_data(
2319 jwplayer_data, video_id, *args, **kwargs)
2320
2321 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2322 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2323 # JWPlayer backward compatibility: flattened playlists
2324 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2325 if 'playlist' not in jwplayer_data:
2326 jwplayer_data = {'playlist': [jwplayer_data]}
2327
2328 entries = []
2329
2330 # JWPlayer backward compatibility: single playlist item
2331 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2332 if not isinstance(jwplayer_data['playlist'], list):
2333 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2334
2335 for video_data in jwplayer_data['playlist']:
2336 # JWPlayer backward compatibility: flattened sources
2337 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2338 if 'sources' not in video_data:
2339 video_data['sources'] = [video_data]
2340
2341 this_video_id = video_id or video_data['mediaid']
2342
2343 formats = self._parse_jwplayer_formats(
2344 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2345 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2346
2347 subtitles = {}
2348 tracks = video_data.get('tracks')
2349 if tracks and isinstance(tracks, list):
2350 for track in tracks:
2351 if not isinstance(track, dict):
2352 continue
2353 if track.get('kind') != 'captions':
2354 continue
2355 track_url = urljoin(base_url, track.get('file'))
2356 if not track_url:
2357 continue
2358 subtitles.setdefault(track.get('label') or 'en', []).append({
2359 'url': self._proto_relative_url(track_url)
2360 })
2361
2362 entry = {
2363 'id': this_video_id,
2364 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2365 'description': video_data.get('description'),
2366 'thumbnail': self._proto_relative_url(video_data.get('image')),
2367 'timestamp': int_or_none(video_data.get('pubdate')),
2368 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2369 'subtitles': subtitles,
2370 }
2371 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2372 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2373 entry.update({
2374 '_type': 'url_transparent',
2375 'url': formats[0]['url'],
2376 })
2377 else:
2378 self._sort_formats(formats)
2379 entry['formats'] = formats
2380 entries.append(entry)
2381 if len(entries) == 1:
2382 return entries[0]
2383 else:
2384 return self.playlist_result(entries)
2385
2386 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2387 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2388 urls = []
2389 formats = []
2390 for source in jwplayer_sources_data:
2391 if not isinstance(source, dict):
2392 continue
2393 source_url = self._proto_relative_url(source.get('file'))
2394 if not source_url:
2395 continue
2396 if base_url:
2397 source_url = compat_urlparse.urljoin(base_url, source_url)
2398 if source_url in urls:
2399 continue
2400 urls.append(source_url)
2401 source_type = source.get('type') or ''
2402 ext = mimetype2ext(source_type) or determine_ext(source_url)
2403 if source_type == 'hls' or ext == 'm3u8':
2404 formats.extend(self._extract_m3u8_formats(
2405 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2406 m3u8_id=m3u8_id, fatal=False))
2407 elif ext == 'mpd':
2408 formats.extend(self._extract_mpd_formats(
2409 source_url, video_id, mpd_id=mpd_id, fatal=False))
2410 elif ext == 'smil':
2411 formats.extend(self._extract_smil_formats(
2412 source_url, video_id, fatal=False))
2413 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2414 elif source_type.startswith('audio') or ext in (
2415 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2416 formats.append({
2417 'url': source_url,
2418 'vcodec': 'none',
2419 'ext': ext,
2420 })
2421 else:
2422 height = int_or_none(source.get('height'))
2423 if height is None:
2424 # Often no height is provided but there is a label in
2425 # format like "1080p", "720p SD", or 1080.
2426 height = int_or_none(self._search_regex(
2427 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2428 'height', default=None))
2429 a_format = {
2430 'url': source_url,
2431 'width': int_or_none(source.get('width')),
2432 'height': height,
2433 'tbr': int_or_none(source.get('bitrate')),
2434 'ext': ext,
2435 }
2436 if source_url.startswith('rtmp'):
2437 a_format['ext'] = 'flv'
2438 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2439 # of jwplayer.flash.swf
2440 rtmp_url_parts = re.split(
2441 r'((?:mp4|mp3|flv):)', source_url, 1)
2442 if len(rtmp_url_parts) == 3:
2443 rtmp_url, prefix, play_path = rtmp_url_parts
2444 a_format.update({
2445 'url': rtmp_url,
2446 'play_path': prefix + play_path,
2447 })
2448 if rtmp_params:
2449 a_format.update(rtmp_params)
2450 formats.append(a_format)
2451 return formats
2452
2453 def _live_title(self, name):
2454 """ Generate the title for a live video """
2455 now = datetime.datetime.now()
2456 now_str = now.strftime('%Y-%m-%d %H:%M')
2457 return name + ' ' + now_str
2458
2459 def _int(self, v, name, fatal=False, **kwargs):
2460 res = int_or_none(v, **kwargs)
2461 if 'get_attr' in kwargs:
2462 print(getattr(v, kwargs['get_attr']))
2463 if res is None:
2464 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2465 if fatal:
2466 raise ExtractorError(msg)
2467 else:
2468 self._downloader.report_warning(msg)
2469 return res
2470
2471 def _float(self, v, name, fatal=False, **kwargs):
2472 res = float_or_none(v, **kwargs)
2473 if res is None:
2474 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2475 if fatal:
2476 raise ExtractorError(msg)
2477 else:
2478 self._downloader.report_warning(msg)
2479 return res
2480
2481 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2482 path='/', secure=False, discard=False, rest={}, **kwargs):
2483 cookie = compat_cookiejar.Cookie(
2484 0, name, value, port, port is not None, domain, True,
2485 domain.startswith('.'), path, True, secure, expire_time,
2486 discard, None, None, rest)
2487 self._downloader.cookiejar.set_cookie(cookie)
2488
2489 def _get_cookies(self, url):
2490 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2491 req = sanitized_Request(url)
2492 self._downloader.cookiejar.add_cookie_header(req)
2493 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2494
2495 def get_testcases(self, include_onlymatching=False):
2496 t = getattr(self, '_TEST', None)
2497 if t:
2498 assert not hasattr(self, '_TESTS'), \
2499 '%s has _TEST and _TESTS' % type(self).__name__
2500 tests = [t]
2501 else:
2502 tests = getattr(self, '_TESTS', [])
2503 for t in tests:
2504 if not include_onlymatching and t.get('only_matching', False):
2505 continue
2506 t['name'] = type(self).__name__[:-len('IE')]
2507 yield t
2508
2509 def is_suitable(self, age_limit):
2510 """ Test whether the extractor is generally suitable for the given
2511 age limit (i.e. pornographic sites are not, all others usually are) """
2512
2513 any_restricted = False
2514 for tc in self.get_testcases(include_onlymatching=False):
2515 if tc.get('playlist', []):
2516 tc = tc['playlist'][0]
2517 is_restricted = age_restricted(
2518 tc.get('info_dict', {}).get('age_limit'), age_limit)
2519 if not is_restricted:
2520 return True
2521 any_restricted = any_restricted or is_restricted
2522 return not any_restricted
2523
2524 def extract_subtitles(self, *args, **kwargs):
2525 if (self._downloader.params.get('writesubtitles', False) or
2526 self._downloader.params.get('listsubtitles')):
2527 return self._get_subtitles(*args, **kwargs)
2528 return {}
2529
2530 def _get_subtitles(self, *args, **kwargs):
2531 raise NotImplementedError('This method must be implemented by subclasses')
2532
2533 @staticmethod
2534 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2535 """ Merge subtitle items for one language. Items with duplicated URLs
2536 will be dropped. """
2537 list1_urls = set([item['url'] for item in subtitle_list1])
2538 ret = list(subtitle_list1)
2539 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2540 return ret
2541
2542 @classmethod
2543 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2544 """ Merge two subtitle dictionaries, language by language. """
2545 ret = dict(subtitle_dict1)
2546 for lang in subtitle_dict2:
2547 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2548 return ret
2549
2550 def extract_automatic_captions(self, *args, **kwargs):
2551 if (self._downloader.params.get('writeautomaticsub', False) or
2552 self._downloader.params.get('listsubtitles')):
2553 return self._get_automatic_captions(*args, **kwargs)
2554 return {}
2555
2556 def _get_automatic_captions(self, *args, **kwargs):
2557 raise NotImplementedError('This method must be implemented by subclasses')
2558
2559 def mark_watched(self, *args, **kwargs):
2560 if (self._downloader.params.get('mark_watched', False) and
2561 (self._get_login_info()[0] is not None or
2562 self._downloader.params.get('cookiefile') is not None)):
2563 self._mark_watched(*args, **kwargs)
2564
2565 def _mark_watched(self, *args, **kwargs):
2566 raise NotImplementedError('This method must be implemented by subclasses')
2567
2568 def geo_verification_headers(self):
2569 headers = {}
2570 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2571 if geo_verification_proxy:
2572 headers['Ytdl-request-proxy'] = geo_verification_proxy
2573 return headers
2574
2575 def _generic_id(self, url):
2576 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2577
2578 def _generic_title(self, url):
2579 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2580
2581
2582 class SearchInfoExtractor(InfoExtractor):
2583 """
2584 Base class for paged search queries extractors.
2585 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2586 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2587 """
2588
2589 @classmethod
2590 def _make_valid_url(cls):
2591 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2592
2593 @classmethod
2594 def suitable(cls, url):
2595 return re.match(cls._make_valid_url(), url) is not None
2596
2597 def _real_extract(self, query):
2598 mobj = re.match(self._make_valid_url(), query)
2599 if mobj is None:
2600 raise ExtractorError('Invalid search query "%s"' % query)
2601
2602 prefix = mobj.group('prefix')
2603 query = mobj.group('query')
2604 if prefix == '':
2605 return self._get_n_results(query, 1)
2606 elif prefix == 'all':
2607 return self._get_n_results(query, self._MAX_RESULTS)
2608 else:
2609 n = int(prefix)
2610 if n <= 0:
2611 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2612 elif n > self._MAX_RESULTS:
2613 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2614 n = self._MAX_RESULTS
2615 return self._get_n_results(query, n)
2616
2617 def _get_n_results(self, query, n):
2618 """Get a specified number of results for a query"""
2619 raise NotImplementedError('This method must be implemented by subclasses')
2620
2621 @property
2622 def SEARCH_KEY(self):
2623 return self._SEARCH_KEY