]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/common.py
a692406931d4b63711609c7b8355635235849275
[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 remove_encrypted_media
33 from ..utils import (
34 NO_DEFAULT,
35 age_restricted,
36 base_url,
37 bug_reports_message,
38 clean_html,
39 compiled_regex_type,
40 determine_ext,
41 determine_protocol,
42 error_to_compat_str,
43 ExtractorError,
44 extract_attributes,
45 fix_xml_ampersands,
46 float_or_none,
47 GeoRestrictedError,
48 GeoUtils,
49 int_or_none,
50 js_to_json,
51 mimetype2ext,
52 orderedSet,
53 parse_codecs,
54 parse_duration,
55 parse_iso8601,
56 parse_m3u8_attributes,
57 RegexNotFoundError,
58 sanitized_Request,
59 sanitize_filename,
60 unescapeHTML,
61 unified_strdate,
62 unified_timestamp,
63 update_Request,
64 update_url_query,
65 urljoin,
66 url_basename,
67 xpath_element,
68 xpath_text,
69 xpath_with_ns,
70 )
71
72
73 class InfoExtractor(object):
74 """Information Extractor class.
75
76 Information extractors are the classes that, given a URL, extract
77 information about the video (or videos) the URL refers to. This
78 information includes the real video URL, the video title, author and
79 others. The information is stored in a dictionary which is then
80 passed to the YoutubeDL. The YoutubeDL processes this
81 information possibly downloading the video to the file system, among
82 other possible outcomes.
83
84 The type field determines the type of the result.
85 By far the most common value (and the default if _type is missing) is
86 "video", which indicates a single video.
87
88 For a video, the dictionaries must include the following fields:
89
90 id: Video identifier.
91 title: Video title, unescaped.
92
93 Additionally, it must contain either a formats entry or a url one:
94
95 formats: A list of dictionaries for each format available, ordered
96 from worst to best quality.
97
98 Potential fields:
99 * url Mandatory. The URL of the video file
100 * manifest_url
101 The URL of the manifest file in case of
102 fragmented media (DASH, hls, hds)
103 * ext Will be calculated from URL if missing
104 * format A human-readable description of the format
105 ("mp4 container with h264/opus").
106 Calculated from the format_id, width, height.
107 and format_note fields if missing.
108 * format_id A short description of the format
109 ("mp4_h264_opus" or "19").
110 Technically optional, but strongly recommended.
111 * format_note Additional info about the format
112 ("3D" or "DASH video")
113 * width Width of the video, if known
114 * height Height of the video, if known
115 * resolution Textual description of width and height
116 * tbr Average bitrate of audio and video in KBit/s
117 * abr Average audio bitrate in KBit/s
118 * acodec Name of the audio codec in use
119 * asr Audio sampling rate in Hertz
120 * vbr Average video bitrate in KBit/s
121 * fps Frame rate
122 * vcodec Name of the video codec in use
123 * container Name of the container format
124 * filesize The number of bytes, if known in advance
125 * filesize_approx An estimate for the number of bytes
126 * player_url SWF Player URL (used for rtmpdump).
127 * protocol The protocol that will be used for the actual
128 download, lower-case.
129 "http", "https", "rtsp", "rtmp", "rtmpe",
130 "m3u8", "m3u8_native" or "http_dash_segments".
131 * fragment_base_url
132 Base URL for fragments. Each fragment's path
133 value (if present) will be relative to
134 this URL.
135 * fragments A list of fragments of a fragmented media.
136 Each fragment entry must contain either an url
137 or a path. If an url is present it should be
138 considered by a client. Otherwise both path and
139 fragment_base_url must be present. Here is
140 the list of all potential fields:
141 * "url" - fragment's URL
142 * "path" - fragment's path relative to
143 fragment_base_url
144 * "duration" (optional, int or float)
145 * "filesize" (optional, int)
146 * preference Order number of this format. If this field is
147 present and not None, the formats get sorted
148 by this field, regardless of all other values.
149 -1 for default (order by other properties),
150 -2 or smaller for less than default.
151 < -1000 to hide the format (if there is
152 another one which is strictly better)
153 * language Language code, e.g. "de" or "en-US".
154 * language_preference Is this in the language mentioned in
155 the URL?
156 10 if it's what the URL is about,
157 -1 for default (don't know),
158 -10 otherwise, other values reserved for now.
159 * quality Order number of the video quality of this
160 format, irrespective of the file format.
161 -1 for default (order by other properties),
162 -2 or smaller for less than default.
163 * source_preference Order number for this video source
164 (quality takes higher priority)
165 -1 for default (order by other properties),
166 -2 or smaller for less than default.
167 * http_headers A dictionary of additional HTTP headers
168 to add to the request.
169 * stretched_ratio If given and not 1, indicates that the
170 video's pixels are not square.
171 width : height ratio as float.
172 * no_resume The server does not support resuming the
173 (HTTP or RTMP) download. Boolean.
174
175 url: Final video URL.
176 ext: Video filename extension.
177 format: The video format, defaults to ext (used for --get-format)
178 player_url: SWF Player URL (used for rtmpdump).
179
180 The following fields are optional:
181
182 alt_title: A secondary title of the video.
183 display_id An alternative identifier for the video, not necessarily
184 unique, but available before title. Typically, id is
185 something like "4234987", title "Dancing naked mole rats",
186 and display_id "dancing-naked-mole-rats"
187 thumbnails: A list of dictionaries, with the following entries:
188 * "id" (optional, string) - Thumbnail format ID
189 * "url"
190 * "preference" (optional, int) - quality of the image
191 * "width" (optional, int)
192 * "height" (optional, int)
193 * "resolution" (optional, string "{width}x{height"},
194 deprecated)
195 * "filesize" (optional, int)
196 thumbnail: Full URL to a video thumbnail image.
197 description: Full video description.
198 uploader: Full name of the video uploader.
199 license: License name the video is licensed under.
200 creator: The creator of the video.
201 release_date: The date (YYYYMMDD) when the video was released.
202 timestamp: UNIX timestamp of the moment the video became available.
203 upload_date: Video upload date (YYYYMMDD).
204 If not explicitly set, calculated from timestamp.
205 uploader_id: Nickname or id of the video uploader.
206 uploader_url: Full URL to a personal webpage of the video uploader.
207 location: Physical location where the video was filmed.
208 subtitles: The available subtitles as a dictionary in the format
209 {tag: subformats}. "tag" is usually a language code, and
210 "subformats" is a list sorted from lower to higher
211 preference, each element is a dictionary with the "ext"
212 entry and one of:
213 * "data": The subtitles file contents
214 * "url": A URL pointing to the subtitles file
215 "ext" will be calculated from URL if missing
216 automatic_captions: Like 'subtitles', used by the YoutubeIE for
217 automatically generated captions
218 duration: Length of the video in seconds, as an integer or float.
219 view_count: How many users have watched the video on the platform.
220 like_count: Number of positive ratings of the video
221 dislike_count: Number of negative ratings of the video
222 repost_count: Number of reposts of the video
223 average_rating: Average rating give by users, the scale used depends on the webpage
224 comment_count: Number of comments on the video
225 comments: A list of comments, each with one or more of the following
226 properties (all but one of text or html optional):
227 * "author" - human-readable name of the comment author
228 * "author_id" - user ID of the comment author
229 * "id" - Comment ID
230 * "html" - Comment as HTML
231 * "text" - Plain text of the comment
232 * "timestamp" - UNIX timestamp of comment
233 * "parent" - ID of the comment this one is replying to.
234 Set to "root" to indicate that this is a
235 comment to the original video.
236 age_limit: Age restriction for the video, as an integer (years)
237 webpage_url: The URL to the video webpage, if given to youtube-dl it
238 should allow to get the same result again. (It will be set
239 by YoutubeDL if it's missing)
240 categories: A list of categories that the video falls in, for example
241 ["Sports", "Berlin"]
242 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
243 is_live: True, False, or None (=unknown). Whether this video is a
244 live stream that goes on instead of a fixed-length video.
245 start_time: Time in seconds where the reproduction should start, as
246 specified in the URL.
247 end_time: Time in seconds where the reproduction should end, as
248 specified in the URL.
249 chapters: A list of dictionaries, with the following entries:
250 * "start_time" - The start time of the chapter in seconds
251 * "end_time" - The end time of the chapter in seconds
252 * "title" (optional, string)
253
254 The following fields should only be used when the video belongs to some logical
255 chapter or section:
256
257 chapter: Name or title of the chapter the video belongs to.
258 chapter_number: Number of the chapter the video belongs to, as an integer.
259 chapter_id: Id of the chapter the video belongs to, as a unicode string.
260
261 The following fields should only be used when the video is an episode of some
262 series, programme or podcast:
263
264 series: Title of the series or programme the video episode belongs to.
265 season: Title of the season the video episode belongs to.
266 season_number: Number of the season the video episode belongs to, as an integer.
267 season_id: Id of the season the video episode belongs to, as a unicode string.
268 episode: Title of the video episode. Unlike mandatory video title field,
269 this field should denote the exact title of the video episode
270 without any kind of decoration.
271 episode_number: Number of the video episode within a season, as an integer.
272 episode_id: Id of the video episode, as a unicode string.
273
274 The following fields should only be used when the media is a track or a part of
275 a music album:
276
277 track: Title of the track.
278 track_number: Number of the track within an album or a disc, as an integer.
279 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
280 as a unicode string.
281 artist: Artist(s) of the track.
282 genre: Genre(s) of the track.
283 album: Title of the album the track belongs to.
284 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
285 album_artist: List of all artists appeared on the album (e.g.
286 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
287 and compilations).
288 disc_number: Number of the disc or other physical medium the track belongs to,
289 as an integer.
290 release_year: Year (YYYY) when the album was released.
291
292 Unless mentioned otherwise, the fields should be Unicode strings.
293
294 Unless mentioned otherwise, None is equivalent to absence of information.
295
296
297 _type "playlist" indicates multiple videos.
298 There must be a key "entries", which is a list, an iterable, or a PagedList
299 object, each element of which is a valid dictionary by this specification.
300
301 Additionally, playlists can have "title", "description" and "id" attributes
302 with the same semantics as videos (see above).
303
304
305 _type "multi_video" indicates that there are multiple videos that
306 form a single show, for examples multiple acts of an opera or TV episode.
307 It must have an entries key like a playlist and contain all the keys
308 required for a video at the same time.
309
310
311 _type "url" indicates that the video must be extracted from another
312 location, possibly by a different extractor. Its only required key is:
313 "url" - the next URL to extract.
314 The key "ie_key" can be set to the class name (minus the trailing "IE",
315 e.g. "Youtube") if the extractor class is known in advance.
316 Additionally, the dictionary may have any properties of the resolved entity
317 known in advance, for example "title" if the title of the referred video is
318 known ahead of time.
319
320
321 _type "url_transparent" entities have the same specification as "url", but
322 indicate that the given additional information is more precise than the one
323 associated with the resolved URL.
324 This is useful when a site employs a video service that hosts the video and
325 its technical metadata, but that video service does not embed a useful
326 title, description etc.
327
328
329 Subclasses of this one should re-define the _real_initialize() and
330 _real_extract() methods and define a _VALID_URL regexp.
331 Probably, they should also be added to the list of extractors.
332
333 _GEO_BYPASS attribute may be set to False in order to disable
334 geo restriction bypass mechanisms for a particular extractor.
335 Though it won't disable explicit geo restriction bypass based on
336 country code provided with geo_bypass_country. (experimental)
337
338 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
339 countries for this extractor. One of these countries will be used by
340 geo restriction bypass mechanism right away in order to bypass
341 geo restriction, of course, if the mechanism is not disabled. (experimental)
342
343 NB: both these geo attributes are experimental and may change in future
344 or be completely removed.
345
346 Finally, the _WORKING attribute should be set to False for broken IEs
347 in order to warn the users and skip the tests.
348 """
349
350 _ready = False
351 _downloader = None
352 _x_forwarded_for_ip = None
353 _GEO_BYPASS = True
354 _GEO_COUNTRIES = None
355 _WORKING = True
356
357 def __init__(self, downloader=None):
358 """Constructor. Receives an optional downloader."""
359 self._ready = False
360 self._x_forwarded_for_ip = None
361 self.set_downloader(downloader)
362
363 @classmethod
364 def suitable(cls, url):
365 """Receives a URL and returns True if suitable for this IE."""
366
367 # This does not use has/getattr intentionally - we want to know whether
368 # we have cached the regexp for *this* class, whereas getattr would also
369 # match the superclass
370 if '_VALID_URL_RE' not in cls.__dict__:
371 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
372 return cls._VALID_URL_RE.match(url) is not None
373
374 @classmethod
375 def _match_id(cls, url):
376 if '_VALID_URL_RE' not in cls.__dict__:
377 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
378 m = cls._VALID_URL_RE.match(url)
379 assert m
380 return compat_str(m.group('id'))
381
382 @classmethod
383 def working(cls):
384 """Getter method for _WORKING."""
385 return cls._WORKING
386
387 def initialize(self):
388 """Initializes an instance (authentication, etc)."""
389 self._initialize_geo_bypass(self._GEO_COUNTRIES)
390 if not self._ready:
391 self._real_initialize()
392 self._ready = True
393
394 def _initialize_geo_bypass(self, countries):
395 """
396 Initialize geo restriction bypass mechanism.
397
398 This method is used to initialize geo bypass mechanism based on faking
399 X-Forwarded-For HTTP header. A random country from provided country list
400 is selected and a random IP belonging to this country is generated. This
401 IP will be passed as X-Forwarded-For HTTP header in all subsequent
402 HTTP requests.
403
404 This method will be used for initial geo bypass mechanism initialization
405 during the instance initialization with _GEO_COUNTRIES.
406
407 You may also manually call it from extractor's code if geo countries
408 information is not available beforehand (e.g. obtained during
409 extraction) or due to some another reason.
410 """
411 if not self._x_forwarded_for_ip:
412 country_code = self._downloader.params.get('geo_bypass_country', None)
413 # If there is no explicit country for geo bypass specified and
414 # the extractor is known to be geo restricted let's fake IP
415 # as X-Forwarded-For right away.
416 if (not country_code and
417 self._GEO_BYPASS and
418 self._downloader.params.get('geo_bypass', True) and
419 countries):
420 country_code = random.choice(countries)
421 if country_code:
422 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
423 if self._downloader.params.get('verbose', False):
424 self._downloader.to_screen(
425 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
426 % (self._x_forwarded_for_ip, country_code.upper()))
427
428 def extract(self, url):
429 """Extracts URL information and returns it in list of dicts."""
430 try:
431 for _ in range(2):
432 try:
433 self.initialize()
434 ie_result = self._real_extract(url)
435 if self._x_forwarded_for_ip:
436 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
437 return ie_result
438 except GeoRestrictedError as e:
439 if self.__maybe_fake_ip_and_retry(e.countries):
440 continue
441 raise
442 except ExtractorError:
443 raise
444 except compat_http_client.IncompleteRead as e:
445 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
446 except (KeyError, StopIteration) as e:
447 raise ExtractorError('An extractor error has occurred.', cause=e)
448
449 def __maybe_fake_ip_and_retry(self, countries):
450 if (not self._downloader.params.get('geo_bypass_country', None) and
451 self._GEO_BYPASS and
452 self._downloader.params.get('geo_bypass', True) and
453 not self._x_forwarded_for_ip and
454 countries):
455 country_code = random.choice(countries)
456 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
457 if self._x_forwarded_for_ip:
458 self.report_warning(
459 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
460 % (self._x_forwarded_for_ip, country_code.upper()))
461 return True
462 return False
463
464 def set_downloader(self, downloader):
465 """Sets the downloader for this IE."""
466 self._downloader = downloader
467
468 def _real_initialize(self):
469 """Real initialization process. Redefine in subclasses."""
470 pass
471
472 def _real_extract(self, url):
473 """Real extraction process. Redefine in subclasses."""
474 pass
475
476 @classmethod
477 def ie_key(cls):
478 """A string for getting the InfoExtractor with get_info_extractor"""
479 return compat_str(cls.__name__[:-2])
480
481 @property
482 def IE_NAME(self):
483 return compat_str(type(self).__name__[:-2])
484
485 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
486 """ Returns the response handle """
487 if note is None:
488 self.report_download_webpage(video_id)
489 elif note is not False:
490 if video_id is None:
491 self.to_screen('%s' % (note,))
492 else:
493 self.to_screen('%s: %s' % (video_id, note))
494 if isinstance(url_or_request, compat_urllib_request.Request):
495 url_or_request = update_Request(
496 url_or_request, data=data, headers=headers, query=query)
497 else:
498 if query:
499 url_or_request = update_url_query(url_or_request, query)
500 if data is not None or headers:
501 url_or_request = sanitized_Request(url_or_request, data, headers)
502 try:
503 return self._downloader.urlopen(url_or_request)
504 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
505 if errnote is False:
506 return False
507 if errnote is None:
508 errnote = 'Unable to download webpage'
509
510 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
511 if fatal:
512 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
513 else:
514 self._downloader.report_warning(errmsg)
515 return False
516
517 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
518 """ Returns a tuple (page content as string, URL handle) """
519 # Strip hashes from the URL (#1038)
520 if isinstance(url_or_request, (compat_str, str)):
521 url_or_request = url_or_request.partition('#')[0]
522
523 # Some sites check X-Forwarded-For HTTP header in order to figure out
524 # the origin of the client behind proxy. This allows bypassing geo
525 # restriction by faking this header's value to IP that belongs to some
526 # geo unrestricted country. We will do so once we encounter any
527 # geo restriction error.
528 if self._x_forwarded_for_ip:
529 if 'X-Forwarded-For' not in headers:
530 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
531
532 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
533 if urlh is False:
534 assert not fatal
535 return False
536 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
537 return (content, urlh)
538
539 @staticmethod
540 def _guess_encoding_from_content(content_type, webpage_bytes):
541 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
542 if m:
543 encoding = m.group(1)
544 else:
545 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
546 webpage_bytes[:1024])
547 if m:
548 encoding = m.group(1).decode('ascii')
549 elif webpage_bytes.startswith(b'\xff\xfe'):
550 encoding = 'utf-16'
551 else:
552 encoding = 'utf-8'
553
554 return encoding
555
556 def __check_blocked(self, content):
557 first_block = content[:512]
558 if ('<title>Access to this site is blocked</title>' in content and
559 'Websense' in first_block):
560 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
561 blocked_iframe = self._html_search_regex(
562 r'<iframe src="([^"]+)"', content,
563 'Websense information URL', default=None)
564 if blocked_iframe:
565 msg += ' Visit %s for more details' % blocked_iframe
566 raise ExtractorError(msg, expected=True)
567 if '<title>The URL you requested has been blocked</title>' in first_block:
568 msg = (
569 'Access to this webpage has been blocked by Indian censorship. '
570 'Use a VPN or proxy server (with --proxy) to route around it.')
571 block_msg = self._html_search_regex(
572 r'</h1><p>(.*?)</p>',
573 content, 'block message', default=None)
574 if block_msg:
575 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
576 raise ExtractorError(msg, expected=True)
577 if ('<title>TTK :: Š”Š¾ŃŃ‚ŃƒŠæ Šŗ рŠµŃŃƒŃ€ŃŃƒ Š¾Š³Ń€Š°Š½ŠøчŠµŠ½</title>' in content and
578 'blocklist.rkn.gov.ru' in content):
579 raise ExtractorError(
580 'Access to this webpage has been blocked by decision of the Russian government. '
581 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
582 expected=True)
583
584 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
585 content_type = urlh.headers.get('Content-Type', '')
586 webpage_bytes = urlh.read()
587 if prefix is not None:
588 webpage_bytes = prefix + webpage_bytes
589 if not encoding:
590 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
591 if self._downloader.params.get('dump_intermediate_pages', False):
592 try:
593 url = url_or_request.get_full_url()
594 except AttributeError:
595 url = url_or_request
596 self.to_screen('Dumping request to ' + url)
597 dump = base64.b64encode(webpage_bytes).decode('ascii')
598 self._downloader.to_screen(dump)
599 if self._downloader.params.get('write_pages', False):
600 try:
601 url = url_or_request.get_full_url()
602 except AttributeError:
603 url = url_or_request
604 basen = '%s_%s' % (video_id, url)
605 if len(basen) > 240:
606 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
607 basen = basen[:240 - len(h)] + h
608 raw_filename = basen + '.dump'
609 filename = sanitize_filename(raw_filename, restricted=True)
610 self.to_screen('Saving request to ' + filename)
611 # Working around MAX_PATH limitation on Windows (see
612 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
613 if compat_os_name == 'nt':
614 absfilepath = os.path.abspath(filename)
615 if len(absfilepath) > 259:
616 filename = '\\\\?\\' + absfilepath
617 with open(filename, 'wb') as outf:
618 outf.write(webpage_bytes)
619
620 try:
621 content = webpage_bytes.decode(encoding, 'replace')
622 except LookupError:
623 content = webpage_bytes.decode('utf-8', 'replace')
624
625 self.__check_blocked(content)
626
627 return content
628
629 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={}):
630 """ Returns the data of the page as a string """
631 success = False
632 try_count = 0
633 while success is False:
634 try:
635 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
636 success = True
637 except compat_http_client.IncompleteRead as e:
638 try_count += 1
639 if try_count >= tries:
640 raise e
641 self._sleep(timeout, video_id)
642 if res is False:
643 return res
644 else:
645 content, _ = res
646 return content
647
648 def _download_xml(self, url_or_request, video_id,
649 note='Downloading XML', errnote='Unable to download XML',
650 transform_source=None, fatal=True, encoding=None,
651 data=None, headers={}, query={}):
652 """Return the xml as an xml.etree.ElementTree.Element"""
653 xml_string = self._download_webpage(
654 url_or_request, video_id, note, errnote, fatal=fatal,
655 encoding=encoding, data=data, headers=headers, query=query)
656 if xml_string is False:
657 return xml_string
658 return self._parse_xml(
659 xml_string, video_id, transform_source=transform_source,
660 fatal=fatal)
661
662 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
663 if transform_source:
664 xml_string = transform_source(xml_string)
665 try:
666 return compat_etree_fromstring(xml_string.encode('utf-8'))
667 except compat_xml_parse_error as ve:
668 errmsg = '%s: Failed to parse XML ' % video_id
669 if fatal:
670 raise ExtractorError(errmsg, cause=ve)
671 else:
672 self.report_warning(errmsg + str(ve))
673
674 def _download_json(self, url_or_request, video_id,
675 note='Downloading JSON metadata',
676 errnote='Unable to download JSON metadata',
677 transform_source=None,
678 fatal=True, encoding=None, data=None, headers={}, query={}):
679 json_string = self._download_webpage(
680 url_or_request, video_id, note, errnote, fatal=fatal,
681 encoding=encoding, data=data, headers=headers, query=query)
682 if (not fatal) and json_string is False:
683 return None
684 return self._parse_json(
685 json_string, video_id, transform_source=transform_source, fatal=fatal)
686
687 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
688 if transform_source:
689 json_string = transform_source(json_string)
690 try:
691 return json.loads(json_string)
692 except ValueError as ve:
693 errmsg = '%s: Failed to parse JSON ' % video_id
694 if fatal:
695 raise ExtractorError(errmsg, cause=ve)
696 else:
697 self.report_warning(errmsg + str(ve))
698
699 def report_warning(self, msg, video_id=None):
700 idstr = '' if video_id is None else '%s: ' % video_id
701 self._downloader.report_warning(
702 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
703
704 def to_screen(self, msg):
705 """Print msg to screen, prefixing it with '[ie_name]'"""
706 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
707
708 def report_extraction(self, id_or_name):
709 """Report information extraction."""
710 self.to_screen('%s: Extracting information' % id_or_name)
711
712 def report_download_webpage(self, video_id):
713 """Report webpage download."""
714 self.to_screen('%s: Downloading webpage' % video_id)
715
716 def report_age_confirmation(self):
717 """Report attempt to confirm age."""
718 self.to_screen('Confirming age')
719
720 def report_login(self):
721 """Report attempt to log in."""
722 self.to_screen('Logging in')
723
724 @staticmethod
725 def raise_login_required(msg='This video is only available for registered users'):
726 raise ExtractorError(
727 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
728 expected=True)
729
730 @staticmethod
731 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
732 raise GeoRestrictedError(msg, countries=countries)
733
734 # Methods for following #608
735 @staticmethod
736 def url_result(url, ie=None, video_id=None, video_title=None):
737 """Returns a URL that points to a page that should be processed"""
738 # TODO: ie should be the class used for getting the info
739 video_info = {'_type': 'url',
740 'url': url,
741 'ie_key': ie}
742 if video_id is not None:
743 video_info['id'] = video_id
744 if video_title is not None:
745 video_info['title'] = video_title
746 return video_info
747
748 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
749 urls = orderedSet(
750 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
751 for m in matches)
752 return self.playlist_result(
753 urls, playlist_id=playlist_id, playlist_title=playlist_title)
754
755 @staticmethod
756 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
757 """Returns a playlist"""
758 video_info = {'_type': 'playlist',
759 'entries': entries}
760 if playlist_id:
761 video_info['id'] = playlist_id
762 if playlist_title:
763 video_info['title'] = playlist_title
764 if playlist_description:
765 video_info['description'] = playlist_description
766 return video_info
767
768 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
769 """
770 Perform a regex search on the given string, using a single or a list of
771 patterns returning the first matching group.
772 In case of failure return a default value or raise a WARNING or a
773 RegexNotFoundError, depending on fatal, specifying the field name.
774 """
775 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
776 mobj = re.search(pattern, string, flags)
777 else:
778 for p in pattern:
779 mobj = re.search(p, string, flags)
780 if mobj:
781 break
782
783 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
784 _name = '\033[0;34m%s\033[0m' % name
785 else:
786 _name = name
787
788 if mobj:
789 if group is None:
790 # return the first matching group
791 return next(g for g in mobj.groups() if g is not None)
792 else:
793 return mobj.group(group)
794 elif default is not NO_DEFAULT:
795 return default
796 elif fatal:
797 raise RegexNotFoundError('Unable to extract %s' % _name)
798 else:
799 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
800 return None
801
802 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
803 """
804 Like _search_regex, but strips HTML tags and unescapes entities.
805 """
806 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
807 if res:
808 return clean_html(res).strip()
809 else:
810 return res
811
812 def _get_netrc_login_info(self, netrc_machine=None):
813 username = None
814 password = None
815 netrc_machine = netrc_machine or self._NETRC_MACHINE
816
817 if self._downloader.params.get('usenetrc', False):
818 try:
819 info = netrc.netrc().authenticators(netrc_machine)
820 if info is not None:
821 username = info[0]
822 password = info[2]
823 else:
824 raise netrc.NetrcParseError(
825 'No authenticators for %s' % netrc_machine)
826 except (IOError, netrc.NetrcParseError) as err:
827 self._downloader.report_warning(
828 'parsing .netrc: %s' % error_to_compat_str(err))
829
830 return username, password
831
832 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
833 """
834 Get the login info as (username, password)
835 First look for the manually specified credentials using username_option
836 and password_option as keys in params dictionary. If no such credentials
837 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
838 value.
839 If there's no info available, return (None, None)
840 """
841 if self._downloader is None:
842 return (None, None)
843
844 downloader_params = self._downloader.params
845
846 # Attempt to use provided username and password or .netrc data
847 if downloader_params.get(username_option) is not None:
848 username = downloader_params[username_option]
849 password = downloader_params[password_option]
850 else:
851 username, password = self._get_netrc_login_info(netrc_machine)
852
853 return username, password
854
855 def _get_tfa_info(self, note='two-factor verification code'):
856 """
857 Get the two-factor authentication info
858 TODO - asking the user will be required for sms/phone verify
859 currently just uses the command line option
860 If there's no info available, return None
861 """
862 if self._downloader is None:
863 return None
864 downloader_params = self._downloader.params
865
866 if downloader_params.get('twofactor') is not None:
867 return downloader_params['twofactor']
868
869 return compat_getpass('Type %s and press [Return]: ' % note)
870
871 # Helper functions for extracting OpenGraph info
872 @staticmethod
873 def _og_regexes(prop):
874 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
875 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
876 % {'prop': re.escape(prop)})
877 template = r'<meta[^>]+?%s[^>]+?%s'
878 return [
879 template % (property_re, content_re),
880 template % (content_re, property_re),
881 ]
882
883 @staticmethod
884 def _meta_regex(prop):
885 return r'''(?isx)<meta
886 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
887 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
888
889 def _og_search_property(self, prop, html, name=None, **kargs):
890 if not isinstance(prop, (list, tuple)):
891 prop = [prop]
892 if name is None:
893 name = 'OpenGraph %s' % prop[0]
894 og_regexes = []
895 for p in prop:
896 og_regexes.extend(self._og_regexes(p))
897 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
898 if escaped is None:
899 return None
900 return unescapeHTML(escaped)
901
902 def _og_search_thumbnail(self, html, **kargs):
903 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
904
905 def _og_search_description(self, html, **kargs):
906 return self._og_search_property('description', html, fatal=False, **kargs)
907
908 def _og_search_title(self, html, **kargs):
909 return self._og_search_property('title', html, **kargs)
910
911 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
912 regexes = self._og_regexes('video') + self._og_regexes('video:url')
913 if secure:
914 regexes = self._og_regexes('video:secure_url') + regexes
915 return self._html_search_regex(regexes, html, name, **kargs)
916
917 def _og_search_url(self, html, **kargs):
918 return self._og_search_property('url', html, **kargs)
919
920 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
921 if not isinstance(name, (list, tuple)):
922 name = [name]
923 if display_name is None:
924 display_name = name[0]
925 return self._html_search_regex(
926 [self._meta_regex(n) for n in name],
927 html, display_name, fatal=fatal, group='content', **kwargs)
928
929 def _dc_search_uploader(self, html):
930 return self._html_search_meta('dc.creator', html, 'uploader')
931
932 def _rta_search(self, html):
933 # See http://www.rtalabel.org/index.php?content=howtofaq#single
934 if re.search(r'(?ix)<meta\s+name="rating"\s+'
935 r' content="RTA-5042-1996-1400-1577-RTA"',
936 html):
937 return 18
938 return 0
939
940 def _media_rating_search(self, html):
941 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
942 rating = self._html_search_meta('rating', html)
943
944 if not rating:
945 return None
946
947 RATING_TABLE = {
948 'safe for kids': 0,
949 'general': 8,
950 '14 years': 14,
951 'mature': 17,
952 'restricted': 19,
953 }
954 return RATING_TABLE.get(rating.lower())
955
956 def _family_friendly_search(self, html):
957 # See http://schema.org/VideoObject
958 family_friendly = self._html_search_meta(
959 'isFamilyFriendly', html, default=None)
960
961 if not family_friendly:
962 return None
963
964 RATING_TABLE = {
965 '1': 0,
966 'true': 0,
967 '0': 18,
968 'false': 18,
969 }
970 return RATING_TABLE.get(family_friendly.lower())
971
972 def _twitter_search_player(self, html):
973 return self._html_search_meta('twitter:player', html,
974 'twitter card player')
975
976 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
977 json_ld = self._search_regex(
978 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
979 html, 'JSON-LD', group='json_ld', **kwargs)
980 default = kwargs.get('default', NO_DEFAULT)
981 if not json_ld:
982 return default if default is not NO_DEFAULT else {}
983 # JSON-LD may be malformed and thus `fatal` should be respected.
984 # At the same time `default` may be passed that assumes `fatal=False`
985 # for _search_regex. Let's simulate the same behavior here as well.
986 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
987 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
988
989 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
990 if isinstance(json_ld, compat_str):
991 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
992 if not json_ld:
993 return {}
994 info = {}
995 if not isinstance(json_ld, (list, tuple, dict)):
996 return info
997 if isinstance(json_ld, dict):
998 json_ld = [json_ld]
999
1000 def extract_video_object(e):
1001 assert e['@type'] == 'VideoObject'
1002 info.update({
1003 'url': e.get('contentUrl'),
1004 'title': unescapeHTML(e.get('name')),
1005 'description': unescapeHTML(e.get('description')),
1006 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
1007 'duration': parse_duration(e.get('duration')),
1008 'timestamp': unified_timestamp(e.get('uploadDate')),
1009 'filesize': float_or_none(e.get('contentSize')),
1010 'tbr': int_or_none(e.get('bitrate')),
1011 'width': int_or_none(e.get('width')),
1012 'height': int_or_none(e.get('height')),
1013 'view_count': int_or_none(e.get('interactionCount')),
1014 })
1015
1016 for e in json_ld:
1017 if e.get('@context') == 'http://schema.org':
1018 item_type = e.get('@type')
1019 if expected_type is not None and expected_type != item_type:
1020 return info
1021 if item_type in ('TVEpisode', 'Episode'):
1022 info.update({
1023 'episode': unescapeHTML(e.get('name')),
1024 'episode_number': int_or_none(e.get('episodeNumber')),
1025 'description': unescapeHTML(e.get('description')),
1026 })
1027 part_of_season = e.get('partOfSeason')
1028 if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1029 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
1030 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
1031 if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1032 info['series'] = unescapeHTML(part_of_series.get('name'))
1033 elif item_type == 'Article':
1034 info.update({
1035 'timestamp': parse_iso8601(e.get('datePublished')),
1036 'title': unescapeHTML(e.get('headline')),
1037 'description': unescapeHTML(e.get('articleBody')),
1038 })
1039 elif item_type == 'VideoObject':
1040 extract_video_object(e)
1041 continue
1042 video = e.get('video')
1043 if isinstance(video, dict) and video.get('@type') == 'VideoObject':
1044 extract_video_object(video)
1045 break
1046 return dict((k, v) for k, v in info.items() if v is not None)
1047
1048 @staticmethod
1049 def _hidden_inputs(html):
1050 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1051 hidden_inputs = {}
1052 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1053 attrs = extract_attributes(input)
1054 if not input:
1055 continue
1056 if attrs.get('type') not in ('hidden', 'submit'):
1057 continue
1058 name = attrs.get('name') or attrs.get('id')
1059 value = attrs.get('value')
1060 if name and value is not None:
1061 hidden_inputs[name] = value
1062 return hidden_inputs
1063
1064 def _form_hidden_inputs(self, form_id, html):
1065 form = self._search_regex(
1066 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1067 html, '%s form' % form_id, group='form')
1068 return self._hidden_inputs(form)
1069
1070 def _sort_formats(self, formats, field_preference=None):
1071 if not formats:
1072 raise ExtractorError('No video formats found')
1073
1074 for f in formats:
1075 # Automatically determine tbr when missing based on abr and vbr (improves
1076 # formats sorting in some cases)
1077 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1078 f['tbr'] = f['abr'] + f['vbr']
1079
1080 def _formats_key(f):
1081 # TODO remove the following workaround
1082 from ..utils import determine_ext
1083 if not f.get('ext') and 'url' in f:
1084 f['ext'] = determine_ext(f['url'])
1085
1086 if isinstance(field_preference, (list, tuple)):
1087 return tuple(
1088 f.get(field)
1089 if f.get(field) is not None
1090 else ('' if field == 'format_id' else -1)
1091 for field in field_preference)
1092
1093 preference = f.get('preference')
1094 if preference is None:
1095 preference = 0
1096 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1097 preference -= 0.5
1098
1099 protocol = f.get('protocol') or determine_protocol(f)
1100 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1101
1102 if f.get('vcodec') == 'none': # audio only
1103 preference -= 50
1104 if self._downloader.params.get('prefer_free_formats'):
1105 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1106 else:
1107 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1108 ext_preference = 0
1109 try:
1110 audio_ext_preference = ORDER.index(f['ext'])
1111 except ValueError:
1112 audio_ext_preference = -1
1113 else:
1114 if f.get('acodec') == 'none': # video only
1115 preference -= 40
1116 if self._downloader.params.get('prefer_free_formats'):
1117 ORDER = ['flv', 'mp4', 'webm']
1118 else:
1119 ORDER = ['webm', 'flv', 'mp4']
1120 try:
1121 ext_preference = ORDER.index(f['ext'])
1122 except ValueError:
1123 ext_preference = -1
1124 audio_ext_preference = 0
1125
1126 return (
1127 preference,
1128 f.get('language_preference') if f.get('language_preference') is not None else -1,
1129 f.get('quality') if f.get('quality') is not None else -1,
1130 f.get('tbr') if f.get('tbr') is not None else -1,
1131 f.get('filesize') if f.get('filesize') is not None else -1,
1132 f.get('vbr') if f.get('vbr') is not None else -1,
1133 f.get('height') if f.get('height') is not None else -1,
1134 f.get('width') if f.get('width') is not None else -1,
1135 proto_preference,
1136 ext_preference,
1137 f.get('abr') if f.get('abr') is not None else -1,
1138 audio_ext_preference,
1139 f.get('fps') if f.get('fps') is not None else -1,
1140 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1141 f.get('source_preference') if f.get('source_preference') is not None else -1,
1142 f.get('format_id') if f.get('format_id') is not None else '',
1143 )
1144 formats.sort(key=_formats_key)
1145
1146 def _check_formats(self, formats, video_id):
1147 if formats:
1148 formats[:] = filter(
1149 lambda f: self._is_valid_url(
1150 f['url'], video_id,
1151 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1152 formats)
1153
1154 @staticmethod
1155 def _remove_duplicate_formats(formats):
1156 format_urls = set()
1157 unique_formats = []
1158 for f in formats:
1159 if f['url'] not in format_urls:
1160 format_urls.add(f['url'])
1161 unique_formats.append(f)
1162 formats[:] = unique_formats
1163
1164 def _is_valid_url(self, url, video_id, item='video', headers={}):
1165 url = self._proto_relative_url(url, scheme='http:')
1166 # For now assume non HTTP(S) URLs always valid
1167 if not (url.startswith('http://') or url.startswith('https://')):
1168 return True
1169 try:
1170 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1171 return True
1172 except ExtractorError as e:
1173 if isinstance(e.cause, compat_urllib_error.URLError):
1174 self.to_screen(
1175 '%s: %s URL is invalid, skipping' % (video_id, item))
1176 return False
1177 raise
1178
1179 def http_scheme(self):
1180 """ Either "http:" or "https:", depending on the user's preferences """
1181 return (
1182 'http:'
1183 if self._downloader.params.get('prefer_insecure', False)
1184 else 'https:')
1185
1186 def _proto_relative_url(self, url, scheme=None):
1187 if url is None:
1188 return url
1189 if url.startswith('//'):
1190 if scheme is None:
1191 scheme = self.http_scheme()
1192 return scheme + url
1193 else:
1194 return url
1195
1196 def _sleep(self, timeout, video_id, msg_template=None):
1197 if msg_template is None:
1198 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1199 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1200 self.to_screen(msg)
1201 time.sleep(timeout)
1202
1203 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1204 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1205 fatal=True, m3u8_id=None):
1206 manifest = self._download_xml(
1207 manifest_url, video_id, 'Downloading f4m manifest',
1208 'Unable to download f4m manifest',
1209 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1210 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1211 transform_source=transform_source,
1212 fatal=fatal)
1213
1214 if manifest is False:
1215 return []
1216
1217 return self._parse_f4m_formats(
1218 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1219 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1220
1221 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1222 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1223 fatal=True, m3u8_id=None):
1224 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1225 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1226 if akamai_pv is not None and ';' in akamai_pv.text:
1227 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1228 if playerVerificationChallenge.strip() != '':
1229 return []
1230
1231 formats = []
1232 manifest_version = '1.0'
1233 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1234 if not media_nodes:
1235 manifest_version = '2.0'
1236 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1237 # Remove unsupported DRM protected media from final formats
1238 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1239 media_nodes = remove_encrypted_media(media_nodes)
1240 if not media_nodes:
1241 return formats
1242 base_url = xpath_text(
1243 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1244 'base URL', default=None)
1245 if base_url:
1246 base_url = base_url.strip()
1247
1248 bootstrap_info = xpath_element(
1249 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1250 'bootstrap info', default=None)
1251
1252 vcodec = None
1253 mime_type = xpath_text(
1254 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1255 'base URL', default=None)
1256 if mime_type and mime_type.startswith('audio/'):
1257 vcodec = 'none'
1258
1259 for i, media_el in enumerate(media_nodes):
1260 tbr = int_or_none(media_el.attrib.get('bitrate'))
1261 width = int_or_none(media_el.attrib.get('width'))
1262 height = int_or_none(media_el.attrib.get('height'))
1263 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1264 # If <bootstrapInfo> is present, the specified f4m is a
1265 # stream-level manifest, and only set-level manifests may refer to
1266 # external resources. See section 11.4 and section 4 of F4M spec
1267 if bootstrap_info is None:
1268 media_url = None
1269 # @href is introduced in 2.0, see section 11.6 of F4M spec
1270 if manifest_version == '2.0':
1271 media_url = media_el.attrib.get('href')
1272 if media_url is None:
1273 media_url = media_el.attrib.get('url')
1274 if not media_url:
1275 continue
1276 manifest_url = (
1277 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1278 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1279 # If media_url is itself a f4m manifest do the recursive extraction
1280 # since bitrates in parent manifest (this one) and media_url manifest
1281 # may differ leading to inability to resolve the format by requested
1282 # bitrate in f4m downloader
1283 ext = determine_ext(manifest_url)
1284 if ext == 'f4m':
1285 f4m_formats = self._extract_f4m_formats(
1286 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1287 transform_source=transform_source, fatal=fatal)
1288 # Sometimes stream-level manifest contains single media entry that
1289 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1290 # At the same time parent's media entry in set-level manifest may
1291 # contain it. We will copy it from parent in such cases.
1292 if len(f4m_formats) == 1:
1293 f = f4m_formats[0]
1294 f.update({
1295 'tbr': f.get('tbr') or tbr,
1296 'width': f.get('width') or width,
1297 'height': f.get('height') or height,
1298 'format_id': f.get('format_id') if not tbr else format_id,
1299 'vcodec': vcodec,
1300 })
1301 formats.extend(f4m_formats)
1302 continue
1303 elif ext == 'm3u8':
1304 formats.extend(self._extract_m3u8_formats(
1305 manifest_url, video_id, 'mp4', preference=preference,
1306 m3u8_id=m3u8_id, fatal=fatal))
1307 continue
1308 formats.append({
1309 'format_id': format_id,
1310 'url': manifest_url,
1311 'manifest_url': manifest_url,
1312 'ext': 'flv' if bootstrap_info is not None else None,
1313 'tbr': tbr,
1314 'width': width,
1315 'height': height,
1316 'vcodec': vcodec,
1317 'preference': preference,
1318 })
1319 return formats
1320
1321 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1322 return {
1323 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1324 'url': m3u8_url,
1325 'ext': ext,
1326 'protocol': 'm3u8',
1327 'preference': preference - 100 if preference else -100,
1328 'resolution': 'multiple',
1329 'format_note': 'Quality selection URL',
1330 }
1331
1332 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1333 entry_protocol='m3u8', preference=None,
1334 m3u8_id=None, note=None, errnote=None,
1335 fatal=True, live=False):
1336 res = self._download_webpage_handle(
1337 m3u8_url, video_id,
1338 note=note or 'Downloading m3u8 information',
1339 errnote=errnote or 'Failed to download m3u8 information',
1340 fatal=fatal)
1341
1342 if res is False:
1343 return []
1344
1345 m3u8_doc, urlh = res
1346 m3u8_url = urlh.geturl()
1347
1348 return self._parse_m3u8_formats(
1349 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1350 preference=preference, m3u8_id=m3u8_id, live=live)
1351
1352 def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
1353 entry_protocol='m3u8', preference=None,
1354 m3u8_id=None, live=False):
1355 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1356 return []
1357
1358 formats = []
1359
1360 format_url = lambda u: (
1361 u
1362 if re.match(r'^https?://', u)
1363 else compat_urlparse.urljoin(m3u8_url, u))
1364
1365 # References:
1366 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1367 # 2. https://github.com/rg3/youtube-dl/issues/12211
1368
1369 # We should try extracting formats only from master playlists [1, 4.3.4],
1370 # i.e. playlists that describe available qualities. On the other hand
1371 # media playlists [1, 4.3.3] should be returned as is since they contain
1372 # just the media without qualities renditions.
1373 # Fortunately, master playlist can be easily distinguished from media
1374 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1375 # master playlist tags MUST NOT appear in a media playist and vice versa.
1376 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1377 # media playlist and MUST NOT appear in master playlist thus we can
1378 # clearly detect media playlist with this criterion.
1379
1380 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1381 return [{
1382 'url': m3u8_url,
1383 'format_id': m3u8_id,
1384 'ext': ext,
1385 'protocol': entry_protocol,
1386 'preference': preference,
1387 }]
1388
1389 groups = {}
1390 last_stream_inf = {}
1391
1392 def extract_media(x_media_line):
1393 media = parse_m3u8_attributes(x_media_line)
1394 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
1395 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
1396 if not (media_type and group_id and name):
1397 return
1398 groups.setdefault(group_id, []).append(media)
1399 if media_type not in ('VIDEO', 'AUDIO'):
1400 return
1401 media_url = media.get('URI')
1402 if media_url:
1403 format_id = []
1404 for v in (group_id, name):
1405 if v:
1406 format_id.append(v)
1407 f = {
1408 'format_id': '-'.join(format_id),
1409 'url': format_url(media_url),
1410 'manifest_url': m3u8_url,
1411 'language': media.get('LANGUAGE'),
1412 'ext': ext,
1413 'protocol': entry_protocol,
1414 'preference': preference,
1415 }
1416 if media_type == 'AUDIO':
1417 f['vcodec'] = 'none'
1418 formats.append(f)
1419
1420 def build_stream_name():
1421 # Despite specification does not mention NAME attribute for
1422 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
1423 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
1424 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
1425 stream_name = last_stream_inf.get('NAME')
1426 if stream_name:
1427 return stream_name
1428 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
1429 # from corresponding rendition group
1430 stream_group_id = last_stream_inf.get('VIDEO')
1431 if not stream_group_id:
1432 return
1433 stream_group = groups.get(stream_group_id)
1434 if not stream_group:
1435 return stream_group_id
1436 rendition = stream_group[0]
1437 return rendition.get('NAME') or stream_group_id
1438
1439 for line in m3u8_doc.splitlines():
1440 if line.startswith('#EXT-X-STREAM-INF:'):
1441 last_stream_inf = parse_m3u8_attributes(line)
1442 elif line.startswith('#EXT-X-MEDIA:'):
1443 extract_media(line)
1444 elif line.startswith('#') or not line.strip():
1445 continue
1446 else:
1447 tbr = float_or_none(
1448 last_stream_inf.get('AVERAGE-BANDWIDTH') or
1449 last_stream_inf.get('BANDWIDTH'), scale=1000)
1450 format_id = []
1451 if m3u8_id:
1452 format_id.append(m3u8_id)
1453 stream_name = build_stream_name()
1454 # Bandwidth of live streams may differ over time thus making
1455 # format_id unpredictable. So it's better to keep provided
1456 # format_id intact.
1457 if not live:
1458 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1459 manifest_url = format_url(line.strip())
1460 f = {
1461 'format_id': '-'.join(format_id),
1462 'url': manifest_url,
1463 'manifest_url': m3u8_url,
1464 'tbr': tbr,
1465 'ext': ext,
1466 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
1467 'protocol': entry_protocol,
1468 'preference': preference,
1469 }
1470 resolution = last_stream_inf.get('RESOLUTION')
1471 if resolution:
1472 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1473 if mobj:
1474 f['width'] = int(mobj.group('width'))
1475 f['height'] = int(mobj.group('height'))
1476 # Unified Streaming Platform
1477 mobj = re.search(
1478 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1479 if mobj:
1480 abr, vbr = mobj.groups()
1481 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1482 f.update({
1483 'vbr': vbr,
1484 'abr': abr,
1485 })
1486 codecs = parse_codecs(last_stream_inf.get('CODECS'))
1487 f.update(codecs)
1488 audio_group_id = last_stream_inf.get('AUDIO')
1489 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
1490 # references a rendition group MUST have a CODECS attribute.
1491 # However, this is not always respected, for example, [2]
1492 # contains EXT-X-STREAM-INF tag which references AUDIO
1493 # rendition group but does not have CODECS and despite
1494 # referencing audio group an audio group, it represents
1495 # a complete (with audio and video) format. So, for such cases
1496 # we will ignore references to rendition groups and treat them
1497 # as complete formats.
1498 if audio_group_id and codecs and f.get('vcodec') != 'none':
1499 audio_group = groups.get(audio_group_id)
1500 if audio_group and audio_group[0].get('URI'):
1501 # TODO: update acodec for audio only formats with
1502 # the same GROUP-ID
1503 f['acodec'] = 'none'
1504 formats.append(f)
1505 last_stream_inf = {}
1506 return formats
1507
1508 @staticmethod
1509 def _xpath_ns(path, namespace=None):
1510 if not namespace:
1511 return path
1512 out = []
1513 for c in path.split('/'):
1514 if not c or c == '.':
1515 out.append(c)
1516 else:
1517 out.append('{%s}%s' % (namespace, c))
1518 return '/'.join(out)
1519
1520 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1521 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1522
1523 if smil is False:
1524 assert not fatal
1525 return []
1526
1527 namespace = self._parse_smil_namespace(smil)
1528
1529 return self._parse_smil_formats(
1530 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1531
1532 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1533 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1534 if smil is False:
1535 return {}
1536 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1537
1538 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1539 return self._download_xml(
1540 smil_url, video_id, 'Downloading SMIL file',
1541 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1542
1543 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1544 namespace = self._parse_smil_namespace(smil)
1545
1546 formats = self._parse_smil_formats(
1547 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1548 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1549
1550 video_id = os.path.splitext(url_basename(smil_url))[0]
1551 title = None
1552 description = None
1553 upload_date = None
1554 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1555 name = meta.attrib.get('name')
1556 content = meta.attrib.get('content')
1557 if not name or not content:
1558 continue
1559 if not title and name == 'title':
1560 title = content
1561 elif not description and name in ('description', 'abstract'):
1562 description = content
1563 elif not upload_date and name == 'date':
1564 upload_date = unified_strdate(content)
1565
1566 thumbnails = [{
1567 'id': image.get('type'),
1568 'url': image.get('src'),
1569 'width': int_or_none(image.get('width')),
1570 'height': int_or_none(image.get('height')),
1571 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1572
1573 return {
1574 'id': video_id,
1575 'title': title or video_id,
1576 'description': description,
1577 'upload_date': upload_date,
1578 'thumbnails': thumbnails,
1579 'formats': formats,
1580 'subtitles': subtitles,
1581 }
1582
1583 def _parse_smil_namespace(self, smil):
1584 return self._search_regex(
1585 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1586
1587 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1588 base = smil_url
1589 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1590 b = meta.get('base') or meta.get('httpBase')
1591 if b:
1592 base = b
1593 break
1594
1595 formats = []
1596 rtmp_count = 0
1597 http_count = 0
1598 m3u8_count = 0
1599
1600 srcs = []
1601 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1602 for medium in media:
1603 src = medium.get('src')
1604 if not src or src in srcs:
1605 continue
1606 srcs.append(src)
1607
1608 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1609 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1610 width = int_or_none(medium.get('width'))
1611 height = int_or_none(medium.get('height'))
1612 proto = medium.get('proto')
1613 ext = medium.get('ext')
1614 src_ext = determine_ext(src)
1615 streamer = medium.get('streamer') or base
1616
1617 if proto == 'rtmp' or streamer.startswith('rtmp'):
1618 rtmp_count += 1
1619 formats.append({
1620 'url': streamer,
1621 'play_path': src,
1622 'ext': 'flv',
1623 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1624 'tbr': bitrate,
1625 'filesize': filesize,
1626 'width': width,
1627 'height': height,
1628 })
1629 if transform_rtmp_url:
1630 streamer, src = transform_rtmp_url(streamer, src)
1631 formats[-1].update({
1632 'url': streamer,
1633 'play_path': src,
1634 })
1635 continue
1636
1637 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1638 src_url = src_url.strip()
1639
1640 if proto == 'm3u8' or src_ext == 'm3u8':
1641 m3u8_formats = self._extract_m3u8_formats(
1642 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1643 if len(m3u8_formats) == 1:
1644 m3u8_count += 1
1645 m3u8_formats[0].update({
1646 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1647 'tbr': bitrate,
1648 'width': width,
1649 'height': height,
1650 })
1651 formats.extend(m3u8_formats)
1652 continue
1653
1654 if src_ext == 'f4m':
1655 f4m_url = src_url
1656 if not f4m_params:
1657 f4m_params = {
1658 'hdcore': '3.2.0',
1659 'plugin': 'flowplayer-3.2.0.1',
1660 }
1661 f4m_url += '&' if '?' in f4m_url else '?'
1662 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1663 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1664 continue
1665
1666 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1667 http_count += 1
1668 formats.append({
1669 'url': src_url,
1670 'ext': ext or src_ext or 'flv',
1671 'format_id': 'http-%d' % (bitrate or http_count),
1672 'tbr': bitrate,
1673 'filesize': filesize,
1674 'width': width,
1675 'height': height,
1676 })
1677 continue
1678
1679 return formats
1680
1681 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1682 urls = []
1683 subtitles = {}
1684 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1685 src = textstream.get('src')
1686 if not src or src in urls:
1687 continue
1688 urls.append(src)
1689 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1690 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1691 subtitles.setdefault(lang, []).append({
1692 'url': src,
1693 'ext': ext,
1694 })
1695 return subtitles
1696
1697 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1698 xspf = self._download_xml(
1699 playlist_url, playlist_id, 'Downloading xpsf playlist',
1700 'Unable to download xspf manifest', fatal=fatal)
1701 if xspf is False:
1702 return []
1703 return self._parse_xspf(xspf, playlist_id)
1704
1705 def _parse_xspf(self, playlist, playlist_id):
1706 NS_MAP = {
1707 'xspf': 'http://xspf.org/ns/0/',
1708 's1': 'http://static.streamone.nl/player/ns/0',
1709 }
1710
1711 entries = []
1712 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1713 title = xpath_text(
1714 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1715 description = xpath_text(
1716 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1717 thumbnail = xpath_text(
1718 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1719 duration = float_or_none(
1720 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1721
1722 formats = [{
1723 'url': location.text,
1724 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1725 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1726 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1727 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1728 self._sort_formats(formats)
1729
1730 entries.append({
1731 'id': playlist_id,
1732 'title': title,
1733 'description': description,
1734 'thumbnail': thumbnail,
1735 'duration': duration,
1736 'formats': formats,
1737 })
1738 return entries
1739
1740 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1741 res = self._download_webpage_handle(
1742 mpd_url, video_id,
1743 note=note or 'Downloading MPD manifest',
1744 errnote=errnote or 'Failed to download MPD manifest',
1745 fatal=fatal)
1746 if res is False:
1747 return []
1748 mpd, urlh = res
1749 mpd_base_url = base_url(urlh.geturl())
1750
1751 return self._parse_mpd_formats(
1752 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1753 formats_dict=formats_dict, mpd_url=mpd_url)
1754
1755 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1756 """
1757 Parse formats from MPD manifest.
1758 References:
1759 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1760 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1761 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1762 """
1763 if mpd_doc.get('type') == 'dynamic':
1764 return []
1765
1766 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1767
1768 def _add_ns(path):
1769 return self._xpath_ns(path, namespace)
1770
1771 def is_drm_protected(element):
1772 return element.find(_add_ns('ContentProtection')) is not None
1773
1774 def extract_multisegment_info(element, ms_parent_info):
1775 ms_info = ms_parent_info.copy()
1776
1777 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1778 # common attributes and elements. We will only extract relevant
1779 # for us.
1780 def extract_common(source):
1781 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1782 if segment_timeline is not None:
1783 s_e = segment_timeline.findall(_add_ns('S'))
1784 if s_e:
1785 ms_info['total_number'] = 0
1786 ms_info['s'] = []
1787 for s in s_e:
1788 r = int(s.get('r', 0))
1789 ms_info['total_number'] += 1 + r
1790 ms_info['s'].append({
1791 't': int(s.get('t', 0)),
1792 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1793 'd': int(s.attrib['d']),
1794 'r': r,
1795 })
1796 start_number = source.get('startNumber')
1797 if start_number:
1798 ms_info['start_number'] = int(start_number)
1799 timescale = source.get('timescale')
1800 if timescale:
1801 ms_info['timescale'] = int(timescale)
1802 segment_duration = source.get('duration')
1803 if segment_duration:
1804 ms_info['segment_duration'] = float(segment_duration)
1805
1806 def extract_Initialization(source):
1807 initialization = source.find(_add_ns('Initialization'))
1808 if initialization is not None:
1809 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1810
1811 segment_list = element.find(_add_ns('SegmentList'))
1812 if segment_list is not None:
1813 extract_common(segment_list)
1814 extract_Initialization(segment_list)
1815 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1816 if segment_urls_e:
1817 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1818 else:
1819 segment_template = element.find(_add_ns('SegmentTemplate'))
1820 if segment_template is not None:
1821 extract_common(segment_template)
1822 media = segment_template.get('media')
1823 if media:
1824 ms_info['media'] = media
1825 initialization = segment_template.get('initialization')
1826 if initialization:
1827 ms_info['initialization'] = initialization
1828 else:
1829 extract_Initialization(segment_template)
1830 return ms_info
1831
1832 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1833 formats = []
1834 for period in mpd_doc.findall(_add_ns('Period')):
1835 period_duration = parse_duration(period.get('duration')) or mpd_duration
1836 period_ms_info = extract_multisegment_info(period, {
1837 'start_number': 1,
1838 'timescale': 1,
1839 })
1840 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1841 if is_drm_protected(adaptation_set):
1842 continue
1843 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1844 for representation in adaptation_set.findall(_add_ns('Representation')):
1845 if is_drm_protected(representation):
1846 continue
1847 representation_attrib = adaptation_set.attrib.copy()
1848 representation_attrib.update(representation.attrib)
1849 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1850 mime_type = representation_attrib['mimeType']
1851 content_type = mime_type.split('/')[0]
1852 if content_type == 'text':
1853 # TODO implement WebVTT downloading
1854 pass
1855 elif content_type in ('video', 'audio'):
1856 base_url = ''
1857 for element in (representation, adaptation_set, period, mpd_doc):
1858 base_url_e = element.find(_add_ns('BaseURL'))
1859 if base_url_e is not None:
1860 base_url = base_url_e.text + base_url
1861 if re.match(r'^https?://', base_url):
1862 break
1863 if mpd_base_url and not re.match(r'^https?://', base_url):
1864 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1865 mpd_base_url += '/'
1866 base_url = mpd_base_url + base_url
1867 representation_id = representation_attrib.get('id')
1868 lang = representation_attrib.get('lang')
1869 url_el = representation.find(_add_ns('BaseURL'))
1870 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1871 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1872 f = {
1873 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1874 'url': base_url,
1875 'manifest_url': mpd_url,
1876 'ext': mimetype2ext(mime_type),
1877 'width': int_or_none(representation_attrib.get('width')),
1878 'height': int_or_none(representation_attrib.get('height')),
1879 'tbr': float_or_none(bandwidth, 1000),
1880 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1881 'fps': int_or_none(representation_attrib.get('frameRate')),
1882 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1883 'format_note': 'DASH %s' % content_type,
1884 'filesize': filesize,
1885 }
1886 f.update(parse_codecs(representation_attrib.get('codecs')))
1887 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1888
1889 def prepare_template(template_name, identifiers):
1890 t = representation_ms_info[template_name]
1891 t = t.replace('$RepresentationID$', representation_id)
1892 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1893 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1894 t.replace('$$', '$')
1895 return t
1896
1897 # @initialization is a regular template like @media one
1898 # so it should be handled just the same way (see
1899 # https://github.com/rg3/youtube-dl/issues/11605)
1900 if 'initialization' in representation_ms_info:
1901 initialization_template = prepare_template(
1902 'initialization',
1903 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1904 # $Time$ shall not be included for @initialization thus
1905 # only $Bandwidth$ remains
1906 ('Bandwidth', ))
1907 representation_ms_info['initialization_url'] = initialization_template % {
1908 'Bandwidth': bandwidth,
1909 }
1910
1911 def location_key(location):
1912 return 'url' if re.match(r'^https?://', location) else 'path'
1913
1914 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1915
1916 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1917 media_location_key = location_key(media_template)
1918
1919 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1920 # can't be used at the same time
1921 if '%(Number' in media_template and 's' not in representation_ms_info:
1922 segment_duration = None
1923 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
1924 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1925 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1926 representation_ms_info['fragments'] = [{
1927 media_location_key: media_template % {
1928 'Number': segment_number,
1929 'Bandwidth': bandwidth,
1930 },
1931 'duration': segment_duration,
1932 } for segment_number in range(
1933 representation_ms_info['start_number'],
1934 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1935 else:
1936 # $Number*$ or $Time$ in media template with S list available
1937 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1938 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1939 representation_ms_info['fragments'] = []
1940 segment_time = 0
1941 segment_d = None
1942 segment_number = representation_ms_info['start_number']
1943
1944 def add_segment_url():
1945 segment_url = media_template % {
1946 'Time': segment_time,
1947 'Bandwidth': bandwidth,
1948 'Number': segment_number,
1949 }
1950 representation_ms_info['fragments'].append({
1951 media_location_key: segment_url,
1952 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1953 })
1954
1955 for num, s in enumerate(representation_ms_info['s']):
1956 segment_time = s.get('t') or segment_time
1957 segment_d = s['d']
1958 add_segment_url()
1959 segment_number += 1
1960 for r in range(s.get('r', 0)):
1961 segment_time += segment_d
1962 add_segment_url()
1963 segment_number += 1
1964 segment_time += segment_d
1965 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1966 # No media template
1967 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1968 # or any YouTube dashsegments video
1969 fragments = []
1970 segment_index = 0
1971 timescale = representation_ms_info['timescale']
1972 for s in representation_ms_info['s']:
1973 duration = float_or_none(s['d'], timescale)
1974 for r in range(s.get('r', 0) + 1):
1975 segment_uri = representation_ms_info['segment_urls'][segment_index]
1976 fragments.append({
1977 location_key(segment_uri): segment_uri,
1978 'duration': duration,
1979 })
1980 segment_index += 1
1981 representation_ms_info['fragments'] = fragments
1982 # NB: MPD manifest may contain direct URLs to unfragmented media.
1983 # No fragments key is present in this case.
1984 if 'fragments' in representation_ms_info:
1985 f.update({
1986 'fragment_base_url': base_url,
1987 'fragments': [],
1988 'protocol': 'http_dash_segments',
1989 })
1990 if 'initialization_url' in representation_ms_info:
1991 initialization_url = representation_ms_info['initialization_url']
1992 if not f.get('url'):
1993 f['url'] = initialization_url
1994 f['fragments'].append({location_key(initialization_url): initialization_url})
1995 f['fragments'].extend(representation_ms_info['fragments'])
1996 try:
1997 existing_format = next(
1998 fo for fo in formats
1999 if fo['format_id'] == representation_id)
2000 except StopIteration:
2001 full_info = formats_dict.get(representation_id, {}).copy()
2002 full_info.update(f)
2003 formats.append(full_info)
2004 else:
2005 existing_format.update(f)
2006 else:
2007 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2008 return formats
2009
2010 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
2011 res = self._download_webpage_handle(
2012 ism_url, video_id,
2013 note=note or 'Downloading ISM manifest',
2014 errnote=errnote or 'Failed to download ISM manifest',
2015 fatal=fatal)
2016 if res is False:
2017 return []
2018 ism, urlh = res
2019
2020 return self._parse_ism_formats(
2021 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
2022
2023 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
2024 """
2025 Parse formats from ISM manifest.
2026 References:
2027 1. [MS-SSTR]: Smooth Streaming Protocol,
2028 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2029 """
2030 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
2031 return []
2032
2033 duration = int(ism_doc.attrib['Duration'])
2034 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2035
2036 formats = []
2037 for stream in ism_doc.findall('StreamIndex'):
2038 stream_type = stream.get('Type')
2039 if stream_type not in ('video', 'audio'):
2040 continue
2041 url_pattern = stream.attrib['Url']
2042 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2043 stream_name = stream.get('Name')
2044 for track in stream.findall('QualityLevel'):
2045 fourcc = track.get('FourCC')
2046 # TODO: add support for WVC1 and WMAP
2047 if fourcc not in ('H264', 'AVC1', 'AACL'):
2048 self.report_warning('%s is not a supported codec' % fourcc)
2049 continue
2050 tbr = int(track.attrib['Bitrate']) // 1000
2051 # [1] does not mention Width and Height attributes. However,
2052 # they're often present while MaxWidth and MaxHeight are
2053 # missing, so should be used as fallbacks
2054 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2055 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2056 sampling_rate = int_or_none(track.get('SamplingRate'))
2057
2058 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2059 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2060
2061 fragments = []
2062 fragment_ctx = {
2063 'time': 0,
2064 }
2065 stream_fragments = stream.findall('c')
2066 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2067 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2068 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2069 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2070 if not fragment_ctx['duration']:
2071 try:
2072 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2073 except IndexError:
2074 next_fragment_time = duration
2075 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2076 for _ in range(fragment_repeat):
2077 fragments.append({
2078 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2079 'duration': fragment_ctx['duration'] / stream_timescale,
2080 })
2081 fragment_ctx['time'] += fragment_ctx['duration']
2082
2083 format_id = []
2084 if ism_id:
2085 format_id.append(ism_id)
2086 if stream_name:
2087 format_id.append(stream_name)
2088 format_id.append(compat_str(tbr))
2089
2090 formats.append({
2091 'format_id': '-'.join(format_id),
2092 'url': ism_url,
2093 'manifest_url': ism_url,
2094 'ext': 'ismv' if stream_type == 'video' else 'isma',
2095 'width': width,
2096 'height': height,
2097 'tbr': tbr,
2098 'asr': sampling_rate,
2099 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2100 'acodec': 'none' if stream_type == 'video' else fourcc,
2101 'protocol': 'ism',
2102 'fragments': fragments,
2103 '_download_params': {
2104 'duration': duration,
2105 'timescale': stream_timescale,
2106 'width': width or 0,
2107 'height': height or 0,
2108 'fourcc': fourcc,
2109 'codec_private_data': track.get('CodecPrivateData'),
2110 'sampling_rate': sampling_rate,
2111 'channels': int_or_none(track.get('Channels', 2)),
2112 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2113 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2114 },
2115 })
2116 return formats
2117
2118 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2119 def absolute_url(video_url):
2120 return compat_urlparse.urljoin(base_url, video_url)
2121
2122 def parse_content_type(content_type):
2123 if not content_type:
2124 return {}
2125 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2126 if ctr:
2127 mimetype, codecs = ctr.groups()
2128 f = parse_codecs(codecs)
2129 f['ext'] = mimetype2ext(mimetype)
2130 return f
2131 return {}
2132
2133 def _media_formats(src, cur_media_type, type_info={}):
2134 full_url = absolute_url(src)
2135 ext = type_info.get('ext') or determine_ext(full_url)
2136 if ext == 'm3u8':
2137 is_plain_url = False
2138 formats = self._extract_m3u8_formats(
2139 full_url, video_id, ext='mp4',
2140 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2141 preference=preference, fatal=False)
2142 elif ext == 'mpd':
2143 is_plain_url = False
2144 formats = self._extract_mpd_formats(
2145 full_url, video_id, mpd_id=mpd_id, fatal=False)
2146 else:
2147 is_plain_url = True
2148 formats = [{
2149 'url': full_url,
2150 'vcodec': 'none' if cur_media_type == 'audio' else None,
2151 }]
2152 return is_plain_url, formats
2153
2154 entries = []
2155 # amp-video and amp-audio are very similar to their HTML5 counterparts
2156 # so we wll include them right here (see
2157 # https://www.ampproject.org/docs/reference/components/amp-video)
2158 media_tags = [(media_tag, media_type, '')
2159 for media_tag, media_type
2160 in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
2161 media_tags.extend(re.findall(
2162 # We only allow video|audio followed by a whitespace or '>'.
2163 # Allowing more characters may end up in significant slow down (see
2164 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2165 # http://www.porntrex.com/maps/videositemap.xml).
2166 r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2167 for media_tag, media_type, media_content in media_tags:
2168 media_info = {
2169 'formats': [],
2170 'subtitles': {},
2171 }
2172 media_attributes = extract_attributes(media_tag)
2173 src = media_attributes.get('src')
2174 if src:
2175 _, formats = _media_formats(src, media_type)
2176 media_info['formats'].extend(formats)
2177 media_info['thumbnail'] = media_attributes.get('poster')
2178 if media_content:
2179 for source_tag in re.findall(r'<source[^>]+>', media_content):
2180 source_attributes = extract_attributes(source_tag)
2181 src = source_attributes.get('src')
2182 if not src:
2183 continue
2184 f = parse_content_type(source_attributes.get('type'))
2185 is_plain_url, formats = _media_formats(src, media_type, f)
2186 if is_plain_url:
2187 # res attribute is not standard but seen several times
2188 # in the wild
2189 f.update({
2190 'height': int_or_none(source_attributes.get('res')),
2191 'format_id': source_attributes.get('label'),
2192 })
2193 f.update(formats[0])
2194 media_info['formats'].append(f)
2195 else:
2196 media_info['formats'].extend(formats)
2197 for track_tag in re.findall(r'<track[^>]+>', media_content):
2198 track_attributes = extract_attributes(track_tag)
2199 kind = track_attributes.get('kind')
2200 if not kind or kind in ('subtitles', 'captions'):
2201 src = track_attributes.get('src')
2202 if not src:
2203 continue
2204 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2205 media_info['subtitles'].setdefault(lang, []).append({
2206 'url': absolute_url(src),
2207 })
2208 if media_info['formats'] or media_info['subtitles']:
2209 entries.append(media_info)
2210 return entries
2211
2212 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2213 formats = []
2214 hdcore_sign = 'hdcore=3.7.0'
2215 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2216 hds_host = hosts.get('hds')
2217 if hds_host:
2218 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2219 if 'hdcore=' not in f4m_url:
2220 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2221 f4m_formats = self._extract_f4m_formats(
2222 f4m_url, video_id, f4m_id='hds', fatal=False)
2223 for entry in f4m_formats:
2224 entry.update({'extra_param_to_segment_url': hdcore_sign})
2225 formats.extend(f4m_formats)
2226 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2227 hls_host = hosts.get('hls')
2228 if hls_host:
2229 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2230 formats.extend(self._extract_m3u8_formats(
2231 m3u8_url, video_id, 'mp4', 'm3u8_native',
2232 m3u8_id='hls', fatal=False))
2233 return formats
2234
2235 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2236 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2237 url_base = self._search_regex(
2238 r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url')
2239 http_base_url = '%s:%s' % ('http', url_base)
2240 formats = []
2241 if 'm3u8' not in skip_protocols:
2242 formats.extend(self._extract_m3u8_formats(
2243 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2244 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2245 if 'f4m' not in skip_protocols:
2246 formats.extend(self._extract_f4m_formats(
2247 http_base_url + '/manifest.f4m',
2248 video_id, f4m_id='hds', fatal=False))
2249 if 'dash' not in skip_protocols:
2250 formats.extend(self._extract_mpd_formats(
2251 http_base_url + '/manifest.mpd',
2252 video_id, mpd_id='dash', fatal=False))
2253 if re.search(r'(?:/smil:|\.smil)', url_base):
2254 if 'smil' not in skip_protocols:
2255 rtmp_formats = self._extract_smil_formats(
2256 http_base_url + '/jwplayer.smil',
2257 video_id, fatal=False)
2258 for rtmp_format in rtmp_formats:
2259 rtsp_format = rtmp_format.copy()
2260 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2261 del rtsp_format['play_path']
2262 del rtsp_format['ext']
2263 rtsp_format.update({
2264 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2265 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2266 'protocol': 'rtsp',
2267 })
2268 formats.extend([rtmp_format, rtsp_format])
2269 else:
2270 for protocol in ('rtmp', 'rtsp'):
2271 if protocol not in skip_protocols:
2272 formats.append({
2273 'url': '%s:%s' % (protocol, url_base),
2274 'format_id': protocol,
2275 'protocol': protocol,
2276 })
2277 return formats
2278
2279 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2280 mobj = re.search(
2281 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
2282 webpage)
2283 if mobj:
2284 try:
2285 jwplayer_data = self._parse_json(mobj.group('options'),
2286 video_id=video_id,
2287 transform_source=transform_source)
2288 except ExtractorError:
2289 pass
2290 else:
2291 if isinstance(jwplayer_data, dict):
2292 return jwplayer_data
2293
2294 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2295 jwplayer_data = self._find_jwplayer_data(
2296 webpage, video_id, transform_source=js_to_json)
2297 return self._parse_jwplayer_data(
2298 jwplayer_data, video_id, *args, **kwargs)
2299
2300 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2301 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2302 # JWPlayer backward compatibility: flattened playlists
2303 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2304 if 'playlist' not in jwplayer_data:
2305 jwplayer_data = {'playlist': [jwplayer_data]}
2306
2307 entries = []
2308
2309 # JWPlayer backward compatibility: single playlist item
2310 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2311 if not isinstance(jwplayer_data['playlist'], list):
2312 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2313
2314 for video_data in jwplayer_data['playlist']:
2315 # JWPlayer backward compatibility: flattened sources
2316 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2317 if 'sources' not in video_data:
2318 video_data['sources'] = [video_data]
2319
2320 this_video_id = video_id or video_data['mediaid']
2321
2322 formats = self._parse_jwplayer_formats(
2323 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2324 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2325
2326 subtitles = {}
2327 tracks = video_data.get('tracks')
2328 if tracks and isinstance(tracks, list):
2329 for track in tracks:
2330 if not isinstance(track, dict):
2331 continue
2332 if track.get('kind') != 'captions':
2333 continue
2334 track_url = urljoin(base_url, track.get('file'))
2335 if not track_url:
2336 continue
2337 subtitles.setdefault(track.get('label') or 'en', []).append({
2338 'url': self._proto_relative_url(track_url)
2339 })
2340
2341 entry = {
2342 'id': this_video_id,
2343 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
2344 'description': video_data.get('description'),
2345 'thumbnail': self._proto_relative_url(video_data.get('image')),
2346 'timestamp': int_or_none(video_data.get('pubdate')),
2347 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2348 'subtitles': subtitles,
2349 }
2350 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
2351 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
2352 entry.update({
2353 '_type': 'url_transparent',
2354 'url': formats[0]['url'],
2355 })
2356 else:
2357 self._sort_formats(formats)
2358 entry['formats'] = formats
2359 entries.append(entry)
2360 if len(entries) == 1:
2361 return entries[0]
2362 else:
2363 return self.playlist_result(entries)
2364
2365 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2366 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2367 urls = []
2368 formats = []
2369 for source in jwplayer_sources_data:
2370 if not isinstance(source, dict):
2371 continue
2372 source_url = self._proto_relative_url(source.get('file'))
2373 if not source_url:
2374 continue
2375 if base_url:
2376 source_url = compat_urlparse.urljoin(base_url, source_url)
2377 if source_url in urls:
2378 continue
2379 urls.append(source_url)
2380 source_type = source.get('type') or ''
2381 ext = mimetype2ext(source_type) or determine_ext(source_url)
2382 if source_type == 'hls' or ext == 'm3u8':
2383 formats.extend(self._extract_m3u8_formats(
2384 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2385 m3u8_id=m3u8_id, fatal=False))
2386 elif ext == 'mpd':
2387 formats.extend(self._extract_mpd_formats(
2388 source_url, video_id, mpd_id=mpd_id, fatal=False))
2389 elif ext == 'smil':
2390 formats.extend(self._extract_smil_formats(
2391 source_url, video_id, fatal=False))
2392 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2393 elif source_type.startswith('audio') or ext in (
2394 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2395 formats.append({
2396 'url': source_url,
2397 'vcodec': 'none',
2398 'ext': ext,
2399 })
2400 else:
2401 height = int_or_none(source.get('height'))
2402 if height is None:
2403 # Often no height is provided but there is a label in
2404 # format like "1080p", "720p SD", or 1080.
2405 height = int_or_none(self._search_regex(
2406 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2407 'height', default=None))
2408 a_format = {
2409 'url': source_url,
2410 'width': int_or_none(source.get('width')),
2411 'height': height,
2412 'tbr': int_or_none(source.get('bitrate')),
2413 'ext': ext,
2414 }
2415 if source_url.startswith('rtmp'):
2416 a_format['ext'] = 'flv'
2417 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2418 # of jwplayer.flash.swf
2419 rtmp_url_parts = re.split(
2420 r'((?:mp4|mp3|flv):)', source_url, 1)
2421 if len(rtmp_url_parts) == 3:
2422 rtmp_url, prefix, play_path = rtmp_url_parts
2423 a_format.update({
2424 'url': rtmp_url,
2425 'play_path': prefix + play_path,
2426 })
2427 if rtmp_params:
2428 a_format.update(rtmp_params)
2429 formats.append(a_format)
2430 return formats
2431
2432 def _live_title(self, name):
2433 """ Generate the title for a live video """
2434 now = datetime.datetime.now()
2435 now_str = now.strftime('%Y-%m-%d %H:%M')
2436 return name + ' ' + now_str
2437
2438 def _int(self, v, name, fatal=False, **kwargs):
2439 res = int_or_none(v, **kwargs)
2440 if 'get_attr' in kwargs:
2441 print(getattr(v, kwargs['get_attr']))
2442 if res is None:
2443 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2444 if fatal:
2445 raise ExtractorError(msg)
2446 else:
2447 self._downloader.report_warning(msg)
2448 return res
2449
2450 def _float(self, v, name, fatal=False, **kwargs):
2451 res = float_or_none(v, **kwargs)
2452 if res is None:
2453 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2454 if fatal:
2455 raise ExtractorError(msg)
2456 else:
2457 self._downloader.report_warning(msg)
2458 return res
2459
2460 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
2461 path='/', secure=False, discard=False, rest={}, **kwargs):
2462 cookie = compat_cookiejar.Cookie(
2463 0, name, value, port, port is not None, domain, True,
2464 domain.startswith('.'), path, True, secure, expire_time,
2465 discard, None, None, rest)
2466 self._downloader.cookiejar.set_cookie(cookie)
2467
2468 def _get_cookies(self, url):
2469 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2470 req = sanitized_Request(url)
2471 self._downloader.cookiejar.add_cookie_header(req)
2472 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2473
2474 def get_testcases(self, include_onlymatching=False):
2475 t = getattr(self, '_TEST', None)
2476 if t:
2477 assert not hasattr(self, '_TESTS'), \
2478 '%s has _TEST and _TESTS' % type(self).__name__
2479 tests = [t]
2480 else:
2481 tests = getattr(self, '_TESTS', [])
2482 for t in tests:
2483 if not include_onlymatching and t.get('only_matching', False):
2484 continue
2485 t['name'] = type(self).__name__[:-len('IE')]
2486 yield t
2487
2488 def is_suitable(self, age_limit):
2489 """ Test whether the extractor is generally suitable for the given
2490 age limit (i.e. pornographic sites are not, all others usually are) """
2491
2492 any_restricted = False
2493 for tc in self.get_testcases(include_onlymatching=False):
2494 if tc.get('playlist', []):
2495 tc = tc['playlist'][0]
2496 is_restricted = age_restricted(
2497 tc.get('info_dict', {}).get('age_limit'), age_limit)
2498 if not is_restricted:
2499 return True
2500 any_restricted = any_restricted or is_restricted
2501 return not any_restricted
2502
2503 def extract_subtitles(self, *args, **kwargs):
2504 if (self._downloader.params.get('writesubtitles', False) or
2505 self._downloader.params.get('listsubtitles')):
2506 return self._get_subtitles(*args, **kwargs)
2507 return {}
2508
2509 def _get_subtitles(self, *args, **kwargs):
2510 raise NotImplementedError('This method must be implemented by subclasses')
2511
2512 @staticmethod
2513 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2514 """ Merge subtitle items for one language. Items with duplicated URLs
2515 will be dropped. """
2516 list1_urls = set([item['url'] for item in subtitle_list1])
2517 ret = list(subtitle_list1)
2518 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2519 return ret
2520
2521 @classmethod
2522 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2523 """ Merge two subtitle dictionaries, language by language. """
2524 ret = dict(subtitle_dict1)
2525 for lang in subtitle_dict2:
2526 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2527 return ret
2528
2529 def extract_automatic_captions(self, *args, **kwargs):
2530 if (self._downloader.params.get('writeautomaticsub', False) or
2531 self._downloader.params.get('listsubtitles')):
2532 return self._get_automatic_captions(*args, **kwargs)
2533 return {}
2534
2535 def _get_automatic_captions(self, *args, **kwargs):
2536 raise NotImplementedError('This method must be implemented by subclasses')
2537
2538 def mark_watched(self, *args, **kwargs):
2539 if (self._downloader.params.get('mark_watched', False) and
2540 (self._get_login_info()[0] is not None or
2541 self._downloader.params.get('cookiefile') is not None)):
2542 self._mark_watched(*args, **kwargs)
2543
2544 def _mark_watched(self, *args, **kwargs):
2545 raise NotImplementedError('This method must be implemented by subclasses')
2546
2547 def geo_verification_headers(self):
2548 headers = {}
2549 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2550 if geo_verification_proxy:
2551 headers['Ytdl-request-proxy'] = geo_verification_proxy
2552 return headers
2553
2554 def _generic_id(self, url):
2555 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2556
2557 def _generic_title(self, url):
2558 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2559
2560
2561 class SearchInfoExtractor(InfoExtractor):
2562 """
2563 Base class for paged search queries extractors.
2564 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2565 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2566 """
2567
2568 @classmethod
2569 def _make_valid_url(cls):
2570 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2571
2572 @classmethod
2573 def suitable(cls, url):
2574 return re.match(cls._make_valid_url(), url) is not None
2575
2576 def _real_extract(self, query):
2577 mobj = re.match(self._make_valid_url(), query)
2578 if mobj is None:
2579 raise ExtractorError('Invalid search query "%s"' % query)
2580
2581 prefix = mobj.group('prefix')
2582 query = mobj.group('query')
2583 if prefix == '':
2584 return self._get_n_results(query, 1)
2585 elif prefix == 'all':
2586 return self._get_n_results(query, self._MAX_RESULTS)
2587 else:
2588 n = int(prefix)
2589 if n <= 0:
2590 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2591 elif n > self._MAX_RESULTS:
2592 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2593 n = self._MAX_RESULTS
2594 return self._get_n_results(query, n)
2595
2596 def _get_n_results(self, query, n):
2597 """Get a specified number of results for a query"""
2598 raise NotImplementedError('This method must be implemented by subclasses')
2599
2600 @property
2601 def SEARCH_KEY(self):
2602 return self._SEARCH_KEY